From ed69becf3f95d1f841d3504f8b50e02b80bd04d6 Mon Sep 17 00:00:00 2001 From: Sayan Shaw Date: Tue, 14 May 2024 15:28:35 -0700 Subject: [PATCH 1/3] update opencv build to just include libspng and libjpeg libraries --- cmake/externals/cvout/.gitkeep | 0 cmake/externals/libjpeg/CMakeLists.txt | 30 + cmake/externals/libjpeg/README | 371 ++ cmake/externals/libjpeg/change.log | 502 ++ cmake/externals/libjpeg/jaricom.c | 153 + cmake/externals/libjpeg/jcapimin.c | 288 + cmake/externals/libjpeg/jcapistd.c | 162 + cmake/externals/libjpeg/jcarith.c | 945 ++++ cmake/externals/libjpeg/jccoefct.c | 456 ++ cmake/externals/libjpeg/jccolor.c | 598 ++ cmake/externals/libjpeg/jcdctmgr.c | 466 ++ cmake/externals/libjpeg/jchuff.c | 1656 ++++++ cmake/externals/libjpeg/jcinit.c | 249 + cmake/externals/libjpeg/jcmainct.c | 297 + cmake/externals/libjpeg/jcmarker.c | 717 +++ cmake/externals/libjpeg/jcmaster.c | 675 +++ cmake/externals/libjpeg/jcomapi.c | 244 + cmake/externals/libjpeg/jconfig.h | 64 + cmake/externals/libjpeg/jcparam.c | 591 ++ cmake/externals/libjpeg/jcprepct.c | 358 ++ cmake/externals/libjpeg/jcsample.c | 545 ++ cmake/externals/libjpeg/jctrans.c | 399 ++ cmake/externals/libjpeg/jdapimin.c | 412 ++ cmake/externals/libjpeg/jdapistd.c | 276 + cmake/externals/libjpeg/jdarith.c | 796 +++ cmake/externals/libjpeg/jdatadst.c | 263 + cmake/externals/libjpeg/jdatasrc.c | 271 + cmake/externals/libjpeg/jdcoefct.c | 744 +++ cmake/externals/libjpeg/jdcolor.c | 769 +++ cmake/externals/libjpeg/jdct.h | 409 ++ cmake/externals/libjpeg/jddctmgr.c | 384 ++ cmake/externals/libjpeg/jdhuff.c | 1559 ++++++ cmake/externals/libjpeg/jdinput.c | 657 +++ cmake/externals/libjpeg/jdmainct.c | 511 ++ cmake/externals/libjpeg/jdmarker.c | 1505 +++++ cmake/externals/libjpeg/jdmaster.c | 532 ++ cmake/externals/libjpeg/jdmerge.c | 437 ++ cmake/externals/libjpeg/jdpostct.c | 290 + cmake/externals/libjpeg/jdsample.c | 341 ++ cmake/externals/libjpeg/jdtrans.c | 140 + cmake/externals/libjpeg/jerror.c | 253 + cmake/externals/libjpeg/jerror.h | 304 ++ cmake/externals/libjpeg/jfdctflt.c | 176 + cmake/externals/libjpeg/jfdctfst.c | 232 + cmake/externals/libjpeg/jfdctint.c | 4415 +++++++++++++++ cmake/externals/libjpeg/jidctflt.c | 238 + cmake/externals/libjpeg/jidctfst.c | 351 ++ cmake/externals/libjpeg/jidctint.c | 5240 ++++++++++++++++++ cmake/externals/libjpeg/jinclude.h | 157 + cmake/externals/libjpeg/jmemansi.c | 167 + cmake/externals/libjpeg/jmemmgr.c | 1115 ++++ cmake/externals/libjpeg/jmemnobs.c | 113 + cmake/externals/libjpeg/jmemsys.h | 198 + cmake/externals/libjpeg/jmorecfg.h | 457 ++ cmake/externals/libjpeg/jpegint.h | 445 ++ cmake/externals/libjpeg/jpeglib.h | 1183 ++++ cmake/externals/libjpeg/jquant1.c | 851 +++ cmake/externals/libjpeg/jquant2.c | 1311 +++++ cmake/externals/libjpeg/jutils.c | 224 + cmake/externals/libjpeg/jversion.h | 14 + cmake/externals/libspng/CMakeLists.txt | 39 + cmake/externals/libspng/LICENSE | 25 + cmake/externals/libspng/spng.c | 6980 ++++++++++++++++++++++++ cmake/externals/libspng/spng.h | 537 ++ cmake/externals/opencv.cmake | 175 +- 65 files changed, 45099 insertions(+), 163 deletions(-) create mode 100644 cmake/externals/cvout/.gitkeep create mode 100644 cmake/externals/libjpeg/CMakeLists.txt create mode 100644 cmake/externals/libjpeg/README create mode 100644 cmake/externals/libjpeg/change.log create mode 100644 cmake/externals/libjpeg/jaricom.c create mode 100644 cmake/externals/libjpeg/jcapimin.c create mode 100644 cmake/externals/libjpeg/jcapistd.c create mode 100644 cmake/externals/libjpeg/jcarith.c create mode 100644 cmake/externals/libjpeg/jccoefct.c create mode 100644 cmake/externals/libjpeg/jccolor.c create mode 100644 cmake/externals/libjpeg/jcdctmgr.c create mode 100644 cmake/externals/libjpeg/jchuff.c create mode 100644 cmake/externals/libjpeg/jcinit.c create mode 100644 cmake/externals/libjpeg/jcmainct.c create mode 100644 cmake/externals/libjpeg/jcmarker.c create mode 100644 cmake/externals/libjpeg/jcmaster.c create mode 100644 cmake/externals/libjpeg/jcomapi.c create mode 100644 cmake/externals/libjpeg/jconfig.h create mode 100644 cmake/externals/libjpeg/jcparam.c create mode 100644 cmake/externals/libjpeg/jcprepct.c create mode 100644 cmake/externals/libjpeg/jcsample.c create mode 100644 cmake/externals/libjpeg/jctrans.c create mode 100644 cmake/externals/libjpeg/jdapimin.c create mode 100644 cmake/externals/libjpeg/jdapistd.c create mode 100644 cmake/externals/libjpeg/jdarith.c create mode 100644 cmake/externals/libjpeg/jdatadst.c create mode 100644 cmake/externals/libjpeg/jdatasrc.c create mode 100644 cmake/externals/libjpeg/jdcoefct.c create mode 100644 cmake/externals/libjpeg/jdcolor.c create mode 100644 cmake/externals/libjpeg/jdct.h create mode 100644 cmake/externals/libjpeg/jddctmgr.c create mode 100644 cmake/externals/libjpeg/jdhuff.c create mode 100644 cmake/externals/libjpeg/jdinput.c create mode 100644 cmake/externals/libjpeg/jdmainct.c create mode 100644 cmake/externals/libjpeg/jdmarker.c create mode 100644 cmake/externals/libjpeg/jdmaster.c create mode 100644 cmake/externals/libjpeg/jdmerge.c create mode 100644 cmake/externals/libjpeg/jdpostct.c create mode 100644 cmake/externals/libjpeg/jdsample.c create mode 100644 cmake/externals/libjpeg/jdtrans.c create mode 100644 cmake/externals/libjpeg/jerror.c create mode 100644 cmake/externals/libjpeg/jerror.h create mode 100644 cmake/externals/libjpeg/jfdctflt.c create mode 100644 cmake/externals/libjpeg/jfdctfst.c create mode 100644 cmake/externals/libjpeg/jfdctint.c create mode 100644 cmake/externals/libjpeg/jidctflt.c create mode 100644 cmake/externals/libjpeg/jidctfst.c create mode 100644 cmake/externals/libjpeg/jidctint.c create mode 100644 cmake/externals/libjpeg/jinclude.h create mode 100644 cmake/externals/libjpeg/jmemansi.c create mode 100644 cmake/externals/libjpeg/jmemmgr.c create mode 100644 cmake/externals/libjpeg/jmemnobs.c create mode 100644 cmake/externals/libjpeg/jmemsys.h create mode 100644 cmake/externals/libjpeg/jmorecfg.h create mode 100644 cmake/externals/libjpeg/jpegint.h create mode 100644 cmake/externals/libjpeg/jpeglib.h create mode 100644 cmake/externals/libjpeg/jquant1.c create mode 100644 cmake/externals/libjpeg/jquant2.c create mode 100644 cmake/externals/libjpeg/jutils.c create mode 100644 cmake/externals/libjpeg/jversion.h create mode 100644 cmake/externals/libspng/CMakeLists.txt create mode 100644 cmake/externals/libspng/LICENSE create mode 100644 cmake/externals/libspng/spng.c create mode 100644 cmake/externals/libspng/spng.h diff --git a/cmake/externals/cvout/.gitkeep b/cmake/externals/cvout/.gitkeep new file mode 100644 index 000000000..e69de29bb diff --git a/cmake/externals/libjpeg/CMakeLists.txt b/cmake/externals/libjpeg/CMakeLists.txt new file mode 100644 index 000000000..4e88164cc --- /dev/null +++ b/cmake/externals/libjpeg/CMakeLists.txt @@ -0,0 +1,30 @@ +# ---------------------------------------------------------------------------- +# CMake file for libjpeg. See root CMakeLists.txt +# +# ---------------------------------------------------------------------------- +project(JPEG_LIBRARY) + +file(GLOB lib_srcs *.c) +file(GLOB lib_hdrs *.h) + +# ---------------------------------------------------------------------------------- +# Define the library target: +# ---------------------------------------------------------------------------------- + +add_library(JPEG_LIBRARY STATIC ${OPENCV_3RDPARTY_EXCLUDE_FROM_ALL} ${lib_srcs} ${lib_hdrs}) + +if(CV_GCC OR CV_CLANG) + set_source_files_properties(jcdctmgr.c PROPERTIES COMPILE_FLAGS "-O1") +endif() + +set_target_properties(JPEG_LIBRARY + PROPERTIES OUTPUT_NAME JPEG_LIBRARY + DEBUG_POSTFIX "${OPENCV_DEBUG_POSTFIX}" + COMPILE_PDB_NAME JPEG_LIBRARY + COMPILE_PDB_NAME_DEBUG "JPEG_LIBRARY${OPENCV_DEBUG_POSTFIX}" + ARCHIVE_OUTPUT_DIRECTORY "${3P_LIBRARY_OUTPUT_PATH}" + ) + +if(ENABLE_SOLUTION_FOLDERS) + set_target_properties(JPEG_LIBRARY PROPERTIES FOLDER "3rdparty") +endif() diff --git a/cmake/externals/libjpeg/README b/cmake/externals/libjpeg/README new file mode 100644 index 000000000..13b245c5c --- /dev/null +++ b/cmake/externals/libjpeg/README @@ -0,0 +1,371 @@ +The Independent JPEG Group's JPEG software +========================================== + +README for release 9d of 12-Jan-2020 +==================================== + +This distribution contains the ninth public release of the Independent JPEG +Group's free JPEG software. You are welcome to redistribute this software and +to use it for any purpose, subject to the conditions under LEGAL ISSUES, below. + +This software is the work of Tom Lane, Guido Vollbeding, Philip Gladstone, +Bill Allombert, Jim Boucher, Lee Crocker, Bob Friesenhahn, Ben Jackson, +John Korejwa, Julian Minguillon, Luis Ortiz, George Phillips, Davide Rossi, +Ge' Weijers, and other members of the Independent JPEG Group. + +IJG is not affiliated with the ISO/IEC JTC1/SC29/WG1 standards committee +(previously known as JPEG, together with ITU-T SG16). + + +DOCUMENTATION ROADMAP +===================== + +This file contains the following sections: + +OVERVIEW General description of JPEG and the IJG software. +LEGAL ISSUES Copyright, lack of warranty, terms of distribution. +REFERENCES Where to learn more about JPEG. +ARCHIVE LOCATIONS Where to find newer versions of this software. +ACKNOWLEDGMENTS Special thanks. +FILE FORMAT WARS Software *not* to get. +TO DO Plans for future IJG releases. + +Other documentation files in the distribution are: + +User documentation: + install.txt How to configure and install the IJG software. + usage.txt Usage instructions for cjpeg, djpeg, jpegtran, + rdjpgcom, and wrjpgcom. + *.1 Unix-style man pages for programs (same info as usage.txt). + wizard.txt Advanced usage instructions for JPEG wizards only. + change.log Version-to-version change highlights. +Programmer and internal documentation: + libjpeg.txt How to use the JPEG library in your own programs. + example.c Sample code for calling the JPEG library. + structure.txt Overview of the JPEG library's internal structure. + filelist.txt Road map of IJG files. + coderules.txt Coding style rules --- please read if you contribute code. + +Please read at least the files install.txt and usage.txt. Some information +can also be found in the JPEG FAQ (Frequently Asked Questions) article. See +ARCHIVE LOCATIONS below to find out where to obtain the FAQ article. + +If you want to understand how the JPEG code works, we suggest reading one or +more of the REFERENCES, then looking at the documentation files (in roughly +the order listed) before diving into the code. + + +OVERVIEW +======== + +This package contains C software to implement JPEG image encoding, decoding, +and transcoding. JPEG (pronounced "jay-peg") is a standardized compression +method for full-color and grayscale images. + +This software implements JPEG baseline, extended-sequential, and progressive +compression processes. Provision is made for supporting all variants of these +processes, although some uncommon parameter settings aren't implemented yet. +We have made no provision for supporting the hierarchical or lossless +processes defined in the standard. + +We provide a set of library routines for reading and writing JPEG image files, +plus two sample applications "cjpeg" and "djpeg", which use the library to +perform conversion between JPEG and some other popular image file formats. +The library is intended to be reused in other applications. + +In order to support file conversion and viewing software, we have included +considerable functionality beyond the bare JPEG coding/decoding capability; +for example, the color quantization modules are not strictly part of JPEG +decoding, but they are essential for output to colormapped file formats or +colormapped displays. These extra functions can be compiled out of the +library if not required for a particular application. + +We have also included "jpegtran", a utility for lossless transcoding between +different JPEG processes, and "rdjpgcom" and "wrjpgcom", two simple +applications for inserting and extracting textual comments in JFIF files. + +The emphasis in designing this software has been on achieving portability and +flexibility, while also making it fast enough to be useful. In particular, +the software is not intended to be read as a tutorial on JPEG. (See the +REFERENCES section for introductory material.) Rather, it is intended to +be reliable, portable, industrial-strength code. We do not claim to have +achieved that goal in every aspect of the software, but we strive for it. + +We welcome the use of this software as a component of commercial products. +No royalty is required, but we do ask for an acknowledgement in product +documentation, as described under LEGAL ISSUES. + + +LEGAL ISSUES +============ + +In plain English: + +1. We don't promise that this software works. (But if you find any bugs, + please let us know!) +2. You can use this software for whatever you want. You don't have to pay us. +3. You may not pretend that you wrote this software. If you use it in a + program, you must acknowledge somewhere in your documentation that + you've used the IJG code. + +In legalese: + +The authors make NO WARRANTY or representation, either express or implied, +with respect to this software, its quality, accuracy, merchantability, or +fitness for a particular purpose. This software is provided "AS IS", and you, +its user, assume the entire risk as to its quality and accuracy. + +This software is copyright (C) 1991-2020, Thomas G. Lane, Guido Vollbeding. +All Rights Reserved except as specified below. + +Permission is hereby granted to use, copy, modify, and distribute this +software (or portions thereof) for any purpose, without fee, subject to these +conditions: +(1) If any part of the source code for this software is distributed, then this +README file must be included, with this copyright and no-warranty notice +unaltered; and any additions, deletions, or changes to the original files +must be clearly indicated in accompanying documentation. +(2) If only executable code is distributed, then the accompanying +documentation must state that "this software is based in part on the work of +the Independent JPEG Group". +(3) Permission for use of this software is granted only if the user accepts +full responsibility for any undesirable consequences; the authors accept +NO LIABILITY for damages of any kind. + +These conditions apply to any software derived from or based on the IJG code, +not just to the unmodified library. If you use our work, you ought to +acknowledge us. + +Permission is NOT granted for the use of any IJG author's name or company name +in advertising or publicity relating to this software or products derived from +it. This software may be referred to only as "the Independent JPEG Group's +software". + +We specifically permit and encourage the use of this software as the basis of +commercial products, provided that all warranty or liability claims are +assumed by the product vendor. + + +The Unix configuration script "configure" was produced with GNU Autoconf. +It is copyright by the Free Software Foundation but is freely distributable. +The same holds for its supporting scripts (config.guess, config.sub, +ltmain.sh). Another support script, install-sh, is copyright by X Consortium +but is also freely distributable. + + +REFERENCES +========== + +We recommend reading one or more of these references before trying to +understand the innards of the JPEG software. + +The best short technical introduction to the JPEG compression algorithm is + Wallace, Gregory K. "The JPEG Still Picture Compression Standard", + Communications of the ACM, April 1991 (vol. 34 no. 4), pp. 30-44. +(Adjacent articles in that issue discuss MPEG motion picture compression, +applications of JPEG, and related topics.) If you don't have the CACM issue +handy, a PDF file containing a revised version of Wallace's article is +available at http://www.ijg.org/files/Wallace.JPEG.pdf. The file (actually +a preprint for an article that appeared in IEEE Trans. Consumer Electronics) +omits the sample images that appeared in CACM, but it includes corrections +and some added material. Note: the Wallace article is copyright ACM and IEEE, +and it may not be used for commercial purposes. + +A somewhat less technical, more leisurely introduction to JPEG can be found in +"The Data Compression Book" by Mark Nelson and Jean-loup Gailly, published by +M&T Books (New York), 2nd ed. 1996, ISBN 1-55851-434-1. This book provides +good explanations and example C code for a multitude of compression methods +including JPEG. It is an excellent source if you are comfortable reading C +code but don't know much about data compression in general. The book's JPEG +sample code is far from industrial-strength, but when you are ready to look +at a full implementation, you've got one here... + +The best currently available description of JPEG is the textbook "JPEG Still +Image Data Compression Standard" by William B. Pennebaker and Joan L. +Mitchell, published by Van Nostrand Reinhold, 1993, ISBN 0-442-01272-1. +Price US$59.95, 638 pp. The book includes the complete text of the ISO JPEG +standards (DIS 10918-1 and draft DIS 10918-2). +Although this is by far the most detailed and comprehensive exposition of +JPEG publicly available, we point out that it is still missing an explanation +of the most essential properties and algorithms of the underlying DCT +technology. +If you think that you know about DCT-based JPEG after reading this book, +then you are in delusion. The real fundamentals and corresponding potential +of DCT-based JPEG are not publicly known so far, and that is the reason for +all the mistaken developments taking place in the image coding domain. + +The original JPEG standard is divided into two parts, Part 1 being the actual +specification, while Part 2 covers compliance testing methods. Part 1 is +titled "Digital Compression and Coding of Continuous-tone Still Images, +Part 1: Requirements and guidelines" and has document numbers ISO/IEC IS +10918-1, ITU-T T.81. Part 2 is titled "Digital Compression and Coding of +Continuous-tone Still Images, Part 2: Compliance testing" and has document +numbers ISO/IEC IS 10918-2, ITU-T T.83. +IJG JPEG 8 introduced an implementation of the JPEG SmartScale extension +which is specified in two documents: A contributed document at ITU and ISO +with title "ITU-T JPEG-Plus Proposal for Extending ITU-T T.81 for Advanced +Image Coding", April 2006, Geneva, Switzerland. The latest version of this +document is Revision 3. And a contributed document ISO/IEC JTC1/SC29/WG1 N +5799 with title "Evolution of JPEG", June/July 2011, Berlin, Germany. +IJG JPEG 9 introduces a reversible color transform for improved lossless +compression which is described in a contributed document ISO/IEC JTC1/SC29/ +WG1 N 6080 with title "JPEG 9 Lossless Coding", June/July 2012, Paris, +France. + +The JPEG standard does not specify all details of an interchangeable file +format. For the omitted details we follow the "JFIF" conventions, version 2. +JFIF version 1 has been adopted as Recommendation ITU-T T.871 (05/2011) : +Information technology - Digital compression and coding of continuous-tone +still images: JPEG File Interchange Format (JFIF). It is available as a +free download in PDF file format from http://www.itu.int/rec/T-REC-T.871. +A PDF file of the older JFIF document is available at +http://www.w3.org/Graphics/JPEG/jfif3.pdf. + +The TIFF 6.0 file format specification can be obtained by FTP from +ftp://ftp.sgi.com/graphics/tiff/TIFF6.ps.gz. The JPEG incorporation scheme +found in the TIFF 6.0 spec of 3-June-92 has a number of serious problems. +IJG does not recommend use of the TIFF 6.0 design (TIFF Compression tag 6). +Instead, we recommend the JPEG design proposed by TIFF Technical Note #2 +(Compression tag 7). Copies of this Note can be obtained from +http://www.ijg.org/files/. It is expected that the next revision +of the TIFF spec will replace the 6.0 JPEG design with the Note's design. +Although IJG's own code does not support TIFF/JPEG, the free libtiff library +uses our library to implement TIFF/JPEG per the Note. + + +ARCHIVE LOCATIONS +================= + +The "official" archive site for this software is www.ijg.org. +The most recent released version can always be found there in +directory "files". This particular version will be archived as +http://www.ijg.org/files/jpegsrc.v9d.tar.gz, and in Windows-compatible +"zip" archive format as http://www.ijg.org/files/jpegsr9d.zip. + +The JPEG FAQ (Frequently Asked Questions) article is a source of some +general information about JPEG. +It is available on the World Wide Web at http://www.faqs.org/faqs/jpeg-faq/ +and other news.answers archive sites, including the official news.answers +archive at rtfm.mit.edu: ftp://rtfm.mit.edu/pub/usenet/news.answers/jpeg-faq/. +If you don't have Web or FTP access, send e-mail to mail-server@rtfm.mit.edu +with body + send usenet/news.answers/jpeg-faq/part1 + send usenet/news.answers/jpeg-faq/part2 + + +ACKNOWLEDGMENTS +=============== + +Thank to Juergen Bruder for providing me with a copy of the common DCT +algorithm article, only to find out that I had come to the same result +in a more direct and comprehensible way with a more generative approach. + +Thank to Istvan Sebestyen and Joan L. Mitchell for inviting me to the +ITU JPEG (Study Group 16) meeting in Geneva, Switzerland. + +Thank to Thomas Wiegand and Gary Sullivan for inviting me to the +Joint Video Team (MPEG & ITU) meeting in Geneva, Switzerland. + +Thank to Thomas Richter and Daniel Lee for inviting me to the +ISO/IEC JTC1/SC29/WG1 (previously known as JPEG, together with ITU-T SG16) +meeting in Berlin, Germany. + +Thank to John Korejwa and Massimo Ballerini for inviting me to +fruitful consultations in Boston, MA and Milan, Italy. + +Thank to Hendrik Elstner, Roland Fassauer, Simone Zuck, Guenther +Maier-Gerber, Walter Stoeber, Fred Schmitz, and Norbert Braunagel +for corresponding business development. + +Thank to Nico Zschach and Dirk Stelling of the technical support team +at the Digital Images company in Halle for providing me with extra +equipment for configuration tests. + +Thank to Richard F. Lyon (then of Foveon Inc.) for fruitful +communication about JPEG configuration in Sigma Photo Pro software. + +Thank to Andrew Finkenstadt for hosting the ijg.org site. + +Thank to Thomas G. Lane for the original design and development of +this singular software package. + +Thank to Lars Goehler, Andreas Heinecke, Sebastian Fuss, Yvonne Roebert, +Andrej Werner, and Ulf-Dietrich Braumann for support and public relations. + + +FILE FORMAT WARS +================ + +The ISO/IEC JTC1/SC29/WG1 standards committee (previously known as JPEG, +together with ITU-T SG16) currently promotes different formats containing +the name "JPEG" which is misleading because these formats are incompatible +with original DCT-based JPEG and are based on faulty technologies. +IJG therefore does not and will not support such momentary mistakes +(see REFERENCES). +There exist also distributions under the name "OpenJPEG" promoting such +kind of formats which is misleading because they don't support original +JPEG images. +We have no sympathy for the promotion of inferior formats. Indeed, one of +the original reasons for developing this free software was to help force +convergence on common, interoperable format standards for JPEG files. +Don't use an incompatible file format! +(In any case, our decoder will remain capable of reading existing JPEG +image files indefinitely.) + +The ISO committee pretends to be "responsible for the popular JPEG" in their +public reports which is not true because they don't respond to actual +requirements for the maintenance of the original JPEG specification. +Furthermore, the ISO committee pretends to "ensure interoperability" with +their standards which is not true because their "standards" support only +application-specific and proprietary use cases and contain mathematically +incorrect code. + +There are currently different distributions in circulation containing the +name "libjpeg" which is misleading because they don't have the features and +are incompatible with formats supported by actual IJG libjpeg distributions. +One of those fakes is released by members of the ISO committee and just uses +the name of libjpeg for misdirection of people, similar to the abuse of the +name JPEG as described above, while having nothing in common with actual IJG +libjpeg distributions and containing mathematically incorrect code. +The other one claims to be a "derivative" or "fork" of the original libjpeg, +but violates the license conditions as described under LEGAL ISSUES above +and violates basic C programming properties. +We have no sympathy for the release of misleading, incorrect and illegal +distributions derived from obsolete code bases. +Don't use an obsolete code base! + +According to the UCC (Uniform Commercial Code) law, IJG has the lawful and +legal right to foreclose on certain standardization bodies and other +institutions or corporations that knowingly perform substantial and +systematic deceptive acts and practices, fraud, theft, and damaging of the +value of the people of this planet without their knowing, willing and +intentional consent. +The titles, ownership, and rights of these institutions and all their assets +are now duly secured and held in trust for the free people of this planet. +People of the planet, on every country, may have a financial interest in +the assets of these former principals, agents, and beneficiaries of the +foreclosed institutions and corporations. +IJG asserts what is: that each man, woman, and child has unalienable value +and rights granted and deposited in them by the Creator and not any one of +the people is subordinate to any artificial principality, corporate fiction +or the special interest of another without their appropriate knowing, +willing and intentional consent made by contract or accommodation agreement. +IJG expresses that which already was. +The people have already determined and demanded that public administration +entities, national governments, and their supporting judicial systems must +be fully transparent, accountable, and liable. +IJG has secured the value for all concerned free people of the planet. + +A partial list of foreclosed institutions and corporations ("Hall of Shame") +is currently prepared and will be published later. + + +TO DO +===== + +Version 9 is the second release of a new generation JPEG standard +to overcome the limitations of the original JPEG specification, +and is the first true source reference JPEG codec. +More features are being prepared for coming releases... + +Please send bug reports, offers of help, etc. to jpeg-info@jpegclub.org. diff --git a/cmake/externals/libjpeg/change.log b/cmake/externals/libjpeg/change.log new file mode 100644 index 000000000..ed8ec6afc --- /dev/null +++ b/cmake/externals/libjpeg/change.log @@ -0,0 +1,502 @@ +CHANGE LOG for Independent JPEG Group's JPEG software + + +Version 9d 12-Jan-2020 +----------------------- + +Optimize the optimal Huffman code table generation to produce +slightly smaller files. Thank to John Korejwa for suggestion. +Note: Requires rebuild of testimgp.jpg. + +Decoding Huffman: Use default tables if tables are not defined. +Thank to Simone Azzalin for report (Motion JPEG), +and to Martin Strunz for hint. + +Add sanity check in optimal Huffman code table generation. +Thank to Adam Farley for suggestion. + +rdtarga.c: use read_byte(), with EOF check, instead of getc() +in read_*_pixel(). +Thank to Chijin Zhou for cjpeg potential vulnerability report. + +jmemnobs.c: respect the max_memory_to_use setting in +jpeg_mem_available() computation. Thank to Sheng Shu and +Dongdong She for djpeg potential vulnerability report. + +jdarith.c, jdhuff.c: avoid left shift of negative value +compiler warning in decode_mcu_AC_refine(). +Thank to Indu Bhagat for suggestion. + +Add x64 (64-bit) platform support, avoid compiler warnings. +Thank to Jonathan Potter, Feiyun Wang, and Sheng Shu for suggestion. + +Adjust libjpeg version specification for pkg-config file. +Thank to Chen Chen for suggestion. + +Restore GIF read and write support from libjpeg version 6a. +Thank to Wolfgang Werner (W.W.) Heinz for suggestion. + +Improve consistency in raw (downsampled) image data processing mode. +Thank to Zhongyuan Zhou for hint. + +Avoid out of bounds array read (AC derived table pointers) +in start pass in jdhuff.c. Thank to Peng Li for report. + +Improve code sanity (jdhuff.c). +Thank to Reza Mirzazade farkhani for reports. + +Add jpegtran -drop option; add options to the crop extension and wipe +to fill the extra area with content from the source image region, +instead of gray out. + + +Version 9c 14-Jan-2018 +----------------------- + +jpegtran: add an option to the -wipe switch to fill the region +with the average of adjacent blocks, instead of gray out. +Thank to Caitlyn Feddock and Maddie Ziegler for inspiration. + +Make range extension bits adjustable (in jpegint.h). +Thank to Robin Watts for suggestion. + +Provide macros for fflush() and ferror() in jinclude.h in order +to facilitate adaption by applications using an own FILE class. +Thank to Gerhard Huber for suggestion. + +Add libjpeg pkg-config file. Thank to Mark Lavi, Vincent Torri, +Patrick McMunn, and Huw Davies for suggestion. + +Add sanity checks in cjpeg image reader modules. +Thank to Bingchang, Liu for reports. + + +Version 9b 17-Jan-2016 +----------------------- + +Improvements and optimizations in DCT and color calculations. +Normalize range limit array composition and access pattern. +Thank to Sia Furler and Maddie Ziegler for inspiration. + +Use merged upsample with scaled DCT sizes larger than 8. +Thank to Taylor Hatala for inspiration. + +Check for excessive comment lengths in argument parsing in wrjpgcom.c. +Thank to Julian Cohen for hint. + +Add makefile.b32 for use with Borland C++ 32-bit (bcc32). +Thank to Joe Slater for contribution. + +Document 'f' specifier for jpegtran -crop specification. +Thank to Michele Martone for suggestion. + +Use defined value from header instead of hardwired number in rdswitch.c. +Thank to Robert Sprowson for hint. + + +Version 9a 19-Jan-2014 +----------------------- + +Add support for wide gamut color spaces (JFIF version 2). +Improve clarity and accuracy in color conversion modules. +Note: Requires rebuild of test images. + +Extend the bit depth support to all values from 8 to 12 +(BITS_IN_JSAMPLE configuration option in jmorecfg.h). +jpegtran now supports N bits sample data precision with all N from 8 to 12 +in a single instance. Thank to Roland Fassauer for inspiration. + +Try to resolve issues with new boolean type definition. +Thank also to v4hn for suggestion. + +Enable option to use default Huffman tables for lossless compression +(for hardware solution), and in this case improve lossless RGB compression +with reversible color transform. Thank to Benny Alexandar for hint. + +Extend the entropy decoding structure, so that extraneous bytes between +compressed scan data and following marker can be reported correctly. +Thank to Nigel Tao for hint. + +Add jpegtran -wipe option and extension for -crop. +Thank to Andrew Senior, David Clunie, and Josef Schmid for suggestion. + + +Version 9 13-Jan-2013 +---------------------- + +Add cjpeg -rgb1 option to create an RGB JPEG file, and insert +a simple reversible color transform into the processing which +significantly improves the compression. +The recommended command for lossless coding of RGB images is now +cjpeg -rgb1 -block 1 -arithmetic. +As said, this option improves the compression significantly, but +the files are not compatible with JPEG decoders prior to IJG v9 +due to the included color transform. +The used color transform and marker signaling is compatible with +other JPEG standards (e.g., JPEG-LS part 2). + +Remove the automatic de-ANSI-fication support (Automake 1.12). +Thank also to Nitin A Kamble for suggestion. + +Add remark for jpeg_mem_dest() in jdatadst.c. +Thank to Elie-Gregoire Khoury for the hint. + +Support files with invalid component identifiers (created +by Adobe PDF). Thank to Robin Watts for the suggestion. + +Adapt full buffer case in jcmainct.c for use with scaled DCT. +Thank to Sergii Biloshytskyi for the suggestion. + +Add type identifier for declaration of noreturn functions. +Thank to Brett L. Moore for the suggestion. + +Correct argument type in format string, avoid compiler warnings. +Thank to Vincent Torri for hint. + +Add missing #include directives in configuration checks, avoid +configuration errors. Thank to John Spencer for the hint. + + +Version 8d 15-Jan-2012 +----------------------- + +Add cjpeg -rgb option to create RGB JPEG files. +Using this switch suppresses the conversion from RGB +colorspace input to the default YCbCr JPEG colorspace. +This feature allows true lossless JPEG coding of RGB color images. +The recommended command for this purpose is currently +cjpeg -rgb -block 1 -arithmetic. +SmartScale capable decoder (introduced with IJG JPEG 8) required. +Thank to Michael Koch for the initial suggestion. + +Add option to disable the region adjustment in the transupp crop code. +Thank to Jeffrey Friedl for the suggestion. + +Thank to Richard Jones and Edd Dawson for various minor corrections. + +Thank to Akim Demaille for configure.ac cleanup. + + +Version 8c 16-Jan-2011 +----------------------- + +Add option to compression library and cjpeg (-block N) to use +different DCT block size. +All N from 1 to 16 are possible. Default is 8 (baseline format). +Larger values produce higher compression, +smaller values produce higher quality. +SmartScale capable decoder (introduced with IJG JPEG 8) required. + + +Version 8b 16-May-2010 +----------------------- + +Repair problem in new memory source manager with corrupt JPEG data. +Thank to Ted Campbell and Samuel Chun for the report. + +Repair problem in Makefile.am test target. +Thank to anonymous user for the report. + +Support MinGW installation with automatic configure. +Thank to Volker Grabsch for the suggestion. + + +Version 8a 28-Feb-2010 +----------------------- + +Writing tables-only datastreams via jpeg_write_tables works again. + +Support 32-bit BMPs (RGB image with Alpha channel) for read in cjpeg. +Thank to Brett Blackham for the suggestion. + +Improve accuracy in floating point IDCT calculation. +Thank to Robert Hooke for the hint. + + +Version 8 10-Jan-2010 +---------------------- + +jpegtran now supports the same -scale option as djpeg for "lossless" resize. +An implementation of the JPEG SmartScale extension is required for this +feature. A (draft) specification of the JPEG SmartScale extension is +available as a contributed document at ITU and ISO. Revision 2 or later +of the document is required (latest document version is Revision 3). +The SmartScale extension will enable more features beside lossless resize +in future implementations, as described in the document (new compression +options). + +Add sanity check in BMP reader module to avoid cjpeg crash for empty input +image (thank to Isaev Ildar of ISP RAS, Moscow, RU for reporting this error). + +Add data source and destination managers for read from and write to +memory buffers. New API functions jpeg_mem_src and jpeg_mem_dest. +Thank to Roberto Boni from Italy for the suggestion. + + +Version 7 27-Jun-2009 +---------------------- + +New scaled DCTs implemented. +djpeg now supports scalings N/8 with all N from 1 to 16. +cjpeg now supports scalings 8/N with all N from 1 to 16. +Scaled DCTs with size larger than 8 are now also used for resolving the +common 2x2 chroma subsampling case without additional spatial resampling. +Separate spatial resampling for those kind of files is now only necessary +for N>8 scaling cases. +Furthermore, separate scaled DCT functions are provided for direct resolving +of the common asymmetric subsampling cases (2x1 and 1x2) without additional +spatial resampling. + +cjpeg -quality option has been extended for support of separate quality +settings for luminance and chrominance (or in general, for every provided +quantization table slot). +New API function jpeg_default_qtables() and q_scale_factor array in library. + +Added -nosmooth option to cjpeg, complementary to djpeg. +New variable "do_fancy_downsampling" in library, complement to fancy +upsampling. Fancy upsampling now uses direct DCT scaling with sizes +larger than 8. The old method is not reversible and has been removed. + +Support arithmetic entropy encoding and decoding. +Added files jaricom.c, jcarith.c, jdarith.c. + +Straighten the file structure: +Removed files jidctred.c, jcphuff.c, jchuff.h, jdphuff.c, jdhuff.h. + +jpegtran has a new "lossless" cropping feature. + +Implement -perfect option in jpegtran, new API function +jtransform_perfect_transform() in transupp. (DP 204_perfect.dpatch) + +Better error messages for jpegtran fopen failure. +(DP 203_jpegtran_errmsg.dpatch) + +Fix byte order issue with 16bit PPM/PGM files in rdppm.c/wrppm.c: +according to Netpbm, the de facto standard implementation of the PNM formats, +the most significant byte is first. (DP 203_rdppm.dpatch) + +Add -raw option to rdjpgcom not to mangle the output. +(DP 205_rdjpgcom_raw.dpatch) + +Make rdjpgcom locale aware. (DP 201_rdjpgcom_locale.dpatch) + +Add extern "C" to jpeglib.h. +This avoids the need to put extern "C" { ... } around #include "jpeglib.h" +in your C++ application. Defining the symbol DONT_USE_EXTERN_C in the +configuration prevents this. (DP 202_jpeglib.h_c++.dpatch) + + +Version 6b 27-Mar-1998 +----------------------- + +jpegtran has new features for lossless image transformations (rotation +and flipping) as well as "lossless" reduction to grayscale. + +jpegtran now copies comments by default; it has a -copy switch to enable +copying all APPn blocks as well, or to suppress comments. (Formerly it +always suppressed comments and APPn blocks.) jpegtran now also preserves +JFIF version and resolution information. + +New decompressor library feature: COM and APPn markers found in the input +file can be saved in memory for later use by the application. (Before, +you had to code this up yourself with a custom marker processor.) + +There is an unused field "void * client_data" now in compress and decompress +parameter structs; this may be useful in some applications. + +JFIF version number information is now saved by the decoder and accepted by +the encoder. jpegtran uses this to copy the source file's version number, +to ensure "jpegtran -copy all" won't create bogus files that contain JFXX +extensions but claim to be version 1.01. Applications that generate their +own JFXX extension markers also (finally) have a supported way to cause the +encoder to emit JFIF version number 1.02. + +djpeg's trace mode reports JFIF 1.02 thumbnail images as such, rather +than as unknown APP0 markers. + +In -verbose mode, djpeg and rdjpgcom will try to print the contents of +APP12 markers as text. Some digital cameras store useful text information +in APP12 markers. + +Handling of truncated data streams is more robust: blocks beyond the one in +which the error occurs will be output as uniform gray, or left unchanged +if decoding a progressive JPEG. The appearance no longer depends on the +Huffman tables being used. + +Huffman tables are checked for validity much more carefully than before. + +To avoid the Unisys LZW patent, djpeg's GIF output capability has been +changed to produce "uncompressed GIFs", and cjpeg's GIF input capability +has been removed altogether. We're not happy about it either, but there +seems to be no good alternative. + +The configure script now supports building libjpeg as a shared library +on many flavors of Unix (all the ones that GNU libtool knows how to +build shared libraries for). Use "./configure --enable-shared" to +try this out. + +New jconfig file and makefiles for Microsoft Visual C++ and Developer Studio. +Also, a jconfig file and a build script for Metrowerks CodeWarrior +on Apple Macintosh. makefile.dj has been updated for DJGPP v2, and there +are miscellaneous other minor improvements in the makefiles. + +jmemmac.c now knows how to create temporary files following Mac System 7 +conventions. + +djpeg's -map switch is now able to read raw-format PPM files reliably. + +cjpeg -progressive -restart no longer generates any unnecessary DRI markers. + +Multiple calls to jpeg_simple_progression for a single JPEG object +no longer leak memory. + + +Version 6a 7-Feb-96 +-------------------- + +Library initialization sequence modified to detect version mismatches +and struct field packing mismatches between library and calling application. +This change requires applications to be recompiled, but does not require +any application source code change. + +All routine declarations changed to the style "GLOBAL(type) name ...", +that is, GLOBAL, LOCAL, METHODDEF, EXTERN are now macros taking the +routine's return type as an argument. This makes it possible to add +Microsoft-style linkage keywords to all the routines by changing just +these macros. Note that any application code that was using these macros +will have to be changed. + +DCT coefficient quantization tables are now stored in normal array order +rather than zigzag order. Application code that calls jpeg_add_quant_table, +or otherwise manipulates quantization tables directly, will need to be +changed. If you need to make such code work with either older or newer +versions of the library, a test like "#if JPEG_LIB_VERSION >= 61" is +recommended. + +djpeg's trace capability now dumps DQT tables in natural order, not zigzag +order. This allows the trace output to be made into a "-qtables" file +more easily. + +New system-dependent memory manager module for use on Apple Macintosh. + +Fix bug in cjpeg's -smooth option: last one or two scanlines would be +duplicates of the prior line unless the image height mod 16 was 1 or 2. + +Repair minor problems in VMS, BCC, MC6 makefiles. + +New configure script based on latest GNU Autoconf. + +Correct the list of include files needed by MetroWerks C for ccommand(). + +Numerous small documentation updates. + + +Version 6 2-Aug-95 +------------------- + +Progressive JPEG support: library can read and write full progressive JPEG +files. A "buffered image" mode supports incremental decoding for on-the-fly +display of progressive images. Simply recompiling an existing IJG-v5-based +decoder with v6 should allow it to read progressive files, though of course +without any special progressive display. + +New "jpegtran" application performs lossless transcoding between different +JPEG formats; primarily, it can be used to convert baseline to progressive +JPEG and vice versa. In support of jpegtran, the library now allows lossless +reading and writing of JPEG files as DCT coefficient arrays. This ability +may be of use in other applications. + +Notes for programmers: +* We changed jpeg_start_decompress() to be able to suspend; this makes all +decoding modes available to suspending-input applications. However, +existing applications that use suspending input will need to be changed +to check the return value from jpeg_start_decompress(). You don't need to +do anything if you don't use a suspending data source. +* We changed the interface to the virtual array routines: access_virt_array +routines now take a count of the number of rows to access this time. The +last parameter to request_virt_array routines is now interpreted as the +maximum number of rows that may be accessed at once, but not necessarily +the height of every access. + + +Version 5b 15-Mar-95 +--------------------- + +Correct bugs with grayscale images having v_samp_factor > 1. + +jpeg_write_raw_data() now supports output suspension. + +Correct bugs in "configure" script for case of compiling in +a directory other than the one containing the source files. + +Repair bug in jquant1.c: sometimes didn't use as many colors as it could. + +Borland C makefile and jconfig file work under either MS-DOS or OS/2. + +Miscellaneous improvements to documentation. + + +Version 5a 7-Dec-94 +-------------------- + +Changed color conversion roundoff behavior so that grayscale values are +represented exactly. (This causes test image files to change.) + +Make ordered dither use 16x16 instead of 4x4 pattern for a small quality +improvement. + +New configure script based on latest GNU Autoconf. +Fix configure script to handle CFLAGS correctly. +Rename *.auto files to *.cfg, so that configure script still works if +file names have been truncated for DOS. + +Fix bug in rdbmp.c: didn't allow for extra data between header and image. + +Modify rdppm.c/wrppm.c to handle 2-byte raw PPM/PGM formats for 12-bit data. + +Fix several bugs in rdrle.c. + +NEED_SHORT_EXTERNAL_NAMES option was broken. + +Revise jerror.h/jerror.c for more flexibility in message table. + +Repair oversight in jmemname.c NO_MKTEMP case: file could be there +but unreadable. + + +Version 5 24-Sep-94 +-------------------- + +Version 5 represents a nearly complete redesign and rewrite of the IJG +software. Major user-visible changes include: + * Automatic configuration simplifies installation for most Unix systems. + * A range of speed vs. image quality tradeoffs are supported. + This includes resizing of an image during decompression: scaling down + by a factor of 1/2, 1/4, or 1/8 is handled very efficiently. + * New programs rdjpgcom and wrjpgcom allow insertion and extraction + of text comments in a JPEG file. + +The application programmer's interface to the library has changed completely. +Notable improvements include: + * We have eliminated the use of callback routines for handling the + uncompressed image data. The application now sees the library as a + set of routines that it calls to read or write image data on a + scanline-by-scanline basis. + * The application image data is represented in a conventional interleaved- + pixel format, rather than as a separate array for each color channel. + This can save a copying step in many programs. + * The handling of compressed data has been cleaned up: the application can + supply routines to source or sink the compressed data. It is possible to + suspend processing on source/sink buffer overrun, although this is not + supported in all operating modes. + * All static state has been eliminated from the library, so that multiple + instances of compression or decompression can be active concurrently. + * JPEG abbreviated datastream formats are supported, ie, quantization and + Huffman tables can be stored separately from the image data. + * And not only that, but the documentation of the library has improved + considerably! + + +The last widely used release before the version 5 rewrite was version 4A of +18-Feb-93. Change logs before that point have been discarded, since they +are not of much interest after the rewrite. diff --git a/cmake/externals/libjpeg/jaricom.c b/cmake/externals/libjpeg/jaricom.c new file mode 100644 index 000000000..690068861 --- /dev/null +++ b/cmake/externals/libjpeg/jaricom.c @@ -0,0 +1,153 @@ +/* + * jaricom.c + * + * Developed 1997-2011 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains probability estimation tables for common use in + * arithmetic entropy encoding and decoding routines. + * + * This data represents Table D.3 in the JPEG spec (D.2 in the draft), + * ISO/IEC IS 10918-1 and CCITT Recommendation ITU-T T.81, and Table 24 + * in the JBIG spec, ISO/IEC IS 11544 and CCITT Recommendation ITU-T T.82. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + +/* The following #define specifies the packing of the four components + * into the compact INT32 representation. + * Note that this formula must match the actual arithmetic encoder + * and decoder implementation. The implementation has to be changed + * if this formula is changed. + * The current organization is leaned on Markus Kuhn's JBIG + * implementation (jbig_tab.c). + */ + +#define V(i,a,b,c,d) (((INT32)a << 16) | ((INT32)c << 8) | ((INT32)d << 7) | b) + +const INT32 jpeg_aritab[113+1] = { +/* + * Index, Qe_Value, Next_Index_LPS, Next_Index_MPS, Switch_MPS + */ + V( 0, 0x5a1d, 1, 1, 1 ), + V( 1, 0x2586, 14, 2, 0 ), + V( 2, 0x1114, 16, 3, 0 ), + V( 3, 0x080b, 18, 4, 0 ), + V( 4, 0x03d8, 20, 5, 0 ), + V( 5, 0x01da, 23, 6, 0 ), + V( 6, 0x00e5, 25, 7, 0 ), + V( 7, 0x006f, 28, 8, 0 ), + V( 8, 0x0036, 30, 9, 0 ), + V( 9, 0x001a, 33, 10, 0 ), + V( 10, 0x000d, 35, 11, 0 ), + V( 11, 0x0006, 9, 12, 0 ), + V( 12, 0x0003, 10, 13, 0 ), + V( 13, 0x0001, 12, 13, 0 ), + V( 14, 0x5a7f, 15, 15, 1 ), + V( 15, 0x3f25, 36, 16, 0 ), + V( 16, 0x2cf2, 38, 17, 0 ), + V( 17, 0x207c, 39, 18, 0 ), + V( 18, 0x17b9, 40, 19, 0 ), + V( 19, 0x1182, 42, 20, 0 ), + V( 20, 0x0cef, 43, 21, 0 ), + V( 21, 0x09a1, 45, 22, 0 ), + V( 22, 0x072f, 46, 23, 0 ), + V( 23, 0x055c, 48, 24, 0 ), + V( 24, 0x0406, 49, 25, 0 ), + V( 25, 0x0303, 51, 26, 0 ), + V( 26, 0x0240, 52, 27, 0 ), + V( 27, 0x01b1, 54, 28, 0 ), + V( 28, 0x0144, 56, 29, 0 ), + V( 29, 0x00f5, 57, 30, 0 ), + V( 30, 0x00b7, 59, 31, 0 ), + V( 31, 0x008a, 60, 32, 0 ), + V( 32, 0x0068, 62, 33, 0 ), + V( 33, 0x004e, 63, 34, 0 ), + V( 34, 0x003b, 32, 35, 0 ), + V( 35, 0x002c, 33, 9, 0 ), + V( 36, 0x5ae1, 37, 37, 1 ), + V( 37, 0x484c, 64, 38, 0 ), + V( 38, 0x3a0d, 65, 39, 0 ), + V( 39, 0x2ef1, 67, 40, 0 ), + V( 40, 0x261f, 68, 41, 0 ), + V( 41, 0x1f33, 69, 42, 0 ), + V( 42, 0x19a8, 70, 43, 0 ), + V( 43, 0x1518, 72, 44, 0 ), + V( 44, 0x1177, 73, 45, 0 ), + V( 45, 0x0e74, 74, 46, 0 ), + V( 46, 0x0bfb, 75, 47, 0 ), + V( 47, 0x09f8, 77, 48, 0 ), + V( 48, 0x0861, 78, 49, 0 ), + V( 49, 0x0706, 79, 50, 0 ), + V( 50, 0x05cd, 48, 51, 0 ), + V( 51, 0x04de, 50, 52, 0 ), + V( 52, 0x040f, 50, 53, 0 ), + V( 53, 0x0363, 51, 54, 0 ), + V( 54, 0x02d4, 52, 55, 0 ), + V( 55, 0x025c, 53, 56, 0 ), + V( 56, 0x01f8, 54, 57, 0 ), + V( 57, 0x01a4, 55, 58, 0 ), + V( 58, 0x0160, 56, 59, 0 ), + V( 59, 0x0125, 57, 60, 0 ), + V( 60, 0x00f6, 58, 61, 0 ), + V( 61, 0x00cb, 59, 62, 0 ), + V( 62, 0x00ab, 61, 63, 0 ), + V( 63, 0x008f, 61, 32, 0 ), + V( 64, 0x5b12, 65, 65, 1 ), + V( 65, 0x4d04, 80, 66, 0 ), + V( 66, 0x412c, 81, 67, 0 ), + V( 67, 0x37d8, 82, 68, 0 ), + V( 68, 0x2fe8, 83, 69, 0 ), + V( 69, 0x293c, 84, 70, 0 ), + V( 70, 0x2379, 86, 71, 0 ), + V( 71, 0x1edf, 87, 72, 0 ), + V( 72, 0x1aa9, 87, 73, 0 ), + V( 73, 0x174e, 72, 74, 0 ), + V( 74, 0x1424, 72, 75, 0 ), + V( 75, 0x119c, 74, 76, 0 ), + V( 76, 0x0f6b, 74, 77, 0 ), + V( 77, 0x0d51, 75, 78, 0 ), + V( 78, 0x0bb6, 77, 79, 0 ), + V( 79, 0x0a40, 77, 48, 0 ), + V( 80, 0x5832, 80, 81, 1 ), + V( 81, 0x4d1c, 88, 82, 0 ), + V( 82, 0x438e, 89, 83, 0 ), + V( 83, 0x3bdd, 90, 84, 0 ), + V( 84, 0x34ee, 91, 85, 0 ), + V( 85, 0x2eae, 92, 86, 0 ), + V( 86, 0x299a, 93, 87, 0 ), + V( 87, 0x2516, 86, 71, 0 ), + V( 88, 0x5570, 88, 89, 1 ), + V( 89, 0x4ca9, 95, 90, 0 ), + V( 90, 0x44d9, 96, 91, 0 ), + V( 91, 0x3e22, 97, 92, 0 ), + V( 92, 0x3824, 99, 93, 0 ), + V( 93, 0x32b4, 99, 94, 0 ), + V( 94, 0x2e17, 93, 86, 0 ), + V( 95, 0x56a8, 95, 96, 1 ), + V( 96, 0x4f46, 101, 97, 0 ), + V( 97, 0x47e5, 102, 98, 0 ), + V( 98, 0x41cf, 103, 99, 0 ), + V( 99, 0x3c3d, 104, 100, 0 ), + V( 100, 0x375e, 99, 93, 0 ), + V( 101, 0x5231, 105, 102, 0 ), + V( 102, 0x4c0f, 106, 103, 0 ), + V( 103, 0x4639, 107, 104, 0 ), + V( 104, 0x415e, 103, 99, 0 ), + V( 105, 0x5627, 105, 106, 1 ), + V( 106, 0x50e7, 108, 107, 0 ), + V( 107, 0x4b85, 109, 103, 0 ), + V( 108, 0x5597, 110, 109, 0 ), + V( 109, 0x504f, 111, 107, 0 ), + V( 110, 0x5a10, 110, 111, 1 ), + V( 111, 0x5522, 112, 109, 0 ), + V( 112, 0x59eb, 112, 111, 1 ), +/* + * This last entry is used for fixed probability estimate of 0.5 + * as suggested in Section 10.3 Table 5 of ITU-T Rec. T.851. + */ + V( 113, 0x5a1d, 113, 113, 0 ) +}; diff --git a/cmake/externals/libjpeg/jcapimin.c b/cmake/externals/libjpeg/jcapimin.c new file mode 100644 index 000000000..639ce86f4 --- /dev/null +++ b/cmake/externals/libjpeg/jcapimin.c @@ -0,0 +1,288 @@ +/* + * jcapimin.c + * + * Copyright (C) 1994-1998, Thomas G. Lane. + * Modified 2003-2010 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains application interface code for the compression half + * of the JPEG library. These are the "minimum" API routines that may be + * needed in either the normal full-compression case or the transcoding-only + * case. + * + * Most of the routines intended to be called directly by an application + * are in this file or in jcapistd.c. But also see jcparam.c for + * parameter-setup helper routines, jcomapi.c for routines shared by + * compression and decompression, and jctrans.c for the transcoding case. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* + * Initialization of a JPEG compression object. + * The error manager must already be set up (in case memory manager fails). + */ + +GLOBAL(void) +jpeg_CreateCompress (j_compress_ptr cinfo, int version, size_t structsize) +{ + int i; + + /* Guard against version mismatches between library and caller. */ + cinfo->mem = NULL; /* so jpeg_destroy knows mem mgr not called */ + if (version != JPEG_LIB_VERSION) + ERREXIT2(cinfo, JERR_BAD_LIB_VERSION, JPEG_LIB_VERSION, version); + if (structsize != SIZEOF(struct jpeg_compress_struct)) + ERREXIT2(cinfo, JERR_BAD_STRUCT_SIZE, + (int) SIZEOF(struct jpeg_compress_struct), (int) structsize); + + /* For debugging purposes, we zero the whole master structure. + * But the application has already set the err pointer, and may have set + * client_data, so we have to save and restore those fields. + * Note: if application hasn't set client_data, tools like Purify may + * complain here. + */ + { + struct jpeg_error_mgr * err = cinfo->err; + void * client_data = cinfo->client_data; /* ignore Purify complaint here */ + MEMZERO(cinfo, SIZEOF(struct jpeg_compress_struct)); + cinfo->err = err; + cinfo->client_data = client_data; + } + cinfo->is_decompressor = FALSE; + + /* Initialize a memory manager instance for this object */ + jinit_memory_mgr((j_common_ptr) cinfo); + + /* Zero out pointers to permanent structures. */ + cinfo->progress = NULL; + cinfo->dest = NULL; + + cinfo->comp_info = NULL; + + for (i = 0; i < NUM_QUANT_TBLS; i++) { + cinfo->quant_tbl_ptrs[i] = NULL; + cinfo->q_scale_factor[i] = 100; + } + + for (i = 0; i < NUM_HUFF_TBLS; i++) { + cinfo->dc_huff_tbl_ptrs[i] = NULL; + cinfo->ac_huff_tbl_ptrs[i] = NULL; + } + + /* Must do it here for emit_dqt in case jpeg_write_tables is used */ + cinfo->block_size = DCTSIZE; + cinfo->natural_order = jpeg_natural_order; + cinfo->lim_Se = DCTSIZE2-1; + + cinfo->script_space = NULL; + + cinfo->input_gamma = 1.0; /* in case application forgets */ + + /* OK, I'm ready */ + cinfo->global_state = CSTATE_START; +} + + +/* + * Destruction of a JPEG compression object + */ + +GLOBAL(void) +jpeg_destroy_compress (j_compress_ptr cinfo) +{ + jpeg_destroy((j_common_ptr) cinfo); /* use common routine */ +} + + +/* + * Abort processing of a JPEG compression operation, + * but don't destroy the object itself. + */ + +GLOBAL(void) +jpeg_abort_compress (j_compress_ptr cinfo) +{ + jpeg_abort((j_common_ptr) cinfo); /* use common routine */ +} + + +/* + * Forcibly suppress or un-suppress all quantization and Huffman tables. + * Marks all currently defined tables as already written (if suppress) + * or not written (if !suppress). This will control whether they get emitted + * by a subsequent jpeg_start_compress call. + * + * This routine is exported for use by applications that want to produce + * abbreviated JPEG datastreams. It logically belongs in jcparam.c, but + * since it is called by jpeg_start_compress, we put it here --- otherwise + * jcparam.o would be linked whether the application used it or not. + */ + +GLOBAL(void) +jpeg_suppress_tables (j_compress_ptr cinfo, boolean suppress) +{ + int i; + JQUANT_TBL * qtbl; + JHUFF_TBL * htbl; + + for (i = 0; i < NUM_QUANT_TBLS; i++) { + if ((qtbl = cinfo->quant_tbl_ptrs[i]) != NULL) + qtbl->sent_table = suppress; + } + + for (i = 0; i < NUM_HUFF_TBLS; i++) { + if ((htbl = cinfo->dc_huff_tbl_ptrs[i]) != NULL) + htbl->sent_table = suppress; + if ((htbl = cinfo->ac_huff_tbl_ptrs[i]) != NULL) + htbl->sent_table = suppress; + } +} + + +/* + * Finish JPEG compression. + * + * If a multipass operating mode was selected, this may do a great deal of + * work including most of the actual output. + */ + +GLOBAL(void) +jpeg_finish_compress (j_compress_ptr cinfo) +{ + JDIMENSION iMCU_row; + + if (cinfo->global_state == CSTATE_SCANNING || + cinfo->global_state == CSTATE_RAW_OK) { + /* Terminate first pass */ + if (cinfo->next_scanline < cinfo->image_height) + ERREXIT(cinfo, JERR_TOO_LITTLE_DATA); + (*cinfo->master->finish_pass) (cinfo); + } else if (cinfo->global_state != CSTATE_WRCOEFS) + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + /* Perform any remaining passes */ + while (! cinfo->master->is_last_pass) { + (*cinfo->master->prepare_for_pass) (cinfo); + for (iMCU_row = 0; iMCU_row < cinfo->total_iMCU_rows; iMCU_row++) { + if (cinfo->progress != NULL) { + cinfo->progress->pass_counter = (long) iMCU_row; + cinfo->progress->pass_limit = (long) cinfo->total_iMCU_rows; + (*cinfo->progress->progress_monitor) ((j_common_ptr) cinfo); + } + /* We bypass the main controller and invoke coef controller directly; + * all work is being done from the coefficient buffer. + */ + if (! (*cinfo->coef->compress_data) (cinfo, (JSAMPIMAGE) NULL)) + ERREXIT(cinfo, JERR_CANT_SUSPEND); + } + (*cinfo->master->finish_pass) (cinfo); + } + /* Write EOI, do final cleanup */ + (*cinfo->marker->write_file_trailer) (cinfo); + (*cinfo->dest->term_destination) (cinfo); + /* We can use jpeg_abort to release memory and reset global_state */ + jpeg_abort((j_common_ptr) cinfo); +} + + +/* + * Write a special marker. + * This is only recommended for writing COM or APPn markers. + * Must be called after jpeg_start_compress() and before + * first call to jpeg_write_scanlines() or jpeg_write_raw_data(). + */ + +GLOBAL(void) +jpeg_write_marker (j_compress_ptr cinfo, int marker, + const JOCTET *dataptr, unsigned int datalen) +{ + JMETHOD(void, write_marker_byte, (j_compress_ptr info, int val)); + + if (cinfo->next_scanline != 0 || + (cinfo->global_state != CSTATE_SCANNING && + cinfo->global_state != CSTATE_RAW_OK && + cinfo->global_state != CSTATE_WRCOEFS)) + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + + (*cinfo->marker->write_marker_header) (cinfo, marker, datalen); + write_marker_byte = cinfo->marker->write_marker_byte; /* copy for speed */ + while (datalen--) { + (*write_marker_byte) (cinfo, *dataptr); + dataptr++; + } +} + +/* Same, but piecemeal. */ + +GLOBAL(void) +jpeg_write_m_header (j_compress_ptr cinfo, int marker, unsigned int datalen) +{ + if (cinfo->next_scanline != 0 || + (cinfo->global_state != CSTATE_SCANNING && + cinfo->global_state != CSTATE_RAW_OK && + cinfo->global_state != CSTATE_WRCOEFS)) + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + + (*cinfo->marker->write_marker_header) (cinfo, marker, datalen); +} + +GLOBAL(void) +jpeg_write_m_byte (j_compress_ptr cinfo, int val) +{ + (*cinfo->marker->write_marker_byte) (cinfo, val); +} + + +/* + * Alternate compression function: just write an abbreviated table file. + * Before calling this, all parameters and a data destination must be set up. + * + * To produce a pair of files containing abbreviated tables and abbreviated + * image data, one would proceed as follows: + * + * initialize JPEG object + * set JPEG parameters + * set destination to table file + * jpeg_write_tables(cinfo); + * set destination to image file + * jpeg_start_compress(cinfo, FALSE); + * write data... + * jpeg_finish_compress(cinfo); + * + * jpeg_write_tables has the side effect of marking all tables written + * (same as jpeg_suppress_tables(..., TRUE)). Thus a subsequent start_compress + * will not re-emit the tables unless it is passed write_all_tables=TRUE. + */ + +GLOBAL(void) +jpeg_write_tables (j_compress_ptr cinfo) +{ + if (cinfo->global_state != CSTATE_START) + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + + /* (Re)initialize error mgr and destination modules */ + (*cinfo->err->reset_error_mgr) ((j_common_ptr) cinfo); + (*cinfo->dest->init_destination) (cinfo); + /* Initialize the marker writer ... bit of a crock to do it here. */ + jinit_marker_writer(cinfo); + /* Write them tables! */ + (*cinfo->marker->write_tables_only) (cinfo); + /* And clean up. */ + (*cinfo->dest->term_destination) (cinfo); + /* + * In library releases up through v6a, we called jpeg_abort() here to free + * any working memory allocated by the destination manager and marker + * writer. Some applications had a problem with that: they allocated space + * of their own from the library memory manager, and didn't want it to go + * away during write_tables. So now we do nothing. This will cause a + * memory leak if an app calls write_tables repeatedly without doing a full + * compression cycle or otherwise resetting the JPEG object. However, that + * seems less bad than unexpectedly freeing memory in the normal case. + * An app that prefers the old behavior can call jpeg_abort for itself after + * each call to jpeg_write_tables(). + */ +} diff --git a/cmake/externals/libjpeg/jcapistd.c b/cmake/externals/libjpeg/jcapistd.c new file mode 100644 index 000000000..0917afa97 --- /dev/null +++ b/cmake/externals/libjpeg/jcapistd.c @@ -0,0 +1,162 @@ +/* + * jcapistd.c + * + * Copyright (C) 1994-1996, Thomas G. Lane. + * Modified 2013 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains application interface code for the compression half + * of the JPEG library. These are the "standard" API routines that are + * used in the normal full-compression case. They are not used by a + * transcoding-only application. Note that if an application links in + * jpeg_start_compress, it will end up linking in the entire compressor. + * We thus must separate this file from jcapimin.c to avoid linking the + * whole compression library into a transcoder. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* + * Compression initialization. + * Before calling this, all parameters and a data destination must be set up. + * + * We require a write_all_tables parameter as a failsafe check when writing + * multiple datastreams from the same compression object. Since prior runs + * will have left all the tables marked sent_table=TRUE, a subsequent run + * would emit an abbreviated stream (no tables) by default. This may be what + * is wanted, but for safety's sake it should not be the default behavior: + * programmers should have to make a deliberate choice to emit abbreviated + * images. Therefore the documentation and examples should encourage people + * to pass write_all_tables=TRUE; then it will take active thought to do the + * wrong thing. + */ + +GLOBAL(void) +jpeg_start_compress (j_compress_ptr cinfo, boolean write_all_tables) +{ + if (cinfo->global_state != CSTATE_START) + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + + if (write_all_tables) + jpeg_suppress_tables(cinfo, FALSE); /* mark all tables to be written */ + + /* (Re)initialize error mgr and destination modules */ + (*cinfo->err->reset_error_mgr) ((j_common_ptr) cinfo); + (*cinfo->dest->init_destination) (cinfo); + /* Perform master selection of active modules */ + jinit_compress_master(cinfo); + /* Set up for the first pass */ + (*cinfo->master->prepare_for_pass) (cinfo); + /* Ready for application to drive first pass through jpeg_write_scanlines + * or jpeg_write_raw_data. + */ + cinfo->next_scanline = 0; + cinfo->global_state = (cinfo->raw_data_in ? CSTATE_RAW_OK : CSTATE_SCANNING); +} + + +/* + * Write some scanlines of data to the JPEG compressor. + * + * The return value will be the number of lines actually written. + * This should be less than the supplied num_lines only in case that + * the data destination module has requested suspension of the compressor, + * or if more than image_height scanlines are passed in. + * + * Note: we warn about excess calls to jpeg_write_scanlines() since + * this likely signals an application programmer error. However, + * excess scanlines passed in the last valid call are *silently* ignored, + * so that the application need not adjust num_lines for end-of-image + * when using a multiple-scanline buffer. + */ + +GLOBAL(JDIMENSION) +jpeg_write_scanlines (j_compress_ptr cinfo, JSAMPARRAY scanlines, + JDIMENSION num_lines) +{ + JDIMENSION row_ctr, rows_left; + + if (cinfo->global_state != CSTATE_SCANNING) + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + if (cinfo->next_scanline >= cinfo->image_height) + WARNMS(cinfo, JWRN_TOO_MUCH_DATA); + + /* Call progress monitor hook if present */ + if (cinfo->progress != NULL) { + cinfo->progress->pass_counter = (long) cinfo->next_scanline; + cinfo->progress->pass_limit = (long) cinfo->image_height; + (*cinfo->progress->progress_monitor) ((j_common_ptr) cinfo); + } + + /* Give master control module another chance if this is first call to + * jpeg_write_scanlines. This lets output of the frame/scan headers be + * delayed so that application can write COM, etc, markers between + * jpeg_start_compress and jpeg_write_scanlines. + */ + if (cinfo->master->call_pass_startup) + (*cinfo->master->pass_startup) (cinfo); + + /* Ignore any extra scanlines at bottom of image. */ + rows_left = cinfo->image_height - cinfo->next_scanline; + if (num_lines > rows_left) + num_lines = rows_left; + + row_ctr = 0; + (*cinfo->main->process_data) (cinfo, scanlines, &row_ctr, num_lines); + cinfo->next_scanline += row_ctr; + return row_ctr; +} + + +/* + * Alternate entry point to write raw data. + * Processes exactly one iMCU row per call, unless suspended. + */ + +GLOBAL(JDIMENSION) +jpeg_write_raw_data (j_compress_ptr cinfo, JSAMPIMAGE data, + JDIMENSION num_lines) +{ + JDIMENSION lines_per_iMCU_row; + + if (cinfo->global_state != CSTATE_RAW_OK) + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + if (cinfo->next_scanline >= cinfo->image_height) { + WARNMS(cinfo, JWRN_TOO_MUCH_DATA); + return 0; + } + + /* Call progress monitor hook if present */ + if (cinfo->progress != NULL) { + cinfo->progress->pass_counter = (long) cinfo->next_scanline; + cinfo->progress->pass_limit = (long) cinfo->image_height; + (*cinfo->progress->progress_monitor) ((j_common_ptr) cinfo); + } + + /* Give master control module another chance if this is first call to + * jpeg_write_raw_data. This lets output of the frame/scan headers be + * delayed so that application can write COM, etc, markers between + * jpeg_start_compress and jpeg_write_raw_data. + */ + if (cinfo->master->call_pass_startup) + (*cinfo->master->pass_startup) (cinfo); + + /* Verify that at least one iMCU row has been passed. */ + lines_per_iMCU_row = cinfo->max_v_samp_factor * cinfo->min_DCT_v_scaled_size; + if (num_lines < lines_per_iMCU_row) + ERREXIT(cinfo, JERR_BUFFER_SIZE); + + /* Directly compress the row. */ + if (! (*cinfo->coef->compress_data) (cinfo, data)) { + /* If compressor did not consume the whole row, suspend processing. */ + return 0; + } + + /* OK, we processed one iMCU row. */ + cinfo->next_scanline += lines_per_iMCU_row; + return lines_per_iMCU_row; +} diff --git a/cmake/externals/libjpeg/jcarith.c b/cmake/externals/libjpeg/jcarith.c new file mode 100644 index 000000000..1b45089a3 --- /dev/null +++ b/cmake/externals/libjpeg/jcarith.c @@ -0,0 +1,945 @@ +/* + * jcarith.c + * + * Developed 1997-2020 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains portable arithmetic entropy encoding routines for JPEG + * (implementing the ISO/IEC IS 10918-1 and CCITT Recommendation ITU-T T.81). + * + * Both sequential and progressive modes are supported in this single module. + * + * Suspension is not currently supported in this module. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* Expanded entropy encoder object for arithmetic encoding. */ + +typedef struct { + struct jpeg_entropy_encoder pub; /* public fields */ + + INT32 c; /* C register, base of coding interval, layout as in sec. D.1.3 */ + INT32 a; /* A register, normalized size of coding interval */ + INT32 sc; /* counter for stacked 0xFF values which might overflow */ + INT32 zc; /* counter for pending 0x00 output values which might * + * be discarded at the end ("Pacman" termination) */ + int ct; /* bit shift counter, determines when next byte will be written */ + int buffer; /* buffer for most recent output byte != 0xFF */ + + int last_dc_val[MAX_COMPS_IN_SCAN]; /* last DC coef for each component */ + int dc_context[MAX_COMPS_IN_SCAN]; /* context index for DC conditioning */ + + unsigned int restarts_to_go; /* MCUs left in this restart interval */ + int next_restart_num; /* next restart number to write (0-7) */ + + /* Pointers to statistics areas (these workspaces have image lifespan) */ + unsigned char * dc_stats[NUM_ARITH_TBLS]; + unsigned char * ac_stats[NUM_ARITH_TBLS]; + + /* Statistics bin for coding with fixed probability 0.5 */ + unsigned char fixed_bin[4]; +} arith_entropy_encoder; + +typedef arith_entropy_encoder * arith_entropy_ptr; + +/* The following two definitions specify the allocation chunk size + * for the statistics area. + * According to sections F.1.4.4.1.3 and F.1.4.4.2, we need at least + * 49 statistics bins for DC, and 245 statistics bins for AC coding. + * + * We use a compact representation with 1 byte per statistics bin, + * thus the numbers directly represent byte sizes. + * This 1 byte per statistics bin contains the meaning of the MPS + * (more probable symbol) in the highest bit (mask 0x80), and the + * index into the probability estimation state machine table + * in the lower bits (mask 0x7F). + */ + +#define DC_STAT_BINS 64 +#define AC_STAT_BINS 256 + +/* NOTE: Uncomment the following #define if you want to use the + * given formula for calculating the AC conditioning parameter Kx + * for spectral selection progressive coding in section G.1.3.2 + * of the spec (Kx = Kmin + SRL (8 + Se - Kmin) 4). + * Although the spec and P&M authors claim that this "has proven + * to give good results for 8 bit precision samples", I'm not + * convinced yet that this is really beneficial. + * Early tests gave only very marginal compression enhancements + * (a few - around 5 or so - bytes even for very large files), + * which would turn out rather negative if we'd suppress the + * DAC (Define Arithmetic Conditioning) marker segments for + * the default parameters in the future. + * Note that currently the marker writing module emits 12-byte + * DAC segments for a full-component scan in a color image. + * This is not worth worrying about IMHO. However, since the + * spec defines the default values to be used if the tables + * are omitted (unlike Huffman tables, which are required + * anyway), one might optimize this behaviour in the future, + * and then it would be disadvantageous to use custom tables if + * they don't provide sufficient gain to exceed the DAC size. + * + * On the other hand, I'd consider it as a reasonable result + * that the conditioning has no significant influence on the + * compression performance. This means that the basic + * statistical model is already rather stable. + * + * Thus, at the moment, we use the default conditioning values + * anyway, and do not use the custom formula. + * +#define CALCULATE_SPECTRAL_CONDITIONING + */ + +/* IRIGHT_SHIFT is like RIGHT_SHIFT, but works on int rather than INT32. + * We assume that int right shift is unsigned if INT32 right shift is, + * which should be safe. + */ + +#ifdef RIGHT_SHIFT_IS_UNSIGNED +#define ISHIFT_TEMPS int ishift_temp; +#define IRIGHT_SHIFT(x,shft) \ + ((ishift_temp = (x)) < 0 ? \ + (ishift_temp >> (shft)) | ((~0) << (16-(shft))) : \ + (ishift_temp >> (shft))) +#else +#define ISHIFT_TEMPS +#define IRIGHT_SHIFT(x,shft) ((x) >> (shft)) +#endif + + +LOCAL(void) +emit_byte (int val, j_compress_ptr cinfo) +/* Write next output byte; we do not support suspension in this module. */ +{ + struct jpeg_destination_mgr * dest = cinfo->dest; + + *dest->next_output_byte++ = (JOCTET) val; + if (--dest->free_in_buffer == 0) + if (! (*dest->empty_output_buffer) (cinfo)) + ERREXIT(cinfo, JERR_CANT_SUSPEND); +} + + +/* + * Finish up at the end of an arithmetic-compressed scan. + */ + +METHODDEF(void) +finish_pass (j_compress_ptr cinfo) +{ + arith_entropy_ptr e = (arith_entropy_ptr) cinfo->entropy; + INT32 temp; + + /* Section D.1.8: Termination of encoding */ + + /* Find the e->c in the coding interval with the largest + * number of trailing zero bits */ + if ((temp = (e->a - 1 + e->c) & 0xFFFF0000L) < e->c) + e->c = temp + 0x8000L; + else + e->c = temp; + /* Send remaining bytes to output */ + e->c <<= e->ct; + if (e->c & 0xF8000000L) { + /* One final overflow has to be handled */ + if (e->buffer >= 0) { + if (e->zc) + do emit_byte(0x00, cinfo); + while (--e->zc); + emit_byte(e->buffer + 1, cinfo); + if (e->buffer + 1 == 0xFF) + emit_byte(0x00, cinfo); + } + e->zc += e->sc; /* carry-over converts stacked 0xFF bytes to 0x00 */ + e->sc = 0; + } else { + if (e->buffer == 0) + ++e->zc; + else if (e->buffer >= 0) { + if (e->zc) + do emit_byte(0x00, cinfo); + while (--e->zc); + emit_byte(e->buffer, cinfo); + } + if (e->sc) { + if (e->zc) + do emit_byte(0x00, cinfo); + while (--e->zc); + do { + emit_byte(0xFF, cinfo); + emit_byte(0x00, cinfo); + } while (--e->sc); + } + } + /* Output final bytes only if they are not 0x00 */ + if (e->c & 0x7FFF800L) { + if (e->zc) /* output final pending zero bytes */ + do emit_byte(0x00, cinfo); + while (--e->zc); + emit_byte((int) ((e->c >> 19) & 0xFF), cinfo); + if (((e->c >> 19) & 0xFF) == 0xFF) + emit_byte(0x00, cinfo); + if (e->c & 0x7F800L) { + emit_byte((int) ((e->c >> 11) & 0xFF), cinfo); + if (((e->c >> 11) & 0xFF) == 0xFF) + emit_byte(0x00, cinfo); + } + } +} + + +/* + * The core arithmetic encoding routine (common in JPEG and JBIG). + * This needs to go as fast as possible. + * Machine-dependent optimization facilities + * are not utilized in this portable implementation. + * However, this code should be fairly efficient and + * may be a good base for further optimizations anyway. + * + * Parameter 'val' to be encoded may be 0 or 1 (binary decision). + * + * Note: I've added full "Pacman" termination support to the + * byte output routines, which is equivalent to the optional + * Discard_final_zeros procedure (Figure D.15) in the spec. + * Thus, we always produce the shortest possible output + * stream compliant to the spec (no trailing zero bytes, + * except for FF stuffing). + * + * I've also introduced a new scheme for accessing + * the probability estimation state machine table, + * derived from Markus Kuhn's JBIG implementation. + */ + +LOCAL(void) +arith_encode (j_compress_ptr cinfo, unsigned char *st, int val) +{ + register arith_entropy_ptr e = (arith_entropy_ptr) cinfo->entropy; + register unsigned char nl, nm; + register INT32 qe, temp; + register int sv; + + /* Fetch values from our compact representation of Table D.3(D.2): + * Qe values and probability estimation state machine + */ + sv = *st; + qe = jpeg_aritab[sv & 0x7F]; /* => Qe_Value */ + nl = qe & 0xFF; qe >>= 8; /* Next_Index_LPS + Switch_MPS */ + nm = qe & 0xFF; qe >>= 8; /* Next_Index_MPS */ + + /* Encode & estimation procedures per sections D.1.4 & D.1.5 */ + e->a -= qe; + if (val != (sv >> 7)) { + /* Encode the less probable symbol */ + if (e->a >= qe) { + /* If the interval size (qe) for the less probable symbol (LPS) + * is larger than the interval size for the MPS, then exchange + * the two symbols for coding efficiency, otherwise code the LPS + * as usual: */ + e->c += e->a; + e->a = qe; + } + *st = (sv & 0x80) ^ nl; /* Estimate_after_LPS */ + } else { + /* Encode the more probable symbol */ + if (e->a >= 0x8000L) + return; /* A >= 0x8000 -> ready, no renormalization required */ + if (e->a < qe) { + /* If the interval size (qe) for the less probable symbol (LPS) + * is larger than the interval size for the MPS, then exchange + * the two symbols for coding efficiency: */ + e->c += e->a; + e->a = qe; + } + *st = (sv & 0x80) ^ nm; /* Estimate_after_MPS */ + } + + /* Renormalization & data output per section D.1.6 */ + do { + e->a <<= 1; + e->c <<= 1; + if (--e->ct == 0) { + /* Another byte is ready for output */ + temp = e->c >> 19; + if (temp > 0xFF) { + /* Handle overflow over all stacked 0xFF bytes */ + if (e->buffer >= 0) { + if (e->zc) + do emit_byte(0x00, cinfo); + while (--e->zc); + emit_byte(e->buffer + 1, cinfo); + if (e->buffer + 1 == 0xFF) + emit_byte(0x00, cinfo); + } + e->zc += e->sc; /* carry-over converts stacked 0xFF bytes to 0x00 */ + e->sc = 0; + /* Note: The 3 spacer bits in the C register guarantee + * that the new buffer byte can't be 0xFF here + * (see page 160 in the P&M JPEG book). */ + /* New output byte, might overflow later */ + e->buffer = (int) (temp & 0xFF); + } else if (temp == 0xFF) { + ++e->sc; /* stack 0xFF byte (which might overflow later) */ + } else { + /* Output all stacked 0xFF bytes, they will not overflow any more */ + if (e->buffer == 0) + ++e->zc; + else if (e->buffer >= 0) { + if (e->zc) + do emit_byte(0x00, cinfo); + while (--e->zc); + emit_byte(e->buffer, cinfo); + } + if (e->sc) { + if (e->zc) + do emit_byte(0x00, cinfo); + while (--e->zc); + do { + emit_byte(0xFF, cinfo); + emit_byte(0x00, cinfo); + } while (--e->sc); + } + /* New output byte (can still overflow) */ + e->buffer = (int) (temp & 0xFF); + } + e->c &= 0x7FFFFL; + e->ct += 8; + } + } while (e->a < 0x8000L); +} + + +/* + * Emit a restart marker & resynchronize predictions. + */ + +LOCAL(void) +emit_restart (j_compress_ptr cinfo, int restart_num) +{ + arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; + int ci; + jpeg_component_info * compptr; + + finish_pass(cinfo); + + emit_byte(0xFF, cinfo); + emit_byte(JPEG_RST0 + restart_num, cinfo); + + /* Re-initialize statistics areas */ + for (ci = 0; ci < cinfo->comps_in_scan; ci++) { + compptr = cinfo->cur_comp_info[ci]; + /* DC needs no table for refinement scan */ + if (cinfo->Ss == 0 && cinfo->Ah == 0) { + MEMZERO(entropy->dc_stats[compptr->dc_tbl_no], DC_STAT_BINS); + /* Reset DC predictions to 0 */ + entropy->last_dc_val[ci] = 0; + entropy->dc_context[ci] = 0; + } + /* AC needs no table when not present */ + if (cinfo->Se) { + MEMZERO(entropy->ac_stats[compptr->ac_tbl_no], AC_STAT_BINS); + } + } + + /* Reset arithmetic encoding variables */ + entropy->c = 0; + entropy->a = 0x10000L; + entropy->sc = 0; + entropy->zc = 0; + entropy->ct = 11; + entropy->buffer = -1; /* empty */ +} + + +/* + * MCU encoding for DC initial scan (either spectral selection, + * or first pass of successive approximation). + */ + +METHODDEF(boolean) +encode_mcu_DC_first (j_compress_ptr cinfo, JBLOCKARRAY MCU_data) +{ + arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; + unsigned char *st; + int blkn, ci, tbl; + int v, v2, m; + ISHIFT_TEMPS + + /* Emit restart marker if needed */ + if (cinfo->restart_interval) { + if (entropy->restarts_to_go == 0) { + emit_restart(cinfo, entropy->next_restart_num); + entropy->restarts_to_go = cinfo->restart_interval; + entropy->next_restart_num++; + entropy->next_restart_num &= 7; + } + entropy->restarts_to_go--; + } + + /* Encode the MCU data blocks */ + for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { + ci = cinfo->MCU_membership[blkn]; + tbl = cinfo->cur_comp_info[ci]->dc_tbl_no; + + /* Compute the DC value after the required point transform by Al. + * This is simply an arithmetic right shift. + */ + m = IRIGHT_SHIFT((int) (MCU_data[blkn][0][0]), cinfo->Al); + + /* Sections F.1.4.1 & F.1.4.4.1: Encoding of DC coefficients */ + + /* Table F.4: Point to statistics bin S0 for DC coefficient coding */ + st = entropy->dc_stats[tbl] + entropy->dc_context[ci]; + + /* Figure F.4: Encode_DC_DIFF */ + if ((v = m - entropy->last_dc_val[ci]) == 0) { + arith_encode(cinfo, st, 0); + entropy->dc_context[ci] = 0; /* zero diff category */ + } else { + entropy->last_dc_val[ci] = m; + arith_encode(cinfo, st, 1); + /* Figure F.6: Encoding nonzero value v */ + /* Figure F.7: Encoding the sign of v */ + if (v > 0) { + arith_encode(cinfo, st + 1, 0); /* Table F.4: SS = S0 + 1 */ + st += 2; /* Table F.4: SP = S0 + 2 */ + entropy->dc_context[ci] = 4; /* small positive diff category */ + } else { + v = -v; + arith_encode(cinfo, st + 1, 1); /* Table F.4: SS = S0 + 1 */ + st += 3; /* Table F.4: SN = S0 + 3 */ + entropy->dc_context[ci] = 8; /* small negative diff category */ + } + /* Figure F.8: Encoding the magnitude category of v */ + m = 0; + if (v -= 1) { + arith_encode(cinfo, st, 1); + m = 1; + v2 = v; + st = entropy->dc_stats[tbl] + 20; /* Table F.4: X1 = 20 */ + while (v2 >>= 1) { + arith_encode(cinfo, st, 1); + m <<= 1; + st += 1; + } + } + arith_encode(cinfo, st, 0); + /* Section F.1.4.4.1.2: Establish dc_context conditioning category */ + if (m < (int) ((1L << cinfo->arith_dc_L[tbl]) >> 1)) + entropy->dc_context[ci] = 0; /* zero diff category */ + else if (m > (int) ((1L << cinfo->arith_dc_U[tbl]) >> 1)) + entropy->dc_context[ci] += 8; /* large diff category */ + /* Figure F.9: Encoding the magnitude bit pattern of v */ + st += 14; + while (m >>= 1) + arith_encode(cinfo, st, (m & v) ? 1 : 0); + } + } + + return TRUE; +} + + +/* + * MCU encoding for AC initial scan (either spectral selection, + * or first pass of successive approximation). + */ + +METHODDEF(boolean) +encode_mcu_AC_first (j_compress_ptr cinfo, JBLOCKARRAY MCU_data) +{ + arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; + const int * natural_order; + JBLOCKROW block; + unsigned char *st; + int tbl, k, ke; + int v, v2, m; + + /* Emit restart marker if needed */ + if (cinfo->restart_interval) { + if (entropy->restarts_to_go == 0) { + emit_restart(cinfo, entropy->next_restart_num); + entropy->restarts_to_go = cinfo->restart_interval; + entropy->next_restart_num++; + entropy->next_restart_num &= 7; + } + entropy->restarts_to_go--; + } + + natural_order = cinfo->natural_order; + + /* Encode the MCU data block */ + block = MCU_data[0]; + tbl = cinfo->cur_comp_info[0]->ac_tbl_no; + + /* Sections F.1.4.2 & F.1.4.4.2: Encoding of AC coefficients */ + + /* Establish EOB (end-of-block) index */ + ke = cinfo->Se; + do { + /* We must apply the point transform by Al. For AC coefficients this + * is an integer division with rounding towards 0. To do this portably + * in C, we shift after obtaining the absolute value. + */ + if ((v = (*block)[natural_order[ke]]) >= 0) { + if (v >>= cinfo->Al) break; + } else { + v = -v; + if (v >>= cinfo->Al) break; + } + } while (--ke); + + /* Figure F.5: Encode_AC_Coefficients */ + for (k = cinfo->Ss - 1; k < ke;) { + st = entropy->ac_stats[tbl] + 3 * k; + arith_encode(cinfo, st, 0); /* EOB decision */ + for (;;) { + if ((v = (*block)[natural_order[++k]]) >= 0) { + if (v >>= cinfo->Al) { + arith_encode(cinfo, st + 1, 1); + arith_encode(cinfo, entropy->fixed_bin, 0); + break; + } + } else { + v = -v; + if (v >>= cinfo->Al) { + arith_encode(cinfo, st + 1, 1); + arith_encode(cinfo, entropy->fixed_bin, 1); + break; + } + } + arith_encode(cinfo, st + 1, 0); + st += 3; + } + st += 2; + /* Figure F.8: Encoding the magnitude category of v */ + m = 0; + if (v -= 1) { + arith_encode(cinfo, st, 1); + m = 1; + v2 = v; + if (v2 >>= 1) { + arith_encode(cinfo, st, 1); + m <<= 1; + st = entropy->ac_stats[tbl] + + (k <= cinfo->arith_ac_K[tbl] ? 189 : 217); + while (v2 >>= 1) { + arith_encode(cinfo, st, 1); + m <<= 1; + st += 1; + } + } + } + arith_encode(cinfo, st, 0); + /* Figure F.9: Encoding the magnitude bit pattern of v */ + st += 14; + while (m >>= 1) + arith_encode(cinfo, st, (m & v) ? 1 : 0); + } + /* Encode EOB decision only if k < cinfo->Se */ + if (k < cinfo->Se) { + st = entropy->ac_stats[tbl] + 3 * k; + arith_encode(cinfo, st, 1); + } + + return TRUE; +} + + +/* + * MCU encoding for DC successive approximation refinement scan. + * Note: we assume such scans can be multi-component, + * although the spec is not very clear on the point. + */ + +METHODDEF(boolean) +encode_mcu_DC_refine (j_compress_ptr cinfo, JBLOCKARRAY MCU_data) +{ + arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; + unsigned char *st; + int Al, blkn; + + /* Emit restart marker if needed */ + if (cinfo->restart_interval) { + if (entropy->restarts_to_go == 0) { + emit_restart(cinfo, entropy->next_restart_num); + entropy->restarts_to_go = cinfo->restart_interval; + entropy->next_restart_num++; + entropy->next_restart_num &= 7; + } + entropy->restarts_to_go--; + } + + st = entropy->fixed_bin; /* use fixed probability estimation */ + Al = cinfo->Al; + + /* Encode the MCU data blocks */ + for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { + /* We simply emit the Al'th bit of the DC coefficient value. */ + arith_encode(cinfo, st, (MCU_data[blkn][0][0] >> Al) & 1); + } + + return TRUE; +} + + +/* + * MCU encoding for AC successive approximation refinement scan. + */ + +METHODDEF(boolean) +encode_mcu_AC_refine (j_compress_ptr cinfo, JBLOCKARRAY MCU_data) +{ + arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; + const int * natural_order; + JBLOCKROW block; + unsigned char *st; + int tbl, k, ke, kex; + int v; + + /* Emit restart marker if needed */ + if (cinfo->restart_interval) { + if (entropy->restarts_to_go == 0) { + emit_restart(cinfo, entropy->next_restart_num); + entropy->restarts_to_go = cinfo->restart_interval; + entropy->next_restart_num++; + entropy->next_restart_num &= 7; + } + entropy->restarts_to_go--; + } + + natural_order = cinfo->natural_order; + + /* Encode the MCU data block */ + block = MCU_data[0]; + tbl = cinfo->cur_comp_info[0]->ac_tbl_no; + + /* Section G.1.3.3: Encoding of AC coefficients */ + + /* Establish EOB (end-of-block) index */ + ke = cinfo->Se; + do { + /* We must apply the point transform by Al. For AC coefficients this + * is an integer division with rounding towards 0. To do this portably + * in C, we shift after obtaining the absolute value. + */ + if ((v = (*block)[natural_order[ke]]) >= 0) { + if (v >>= cinfo->Al) break; + } else { + v = -v; + if (v >>= cinfo->Al) break; + } + } while (--ke); + + /* Establish EOBx (previous stage end-of-block) index */ + for (kex = ke; kex > 0; kex--) + if ((v = (*block)[natural_order[kex]]) >= 0) { + if (v >>= cinfo->Ah) break; + } else { + v = -v; + if (v >>= cinfo->Ah) break; + } + + /* Figure G.10: Encode_AC_Coefficients_SA */ + for (k = cinfo->Ss - 1; k < ke;) { + st = entropy->ac_stats[tbl] + 3 * k; + if (k >= kex) + arith_encode(cinfo, st, 0); /* EOB decision */ + for (;;) { + if ((v = (*block)[natural_order[++k]]) >= 0) { + if (v >>= cinfo->Al) { + if (v >> 1) /* previously nonzero coef */ + arith_encode(cinfo, st + 2, (v & 1)); + else { /* newly nonzero coef */ + arith_encode(cinfo, st + 1, 1); + arith_encode(cinfo, entropy->fixed_bin, 0); + } + break; + } + } else { + v = -v; + if (v >>= cinfo->Al) { + if (v >> 1) /* previously nonzero coef */ + arith_encode(cinfo, st + 2, (v & 1)); + else { /* newly nonzero coef */ + arith_encode(cinfo, st + 1, 1); + arith_encode(cinfo, entropy->fixed_bin, 1); + } + break; + } + } + arith_encode(cinfo, st + 1, 0); + st += 3; + } + } + /* Encode EOB decision only if k < cinfo->Se */ + if (k < cinfo->Se) { + st = entropy->ac_stats[tbl] + 3 * k; + arith_encode(cinfo, st, 1); + } + + return TRUE; +} + + +/* + * Encode and output one MCU's worth of arithmetic-compressed coefficients. + */ + +METHODDEF(boolean) +encode_mcu (j_compress_ptr cinfo, JBLOCKARRAY MCU_data) +{ + arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; + const int * natural_order; + JBLOCKROW block; + unsigned char *st; + int tbl, k, ke; + int v, v2, m; + int blkn, ci; + jpeg_component_info * compptr; + + /* Emit restart marker if needed */ + if (cinfo->restart_interval) { + if (entropy->restarts_to_go == 0) { + emit_restart(cinfo, entropy->next_restart_num); + entropy->restarts_to_go = cinfo->restart_interval; + entropy->next_restart_num++; + entropy->next_restart_num &= 7; + } + entropy->restarts_to_go--; + } + + natural_order = cinfo->natural_order; + + /* Encode the MCU data blocks */ + for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { + block = MCU_data[blkn]; + ci = cinfo->MCU_membership[blkn]; + compptr = cinfo->cur_comp_info[ci]; + + /* Sections F.1.4.1 & F.1.4.4.1: Encoding of DC coefficients */ + + tbl = compptr->dc_tbl_no; + + /* Table F.4: Point to statistics bin S0 for DC coefficient coding */ + st = entropy->dc_stats[tbl] + entropy->dc_context[ci]; + + /* Figure F.4: Encode_DC_DIFF */ + if ((v = (*block)[0] - entropy->last_dc_val[ci]) == 0) { + arith_encode(cinfo, st, 0); + entropy->dc_context[ci] = 0; /* zero diff category */ + } else { + entropy->last_dc_val[ci] = (*block)[0]; + arith_encode(cinfo, st, 1); + /* Figure F.6: Encoding nonzero value v */ + /* Figure F.7: Encoding the sign of v */ + if (v > 0) { + arith_encode(cinfo, st + 1, 0); /* Table F.4: SS = S0 + 1 */ + st += 2; /* Table F.4: SP = S0 + 2 */ + entropy->dc_context[ci] = 4; /* small positive diff category */ + } else { + v = -v; + arith_encode(cinfo, st + 1, 1); /* Table F.4: SS = S0 + 1 */ + st += 3; /* Table F.4: SN = S0 + 3 */ + entropy->dc_context[ci] = 8; /* small negative diff category */ + } + /* Figure F.8: Encoding the magnitude category of v */ + m = 0; + if (v -= 1) { + arith_encode(cinfo, st, 1); + m = 1; + v2 = v; + st = entropy->dc_stats[tbl] + 20; /* Table F.4: X1 = 20 */ + while (v2 >>= 1) { + arith_encode(cinfo, st, 1); + m <<= 1; + st += 1; + } + } + arith_encode(cinfo, st, 0); + /* Section F.1.4.4.1.2: Establish dc_context conditioning category */ + if (m < (int) ((1L << cinfo->arith_dc_L[tbl]) >> 1)) + entropy->dc_context[ci] = 0; /* zero diff category */ + else if (m > (int) ((1L << cinfo->arith_dc_U[tbl]) >> 1)) + entropy->dc_context[ci] += 8; /* large diff category */ + /* Figure F.9: Encoding the magnitude bit pattern of v */ + st += 14; + while (m >>= 1) + arith_encode(cinfo, st, (m & v) ? 1 : 0); + } + + /* Sections F.1.4.2 & F.1.4.4.2: Encoding of AC coefficients */ + + if ((ke = cinfo->lim_Se) == 0) continue; + tbl = compptr->ac_tbl_no; + + /* Establish EOB (end-of-block) index */ + do { + if ((*block)[natural_order[ke]]) break; + } while (--ke); + + /* Figure F.5: Encode_AC_Coefficients */ + for (k = 0; k < ke;) { + st = entropy->ac_stats[tbl] + 3 * k; + arith_encode(cinfo, st, 0); /* EOB decision */ + while ((v = (*block)[natural_order[++k]]) == 0) { + arith_encode(cinfo, st + 1, 0); + st += 3; + } + arith_encode(cinfo, st + 1, 1); + /* Figure F.6: Encoding nonzero value v */ + /* Figure F.7: Encoding the sign of v */ + if (v > 0) { + arith_encode(cinfo, entropy->fixed_bin, 0); + } else { + v = -v; + arith_encode(cinfo, entropy->fixed_bin, 1); + } + st += 2; + /* Figure F.8: Encoding the magnitude category of v */ + m = 0; + if (v -= 1) { + arith_encode(cinfo, st, 1); + m = 1; + v2 = v; + if (v2 >>= 1) { + arith_encode(cinfo, st, 1); + m <<= 1; + st = entropy->ac_stats[tbl] + + (k <= cinfo->arith_ac_K[tbl] ? 189 : 217); + while (v2 >>= 1) { + arith_encode(cinfo, st, 1); + m <<= 1; + st += 1; + } + } + } + arith_encode(cinfo, st, 0); + /* Figure F.9: Encoding the magnitude bit pattern of v */ + st += 14; + while (m >>= 1) + arith_encode(cinfo, st, (m & v) ? 1 : 0); + } + /* Encode EOB decision only if k < cinfo->lim_Se */ + if (k < cinfo->lim_Se) { + st = entropy->ac_stats[tbl] + 3 * k; + arith_encode(cinfo, st, 1); + } + } + + return TRUE; +} + + +/* + * Initialize for an arithmetic-compressed scan. + */ + +METHODDEF(void) +start_pass (j_compress_ptr cinfo, boolean gather_statistics) +{ + arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; + int ci, tbl; + jpeg_component_info * compptr; + + if (gather_statistics) + /* Make sure to avoid that in the master control logic! + * We are fully adaptive here and need no extra + * statistics gathering pass! + */ + ERREXIT(cinfo, JERR_NOT_COMPILED); + + /* We assume jcmaster.c already validated the progressive scan parameters. */ + + /* Select execution routines */ + if (cinfo->progressive_mode) { + if (cinfo->Ah == 0) { + if (cinfo->Ss == 0) + entropy->pub.encode_mcu = encode_mcu_DC_first; + else + entropy->pub.encode_mcu = encode_mcu_AC_first; + } else { + if (cinfo->Ss == 0) + entropy->pub.encode_mcu = encode_mcu_DC_refine; + else + entropy->pub.encode_mcu = encode_mcu_AC_refine; + } + } else + entropy->pub.encode_mcu = encode_mcu; + + /* Allocate & initialize requested statistics areas */ + for (ci = 0; ci < cinfo->comps_in_scan; ci++) { + compptr = cinfo->cur_comp_info[ci]; + /* DC needs no table for refinement scan */ + if (cinfo->Ss == 0 && cinfo->Ah == 0) { + tbl = compptr->dc_tbl_no; + if (tbl < 0 || tbl >= NUM_ARITH_TBLS) + ERREXIT1(cinfo, JERR_NO_ARITH_TABLE, tbl); + if (entropy->dc_stats[tbl] == NULL) + entropy->dc_stats[tbl] = (unsigned char *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, DC_STAT_BINS); + MEMZERO(entropy->dc_stats[tbl], DC_STAT_BINS); + /* Initialize DC predictions to 0 */ + entropy->last_dc_val[ci] = 0; + entropy->dc_context[ci] = 0; + } + /* AC needs no table when not present */ + if (cinfo->Se) { + tbl = compptr->ac_tbl_no; + if (tbl < 0 || tbl >= NUM_ARITH_TBLS) + ERREXIT1(cinfo, JERR_NO_ARITH_TABLE, tbl); + if (entropy->ac_stats[tbl] == NULL) + entropy->ac_stats[tbl] = (unsigned char *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, AC_STAT_BINS); + MEMZERO(entropy->ac_stats[tbl], AC_STAT_BINS); +#ifdef CALCULATE_SPECTRAL_CONDITIONING + if (cinfo->progressive_mode) + /* Section G.1.3.2: Set appropriate arithmetic conditioning value Kx */ + cinfo->arith_ac_K[tbl] = cinfo->Ss + ((8 + cinfo->Se - cinfo->Ss) >> 4); +#endif + } + } + + /* Initialize arithmetic encoding variables */ + entropy->c = 0; + entropy->a = 0x10000L; + entropy->sc = 0; + entropy->zc = 0; + entropy->ct = 11; + entropy->buffer = -1; /* empty */ + + /* Initialize restart stuff */ + entropy->restarts_to_go = cinfo->restart_interval; + entropy->next_restart_num = 0; +} + + +/* + * Module initialization routine for arithmetic entropy encoding. + */ + +GLOBAL(void) +jinit_arith_encoder (j_compress_ptr cinfo) +{ + arith_entropy_ptr entropy; + int i; + + entropy = (arith_entropy_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(arith_entropy_encoder)); + cinfo->entropy = &entropy->pub; + entropy->pub.start_pass = start_pass; + entropy->pub.finish_pass = finish_pass; + + /* Mark tables unallocated */ + for (i = 0; i < NUM_ARITH_TBLS; i++) { + entropy->dc_stats[i] = NULL; + entropy->ac_stats[i] = NULL; + } + + /* Initialize index for fixed probability estimation */ + entropy->fixed_bin[0] = 113; +} diff --git a/cmake/externals/libjpeg/jccoefct.c b/cmake/externals/libjpeg/jccoefct.c new file mode 100644 index 000000000..494aa2298 --- /dev/null +++ b/cmake/externals/libjpeg/jccoefct.c @@ -0,0 +1,456 @@ +/* + * jccoefct.c + * + * Copyright (C) 1994-1997, Thomas G. Lane. + * Modified 2003-2022 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains the coefficient buffer controller for compression. + * This controller is the top level of the JPEG compressor proper. + * The coefficient buffer lies between forward-DCT and entropy encoding steps. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* We use a full-image coefficient buffer when doing Huffman optimization, + * and also for writing multiple-scan JPEG files. In all cases, the DCT + * step is run during the first pass, and subsequent passes need only read + * the buffered coefficients. + */ +#ifdef ENTROPY_OPT_SUPPORTED +#define FULL_COEF_BUFFER_SUPPORTED +#else +#ifdef C_MULTISCAN_FILES_SUPPORTED +#define FULL_COEF_BUFFER_SUPPORTED +#endif +#endif + + +/* Private buffer controller object */ + +typedef struct { + struct jpeg_c_coef_controller pub; /* public fields */ + + JDIMENSION iMCU_row_num; /* iMCU row # within image */ + JDIMENSION MCU_ctr; /* counts MCUs processed in current row */ + int MCU_vert_offset; /* counts MCU rows within iMCU row */ + int MCU_rows_per_iMCU_row; /* number of such rows needed */ + + /* For single-pass compression, it's sufficient to buffer just one MCU + * (although this may prove a bit slow in practice). + * We append a workspace of C_MAX_BLOCKS_IN_MCU coefficient blocks, + * and reuse it for each MCU constructed and sent. + * In multi-pass modes, this array points to the current MCU's blocks + * within the virtual arrays. + */ + JBLOCKROW MCU_buffer[C_MAX_BLOCKS_IN_MCU]; + + /* In multi-pass modes, we need a virtual block array for each component. */ + jvirt_barray_ptr whole_image[MAX_COMPONENTS]; + + /* Workspace for single-pass compression (omitted otherwise). */ + JBLOCK blk_buffer[C_MAX_BLOCKS_IN_MCU]; +} my_coef_controller; + +typedef my_coef_controller * my_coef_ptr; + + +/* Forward declarations */ +METHODDEF(boolean) compress_data + JPP((j_compress_ptr cinfo, JSAMPIMAGE input_buf)); +#ifdef FULL_COEF_BUFFER_SUPPORTED +METHODDEF(boolean) compress_first_pass + JPP((j_compress_ptr cinfo, JSAMPIMAGE input_buf)); +METHODDEF(boolean) compress_output + JPP((j_compress_ptr cinfo, JSAMPIMAGE input_buf)); +#endif + + +LOCAL(void) +start_iMCU_row (j_compress_ptr cinfo) +/* Reset within-iMCU-row counters for a new row */ +{ + my_coef_ptr coef = (my_coef_ptr) cinfo->coef; + + /* In an interleaved scan, an MCU row is the same as an iMCU row. + * In a noninterleaved scan, an iMCU row has v_samp_factor MCU rows. + * But at the bottom of the image, process only what's left. + */ + if (cinfo->comps_in_scan > 1) { + coef->MCU_rows_per_iMCU_row = 1; + } else { + if (coef->iMCU_row_num < (cinfo->total_iMCU_rows-1)) + coef->MCU_rows_per_iMCU_row = cinfo->cur_comp_info[0]->v_samp_factor; + else + coef->MCU_rows_per_iMCU_row = cinfo->cur_comp_info[0]->last_row_height; + } + + coef->MCU_ctr = 0; + coef->MCU_vert_offset = 0; +} + + +/* + * Initialize for a processing pass. + */ + +METHODDEF(void) +start_pass_coef (j_compress_ptr cinfo, J_BUF_MODE pass_mode) +{ + my_coef_ptr coef = (my_coef_ptr) cinfo->coef; + + coef->iMCU_row_num = 0; + start_iMCU_row(cinfo); + + switch (pass_mode) { + case JBUF_PASS_THRU: + if (coef->whole_image[0] != NULL) + ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); + coef->pub.compress_data = compress_data; + break; +#ifdef FULL_COEF_BUFFER_SUPPORTED + case JBUF_SAVE_AND_PASS: + if (coef->whole_image[0] == NULL) + ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); + coef->pub.compress_data = compress_first_pass; + break; + case JBUF_CRANK_DEST: + if (coef->whole_image[0] == NULL) + ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); + coef->pub.compress_data = compress_output; + break; +#endif + default: + ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); + } +} + + +/* + * Process some data in the single-pass case. + * We process the equivalent of one fully interleaved MCU row ("iMCU" row) + * per call, ie, v_samp_factor block rows for each component in the image. + * Returns TRUE if the iMCU row is completed, FALSE if suspended. + * + * NB: input_buf contains a plane for each component in image, + * which we index according to the component's SOF position. + */ + +METHODDEF(boolean) +compress_data (j_compress_ptr cinfo, JSAMPIMAGE input_buf) +{ + my_coef_ptr coef = (my_coef_ptr) cinfo->coef; + JDIMENSION MCU_col_num; /* index of current MCU within row */ + JDIMENSION last_MCU_col = cinfo->MCUs_per_row - 1; + JDIMENSION last_iMCU_row = cinfo->total_iMCU_rows - 1; + int ci, xindex, yindex, yoffset, blockcnt; + JBLOCKROW blkp; + JSAMPARRAY input_ptr; + JDIMENSION xpos; + jpeg_component_info *compptr; + forward_DCT_ptr forward_DCT; + + /* Loop to write as much as one whole iMCU row */ + for (yoffset = coef->MCU_vert_offset; yoffset < coef->MCU_rows_per_iMCU_row; + yoffset++) { + for (MCU_col_num = coef->MCU_ctr; MCU_col_num <= last_MCU_col; + MCU_col_num++) { + /* Determine where data comes from in input_buf and do the DCT thing. + * Each call on forward_DCT processes a horizontal row of DCT blocks as + * wide as an MCU. Dummy blocks at the right or bottom edge are filled in + * specially. The data in them does not matter for image reconstruction, + * so we fill them with values that will encode to the smallest amount of + * data, viz: all zeroes in the AC entries, DC entries equal to previous + * block's DC value. (Thanks to Thomas Kinsman for this idea.) + */ + blkp = coef->blk_buffer; /* pointer to current DCT block within MCU */ + for (ci = 0; ci < cinfo->comps_in_scan; ci++) { + compptr = cinfo->cur_comp_info[ci]; + forward_DCT = cinfo->fdct->forward_DCT[compptr->component_index]; + input_ptr = input_buf[compptr->component_index] + + yoffset * compptr->DCT_v_scaled_size; + /* ypos == (yoffset + yindex) * compptr->DCT_v_scaled_size */ + blockcnt = (MCU_col_num < last_MCU_col) ? compptr->MCU_width + : compptr->last_col_width; + xpos = MCU_col_num * compptr->MCU_sample_width; + for (yindex = 0; yindex < compptr->MCU_height; yindex++) { + if (coef->iMCU_row_num < last_iMCU_row || + yoffset + yindex < compptr->last_row_height) { + (*forward_DCT) (cinfo, compptr, input_ptr, blkp, + xpos, (JDIMENSION) blockcnt); + input_ptr += compptr->DCT_v_scaled_size; + blkp += blockcnt; + /* Dummy blocks at right edge */ + if ((xindex = compptr->MCU_width - blockcnt) == 0) + continue; + } else { + /* At bottom of image, need a whole row of dummy blocks */ + xindex = compptr->MCU_width; + } + /* Fill in any dummy blocks needed in this row */ + MEMZERO(blkp, xindex * SIZEOF(JBLOCK)); + do { + blkp[0][0] = blkp[-1][0]; + blkp++; + } while (--xindex); + } + } + /* Try to write the MCU. In event of a suspension failure, we will + * re-DCT the MCU on restart (a bit inefficient, could be fixed...) + */ + if (! (*cinfo->entropy->encode_mcu) (cinfo, coef->MCU_buffer)) { + /* Suspension forced; update state counters and exit */ + coef->MCU_vert_offset = yoffset; + coef->MCU_ctr = MCU_col_num; + return FALSE; + } + } + /* Completed an MCU row, but perhaps not an iMCU row */ + coef->MCU_ctr = 0; + } + /* Completed the iMCU row, advance counters for next one */ + coef->iMCU_row_num++; + start_iMCU_row(cinfo); + return TRUE; +} + + +#ifdef FULL_COEF_BUFFER_SUPPORTED + +/* + * Process some data in the first pass of a multi-pass case. + * We process the equivalent of one fully interleaved MCU row ("iMCU" row) + * per call, ie, v_samp_factor block rows for each component in the image. + * This amount of data is read from the source buffer, DCT'd and quantized, + * and saved into the virtual arrays. We also generate suitable dummy blocks + * as needed at the right and lower edges. (The dummy blocks are constructed + * in the virtual arrays, which have been padded appropriately.) This makes + * it possible for subsequent passes not to worry about real vs. dummy blocks. + * + * We must also emit the data to the entropy encoder. This is conveniently + * done by calling compress_output() after we've loaded the current strip + * of the virtual arrays. + * + * NB: input_buf contains a plane for each component in image. All + * components are DCT'd and loaded into the virtual arrays in this pass. + * However, it may be that only a subset of the components are emitted to + * the entropy encoder during this first pass; be careful about looking + * at the scan-dependent variables (MCU dimensions, etc). + */ + +METHODDEF(boolean) +compress_first_pass (j_compress_ptr cinfo, JSAMPIMAGE input_buf) +{ + my_coef_ptr coef = (my_coef_ptr) cinfo->coef; + JDIMENSION last_iMCU_row = cinfo->total_iMCU_rows - 1; + JDIMENSION blocks_across, MCUs_across, MCUindex; + int bi, ci, h_samp_factor, block_row, block_rows, ndummy; + JCOEF lastDC; + jpeg_component_info *compptr; + JBLOCKARRAY buffer; + JBLOCKROW thisblockrow, lastblockrow; + JSAMPARRAY input_ptr; + forward_DCT_ptr forward_DCT; + + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + /* Align the virtual buffer for this component. */ + buffer = (*cinfo->mem->access_virt_barray) + ((j_common_ptr) cinfo, coef->whole_image[ci], + coef->iMCU_row_num * compptr->v_samp_factor, + (JDIMENSION) compptr->v_samp_factor, TRUE); + /* Count non-dummy DCT block rows in this iMCU row. */ + if (coef->iMCU_row_num < last_iMCU_row) + block_rows = compptr->v_samp_factor; + else { + /* NB: can't use last_row_height here, since may not be set! */ + block_rows = (int) (compptr->height_in_blocks % compptr->v_samp_factor); + if (block_rows == 0) block_rows = compptr->v_samp_factor; + } + blocks_across = compptr->width_in_blocks; + h_samp_factor = compptr->h_samp_factor; + /* Count number of dummy blocks to be added at the right margin. */ + ndummy = (int) (blocks_across % h_samp_factor); + if (ndummy > 0) + ndummy = h_samp_factor - ndummy; + forward_DCT = cinfo->fdct->forward_DCT[ci]; + input_ptr = input_buf[ci]; + /* Perform DCT for all non-dummy blocks in this iMCU row. Each call + * on forward_DCT processes a complete horizontal row of DCT blocks. + */ + for (block_row = 0; block_row < block_rows; block_row++) { + thisblockrow = buffer[block_row]; + (*forward_DCT) (cinfo, compptr, input_ptr, thisblockrow, + (JDIMENSION) 0, blocks_across); + input_ptr += compptr->DCT_v_scaled_size; + if (ndummy > 0) { + /* Create dummy blocks at the right edge of the image. */ + thisblockrow += blocks_across; /* => first dummy block */ + FMEMZERO((void FAR *) thisblockrow, ndummy * SIZEOF(JBLOCK)); + lastDC = thisblockrow[-1][0]; + for (bi = 0; bi < ndummy; bi++) { + thisblockrow[bi][0] = lastDC; + } + } + } + /* If at end of image, create dummy block rows as needed. + * The tricky part here is that within each MCU, we want the DC values + * of the dummy blocks to match the last real block's DC value. + * This squeezes a few more bytes out of the resulting file... + */ + if (block_row < compptr->v_samp_factor) { + blocks_across += ndummy; /* include lower right corner */ + MCUs_across = blocks_across / h_samp_factor; + do { + thisblockrow = buffer[block_row]; + lastblockrow = buffer[block_row-1]; + FMEMZERO((void FAR *) thisblockrow, + (size_t) blocks_across * SIZEOF(JBLOCK)); + for (MCUindex = 0; MCUindex < MCUs_across; MCUindex++) { + lastDC = lastblockrow[h_samp_factor-1][0]; + for (bi = 0; bi < h_samp_factor; bi++) { + thisblockrow[bi][0] = lastDC; + } + thisblockrow += h_samp_factor; /* advance to next MCU in row */ + lastblockrow += h_samp_factor; + } + } while (++block_row < compptr->v_samp_factor); + } + } + /* NB: compress_output will increment iMCU_row_num if successful. + * A suspension return will result in redoing all the work above next time. + */ + + /* Emit data to the entropy encoder, sharing code with subsequent passes */ + return compress_output(cinfo, input_buf); +} + + +/* + * Process some data in subsequent passes of a multi-pass case. + * We process the equivalent of one fully interleaved MCU row ("iMCU" row) + * per call, ie, v_samp_factor block rows for each component in the scan. + * The data is obtained from the virtual arrays and fed to the entropy coder. + * Returns TRUE if the iMCU row is completed, FALSE if suspended. + * + * NB: input_buf is ignored; it is likely to be a NULL pointer. + */ + +METHODDEF(boolean) +compress_output (j_compress_ptr cinfo, JSAMPIMAGE input_buf) +{ + my_coef_ptr coef = (my_coef_ptr) cinfo->coef; + JDIMENSION MCU_col_num; /* index of current MCU within row */ + int ci, xindex, yindex, yoffset; + JDIMENSION start_col; + JBLOCKARRAY blkp; + JBLOCKARRAY buffer[MAX_COMPS_IN_SCAN]; + JBLOCKROW buffer_ptr; + jpeg_component_info *compptr; + + /* Align the virtual buffers for the components used in this scan. + * NB: during first pass, this is safe only because the buffers will + * already be aligned properly, so jmemmgr.c won't need to do any I/O. + */ + for (ci = 0; ci < cinfo->comps_in_scan; ci++) { + compptr = cinfo->cur_comp_info[ci]; + buffer[ci] = (*cinfo->mem->access_virt_barray) + ((j_common_ptr) cinfo, coef->whole_image[compptr->component_index], + coef->iMCU_row_num * compptr->v_samp_factor, + (JDIMENSION) compptr->v_samp_factor, FALSE); + } + + /* Loop to process one whole iMCU row */ + for (yoffset = coef->MCU_vert_offset; yoffset < coef->MCU_rows_per_iMCU_row; + yoffset++) { + for (MCU_col_num = coef->MCU_ctr; MCU_col_num < cinfo->MCUs_per_row; + MCU_col_num++) { + /* Construct list of pointers to DCT blocks belonging to this MCU */ + blkp = coef->MCU_buffer; /* pointer to current DCT block within MCU */ + for (ci = 0; ci < cinfo->comps_in_scan; ci++) { + compptr = cinfo->cur_comp_info[ci]; + start_col = MCU_col_num * compptr->MCU_width; + for (yindex = 0; yindex < compptr->MCU_height; yindex++) { + buffer_ptr = buffer[ci][yoffset + yindex] + start_col; + xindex = compptr->MCU_width; + do { + *blkp++ = buffer_ptr++; + } while (--xindex); + } + } + /* Try to write the MCU. */ + if (! (*cinfo->entropy->encode_mcu) (cinfo, coef->MCU_buffer)) { + /* Suspension forced; update state counters and exit */ + coef->MCU_vert_offset = yoffset; + coef->MCU_ctr = MCU_col_num; + return FALSE; + } + } + /* Completed an MCU row, but perhaps not an iMCU row */ + coef->MCU_ctr = 0; + } + /* Completed the iMCU row, advance counters for next one */ + coef->iMCU_row_num++; + start_iMCU_row(cinfo); + return TRUE; +} + +#endif /* FULL_COEF_BUFFER_SUPPORTED */ + + +/* + * Initialize coefficient buffer controller. + */ + +GLOBAL(void) +jinit_c_coef_controller (j_compress_ptr cinfo, boolean need_full_buffer) +{ + my_coef_ptr coef; + + if (need_full_buffer) { +#ifdef FULL_COEF_BUFFER_SUPPORTED + /* Allocate a full-image virtual array for each component, */ + /* padded to a multiple of samp_factor DCT blocks in each direction. */ + int ci; + jpeg_component_info *compptr; + + coef = (my_coef_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, + SIZEOF(my_coef_controller) - SIZEOF(coef->blk_buffer)); + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + coef->whole_image[ci] = (*cinfo->mem->request_virt_barray) + ((j_common_ptr) cinfo, JPOOL_IMAGE, FALSE, + (JDIMENSION) jround_up((long) compptr->width_in_blocks, + (long) compptr->h_samp_factor), + (JDIMENSION) jround_up((long) compptr->height_in_blocks, + (long) compptr->v_samp_factor), + (JDIMENSION) compptr->v_samp_factor); + } +#else + ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); +#endif + } else { + /* We only need a single-MCU buffer. */ + JBLOCKARRAY blkp; + JBLOCKROW buffer_ptr; + int bi; + + coef = (my_coef_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_coef_controller)); + blkp = coef->MCU_buffer; + buffer_ptr = coef->blk_buffer; + bi = C_MAX_BLOCKS_IN_MCU; + do { + *blkp++ = buffer_ptr++; + } while (--bi); + coef->whole_image[0] = NULL; /* flag for no virtual arrays */ + } + + coef->pub.start_pass = start_pass_coef; + cinfo->coef = &coef->pub; +} diff --git a/cmake/externals/libjpeg/jccolor.c b/cmake/externals/libjpeg/jccolor.c new file mode 100644 index 000000000..c028dd9db --- /dev/null +++ b/cmake/externals/libjpeg/jccolor.c @@ -0,0 +1,598 @@ +/* + * jccolor.c + * + * Copyright (C) 1991-1996, Thomas G. Lane. + * Modified 2011-2023 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains input colorspace conversion routines. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* Private subobject */ + +typedef struct { + struct jpeg_color_converter pub; /* public fields */ + + /* Private state for RGB->YCC conversion */ + INT32 * rgb_ycc_tab; /* => table for RGB to YCbCr conversion */ +} my_color_converter; + +typedef my_color_converter * my_cconvert_ptr; + + +/**************** RGB -> YCbCr conversion: most common case **************/ + +/* + * YCbCr is defined per Recommendation ITU-R BT.601-7 (03/2011), + * previously known as Recommendation CCIR 601-1, except that Cb and Cr + * are normalized to the range 0..MAXJSAMPLE rather than -0.5 .. 0.5. + * sRGB (standard RGB color space) is defined per IEC 61966-2-1:1999. + * sYCC (standard luma-chroma-chroma color space with extended gamut) + * is defined per IEC 61966-2-1:1999 Amendment A1:2003 Annex F. + * bg-sRGB and bg-sYCC (big gamut standard color spaces) + * are defined per IEC 61966-2-1:1999 Amendment A1:2003 Annex G. + * Note that the derived conversion coefficients given in some of these + * documents are imprecise. The general conversion equations are + * Y = Kr * R + (1 - Kr - Kb) * G + Kb * B + * Cb = (B - Y) / (1 - Kb) / K + * Cr = (R - Y) / (1 - Kr) / K + * With Kr = 0.299 and Kb = 0.114 (derived according to SMPTE RP 177-1993 + * from the 1953 FCC NTSC primaries and CIE Illuminant C), K = 2 for sYCC, + * the conversion equations to be implemented are therefore + * Y = 0.299 * R + 0.587 * G + 0.114 * B + * Cb = -0.168735892 * R - 0.331264108 * G + 0.5 * B + CENTERJSAMPLE + * Cr = 0.5 * R - 0.418687589 * G - 0.081312411 * B + CENTERJSAMPLE + * Note: older versions of the IJG code used a zero offset of MAXJSAMPLE/2, + * rather than CENTERJSAMPLE, for Cb and Cr. This gave equal positive and + * negative swings for Cb/Cr, but meant that grayscale values (Cb=Cr=0) + * were not represented exactly. Now we sacrifice exact representation of + * maximum red and maximum blue in order to get exact grayscales. + * + * To avoid floating-point arithmetic, we represent the fractional constants + * as integers scaled up by 2^16 (about 4 digits precision); we have to divide + * the products by 2^16, with appropriate rounding, to get the correct answer. + * + * For even more speed, we avoid doing any multiplications in the inner loop + * by precalculating the constants times R,G,B for all possible values. + * For 8-bit JSAMPLEs this is very reasonable (only 256 entries per table); + * for 9-bit to 12-bit samples it is still acceptable. It's not very + * reasonable for 16-bit samples, but if you want lossless storage + * you shouldn't be changing colorspace anyway. + * The CENTERJSAMPLE offsets and the rounding fudge-factor of 0.5 are included + * in the tables to save adding them separately in the inner loop. + */ + +#define SCALEBITS 16 /* speediest right-shift on some machines */ +#define CBCR_OFFSET ((INT32) CENTERJSAMPLE << SCALEBITS) +#define ONE_HALF ((INT32) 1 << (SCALEBITS-1)) +#define FIX(x) ((INT32) ((x) * (1L< Y section */ +#define G_Y_OFF (1*(MAXJSAMPLE+1)) /* offset to G => Y section */ +#define B_Y_OFF (2*(MAXJSAMPLE+1)) /* etc. */ +#define R_CB_OFF (3*(MAXJSAMPLE+1)) +#define G_CB_OFF (4*(MAXJSAMPLE+1)) +#define B_CB_OFF (5*(MAXJSAMPLE+1)) +#define R_CR_OFF B_CB_OFF /* B=>Cb, R=>Cr are the same */ +#define G_CR_OFF (6*(MAXJSAMPLE+1)) +#define B_CR_OFF (7*(MAXJSAMPLE+1)) +#define TABLE_SIZE (8*(MAXJSAMPLE+1)) + + +/* + * Initialize for RGB->YCC colorspace conversion. + */ + +METHODDEF(void) +rgb_ycc_start (j_compress_ptr cinfo) +{ + my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; + INT32 * rgb_ycc_tab; + INT32 i; + + /* Allocate and fill in the conversion tables. */ + cconvert->rgb_ycc_tab = rgb_ycc_tab = (INT32 *) + (*cinfo->mem->alloc_small) ((j_common_ptr) cinfo, JPOOL_IMAGE, + TABLE_SIZE * SIZEOF(INT32)); + + for (i = 0; i <= MAXJSAMPLE; i++) { + rgb_ycc_tab[i+R_Y_OFF] = FIX(0.299) * i; + rgb_ycc_tab[i+G_Y_OFF] = FIX(0.587) * i; + rgb_ycc_tab[i+B_Y_OFF] = FIX(0.114) * i + ONE_HALF; + rgb_ycc_tab[i+R_CB_OFF] = (- FIX(0.168735892)) * i; + rgb_ycc_tab[i+G_CB_OFF] = (- FIX(0.331264108)) * i; + /* We use a rounding fudge-factor of 0.5-epsilon for Cb and Cr. + * This ensures that the maximum output will round to MAXJSAMPLE + * not MAXJSAMPLE+1, and thus that we don't have to range-limit. + */ + rgb_ycc_tab[i+B_CB_OFF] = (i << (SCALEBITS-1)) + CBCR_OFFSET + ONE_HALF-1; +/* B=>Cb and R=>Cr tables are the same + rgb_ycc_tab[i+R_CR_OFF] = (i << (SCALEBITS-1)) + CBCR_OFFSET + ONE_HALF-1; +*/ + rgb_ycc_tab[i+G_CR_OFF] = (- FIX(0.418687589)) * i; + rgb_ycc_tab[i+B_CR_OFF] = (- FIX(0.081312411)) * i; + } +} + + +/* + * Convert some rows of samples to the JPEG colorspace. + * + * Note that we change from the application's interleaved-pixel format + * to our internal noninterleaved, one-plane-per-component format. The + * input buffer is therefore three times as wide as the output buffer. + * + * A starting row offset is provided only for the output buffer. The + * caller can easily adjust the passed input_buf value to accommodate + * any row offset required on that side. + */ + +METHODDEF(void) +rgb_ycc_convert (j_compress_ptr cinfo, + JSAMPARRAY input_buf, JSAMPIMAGE output_buf, + JDIMENSION output_row, int num_rows) +{ + my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; + register int r, g, b; + register INT32 * ctab = cconvert->rgb_ycc_tab; + register JSAMPROW inptr; + register JSAMPROW outptr0, outptr1, outptr2; + register JDIMENSION col; + JDIMENSION num_cols = cinfo->image_width; + + while (--num_rows >= 0) { + inptr = *input_buf++; + outptr0 = output_buf[0][output_row]; + outptr1 = output_buf[1][output_row]; + outptr2 = output_buf[2][output_row]; + output_row++; + for (col = 0; col < num_cols; col++) { + r = GETJSAMPLE(inptr[RGB_RED]); + g = GETJSAMPLE(inptr[RGB_GREEN]); + b = GETJSAMPLE(inptr[RGB_BLUE]); + inptr += RGB_PIXELSIZE; + /* If the inputs are 0..MAXJSAMPLE, the outputs of these equations + * must be too; we do not need an explicit range-limiting operation. + * Hence the value being shifted is never negative, and we don't + * need the general RIGHT_SHIFT macro. + */ + /* Y */ + outptr0[col] = (JSAMPLE) + ((ctab[r+R_Y_OFF] + ctab[g+G_Y_OFF] + ctab[b+B_Y_OFF]) + >> SCALEBITS); + /* Cb */ + outptr1[col] = (JSAMPLE) + ((ctab[r+R_CB_OFF] + ctab[g+G_CB_OFF] + ctab[b+B_CB_OFF]) + >> SCALEBITS); + /* Cr */ + outptr2[col] = (JSAMPLE) + ((ctab[r+R_CR_OFF] + ctab[g+G_CR_OFF] + ctab[b+B_CR_OFF]) + >> SCALEBITS); + } + } +} + + +/**************** Cases other than RGB -> YCbCr **************/ + + +/* + * Convert some rows of samples to the JPEG colorspace. + * This version handles RGB->grayscale conversion, + * which is the same as the RGB->Y portion of RGB->YCbCr. + * We assume rgb_ycc_start has been called (we only use the Y tables). + */ + +METHODDEF(void) +rgb_gray_convert (j_compress_ptr cinfo, + JSAMPARRAY input_buf, JSAMPIMAGE output_buf, + JDIMENSION output_row, int num_rows) +{ + my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; + register INT32 y; + register INT32 * ctab = cconvert->rgb_ycc_tab; + register JSAMPROW inptr; + register JSAMPROW outptr; + register JDIMENSION col; + JDIMENSION num_cols = cinfo->image_width; + + while (--num_rows >= 0) { + inptr = *input_buf++; + outptr = output_buf[0][output_row++]; + for (col = 0; col < num_cols; col++) { + y = ctab[R_Y_OFF + GETJSAMPLE(inptr[RGB_RED])]; + y += ctab[G_Y_OFF + GETJSAMPLE(inptr[RGB_GREEN])]; + y += ctab[B_Y_OFF + GETJSAMPLE(inptr[RGB_BLUE])]; + inptr += RGB_PIXELSIZE; + outptr[col] = (JSAMPLE) (y >> SCALEBITS); + } + } +} + + +/* + * Convert some rows of samples to the JPEG colorspace. + * This version handles Adobe-style CMYK->YCCK conversion, + * where we convert R=1-C, G=1-M, and B=1-Y to YCbCr using the + * same conversion as above, while passing K (black) unchanged. + * We assume rgb_ycc_start has been called. + */ + +METHODDEF(void) +cmyk_ycck_convert (j_compress_ptr cinfo, + JSAMPARRAY input_buf, JSAMPIMAGE output_buf, + JDIMENSION output_row, int num_rows) +{ + my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; + register int r, g, b; + register INT32 * ctab = cconvert->rgb_ycc_tab; + register JSAMPROW inptr; + register JSAMPROW outptr0, outptr1, outptr2, outptr3; + register JDIMENSION col; + JDIMENSION num_cols = cinfo->image_width; + + while (--num_rows >= 0) { + inptr = *input_buf++; + outptr0 = output_buf[0][output_row]; + outptr1 = output_buf[1][output_row]; + outptr2 = output_buf[2][output_row]; + outptr3 = output_buf[3][output_row]; + output_row++; + for (col = 0; col < num_cols; col++) { + r = MAXJSAMPLE - GETJSAMPLE(inptr[0]); + g = MAXJSAMPLE - GETJSAMPLE(inptr[1]); + b = MAXJSAMPLE - GETJSAMPLE(inptr[2]); + /* K passes through as-is */ + outptr3[col] = inptr[3]; /* don't need GETJSAMPLE here */ + inptr += 4; + /* If the inputs are 0..MAXJSAMPLE, the outputs of these equations + * must be too; we do not need an explicit range-limiting operation. + * Hence the value being shifted is never negative, and we don't + * need the general RIGHT_SHIFT macro. + */ + /* Y */ + outptr0[col] = (JSAMPLE) + ((ctab[r+R_Y_OFF] + ctab[g+G_Y_OFF] + ctab[b+B_Y_OFF]) + >> SCALEBITS); + /* Cb */ + outptr1[col] = (JSAMPLE) + ((ctab[r+R_CB_OFF] + ctab[g+G_CB_OFF] + ctab[b+B_CB_OFF]) + >> SCALEBITS); + /* Cr */ + outptr2[col] = (JSAMPLE) + ((ctab[r+R_CR_OFF] + ctab[g+G_CR_OFF] + ctab[b+B_CR_OFF]) + >> SCALEBITS); + } + } +} + + +/* + * Convert some rows of samples to the JPEG colorspace. + * [R,G,B] to [R-G,G,B-G] conversion with modulo calculation + * (forward reversible color transform). + * This can be seen as an adaption of the general RGB->YCbCr + * conversion equation with Kr = Kb = 0, while replacing the + * normalization by modulo calculation. + */ + +METHODDEF(void) +rgb_rgb1_convert (j_compress_ptr cinfo, + JSAMPARRAY input_buf, JSAMPIMAGE output_buf, + JDIMENSION output_row, int num_rows) +{ + register int r, g, b; + register JSAMPROW inptr; + register JSAMPROW outptr0, outptr1, outptr2; + register JDIMENSION col; + JDIMENSION num_cols = cinfo->image_width; + + while (--num_rows >= 0) { + inptr = *input_buf++; + outptr0 = output_buf[0][output_row]; + outptr1 = output_buf[1][output_row]; + outptr2 = output_buf[2][output_row]; + output_row++; + for (col = 0; col < num_cols; col++) { + r = GETJSAMPLE(inptr[RGB_RED]); + g = GETJSAMPLE(inptr[RGB_GREEN]); + b = GETJSAMPLE(inptr[RGB_BLUE]); + inptr += RGB_PIXELSIZE; + /* Assume that MAXJSAMPLE+1 is a power of 2, so that the MOD + * (modulo) operator is equivalent to the bitmask operator AND. + */ + outptr0[col] = (JSAMPLE) ((r - g + CENTERJSAMPLE) & MAXJSAMPLE); + outptr1[col] = (JSAMPLE) g; + outptr2[col] = (JSAMPLE) ((b - g + CENTERJSAMPLE) & MAXJSAMPLE); + } + } +} + + +/* + * Convert some rows of samples to the JPEG colorspace. + * This version handles grayscale output with no conversion. + * The source can be either plain grayscale or YCC (since Y == gray). + */ + +METHODDEF(void) +grayscale_convert (j_compress_ptr cinfo, + JSAMPARRAY input_buf, JSAMPIMAGE output_buf, + JDIMENSION output_row, int num_rows) +{ + register JSAMPROW inptr; + register JSAMPROW outptr; + register JDIMENSION count; + register int instride = cinfo->input_components; + JDIMENSION num_cols = cinfo->image_width; + + while (--num_rows >= 0) { + inptr = *input_buf++; + outptr = output_buf[0][output_row++]; + for (count = num_cols; count > 0; count--) { + *outptr++ = *inptr; /* don't need GETJSAMPLE() here */ + inptr += instride; + } + } +} + + +/* + * Convert some rows of samples to the JPEG colorspace. + * No colorspace conversion, but change from interleaved + * to separate-planes representation. + */ + +METHODDEF(void) +rgb_convert (j_compress_ptr cinfo, + JSAMPARRAY input_buf, JSAMPIMAGE output_buf, + JDIMENSION output_row, int num_rows) +{ + register JSAMPROW inptr; + register JSAMPROW outptr0, outptr1, outptr2; + register JDIMENSION col; + JDIMENSION num_cols = cinfo->image_width; + + while (--num_rows >= 0) { + inptr = *input_buf++; + outptr0 = output_buf[0][output_row]; + outptr1 = output_buf[1][output_row]; + outptr2 = output_buf[2][output_row]; + output_row++; + for (col = 0; col < num_cols; col++) { + /* We can dispense with GETJSAMPLE() here */ + outptr0[col] = inptr[RGB_RED]; + outptr1[col] = inptr[RGB_GREEN]; + outptr2[col] = inptr[RGB_BLUE]; + inptr += RGB_PIXELSIZE; + } + } +} + + +/* + * Convert some rows of samples to the JPEG colorspace. + * This version handles multi-component colorspaces without conversion. + * We assume input_components == num_components. + */ + +METHODDEF(void) +null_convert (j_compress_ptr cinfo, + JSAMPARRAY input_buf, JSAMPIMAGE output_buf, + JDIMENSION output_row, int num_rows) +{ + register JSAMPROW inptr; + register JSAMPROW outptr; + register JDIMENSION count; + register int num_comps = cinfo->num_components; + JDIMENSION num_cols = cinfo->image_width; + int ci; + + while (--num_rows >= 0) { + /* It seems fastest to make a separate pass for each component. */ + for (ci = 0; ci < num_comps; ci++) { + inptr = input_buf[0] + ci; + outptr = output_buf[ci][output_row]; + for (count = num_cols; count > 0; count--) { + *outptr++ = *inptr; /* don't need GETJSAMPLE() here */ + inptr += num_comps; + } + } + input_buf++; + output_row++; + } +} + + +/* + * Empty method for start_pass. + */ + +METHODDEF(void) +null_method (j_compress_ptr cinfo) +{ + /* no work needed */ +} + + +/* + * Module initialization routine for input colorspace conversion. + */ + +GLOBAL(void) +jinit_color_converter (j_compress_ptr cinfo) +{ + my_cconvert_ptr cconvert; + + cconvert = (my_cconvert_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_color_converter)); + cinfo->cconvert = &cconvert->pub; + /* set start_pass to null method until we find out differently */ + cconvert->pub.start_pass = null_method; + + /* Make sure input_components agrees with in_color_space */ + switch (cinfo->in_color_space) { + case JCS_GRAYSCALE: + if (cinfo->input_components != 1) + ERREXIT(cinfo, JERR_BAD_IN_COLORSPACE); + break; + + case JCS_RGB: + case JCS_BG_RGB: +#if RGB_PIXELSIZE != 3 + if (cinfo->input_components != RGB_PIXELSIZE) + ERREXIT(cinfo, JERR_BAD_IN_COLORSPACE); + break; +#endif /* else share code with YCbCr */ + + case JCS_YCbCr: + case JCS_BG_YCC: + if (cinfo->input_components != 3) + ERREXIT(cinfo, JERR_BAD_IN_COLORSPACE); + break; + + case JCS_CMYK: + case JCS_YCCK: + if (cinfo->input_components != 4) + ERREXIT(cinfo, JERR_BAD_IN_COLORSPACE); + break; + + default: /* JCS_UNKNOWN can be anything */ + if (cinfo->input_components < 1) + ERREXIT(cinfo, JERR_BAD_IN_COLORSPACE); + } + + /* Support color transform only for RGB colorspaces */ + if (cinfo->color_transform && + cinfo->jpeg_color_space != JCS_RGB && + cinfo->jpeg_color_space != JCS_BG_RGB) + ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); + + /* Check num_components, set conversion method based on requested space */ + switch (cinfo->jpeg_color_space) { + case JCS_GRAYSCALE: + if (cinfo->num_components != 1) + ERREXIT(cinfo, JERR_BAD_J_COLORSPACE); + switch (cinfo->in_color_space) { + case JCS_GRAYSCALE: + case JCS_YCbCr: + case JCS_BG_YCC: + cconvert->pub.color_convert = grayscale_convert; + break; + case JCS_RGB: + cconvert->pub.start_pass = rgb_ycc_start; + cconvert->pub.color_convert = rgb_gray_convert; + break; + default: + ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); + } + break; + + case JCS_RGB: + case JCS_BG_RGB: + if (cinfo->num_components != 3) + ERREXIT(cinfo, JERR_BAD_J_COLORSPACE); + if (cinfo->in_color_space != cinfo->jpeg_color_space) + ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); + switch (cinfo->color_transform) { + case JCT_NONE: + cconvert->pub.color_convert = rgb_convert; + break; + case JCT_SUBTRACT_GREEN: + cconvert->pub.color_convert = rgb_rgb1_convert; + break; + default: + ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); + } + break; + + case JCS_YCbCr: + if (cinfo->num_components != 3) + ERREXIT(cinfo, JERR_BAD_J_COLORSPACE); + switch (cinfo->in_color_space) { + case JCS_RGB: + cconvert->pub.start_pass = rgb_ycc_start; + cconvert->pub.color_convert = rgb_ycc_convert; + break; + case JCS_YCbCr: + cconvert->pub.color_convert = null_convert; + break; + default: + ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); + } + break; + + case JCS_BG_YCC: + if (cinfo->num_components != 3) + ERREXIT(cinfo, JERR_BAD_J_COLORSPACE); + switch (cinfo->in_color_space) { + case JCS_RGB: + /* For conversion from normal RGB input to BG_YCC representation, + * the Cb/Cr values are first computed as usual, and then + * quantized further after DCT processing by a factor of + * 2 in reference to the nominal quantization factor. + */ + /* need quantization scale by factor of 2 after DCT */ + cinfo->comp_info[1].component_needed = TRUE; + cinfo->comp_info[2].component_needed = TRUE; + /* compute normal YCC first */ + cconvert->pub.start_pass = rgb_ycc_start; + cconvert->pub.color_convert = rgb_ycc_convert; + break; + case JCS_YCbCr: + /* need quantization scale by factor of 2 after DCT */ + cinfo->comp_info[1].component_needed = TRUE; + cinfo->comp_info[2].component_needed = TRUE; + /*FALLTHROUGH*/ + case JCS_BG_YCC: + /* Pass through for BG_YCC input */ + cconvert->pub.color_convert = null_convert; + break; + default: + ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); + } + break; + + case JCS_CMYK: + if (cinfo->num_components != 4) + ERREXIT(cinfo, JERR_BAD_J_COLORSPACE); + if (cinfo->in_color_space != JCS_CMYK) + ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); + cconvert->pub.color_convert = null_convert; + break; + + case JCS_YCCK: + if (cinfo->num_components != 4) + ERREXIT(cinfo, JERR_BAD_J_COLORSPACE); + switch (cinfo->in_color_space) { + case JCS_CMYK: + cconvert->pub.start_pass = rgb_ycc_start; + cconvert->pub.color_convert = cmyk_ycck_convert; + break; + case JCS_YCCK: + cconvert->pub.color_convert = null_convert; + break; + default: + ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); + } + break; + + default: /* allow null conversion of JCS_UNKNOWN */ + if (cinfo->jpeg_color_space != cinfo->in_color_space || + cinfo->num_components != cinfo->input_components) + ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); + cconvert->pub.color_convert = null_convert; + } +} diff --git a/cmake/externals/libjpeg/jcdctmgr.c b/cmake/externals/libjpeg/jcdctmgr.c new file mode 100644 index 000000000..a48ccd814 --- /dev/null +++ b/cmake/externals/libjpeg/jcdctmgr.c @@ -0,0 +1,466 @@ +/* + * jcdctmgr.c + * + * Copyright (C) 1994-1996, Thomas G. Lane. + * Modified 2003-2020 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains the forward-DCT management logic. + * This code selects a particular DCT implementation to be used, + * and it performs related housekeeping chores including coefficient + * quantization. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" +#include "jdct.h" /* Private declarations for DCT subsystem */ + + +/* Private subobject for this module */ + +typedef struct { + struct jpeg_forward_dct pub; /* public fields */ + + /* Pointer to the DCT routine actually in use */ + forward_DCT_method_ptr do_dct[MAX_COMPONENTS]; + +#ifdef DCT_FLOAT_SUPPORTED + /* Same as above for the floating-point case. */ + float_DCT_method_ptr do_float_dct[MAX_COMPONENTS]; +#endif +} my_fdct_controller; + +typedef my_fdct_controller * my_fdct_ptr; + + +/* The allocated post-DCT divisor tables -- big enough for any + * supported variant and not identical to the quant table entries, + * because of scaling (especially for an unnormalized DCT) -- + * are pointed to by dct_table in the per-component comp_info + * structures. Each table is given in normal array order. + */ + +typedef union { + DCTELEM int_array[DCTSIZE2]; +#ifdef DCT_FLOAT_SUPPORTED + FAST_FLOAT float_array[DCTSIZE2]; +#endif +} divisor_table; + + +/* The current scaled-DCT routines require ISLOW-style divisor tables, + * so be sure to compile that code if either ISLOW or SCALING is requested. + */ +#ifdef DCT_ISLOW_SUPPORTED +#define PROVIDE_ISLOW_TABLES +#else +#ifdef DCT_SCALING_SUPPORTED +#define PROVIDE_ISLOW_TABLES +#endif +#endif + + +/* + * Perform forward DCT on one or more blocks of a component. + * + * The input samples are taken from the sample_data[] array starting at + * position start_col, and moving to the right for any additional blocks. + * The quantized coefficients are returned in coef_blocks[]. + */ + +METHODDEF(void) +forward_DCT (j_compress_ptr cinfo, jpeg_component_info * compptr, + JSAMPARRAY sample_data, JBLOCKROW coef_blocks, + JDIMENSION start_col, JDIMENSION num_blocks) +/* This version is used for integer DCT implementations. */ +{ + /* This routine is heavily used, so it's worth coding it tightly. */ + my_fdct_ptr fdct = (my_fdct_ptr) cinfo->fdct; + forward_DCT_method_ptr do_dct = fdct->do_dct[compptr->component_index]; + DCTELEM * divisors = (DCTELEM *) compptr->dct_table; + DCTELEM workspace[DCTSIZE2]; /* work area for FDCT subroutine */ + JDIMENSION bi; + + for (bi = 0; bi < num_blocks; bi++, start_col += compptr->DCT_h_scaled_size) { + /* Perform the DCT */ + (*do_dct) (workspace, sample_data, start_col); + + /* Quantize/descale the coefficients, and store into coef_blocks[] */ + { register DCTELEM temp, qval; + register int i; + register JCOEFPTR output_ptr = coef_blocks[bi]; + + for (i = 0; i < DCTSIZE2; i++) { + qval = divisors[i]; + temp = workspace[i]; + /* Divide the coefficient value by qval, ensuring proper rounding. + * Since C does not specify the direction of rounding for negative + * quotients, we have to force the dividend positive for portability. + * + * In most files, at least half of the output values will be zero + * (at default quantization settings, more like three-quarters...) + * so we should ensure that this case is fast. On many machines, + * a comparison is enough cheaper than a divide to make a special test + * a win. Since both inputs will be nonnegative, we need only test + * for a < b to discover whether a/b is 0. + * If your machine's division is fast enough, define FAST_DIVIDE. + */ +#ifdef FAST_DIVIDE +#define DIVIDE_BY(a,b) a /= b +#else +#define DIVIDE_BY(a,b) if (a >= b) a /= b; else a = 0 +#endif + if (temp < 0) { + temp = -temp; + temp += qval>>1; /* for rounding */ + DIVIDE_BY(temp, qval); + temp = -temp; + } else { + temp += qval>>1; /* for rounding */ + DIVIDE_BY(temp, qval); + } + output_ptr[i] = (JCOEF) temp; + } + } + } +} + + +#ifdef DCT_FLOAT_SUPPORTED + +METHODDEF(void) +forward_DCT_float (j_compress_ptr cinfo, jpeg_component_info * compptr, + JSAMPARRAY sample_data, JBLOCKROW coef_blocks, + JDIMENSION start_col, JDIMENSION num_blocks) +/* This version is used for floating-point DCT implementations. */ +{ + /* This routine is heavily used, so it's worth coding it tightly. */ + my_fdct_ptr fdct = (my_fdct_ptr) cinfo->fdct; + float_DCT_method_ptr do_dct = fdct->do_float_dct[compptr->component_index]; + FAST_FLOAT * divisors = (FAST_FLOAT *) compptr->dct_table; + FAST_FLOAT workspace[DCTSIZE2]; /* work area for FDCT subroutine */ + JDIMENSION bi; + + for (bi = 0; bi < num_blocks; bi++, start_col += compptr->DCT_h_scaled_size) { + /* Perform the DCT */ + (*do_dct) (workspace, sample_data, start_col); + + /* Quantize/descale the coefficients, and store into coef_blocks[] */ + { register FAST_FLOAT temp; + register int i; + register JCOEFPTR output_ptr = coef_blocks[bi]; + + for (i = 0; i < DCTSIZE2; i++) { + /* Apply the quantization and scaling factor */ + temp = workspace[i] * divisors[i]; + /* Round to nearest integer. + * Since C does not specify the direction of rounding for negative + * quotients, we have to force the dividend positive for portability. + * The maximum coefficient size is +-16K (for 12-bit data), so this + * code should work for either 16-bit or 32-bit ints. + */ + output_ptr[i] = (JCOEF) ((int) (temp + (FAST_FLOAT) 16384.5) - 16384); + } + } + } +} + +#endif /* DCT_FLOAT_SUPPORTED */ + + +/* + * Initialize for a processing pass. + * Verify that all referenced Q-tables are present, and set up + * the divisor table for each one. + * In the current implementation, DCT of all components is done during + * the first pass, even if only some components will be output in the + * first scan. Hence all components should be examined here. + */ + +METHODDEF(void) +start_pass_fdctmgr (j_compress_ptr cinfo) +{ + my_fdct_ptr fdct = (my_fdct_ptr) cinfo->fdct; + int ci, qtblno, i; + jpeg_component_info *compptr; + int method = 0; + JQUANT_TBL * qtbl; + DCTELEM * dtbl; + + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + /* Select the proper DCT routine for this component's scaling */ + switch ((compptr->DCT_h_scaled_size << 8) + compptr->DCT_v_scaled_size) { +#ifdef DCT_SCALING_SUPPORTED + case ((1 << 8) + 1): + fdct->do_dct[ci] = jpeg_fdct_1x1; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((2 << 8) + 2): + fdct->do_dct[ci] = jpeg_fdct_2x2; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((3 << 8) + 3): + fdct->do_dct[ci] = jpeg_fdct_3x3; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((4 << 8) + 4): + fdct->do_dct[ci] = jpeg_fdct_4x4; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((5 << 8) + 5): + fdct->do_dct[ci] = jpeg_fdct_5x5; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((6 << 8) + 6): + fdct->do_dct[ci] = jpeg_fdct_6x6; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((7 << 8) + 7): + fdct->do_dct[ci] = jpeg_fdct_7x7; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((9 << 8) + 9): + fdct->do_dct[ci] = jpeg_fdct_9x9; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((10 << 8) + 10): + fdct->do_dct[ci] = jpeg_fdct_10x10; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((11 << 8) + 11): + fdct->do_dct[ci] = jpeg_fdct_11x11; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((12 << 8) + 12): + fdct->do_dct[ci] = jpeg_fdct_12x12; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((13 << 8) + 13): + fdct->do_dct[ci] = jpeg_fdct_13x13; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((14 << 8) + 14): + fdct->do_dct[ci] = jpeg_fdct_14x14; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((15 << 8) + 15): + fdct->do_dct[ci] = jpeg_fdct_15x15; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((16 << 8) + 16): + fdct->do_dct[ci] = jpeg_fdct_16x16; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((16 << 8) + 8): + fdct->do_dct[ci] = jpeg_fdct_16x8; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((14 << 8) + 7): + fdct->do_dct[ci] = jpeg_fdct_14x7; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((12 << 8) + 6): + fdct->do_dct[ci] = jpeg_fdct_12x6; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((10 << 8) + 5): + fdct->do_dct[ci] = jpeg_fdct_10x5; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((8 << 8) + 4): + fdct->do_dct[ci] = jpeg_fdct_8x4; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((6 << 8) + 3): + fdct->do_dct[ci] = jpeg_fdct_6x3; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((4 << 8) + 2): + fdct->do_dct[ci] = jpeg_fdct_4x2; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((2 << 8) + 1): + fdct->do_dct[ci] = jpeg_fdct_2x1; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((8 << 8) + 16): + fdct->do_dct[ci] = jpeg_fdct_8x16; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((7 << 8) + 14): + fdct->do_dct[ci] = jpeg_fdct_7x14; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((6 << 8) + 12): + fdct->do_dct[ci] = jpeg_fdct_6x12; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((5 << 8) + 10): + fdct->do_dct[ci] = jpeg_fdct_5x10; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((4 << 8) + 8): + fdct->do_dct[ci] = jpeg_fdct_4x8; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((3 << 8) + 6): + fdct->do_dct[ci] = jpeg_fdct_3x6; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((2 << 8) + 4): + fdct->do_dct[ci] = jpeg_fdct_2x4; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; + case ((1 << 8) + 2): + fdct->do_dct[ci] = jpeg_fdct_1x2; + method = JDCT_ISLOW; /* jfdctint uses islow-style table */ + break; +#endif + case ((DCTSIZE << 8) + DCTSIZE): + switch (cinfo->dct_method) { +#ifdef DCT_ISLOW_SUPPORTED + case JDCT_ISLOW: + fdct->do_dct[ci] = jpeg_fdct_islow; + method = JDCT_ISLOW; + break; +#endif +#ifdef DCT_IFAST_SUPPORTED + case JDCT_IFAST: + fdct->do_dct[ci] = jpeg_fdct_ifast; + method = JDCT_IFAST; + break; +#endif +#ifdef DCT_FLOAT_SUPPORTED + case JDCT_FLOAT: + fdct->do_float_dct[ci] = jpeg_fdct_float; + method = JDCT_FLOAT; + break; +#endif + default: + ERREXIT(cinfo, JERR_NOT_COMPILED); + } + break; + default: + ERREXIT2(cinfo, JERR_BAD_DCTSIZE, + compptr->DCT_h_scaled_size, compptr->DCT_v_scaled_size); + } + qtblno = compptr->quant_tbl_no; + /* Make sure specified quantization table is present */ + if (qtblno < 0 || qtblno >= NUM_QUANT_TBLS || + cinfo->quant_tbl_ptrs[qtblno] == NULL) + ERREXIT1(cinfo, JERR_NO_QUANT_TABLE, qtblno); + qtbl = cinfo->quant_tbl_ptrs[qtblno]; + /* Create divisor table from quant table */ + switch (method) { +#ifdef PROVIDE_ISLOW_TABLES + case JDCT_ISLOW: + /* For LL&M IDCT method, divisors are equal to raw quantization + * coefficients multiplied by 8 (to counteract scaling). + */ + dtbl = (DCTELEM *) compptr->dct_table; + for (i = 0; i < DCTSIZE2; i++) { + dtbl[i] = + ((DCTELEM) qtbl->quantval[i]) << (compptr->component_needed ? 4 : 3); + } + fdct->pub.forward_DCT[ci] = forward_DCT; + break; +#endif +#ifdef DCT_IFAST_SUPPORTED + case JDCT_IFAST: + { + /* For AA&N IDCT method, divisors are equal to quantization + * coefficients scaled by scalefactor[row]*scalefactor[col], where + * scalefactor[0] = 1 + * scalefactor[k] = cos(k*PI/16) * sqrt(2) for k=1..7 + * We apply a further scale factor of 8. + */ +#define CONST_BITS 14 + static const INT16 aanscales[DCTSIZE2] = { + /* precomputed values scaled up by 14 bits */ + 16384, 22725, 21407, 19266, 16384, 12873, 8867, 4520, + 22725, 31521, 29692, 26722, 22725, 17855, 12299, 6270, + 21407, 29692, 27969, 25172, 21407, 16819, 11585, 5906, + 19266, 26722, 25172, 22654, 19266, 15137, 10426, 5315, + 16384, 22725, 21407, 19266, 16384, 12873, 8867, 4520, + 12873, 17855, 16819, 15137, 12873, 10114, 6967, 3552, + 8867, 12299, 11585, 10426, 8867, 6967, 4799, 2446, + 4520, 6270, 5906, 5315, 4520, 3552, 2446, 1247 + }; + SHIFT_TEMPS + + dtbl = (DCTELEM *) compptr->dct_table; + for (i = 0; i < DCTSIZE2; i++) { + dtbl[i] = (DCTELEM) + DESCALE(MULTIPLY16V16((INT32) qtbl->quantval[i], + (INT32) aanscales[i]), + compptr->component_needed ? CONST_BITS-4 : CONST_BITS-3); + } + } + fdct->pub.forward_DCT[ci] = forward_DCT; + break; +#endif +#ifdef DCT_FLOAT_SUPPORTED + case JDCT_FLOAT: + { + /* For float AA&N IDCT method, divisors are equal to quantization + * coefficients scaled by scalefactor[row]*scalefactor[col], where + * scalefactor[0] = 1 + * scalefactor[k] = cos(k*PI/16) * sqrt(2) for k=1..7 + * We apply a further scale factor of 8. + * What's actually stored is 1/divisor so that the inner loop can + * use a multiplication rather than a division. + */ + FAST_FLOAT * fdtbl = (FAST_FLOAT *) compptr->dct_table; + int row, col; + static const double aanscalefactor[DCTSIZE] = { + 1.0, 1.387039845, 1.306562965, 1.175875602, + 1.0, 0.785694958, 0.541196100, 0.275899379 + }; + + i = 0; + for (row = 0; row < DCTSIZE; row++) { + for (col = 0; col < DCTSIZE; col++) { + fdtbl[i] = (FAST_FLOAT) + (1.0 / ((double) qtbl->quantval[i] * + aanscalefactor[row] * aanscalefactor[col] * + (compptr->component_needed ? 16.0 : 8.0))); + i++; + } + } + } + fdct->pub.forward_DCT[ci] = forward_DCT_float; + break; +#endif + default: + ERREXIT(cinfo, JERR_NOT_COMPILED); + } + } +} + + +/* + * Initialize FDCT manager. + */ + +GLOBAL(void) +jinit_forward_dct (j_compress_ptr cinfo) +{ + my_fdct_ptr fdct; + int ci; + jpeg_component_info *compptr; + + fdct = (my_fdct_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_fdct_controller)); + cinfo->fdct = &fdct->pub; + fdct->pub.start_pass = start_pass_fdctmgr; + + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + /* Allocate a divisor table for each component */ + compptr->dct_table = (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(divisor_table)); + } +} diff --git a/cmake/externals/libjpeg/jchuff.c b/cmake/externals/libjpeg/jchuff.c new file mode 100644 index 000000000..1f527b218 --- /dev/null +++ b/cmake/externals/libjpeg/jchuff.c @@ -0,0 +1,1656 @@ +/* + * jchuff.c + * + * Copyright (C) 1991-1997, Thomas G. Lane. + * Modified 2006-2023 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains Huffman entropy encoding routines. + * Both sequential and progressive modes are supported in this single module. + * + * Much of the complexity here has to do with supporting output suspension. + * If the data destination module demands suspension, we want to be able to + * back up to the start of the current MCU. To do this, we copy state + * variables into local working storage, and update them back to the + * permanent JPEG objects only upon successful completion of an MCU. + * + * We do not support output suspension for the progressive JPEG mode, since + * the library currently does not allow multiple-scan files to be written + * with output suspension. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* The legal range of a DCT coefficient is + * -1024 .. +1023 for 8-bit sample data precision; + * -16384 .. +16383 for 12-bit sample data precision. + * Hence the magnitude should always fit in sample data precision + 2 bits. + */ + +/* Derived data constructed for each Huffman table */ + +typedef struct { + unsigned int ehufco[256]; /* code for each symbol */ + char ehufsi[256]; /* length of code for each symbol */ + /* If no code has been allocated for a symbol S, ehufsi[S] contains 0 */ +} c_derived_tbl; + + +/* Expanded entropy encoder object for Huffman encoding. + * + * The savable_state subrecord contains fields that change within an MCU, + * but must not be updated permanently until we complete the MCU. + */ + +typedef struct { + INT32 put_buffer; /* current bit-accumulation buffer */ + int put_bits; /* # of bits now in it */ + int last_dc_val[MAX_COMPS_IN_SCAN]; /* last DC coef for each component */ +} savable_state; + +/* This macro is to work around compilers with missing or broken + * structure assignment. You'll need to fix this code if you have + * such a compiler and you change MAX_COMPS_IN_SCAN. + */ + +#ifndef NO_STRUCT_ASSIGN +#define ASSIGN_STATE(dest,src) ((dest) = (src)) +#else +#if MAX_COMPS_IN_SCAN == 4 +#define ASSIGN_STATE(dest,src) \ + ((dest).put_buffer = (src).put_buffer, \ + (dest).put_bits = (src).put_bits, \ + (dest).last_dc_val[0] = (src).last_dc_val[0], \ + (dest).last_dc_val[1] = (src).last_dc_val[1], \ + (dest).last_dc_val[2] = (src).last_dc_val[2], \ + (dest).last_dc_val[3] = (src).last_dc_val[3]) +#endif +#endif + + +typedef struct { + struct jpeg_entropy_encoder pub; /* public fields */ + + savable_state saved; /* Bit buffer & DC state at start of MCU */ + + /* These fields are NOT loaded into local working state. */ + unsigned int restarts_to_go; /* MCUs left in this restart interval */ + int next_restart_num; /* next restart number to write (0-7) */ + + /* Pointers to derived tables (these workspaces have image lifespan) */ + c_derived_tbl * dc_derived_tbls[NUM_HUFF_TBLS]; + c_derived_tbl * ac_derived_tbls[NUM_HUFF_TBLS]; + + /* Statistics tables for optimization */ + long * dc_count_ptrs[NUM_HUFF_TBLS]; + long * ac_count_ptrs[NUM_HUFF_TBLS]; + + /* Following fields used only in progressive mode */ + + /* Mode flag: TRUE for optimization, FALSE for actual data output */ + boolean gather_statistics; + + /* next_output_byte/free_in_buffer are local copies of cinfo->dest fields. + */ + JOCTET * next_output_byte; /* => next byte to write in buffer */ + size_t free_in_buffer; /* # of byte spaces remaining in buffer */ + j_compress_ptr cinfo; /* link to cinfo (needed for dump_buffer) */ + + /* Coding status for AC components */ + int ac_tbl_no; /* the table number of the single component */ + unsigned int EOBRUN; /* run length of EOBs */ + unsigned int BE; /* # of buffered correction bits before MCU */ + char * bit_buffer; /* buffer for correction bits (1 per char) */ + /* packing correction bits tightly would save some space but cost time... */ +} huff_entropy_encoder; + +typedef huff_entropy_encoder * huff_entropy_ptr; + +/* Working state while writing an MCU (sequential mode). + * This struct contains all the fields that are needed by subroutines. + */ + +typedef struct { + JOCTET * next_output_byte; /* => next byte to write in buffer */ + size_t free_in_buffer; /* # of byte spaces remaining in buffer */ + savable_state cur; /* Current bit buffer & DC state */ + j_compress_ptr cinfo; /* dump_buffer needs access to this */ +} working_state; + +/* MAX_CORR_BITS is the number of bits the AC refinement correction-bit + * buffer can hold. Larger sizes may slightly improve compression, but + * 1000 is already well into the realm of overkill. + * The minimum safe size is 64 bits. + */ + +#define MAX_CORR_BITS 1000 /* Max # of correction bits I can buffer */ + +/* IRIGHT_SHIFT is like RIGHT_SHIFT, but works on int rather than INT32. + * We assume that int right shift is unsigned if INT32 right shift is, + * which should be safe. + */ + +#ifdef RIGHT_SHIFT_IS_UNSIGNED +#define ISHIFT_TEMPS int ishift_temp; +#define IRIGHT_SHIFT(x,shft) \ + ((ishift_temp = (x)) < 0 ? \ + (ishift_temp >> (shft)) | ((~0) << (16-(shft))) : \ + (ishift_temp >> (shft))) +#else +#define ISHIFT_TEMPS +#define IRIGHT_SHIFT(x,shft) ((x) >> (shft)) +#endif + + +/* + * Compute the derived values for a Huffman table. + * This routine also performs some validation checks on the table. + */ + +LOCAL(void) +jpeg_make_c_derived_tbl (j_compress_ptr cinfo, boolean isDC, int tblno, + c_derived_tbl ** pdtbl) +{ + JHUFF_TBL *htbl; + c_derived_tbl *dtbl; + int p, i, l, lastp, si, maxsymbol; + char huffsize[257]; + unsigned int huffcode[257]; + unsigned int code; + + /* Note that huffsize[] and huffcode[] are filled in code-length order, + * paralleling the order of the symbols themselves in htbl->huffval[]. + */ + + /* Find the input Huffman table */ + if (tblno < 0 || tblno >= NUM_HUFF_TBLS) + ERREXIT1(cinfo, JERR_NO_HUFF_TABLE, tblno); + htbl = + isDC ? cinfo->dc_huff_tbl_ptrs[tblno] : cinfo->ac_huff_tbl_ptrs[tblno]; + if (htbl == NULL) + htbl = jpeg_std_huff_table((j_common_ptr) cinfo, isDC, tblno); + + /* Allocate a workspace if we haven't already done so. */ + if (*pdtbl == NULL) + *pdtbl = (c_derived_tbl *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(c_derived_tbl)); + dtbl = *pdtbl; + + /* Figure C.1: make table of Huffman code length for each symbol */ + + p = 0; + for (l = 1; l <= 16; l++) { + i = (int) htbl->bits[l]; + if (i < 0 || p + i > 256) /* protect against table overrun */ + ERREXIT(cinfo, JERR_BAD_HUFF_TABLE); + while (i--) + huffsize[p++] = (char) l; + } + huffsize[p] = 0; + lastp = p; + + /* Figure C.2: generate the codes themselves */ + /* We also validate that the counts represent a legal Huffman code tree. */ + + code = 0; + si = huffsize[0]; + p = 0; + while (huffsize[p]) { + while (((int) huffsize[p]) == si) { + huffcode[p++] = code; + code++; + } + /* code is now 1 more than the last code used for codelength si; but + * it must still fit in si bits, since no code is allowed to be all ones. + */ + if (((INT32) code) >= (((INT32) 1) << si)) + ERREXIT(cinfo, JERR_BAD_HUFF_TABLE); + code <<= 1; + si++; + } + + /* Figure C.3: generate encoding tables */ + /* These are code and size indexed by symbol value */ + + /* Set all codeless symbols to have code length 0; + * this lets us detect duplicate VAL entries here, and later + * allows emit_bits to detect any attempt to emit such symbols. + */ + MEMZERO(dtbl->ehufsi, SIZEOF(dtbl->ehufsi)); + + /* This is also a convenient place to check for out-of-range + * and duplicated VAL entries. We allow 0..255 for AC symbols + * but only 0..15 for DC. (We could constrain them further + * based on data depth and mode, but this seems enough.) + */ + maxsymbol = isDC ? 15 : 255; + + for (p = 0; p < lastp; p++) { + i = htbl->huffval[p]; + if (i < 0 || i > maxsymbol || dtbl->ehufsi[i]) + ERREXIT(cinfo, JERR_BAD_HUFF_TABLE); + dtbl->ehufco[i] = huffcode[p]; + dtbl->ehufsi[i] = huffsize[p]; + } +} + + +/* Outputting bytes to the file. + * NB: these must be called only when actually outputting, + * that is, entropy->gather_statistics == FALSE. + */ + +/* Emit a byte, taking 'action' if must suspend. */ +#define emit_byte_s(state,val,action) \ + { *(state)->next_output_byte++ = (JOCTET) (val); \ + if (--(state)->free_in_buffer == 0) \ + if (! dump_buffer_s(state)) \ + { action; } } + +/* Emit a byte */ +#define emit_byte_e(entropy,val) \ + { *(entropy)->next_output_byte++ = (JOCTET) (val); \ + if (--(entropy)->free_in_buffer == 0) \ + dump_buffer_e(entropy); } + + +LOCAL(boolean) +dump_buffer_s (working_state * state) +/* Empty the output buffer; return TRUE if successful, FALSE if must suspend */ +{ + struct jpeg_destination_mgr * dest = state->cinfo->dest; + + if (! (*dest->empty_output_buffer) (state->cinfo)) + return FALSE; + /* After a successful buffer dump, must reset buffer pointers */ + state->next_output_byte = dest->next_output_byte; + state->free_in_buffer = dest->free_in_buffer; + return TRUE; +} + + +LOCAL(void) +dump_buffer_e (huff_entropy_ptr entropy) +/* Empty the output buffer; we do not support suspension in this case. */ +{ + struct jpeg_destination_mgr * dest = entropy->cinfo->dest; + + if (! (*dest->empty_output_buffer) (entropy->cinfo)) + ERREXIT(entropy->cinfo, JERR_CANT_SUSPEND); + /* After a successful buffer dump, must reset buffer pointers */ + entropy->next_output_byte = dest->next_output_byte; + entropy->free_in_buffer = dest->free_in_buffer; +} + + +/* Outputting bits to the file */ + +/* Only the right 24 bits of put_buffer are used; the valid bits are + * left-justified in this part. At most 16 bits can be passed to emit_bits + * in one call, and we never retain more than 7 bits in put_buffer + * between calls, so 24 bits are sufficient. + */ + +INLINE +LOCAL(boolean) +emit_bits_s (working_state * state, unsigned int code, int size) +/* Emit some bits; return TRUE if successful, FALSE if must suspend */ +{ + /* This routine is heavily used, so it's worth coding tightly. */ + register INT32 put_buffer; + register int put_bits; + + /* if size is 0, caller used an invalid Huffman table entry */ + if (size == 0) + ERREXIT(state->cinfo, JERR_HUFF_MISSING_CODE); + + /* mask off any extra bits in code */ + put_buffer = ((INT32) code) & ((((INT32) 1) << size) - 1); + + /* new number of bits in buffer */ + put_bits = size + state->cur.put_bits; + + put_buffer <<= 24 - put_bits; /* align incoming bits */ + + /* and merge with old buffer contents */ + put_buffer |= state->cur.put_buffer; + + while (put_bits >= 8) { + int c = (int) ((put_buffer >> 16) & 0xFF); + + emit_byte_s(state, c, return FALSE); + if (c == 0xFF) { /* need to stuff a zero byte? */ + emit_byte_s(state, 0, return FALSE); + } + put_buffer <<= 8; + put_bits -= 8; + } + + state->cur.put_buffer = put_buffer; /* update state variables */ + state->cur.put_bits = put_bits; + + return TRUE; +} + + +INLINE +LOCAL(void) +emit_bits_e (huff_entropy_ptr entropy, unsigned int code, int size) +/* Emit some bits, unless we are in gather mode */ +{ + /* This routine is heavily used, so it's worth coding tightly. */ + register INT32 put_buffer; + register int put_bits; + + /* if size is 0, caller used an invalid Huffman table entry */ + if (size == 0) + ERREXIT(entropy->cinfo, JERR_HUFF_MISSING_CODE); + + if (entropy->gather_statistics) + return; /* do nothing if we're only getting stats */ + + /* mask off any extra bits in code */ + put_buffer = ((INT32) code) & ((((INT32) 1) << size) - 1); + + /* new number of bits in buffer */ + put_bits = size + entropy->saved.put_bits; + + put_buffer <<= 24 - put_bits; /* align incoming bits */ + + /* and merge with old buffer contents */ + put_buffer |= entropy->saved.put_buffer; + + while (put_bits >= 8) { + int c = (int) ((put_buffer >> 16) & 0xFF); + + emit_byte_e(entropy, c); + if (c == 0xFF) { /* need to stuff a zero byte? */ + emit_byte_e(entropy, 0); + } + put_buffer <<= 8; + put_bits -= 8; + } + + entropy->saved.put_buffer = put_buffer; /* update variables */ + entropy->saved.put_bits = put_bits; +} + + +LOCAL(boolean) +flush_bits_s (working_state * state) +{ + if (! emit_bits_s(state, 0x7F, 7)) /* fill any partial byte with ones */ + return FALSE; + state->cur.put_buffer = 0; /* and reset bit-buffer to empty */ + state->cur.put_bits = 0; + return TRUE; +} + + +LOCAL(void) +flush_bits_e (huff_entropy_ptr entropy) +{ + emit_bits_e(entropy, 0x7F, 7); /* fill any partial byte with ones */ + entropy->saved.put_buffer = 0; /* and reset bit-buffer to empty */ + entropy->saved.put_bits = 0; +} + + +/* + * Emit (or just count) a Huffman symbol. + */ + +INLINE +LOCAL(void) +emit_dc_symbol (huff_entropy_ptr entropy, int tbl_no, int symbol) +{ + if (entropy->gather_statistics) + entropy->dc_count_ptrs[tbl_no][symbol]++; + else { + c_derived_tbl * tbl = entropy->dc_derived_tbls[tbl_no]; + emit_bits_e(entropy, tbl->ehufco[symbol], tbl->ehufsi[symbol]); + } +} + + +INLINE +LOCAL(void) +emit_ac_symbol (huff_entropy_ptr entropy, int tbl_no, int symbol) +{ + if (entropy->gather_statistics) + entropy->ac_count_ptrs[tbl_no][symbol]++; + else { + c_derived_tbl * tbl = entropy->ac_derived_tbls[tbl_no]; + emit_bits_e(entropy, tbl->ehufco[symbol], tbl->ehufsi[symbol]); + } +} + + +/* + * Emit bits from a correction bit buffer. + */ + +LOCAL(void) +emit_buffered_bits (huff_entropy_ptr entropy, char * bufstart, + unsigned int nbits) +{ + if (entropy->gather_statistics) + return; /* no real work */ + + while (nbits > 0) { + emit_bits_e(entropy, (unsigned int) (*bufstart), 1); + bufstart++; + nbits--; + } +} + + +/* + * Emit any pending EOBRUN symbol. + */ + +LOCAL(void) +emit_eobrun (huff_entropy_ptr entropy) +{ + register int temp, nbits; + + if (entropy->EOBRUN > 0) { /* if there is any pending EOBRUN */ + temp = entropy->EOBRUN; + nbits = 0; + while ((temp >>= 1)) + nbits++; + /* safety check: shouldn't happen given limited correction-bit buffer */ + if (nbits > 14) + ERREXIT(entropy->cinfo, JERR_HUFF_MISSING_CODE); + + emit_ac_symbol(entropy, entropy->ac_tbl_no, nbits << 4); + if (nbits) + emit_bits_e(entropy, entropy->EOBRUN, nbits); + + entropy->EOBRUN = 0; + + /* Emit any buffered correction bits */ + emit_buffered_bits(entropy, entropy->bit_buffer, entropy->BE); + entropy->BE = 0; + } +} + + +/* + * Emit a restart marker & resynchronize predictions. + */ + +LOCAL(boolean) +emit_restart_s (working_state * state, int restart_num) +{ + int ci; + + if (! flush_bits_s(state)) + return FALSE; + + emit_byte_s(state, 0xFF, return FALSE); + emit_byte_s(state, JPEG_RST0 + restart_num, return FALSE); + + /* Re-initialize DC predictions to 0 */ + for (ci = 0; ci < state->cinfo->comps_in_scan; ci++) + state->cur.last_dc_val[ci] = 0; + + /* The restart counter is not updated until we successfully write the MCU. */ + + return TRUE; +} + + +LOCAL(void) +emit_restart_e (huff_entropy_ptr entropy, int restart_num) +{ + int ci; + + emit_eobrun(entropy); + + if (! entropy->gather_statistics) { + flush_bits_e(entropy); + emit_byte_e(entropy, 0xFF); + emit_byte_e(entropy, JPEG_RST0 + restart_num); + } + + if (entropy->cinfo->Ss == 0) { + /* Re-initialize DC predictions to 0 */ + for (ci = 0; ci < entropy->cinfo->comps_in_scan; ci++) + entropy->saved.last_dc_val[ci] = 0; + } else { + /* Re-initialize all AC-related fields to 0 */ + entropy->EOBRUN = 0; + entropy->BE = 0; + } +} + + +/* + * MCU encoding for DC initial scan (either spectral selection, + * or first pass of successive approximation). + */ + +METHODDEF(boolean) +encode_mcu_DC_first (j_compress_ptr cinfo, JBLOCKARRAY MCU_data) +{ + huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; + register int temp, temp2; + register int nbits; + int max_coef_bits; + int blkn, ci, tbl; + ISHIFT_TEMPS + + entropy->next_output_byte = cinfo->dest->next_output_byte; + entropy->free_in_buffer = cinfo->dest->free_in_buffer; + + /* Emit restart marker if needed */ + if (cinfo->restart_interval) + if (entropy->restarts_to_go == 0) + emit_restart_e(entropy, entropy->next_restart_num); + + /* Since we're encoding a difference, the range limit is twice as much. */ + max_coef_bits = cinfo->data_precision + 3; + + /* Encode the MCU data blocks */ + for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { + ci = cinfo->MCU_membership[blkn]; + tbl = cinfo->cur_comp_info[ci]->dc_tbl_no; + + /* Compute the DC value after the required point transform by Al. + * This is simply an arithmetic right shift. + */ + temp = IRIGHT_SHIFT((int) (MCU_data[blkn][0][0]), cinfo->Al); + + /* DC differences are figured on the point-transformed values. */ + if ((temp2 = temp - entropy->saved.last_dc_val[ci]) == 0) { + /* Count/emit the Huffman-coded symbol for the number of bits */ + emit_dc_symbol(entropy, tbl, 0); + + continue; + } + + entropy->saved.last_dc_val[ci] = temp; + + /* Encode the DC coefficient difference per section G.1.2.1 */ + if ((temp = temp2) < 0) { + temp = -temp; /* temp is abs value of input */ + /* For a negative input, want temp2 = bitwise complement of abs(input) */ + /* This code assumes we are on a two's complement machine */ + temp2--; + } + + /* Find the number of bits needed for the magnitude of the coefficient */ + nbits = 0; + do nbits++; /* there must be at least one 1 bit */ + while ((temp >>= 1)); + /* Check for out-of-range coefficient values */ + if (nbits > max_coef_bits) + ERREXIT(cinfo, JERR_BAD_DCT_COEF); + + /* Count/emit the Huffman-coded symbol for the number of bits */ + emit_dc_symbol(entropy, tbl, nbits); + + /* Emit that number of bits of the value, if positive, */ + /* or the complement of its magnitude, if negative. */ + emit_bits_e(entropy, (unsigned int) temp2, nbits); + } + + cinfo->dest->next_output_byte = entropy->next_output_byte; + cinfo->dest->free_in_buffer = entropy->free_in_buffer; + + /* Update restart-interval state too */ + if (cinfo->restart_interval) { + if (entropy->restarts_to_go == 0) { + entropy->restarts_to_go = cinfo->restart_interval; + entropy->next_restart_num++; + entropy->next_restart_num &= 7; + } + entropy->restarts_to_go--; + } + + return TRUE; +} + + +/* + * MCU encoding for AC initial scan (either spectral selection, + * or first pass of successive approximation). + */ + +METHODDEF(boolean) +encode_mcu_AC_first (j_compress_ptr cinfo, JBLOCKARRAY MCU_data) +{ + huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; + const int * natural_order; + JBLOCKROW block; + register int temp, temp2; + register int nbits; + register int r, k; + int Se, Al, max_coef_bits; + + entropy->next_output_byte = cinfo->dest->next_output_byte; + entropy->free_in_buffer = cinfo->dest->free_in_buffer; + + /* Emit restart marker if needed */ + if (cinfo->restart_interval) + if (entropy->restarts_to_go == 0) + emit_restart_e(entropy, entropy->next_restart_num); + + Se = cinfo->Se; + Al = cinfo->Al; + natural_order = cinfo->natural_order; + max_coef_bits = cinfo->data_precision + 2; + + /* Encode the MCU data block */ + block = MCU_data[0]; + + /* Encode the AC coefficients per section G.1.2.2, fig. G.3 */ + + r = 0; /* r = run length of zeros */ + + for (k = cinfo->Ss; k <= Se; k++) { + if ((temp = (*block)[natural_order[k]]) == 0) { + r++; + continue; + } + /* We must apply the point transform by Al. For AC coefficients this + * is an integer division with rounding towards 0. To do this portably + * in C, we shift after obtaining the absolute value; so the code is + * interwoven with finding the abs value (temp) and output bits (temp2). + */ + if (temp < 0) { + temp = -temp; /* temp is abs value of input */ + /* Apply the point transform, and watch out for case */ + /* that nonzero coef is zero after point transform. */ + if ((temp >>= Al) == 0) { + r++; + continue; + } + /* For a negative coef, want temp2 = bitwise complement of abs(coef) */ + temp2 = ~temp; + } else { + /* Apply the point transform, and watch out for case */ + /* that nonzero coef is zero after point transform. */ + if ((temp >>= Al) == 0) { + r++; + continue; + } + temp2 = temp; + } + + /* Emit any pending EOBRUN */ + if (entropy->EOBRUN > 0) + emit_eobrun(entropy); + /* if run length > 15, must emit special run-length-16 codes (0xF0) */ + while (r > 15) { + emit_ac_symbol(entropy, entropy->ac_tbl_no, 0xF0); + r -= 16; + } + + /* Find the number of bits needed for the magnitude of the coefficient */ + nbits = 0; + do nbits++; /* there must be at least one 1 bit */ + while ((temp >>= 1)); + /* Check for out-of-range coefficient values */ + if (nbits > max_coef_bits) + ERREXIT(cinfo, JERR_BAD_DCT_COEF); + + /* Count/emit Huffman symbol for run length / number of bits */ + emit_ac_symbol(entropy, entropy->ac_tbl_no, (r << 4) + nbits); + + /* Emit that number of bits of the value, if positive, */ + /* or the complement of its magnitude, if negative. */ + emit_bits_e(entropy, (unsigned int) temp2, nbits); + + r = 0; /* reset zero run length */ + } + + if (r > 0) { /* If there are trailing zeroes, */ + entropy->EOBRUN++; /* count an EOB */ + if (entropy->EOBRUN == 0x7FFF) + emit_eobrun(entropy); /* force it out to avoid overflow */ + } + + cinfo->dest->next_output_byte = entropy->next_output_byte; + cinfo->dest->free_in_buffer = entropy->free_in_buffer; + + /* Update restart-interval state too */ + if (cinfo->restart_interval) { + if (entropy->restarts_to_go == 0) { + entropy->restarts_to_go = cinfo->restart_interval; + entropy->next_restart_num++; + entropy->next_restart_num &= 7; + } + entropy->restarts_to_go--; + } + + return TRUE; +} + + +/* + * MCU encoding for DC successive approximation refinement scan. + * Note: we assume such scans can be multi-component, + * although the spec is not very clear on the point. + */ + +METHODDEF(boolean) +encode_mcu_DC_refine (j_compress_ptr cinfo, JBLOCKARRAY MCU_data) +{ + huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; + int Al, blkn; + + entropy->next_output_byte = cinfo->dest->next_output_byte; + entropy->free_in_buffer = cinfo->dest->free_in_buffer; + + /* Emit restart marker if needed */ + if (cinfo->restart_interval) + if (entropy->restarts_to_go == 0) + emit_restart_e(entropy, entropy->next_restart_num); + + Al = cinfo->Al; + + /* Encode the MCU data blocks */ + for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { + /* We simply emit the Al'th bit of the DC coefficient value. */ + emit_bits_e(entropy, (unsigned int) (MCU_data[blkn][0][0] >> Al), 1); + } + + cinfo->dest->next_output_byte = entropy->next_output_byte; + cinfo->dest->free_in_buffer = entropy->free_in_buffer; + + /* Update restart-interval state too */ + if (cinfo->restart_interval) { + if (entropy->restarts_to_go == 0) { + entropy->restarts_to_go = cinfo->restart_interval; + entropy->next_restart_num++; + entropy->next_restart_num &= 7; + } + entropy->restarts_to_go--; + } + + return TRUE; +} + + +/* + * MCU encoding for AC successive approximation refinement scan. + */ + +METHODDEF(boolean) +encode_mcu_AC_refine (j_compress_ptr cinfo, JBLOCKARRAY MCU_data) +{ + huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; + const int * natural_order; + JBLOCKROW block; + register int temp; + register int r, k; + int Se, Al; + int EOB; + char *BR_buffer; + unsigned int BR; + int absvalues[DCTSIZE2]; + + entropy->next_output_byte = cinfo->dest->next_output_byte; + entropy->free_in_buffer = cinfo->dest->free_in_buffer; + + /* Emit restart marker if needed */ + if (cinfo->restart_interval) + if (entropy->restarts_to_go == 0) + emit_restart_e(entropy, entropy->next_restart_num); + + Se = cinfo->Se; + Al = cinfo->Al; + natural_order = cinfo->natural_order; + + /* Encode the MCU data block */ + block = MCU_data[0]; + + /* It is convenient to make a pre-pass to determine the transformed + * coefficients' absolute values and the EOB position. + */ + EOB = 0; + for (k = cinfo->Ss; k <= Se; k++) { + temp = (*block)[natural_order[k]]; + /* We must apply the point transform by Al. For AC coefficients this + * is an integer division with rounding towards 0. To do this portably + * in C, we shift after obtaining the absolute value. + */ + if (temp < 0) + temp = -temp; /* temp is abs value of input */ + temp >>= Al; /* apply the point transform */ + absvalues[k] = temp; /* save abs value for main pass */ + if (temp == 1) + EOB = k; /* EOB = index of last newly-nonzero coef */ + } + + /* Encode the AC coefficients per section G.1.2.3, fig. G.7 */ + + r = 0; /* r = run length of zeros */ + BR = 0; /* BR = count of buffered bits added now */ + BR_buffer = entropy->bit_buffer + entropy->BE; /* Append bits to buffer */ + + for (k = cinfo->Ss; k <= Se; k++) { + if ((temp = absvalues[k]) == 0) { + r++; + continue; + } + + /* Emit any required ZRLs, but not if they can be folded into EOB */ + while (r > 15 && k <= EOB) { + /* emit any pending EOBRUN and the BE correction bits */ + emit_eobrun(entropy); + /* Emit ZRL */ + emit_ac_symbol(entropy, entropy->ac_tbl_no, 0xF0); + r -= 16; + /* Emit buffered correction bits that must be associated with ZRL */ + emit_buffered_bits(entropy, BR_buffer, BR); + BR_buffer = entropy->bit_buffer; /* BE bits are gone now */ + BR = 0; + } + + /* If the coef was previously nonzero, it only needs a correction bit. + * NOTE: a straight translation of the spec's figure G.7 would suggest + * that we also need to test r > 15. But if r > 15, we can only get here + * if k > EOB, which implies that this coefficient is not 1. + */ + if (temp > 1) { + /* The correction bit is the next bit of the absolute value. */ + BR_buffer[BR++] = (char) (temp & 1); + continue; + } + + /* Emit any pending EOBRUN and the BE correction bits */ + emit_eobrun(entropy); + + /* Count/emit Huffman symbol for run length / number of bits */ + emit_ac_symbol(entropy, entropy->ac_tbl_no, (r << 4) + 1); + + /* Emit output bit for newly-nonzero coef */ + temp = ((*block)[natural_order[k]] < 0) ? 0 : 1; + emit_bits_e(entropy, (unsigned int) temp, 1); + + /* Emit buffered correction bits that must be associated with this code */ + emit_buffered_bits(entropy, BR_buffer, BR); + BR_buffer = entropy->bit_buffer; /* BE bits are gone now */ + BR = 0; + r = 0; /* reset zero run length */ + } + + if (r > 0 || BR > 0) { /* If there are trailing zeroes, */ + entropy->EOBRUN++; /* count an EOB */ + entropy->BE += BR; /* concat my correction bits to older ones */ + /* We force out the EOB if we risk either: + * 1. overflow of the EOB counter; + * 2. overflow of the correction bit buffer during the next MCU. + */ + if (entropy->EOBRUN == 0x7FFF || entropy->BE > (MAX_CORR_BITS-DCTSIZE2+1)) + emit_eobrun(entropy); + } + + cinfo->dest->next_output_byte = entropy->next_output_byte; + cinfo->dest->free_in_buffer = entropy->free_in_buffer; + + /* Update restart-interval state too */ + if (cinfo->restart_interval) { + if (entropy->restarts_to_go == 0) { + entropy->restarts_to_go = cinfo->restart_interval; + entropy->next_restart_num++; + entropy->next_restart_num &= 7; + } + entropy->restarts_to_go--; + } + + return TRUE; +} + + +/* Encode a single block's worth of coefficients */ + +LOCAL(boolean) +encode_one_block (working_state * state, JCOEFPTR block, int last_dc_val, + c_derived_tbl *dctbl, c_derived_tbl *actbl) +{ + register int temp, temp2; + register int nbits; + register int r, k; + int Se = state->cinfo->lim_Se; + int max_coef_bits = state->cinfo->data_precision + 3; + const int * natural_order = state->cinfo->natural_order; + + /* Encode the DC coefficient difference per section F.1.2.1 */ + + if ((temp = block[0] - last_dc_val) == 0) { + /* Emit the Huffman-coded symbol for the number of bits */ + if (! emit_bits_s(state, dctbl->ehufco[0], dctbl->ehufsi[0])) + return FALSE; + } else { + if ((temp2 = temp) < 0) { + temp = -temp; /* temp is abs value of input */ + /* For a negative input, want temp2 = bitwise complement of abs(input) */ + /* This code assumes we are on a two's complement machine */ + temp2--; + } + + /* Find the number of bits needed for the magnitude of the coefficient */ + nbits = 0; + do nbits++; /* there must be at least one 1 bit */ + while ((temp >>= 1)); + /* Check for out-of-range coefficient values. + * Since we're encoding a difference, the range limit is twice as much. + */ + if (nbits > max_coef_bits) + ERREXIT(state->cinfo, JERR_BAD_DCT_COEF); + + /* Emit the Huffman-coded symbol for the number of bits */ + if (! emit_bits_s(state, dctbl->ehufco[nbits], dctbl->ehufsi[nbits])) + return FALSE; + + /* Emit that number of bits of the value, if positive, */ + /* or the complement of its magnitude, if negative. */ + if (! emit_bits_s(state, (unsigned int) temp2, nbits)) + return FALSE; + } + + /* Encode the AC coefficients per section F.1.2.2 */ + + r = 0; /* r = run length of zeros */ + + for (k = 1; k <= Se; k++) { + if ((temp = block[natural_order[k]]) == 0) { + r++; + continue; + } + + /* if run length > 15, must emit special run-length-16 codes (0xF0) */ + while (r > 15) { + if (! emit_bits_s(state, actbl->ehufco[0xF0], actbl->ehufsi[0xF0])) + return FALSE; + r -= 16; + } + + if ((temp2 = temp) < 0) { + temp = -temp; /* temp is abs value of input */ + /* For a negative coef, want temp2 = bitwise complement of abs(coef) */ + /* This code assumes we are on a two's complement machine */ + temp2--; + } + + /* Find the number of bits needed for the magnitude of the coefficient */ + nbits = 0; + do nbits++; /* there must be at least one 1 bit */ + while ((temp >>= 1)); + /* Check for out-of-range coefficient values. + * Use ">=" instead of ">" so can use the + * same one larger limit from DC check here. + */ + if (nbits >= max_coef_bits) + ERREXIT(state->cinfo, JERR_BAD_DCT_COEF); + + /* Emit Huffman symbol for run length / number of bits */ + temp = (r << 4) + nbits; + if (! emit_bits_s(state, actbl->ehufco[temp], actbl->ehufsi[temp])) + return FALSE; + + /* Emit that number of bits of the value, if positive, */ + /* or the complement of its magnitude, if negative. */ + if (! emit_bits_s(state, (unsigned int) temp2, nbits)) + return FALSE; + + r = 0; /* reset zero run length */ + } + + /* If the last coef(s) were zero, emit an end-of-block code */ + if (r > 0) + if (! emit_bits_s(state, actbl->ehufco[0], actbl->ehufsi[0])) + return FALSE; + + return TRUE; +} + + +/* + * Encode and output one MCU's worth of Huffman-compressed coefficients. + */ + +METHODDEF(boolean) +encode_mcu_huff (j_compress_ptr cinfo, JBLOCKARRAY MCU_data) +{ + huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; + working_state state; + int blkn, ci; + jpeg_component_info * compptr; + + /* Load up working state */ + state.next_output_byte = cinfo->dest->next_output_byte; + state.free_in_buffer = cinfo->dest->free_in_buffer; + ASSIGN_STATE(state.cur, entropy->saved); + state.cinfo = cinfo; + + /* Emit restart marker if needed */ + if (cinfo->restart_interval) { + if (entropy->restarts_to_go == 0) + if (! emit_restart_s(&state, entropy->next_restart_num)) + return FALSE; + } + + /* Encode the MCU data blocks */ + for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { + ci = cinfo->MCU_membership[blkn]; + compptr = cinfo->cur_comp_info[ci]; + if (! encode_one_block(&state, + MCU_data[blkn][0], state.cur.last_dc_val[ci], + entropy->dc_derived_tbls[compptr->dc_tbl_no], + entropy->ac_derived_tbls[compptr->ac_tbl_no])) + return FALSE; + /* Update last_dc_val */ + state.cur.last_dc_val[ci] = MCU_data[blkn][0][0]; + } + + /* Completed MCU, so update state */ + cinfo->dest->next_output_byte = state.next_output_byte; + cinfo->dest->free_in_buffer = state.free_in_buffer; + ASSIGN_STATE(entropy->saved, state.cur); + + /* Update restart-interval state too */ + if (cinfo->restart_interval) { + if (entropy->restarts_to_go == 0) { + entropy->restarts_to_go = cinfo->restart_interval; + entropy->next_restart_num++; + entropy->next_restart_num &= 7; + } + entropy->restarts_to_go--; + } + + return TRUE; +} + + +/* + * Finish up at the end of a Huffman-compressed scan. + */ + +METHODDEF(void) +finish_pass_huff (j_compress_ptr cinfo) +{ + huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; + working_state state; + + if (cinfo->progressive_mode) { + entropy->next_output_byte = cinfo->dest->next_output_byte; + entropy->free_in_buffer = cinfo->dest->free_in_buffer; + + /* Flush out any buffered data */ + emit_eobrun(entropy); + flush_bits_e(entropy); + + cinfo->dest->next_output_byte = entropy->next_output_byte; + cinfo->dest->free_in_buffer = entropy->free_in_buffer; + } else { + /* Load up working state ... flush_bits needs it */ + state.next_output_byte = cinfo->dest->next_output_byte; + state.free_in_buffer = cinfo->dest->free_in_buffer; + ASSIGN_STATE(state.cur, entropy->saved); + state.cinfo = cinfo; + + /* Flush out the last data */ + if (! flush_bits_s(&state)) + ERREXIT(cinfo, JERR_CANT_SUSPEND); + + /* Update state */ + cinfo->dest->next_output_byte = state.next_output_byte; + cinfo->dest->free_in_buffer = state.free_in_buffer; + ASSIGN_STATE(entropy->saved, state.cur); + } +} + + +/* + * Huffman coding optimization. + * + * We first scan the supplied data and count the number of uses of each symbol + * that is to be Huffman-coded. (This process MUST agree with the code above.) + * Then we build a Huffman coding tree for the observed counts. + * Symbols which are not needed at all for the particular image are not + * assigned any code, which saves space in the DHT marker as well as in + * the compressed data. + */ + + +/* Process a single block's worth of coefficients */ + +LOCAL(void) +htest_one_block (j_compress_ptr cinfo, JCOEFPTR block, int last_dc_val, + long dc_counts[], long ac_counts[]) +{ + register int temp; + register int nbits; + register int r, k; + int Se = cinfo->lim_Se; + int max_coef_bits = cinfo->data_precision + 3; + const int * natural_order = cinfo->natural_order; + + /* Encode the DC coefficient difference per section F.1.2.1 */ + + if ((temp = block[0] - last_dc_val) == 0) { + /* Count the Huffman symbol for the number of bits */ + dc_counts[0]++; + } else { + if (temp < 0) + temp = -temp; /* temp is abs value of input */ + + /* Find the number of bits needed for the magnitude of the coefficient */ + nbits = 0; + do nbits++; /* there must be at least one 1 bit */ + while ((temp >>= 1)); + /* Check for out-of-range coefficient values. + * Since we're encoding a difference, the range limit is twice as much. + */ + if (nbits > max_coef_bits) + ERREXIT(cinfo, JERR_BAD_DCT_COEF); + + /* Count the Huffman symbol for the number of bits */ + dc_counts[nbits]++; + } + + /* Encode the AC coefficients per section F.1.2.2 */ + + r = 0; /* r = run length of zeros */ + + for (k = 1; k <= Se; k++) { + if ((temp = block[natural_order[k]]) == 0) { + r++; + continue; + } + + /* if run length > 15, must emit special run-length-16 codes (0xF0) */ + while (r > 15) { + ac_counts[0xF0]++; + r -= 16; + } + + if (temp < 0) + temp = -temp; /* temp is abs value of input */ + + /* Find the number of bits needed for the magnitude of the coefficient */ + nbits = 0; + do nbits++; /* there must be at least one 1 bit */ + while ((temp >>= 1)); + /* Check for out-of-range coefficient values. + * Use ">=" instead of ">" so can use the + * same one larger limit from DC check here. + */ + if (nbits >= max_coef_bits) + ERREXIT(cinfo, JERR_BAD_DCT_COEF); + + /* Count Huffman symbol for run length / number of bits */ + ac_counts[(r << 4) + nbits]++; + + r = 0; /* reset zero run length */ + } + + /* If the last coef(s) were zero, emit an end-of-block code */ + if (r > 0) + ac_counts[0]++; +} + + +/* + * Trial-encode one MCU's worth of Huffman-compressed coefficients. + * No data is actually output, so no suspension return is possible. + */ + +METHODDEF(boolean) +encode_mcu_gather (j_compress_ptr cinfo, JBLOCKARRAY MCU_data) +{ + huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; + int blkn, ci; + jpeg_component_info * compptr; + + /* Take care of restart intervals if needed */ + if (cinfo->restart_interval) { + if (entropy->restarts_to_go == 0) { + /* Re-initialize DC predictions to 0 */ + for (ci = 0; ci < cinfo->comps_in_scan; ci++) + entropy->saved.last_dc_val[ci] = 0; + /* Update restart state */ + entropy->restarts_to_go = cinfo->restart_interval; + } + entropy->restarts_to_go--; + } + + for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { + ci = cinfo->MCU_membership[blkn]; + compptr = cinfo->cur_comp_info[ci]; + htest_one_block(cinfo, MCU_data[blkn][0], entropy->saved.last_dc_val[ci], + entropy->dc_count_ptrs[compptr->dc_tbl_no], + entropy->ac_count_ptrs[compptr->ac_tbl_no]); + entropy->saved.last_dc_val[ci] = MCU_data[blkn][0][0]; + } + + return TRUE; +} + + +/* + * Generate the best Huffman code table for the given counts, fill htbl. + * + * The JPEG standard requires that no symbol be assigned a codeword of all + * one bits (so that padding bits added at the end of a compressed segment + * can't look like a valid code). Because of the canonical ordering of + * codewords, this just means that there must be an unused slot in the + * longest codeword length category. Section K.2 of the JPEG spec suggests + * reserving such a slot by pretending that symbol 256 is a valid symbol + * with count 1. In theory that's not optimal; giving it count zero but + * including it in the symbol set anyway should give a better Huffman code. + * But the theoretically better code actually seems to come out worse in + * practice, because it produces more all-ones bytes (which incur stuffed + * zero bytes in the final file). In any case the difference is tiny. + * + * The JPEG standard requires Huffman codes to be no more than 16 bits long. + * If some symbols have a very small but nonzero probability, the Huffman tree + * must be adjusted to meet the code length restriction. We currently use + * the adjustment method suggested in JPEG section K.2. This method is *not* + * optimal; it may not choose the best possible limited-length code. But + * typically only very-low-frequency symbols will be given less-than-optimal + * lengths, so the code is almost optimal. Experimental comparisons against + * an optimal limited-length-code algorithm indicate that the difference is + * microscopic --- usually less than a hundredth of a percent of total size. + * So the extra complexity of an optimal algorithm doesn't seem worthwhile. + */ + +LOCAL(void) +jpeg_gen_optimal_table (j_compress_ptr cinfo, JHUFF_TBL * htbl, long freq[]) +{ +#define MAX_CLEN 32 /* assumed maximum initial code length */ + UINT8 bits[MAX_CLEN+1]; /* bits[k] = # of symbols with code length k */ + int codesize[257]; /* codesize[k] = code length of symbol k */ + int others[257]; /* next symbol in current branch of tree */ + int c1, c2, i, j; + UINT8 *p; + long v; + + freq[256] = 1; /* make sure 256 has a nonzero count */ + /* Including the pseudo-symbol 256 in the Huffman procedure guarantees + * that no real symbol is given code-value of all ones, because 256 + * will be placed last in the largest codeword category. + * In the symbol list build procedure this element serves as sentinel + * for the zero run loop. + */ + +#ifndef DONT_USE_FANCY_HUFF_OPT + + /* Build list of symbols sorted in order of descending frequency */ + /* This approach has several benefits (thank to John Korejwa for the idea): + * 1. + * If a codelength category is split during the length limiting procedure + * below, the feature that more frequent symbols are assigned shorter + * codewords remains valid for the adjusted code. + * 2. + * To reduce consecutive ones in a Huffman data stream (thus reducing the + * number of stuff bytes in JPEG) it is preferable to follow 0 branches + * (and avoid 1 branches) as much as possible. This is easily done by + * assigning symbols to leaves of the Huffman tree in order of decreasing + * frequency, with no secondary sort based on codelengths. + * 3. + * The symbol list can be built independently from the assignment of code + * lengths by the Huffman procedure below. + * Note: The symbol list build procedure must be performed first, because + * the Huffman procedure assigning the codelengths clobbers the frequency + * counts! + */ + + /* Here we use the others array as a linked list of nonzero frequencies + * to be sorted. Already sorted elements are removed from the list. + */ + + /* Building list */ + + /* This item does not correspond to a valid symbol frequency and is used + * as starting index. + */ + j = 256; + + for (i = 0;; i++) { + if (freq[i] == 0) /* skip zero frequencies */ + continue; + if (i > 255) + break; + others[j] = i; /* this symbol value */ + j = i; /* previous symbol value */ + } + others[j] = -1; /* mark end of list */ + + /* Sorting list */ + + p = htbl->huffval; + while ((c1 = others[256]) >= 0) { + v = freq[c1]; + i = c1; /* first symbol value */ + j = 256; /* pseudo symbol value for starting index */ + while ((c2 = others[c1]) >= 0) { + if (freq[c2] > v) { + v = freq[c2]; + i = c2; /* this symbol value */ + j = c1; /* previous symbol value */ + } + c1 = c2; + } + others[j] = others[i]; /* remove this symbol i from list */ + *p++ = (UINT8) i; + } + +#endif /* DONT_USE_FANCY_HUFF_OPT */ + + /* This algorithm is explained in section K.2 of the JPEG standard */ + + MEMZERO(bits, SIZEOF(bits)); + MEMZERO(codesize, SIZEOF(codesize)); + for (i = 0; i < 257; i++) + others[i] = -1; /* init links to empty */ + + /* Huffman's basic algorithm to assign optimal code lengths to symbols */ + + for (;;) { + /* Find the smallest nonzero frequency, set c1 = its symbol */ + /* In case of ties, take the larger symbol number */ + c1 = -1; + v = 1000000000L; + for (i = 0; i <= 256; i++) { + if (freq[i] && freq[i] <= v) { + v = freq[i]; + c1 = i; + } + } + + /* Find the next smallest nonzero frequency, set c2 = its symbol */ + /* In case of ties, take the larger symbol number */ + c2 = -1; + v = 1000000000L; + for (i = 0; i <= 256; i++) { + if (freq[i] && freq[i] <= v && i != c1) { + v = freq[i]; + c2 = i; + } + } + + /* Done if we've merged everything into one frequency */ + if (c2 < 0) + break; + + /* Else merge the two counts/trees */ + freq[c1] += freq[c2]; + freq[c2] = 0; + + /* Increment the codesize of everything in c1's tree branch */ + codesize[c1]++; + while (others[c1] >= 0) { + c1 = others[c1]; + codesize[c1]++; + } + + others[c1] = c2; /* chain c2 onto c1's tree branch */ + + /* Increment the codesize of everything in c2's tree branch */ + codesize[c2]++; + while (others[c2] >= 0) { + c2 = others[c2]; + codesize[c2]++; + } + } + + /* Now count the number of symbols of each code length */ + for (i = 0; i <= 256; i++) { + if (codesize[i]) { + /* The JPEG standard seems to think that this can't happen, */ + /* but I'm paranoid... */ + if (codesize[i] > MAX_CLEN) + ERREXIT(cinfo, JERR_HUFF_CLEN_OUTOFBOUNDS); + + bits[codesize[i]]++; + } + } + + /* JPEG doesn't allow symbols with code lengths over 16 bits, so if the pure + * Huffman procedure assigned any such lengths, we must adjust the coding. + * Here is what the JPEG spec says about how this next bit works: + * Since symbols are paired for the longest Huffman code, the symbols are + * removed from this length category two at a time. The prefix for the pair + * (which is one bit shorter) is allocated to one of the pair; then, + * skipping the BITS entry for that prefix length, a code word from the next + * shortest nonzero BITS entry is converted into a prefix for two code words + * one bit longer. + */ + + for (i = MAX_CLEN; i > 16; i--) { + while (bits[i] > 0) { + j = i - 2; /* find length of new prefix to be used */ + while (bits[j] == 0) { + if (j == 0) + ERREXIT(cinfo, JERR_HUFF_CLEN_OUTOFBOUNDS); + j--; + } + + bits[i] -= 2; /* remove two symbols */ + bits[i-1]++; /* one goes in this length */ + bits[j+1] += 2; /* two new symbols in this length */ + bits[j]--; /* symbol of this length is now a prefix */ + } + } + + /* Remove the count for the pseudo-symbol 256 from the largest codelength */ + while (bits[i] == 0) /* find largest codelength still in use */ + i--; + bits[i]--; + + /* Return final symbol counts (only for lengths 0..16) */ + MEMCOPY(htbl->bits, bits, SIZEOF(htbl->bits)); + +#ifdef DONT_USE_FANCY_HUFF_OPT + + /* Return a list of the symbols sorted by code length */ + /* Note: Due to the codelength changes made above, it can happen + * that more frequent symbols are assigned longer codewords. + */ + p = htbl->huffval; + for (i = 1; i <= MAX_CLEN; i++) { + for (j = 0; j <= 255; j++) { + if (codesize[j] == i) { + *p++ = (UINT8) j; + } + } + } + +#endif /* DONT_USE_FANCY_HUFF_OPT */ + + /* Set sent_table FALSE so updated table will be written to JPEG file. */ + htbl->sent_table = FALSE; +} + + +/* + * Finish up a statistics-gathering pass and create the new Huffman tables. + */ + +METHODDEF(void) +finish_pass_gather (j_compress_ptr cinfo) +{ + huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; + int ci, tbl; + jpeg_component_info * compptr; + JHUFF_TBL **htblptr; + boolean did_dc[NUM_HUFF_TBLS]; + boolean did_ac[NUM_HUFF_TBLS]; + + if (cinfo->progressive_mode) + /* Flush out buffered data (all we care about is counting the EOB symbol) */ + emit_eobrun(entropy); + + /* It's important not to apply jpeg_gen_optimal_table more than once + * per table, because it clobbers the input frequency counts! + */ + MEMZERO(did_dc, SIZEOF(did_dc)); + MEMZERO(did_ac, SIZEOF(did_ac)); + + for (ci = 0; ci < cinfo->comps_in_scan; ci++) { + compptr = cinfo->cur_comp_info[ci]; + /* DC needs no table for refinement scan */ + if (cinfo->Ss == 0 && cinfo->Ah == 0) { + tbl = compptr->dc_tbl_no; + if (! did_dc[tbl]) { + htblptr = & cinfo->dc_huff_tbl_ptrs[tbl]; + if (*htblptr == NULL) + *htblptr = jpeg_alloc_huff_table((j_common_ptr) cinfo); + jpeg_gen_optimal_table(cinfo, *htblptr, entropy->dc_count_ptrs[tbl]); + did_dc[tbl] = TRUE; + } + } + /* AC needs no table when not present */ + if (cinfo->Se) { + tbl = compptr->ac_tbl_no; + if (! did_ac[tbl]) { + htblptr = & cinfo->ac_huff_tbl_ptrs[tbl]; + if (*htblptr == NULL) + *htblptr = jpeg_alloc_huff_table((j_common_ptr) cinfo); + jpeg_gen_optimal_table(cinfo, *htblptr, entropy->ac_count_ptrs[tbl]); + did_ac[tbl] = TRUE; + } + } + } +} + + +/* + * Initialize for a Huffman-compressed scan. + * If gather_statistics is TRUE, we do not output anything during the scan, + * just count the Huffman symbols used and generate Huffman code tables. + */ + +METHODDEF(void) +start_pass_huff (j_compress_ptr cinfo, boolean gather_statistics) +{ + huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; + int ci, tbl; + jpeg_component_info * compptr; + + if (gather_statistics) + entropy->pub.finish_pass = finish_pass_gather; + else + entropy->pub.finish_pass = finish_pass_huff; + + if (cinfo->progressive_mode) { + entropy->cinfo = cinfo; + entropy->gather_statistics = gather_statistics; + + /* We assume jcmaster.c already validated the scan parameters. */ + + /* Select execution routine */ + if (cinfo->Ah == 0) { + if (cinfo->Ss == 0) + entropy->pub.encode_mcu = encode_mcu_DC_first; + else + entropy->pub.encode_mcu = encode_mcu_AC_first; + } else { + if (cinfo->Ss == 0) + entropy->pub.encode_mcu = encode_mcu_DC_refine; + else { + entropy->pub.encode_mcu = encode_mcu_AC_refine; + /* AC refinement needs a correction bit buffer */ + if (entropy->bit_buffer == NULL) + entropy->bit_buffer = (char *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, MAX_CORR_BITS * SIZEOF(char)); + } + } + + /* Initialize AC stuff */ + entropy->ac_tbl_no = cinfo->cur_comp_info[0]->ac_tbl_no; + entropy->EOBRUN = 0; + entropy->BE = 0; + } else { + if (gather_statistics) + entropy->pub.encode_mcu = encode_mcu_gather; + else + entropy->pub.encode_mcu = encode_mcu_huff; + } + + for (ci = 0; ci < cinfo->comps_in_scan; ci++) { + compptr = cinfo->cur_comp_info[ci]; + /* DC needs no table for refinement scan */ + if (cinfo->Ss == 0 && cinfo->Ah == 0) { + tbl = compptr->dc_tbl_no; + if (gather_statistics) { + /* Check for invalid table index */ + /* (make_c_derived_tbl does this in the other path) */ + if (tbl < 0 || tbl >= NUM_HUFF_TBLS) + ERREXIT1(cinfo, JERR_NO_HUFF_TABLE, tbl); + /* Allocate and zero the statistics tables */ + /* Note that jpeg_gen_optimal_table expects 257 entries in each table! */ + if (entropy->dc_count_ptrs[tbl] == NULL) + entropy->dc_count_ptrs[tbl] = (long *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, 257 * SIZEOF(long)); + MEMZERO(entropy->dc_count_ptrs[tbl], 257 * SIZEOF(long)); + } else { + /* Compute derived values for Huffman tables */ + /* We may do this more than once for a table, but it's not expensive */ + jpeg_make_c_derived_tbl(cinfo, TRUE, tbl, + & entropy->dc_derived_tbls[tbl]); + } + /* Initialize DC predictions to 0 */ + entropy->saved.last_dc_val[ci] = 0; + } + /* AC needs no table when not present */ + if (cinfo->Se) { + tbl = compptr->ac_tbl_no; + if (gather_statistics) { + if (tbl < 0 || tbl >= NUM_HUFF_TBLS) + ERREXIT1(cinfo, JERR_NO_HUFF_TABLE, tbl); + if (entropy->ac_count_ptrs[tbl] == NULL) + entropy->ac_count_ptrs[tbl] = (long *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, 257 * SIZEOF(long)); + MEMZERO(entropy->ac_count_ptrs[tbl], 257 * SIZEOF(long)); + } else { + jpeg_make_c_derived_tbl(cinfo, FALSE, tbl, + & entropy->ac_derived_tbls[tbl]); + } + } + } + + /* Initialize bit buffer to empty */ + entropy->saved.put_buffer = 0; + entropy->saved.put_bits = 0; + + /* Initialize restart stuff */ + entropy->restarts_to_go = cinfo->restart_interval; + entropy->next_restart_num = 0; +} + + +/* + * Module initialization routine for Huffman entropy encoding. + */ + +GLOBAL(void) +jinit_huff_encoder (j_compress_ptr cinfo) +{ + huff_entropy_ptr entropy; + int i; + + entropy = (huff_entropy_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(huff_entropy_encoder)); + cinfo->entropy = &entropy->pub; + entropy->pub.start_pass = start_pass_huff; + + /* Mark tables unallocated */ + for (i = 0; i < NUM_HUFF_TBLS; i++) { + entropy->dc_derived_tbls[i] = entropy->ac_derived_tbls[i] = NULL; + entropy->dc_count_ptrs[i] = entropy->ac_count_ptrs[i] = NULL; + } + + if (cinfo->progressive_mode) + entropy->bit_buffer = NULL; /* needed only in AC refinement scan */ +} diff --git a/cmake/externals/libjpeg/jcinit.c b/cmake/externals/libjpeg/jcinit.c new file mode 100644 index 000000000..2aea7ca2a --- /dev/null +++ b/cmake/externals/libjpeg/jcinit.c @@ -0,0 +1,249 @@ +/* + * jcinit.c + * + * Copyright (C) 1991-1997, Thomas G. Lane. + * Modified 2003-2017 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains initialization logic for the JPEG compressor. + * This routine is in charge of selecting the modules to be executed and + * making an initialization call to each one. + * + * Logically, this code belongs in jcmaster.c. It's split out because + * linking this routine implies linking the entire compression library. + * For a transcoding-only application, we want to be able to use jcmaster.c + * without linking in the whole library. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* + * Compute JPEG image dimensions and related values. + * NOTE: this is exported for possible use by application. + * Hence it mustn't do anything that can't be done twice. + */ + +GLOBAL(void) +jpeg_calc_jpeg_dimensions (j_compress_ptr cinfo) +/* Do computations that are needed before master selection phase */ +{ + /* Sanity check on input image dimensions to prevent overflow in + * following calculations. + * We do check jpeg_width and jpeg_height in initial_setup in jcmaster.c, + * but image_width and image_height can come from arbitrary data, + * and we need some space for multiplication by block_size. + */ + if (((long) cinfo->image_width >> 24) || ((long) cinfo->image_height >> 24)) + ERREXIT1(cinfo, JERR_IMAGE_TOO_BIG, (unsigned int) JPEG_MAX_DIMENSION); + +#ifdef DCT_SCALING_SUPPORTED + + /* Compute actual JPEG image dimensions and DCT scaling choices. */ + if (cinfo->scale_num >= cinfo->scale_denom * cinfo->block_size) { + /* Provide block_size/1 scaling */ + cinfo->jpeg_width = cinfo->image_width * cinfo->block_size; + cinfo->jpeg_height = cinfo->image_height * cinfo->block_size; + cinfo->min_DCT_h_scaled_size = 1; + cinfo->min_DCT_v_scaled_size = 1; + } else if (cinfo->scale_num * 2 >= cinfo->scale_denom * cinfo->block_size) { + /* Provide block_size/2 scaling */ + cinfo->jpeg_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 2L); + cinfo->jpeg_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 2L); + cinfo->min_DCT_h_scaled_size = 2; + cinfo->min_DCT_v_scaled_size = 2; + } else if (cinfo->scale_num * 3 >= cinfo->scale_denom * cinfo->block_size) { + /* Provide block_size/3 scaling */ + cinfo->jpeg_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 3L); + cinfo->jpeg_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 3L); + cinfo->min_DCT_h_scaled_size = 3; + cinfo->min_DCT_v_scaled_size = 3; + } else if (cinfo->scale_num * 4 >= cinfo->scale_denom * cinfo->block_size) { + /* Provide block_size/4 scaling */ + cinfo->jpeg_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 4L); + cinfo->jpeg_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 4L); + cinfo->min_DCT_h_scaled_size = 4; + cinfo->min_DCT_v_scaled_size = 4; + } else if (cinfo->scale_num * 5 >= cinfo->scale_denom * cinfo->block_size) { + /* Provide block_size/5 scaling */ + cinfo->jpeg_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 5L); + cinfo->jpeg_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 5L); + cinfo->min_DCT_h_scaled_size = 5; + cinfo->min_DCT_v_scaled_size = 5; + } else if (cinfo->scale_num * 6 >= cinfo->scale_denom * cinfo->block_size) { + /* Provide block_size/6 scaling */ + cinfo->jpeg_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 6L); + cinfo->jpeg_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 6L); + cinfo->min_DCT_h_scaled_size = 6; + cinfo->min_DCT_v_scaled_size = 6; + } else if (cinfo->scale_num * 7 >= cinfo->scale_denom * cinfo->block_size) { + /* Provide block_size/7 scaling */ + cinfo->jpeg_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 7L); + cinfo->jpeg_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 7L); + cinfo->min_DCT_h_scaled_size = 7; + cinfo->min_DCT_v_scaled_size = 7; + } else if (cinfo->scale_num * 8 >= cinfo->scale_denom * cinfo->block_size) { + /* Provide block_size/8 scaling */ + cinfo->jpeg_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 8L); + cinfo->jpeg_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 8L); + cinfo->min_DCT_h_scaled_size = 8; + cinfo->min_DCT_v_scaled_size = 8; + } else if (cinfo->scale_num * 9 >= cinfo->scale_denom * cinfo->block_size) { + /* Provide block_size/9 scaling */ + cinfo->jpeg_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 9L); + cinfo->jpeg_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 9L); + cinfo->min_DCT_h_scaled_size = 9; + cinfo->min_DCT_v_scaled_size = 9; + } else if (cinfo->scale_num * 10 >= cinfo->scale_denom * cinfo->block_size) { + /* Provide block_size/10 scaling */ + cinfo->jpeg_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 10L); + cinfo->jpeg_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 10L); + cinfo->min_DCT_h_scaled_size = 10; + cinfo->min_DCT_v_scaled_size = 10; + } else if (cinfo->scale_num * 11 >= cinfo->scale_denom * cinfo->block_size) { + /* Provide block_size/11 scaling */ + cinfo->jpeg_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 11L); + cinfo->jpeg_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 11L); + cinfo->min_DCT_h_scaled_size = 11; + cinfo->min_DCT_v_scaled_size = 11; + } else if (cinfo->scale_num * 12 >= cinfo->scale_denom * cinfo->block_size) { + /* Provide block_size/12 scaling */ + cinfo->jpeg_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 12L); + cinfo->jpeg_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 12L); + cinfo->min_DCT_h_scaled_size = 12; + cinfo->min_DCT_v_scaled_size = 12; + } else if (cinfo->scale_num * 13 >= cinfo->scale_denom * cinfo->block_size) { + /* Provide block_size/13 scaling */ + cinfo->jpeg_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 13L); + cinfo->jpeg_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 13L); + cinfo->min_DCT_h_scaled_size = 13; + cinfo->min_DCT_v_scaled_size = 13; + } else if (cinfo->scale_num * 14 >= cinfo->scale_denom * cinfo->block_size) { + /* Provide block_size/14 scaling */ + cinfo->jpeg_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 14L); + cinfo->jpeg_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 14L); + cinfo->min_DCT_h_scaled_size = 14; + cinfo->min_DCT_v_scaled_size = 14; + } else if (cinfo->scale_num * 15 >= cinfo->scale_denom * cinfo->block_size) { + /* Provide block_size/15 scaling */ + cinfo->jpeg_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 15L); + cinfo->jpeg_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 15L); + cinfo->min_DCT_h_scaled_size = 15; + cinfo->min_DCT_v_scaled_size = 15; + } else { + /* Provide block_size/16 scaling */ + cinfo->jpeg_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 16L); + cinfo->jpeg_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 16L); + cinfo->min_DCT_h_scaled_size = 16; + cinfo->min_DCT_v_scaled_size = 16; + } + +#else /* !DCT_SCALING_SUPPORTED */ + + /* Hardwire it to "no scaling" */ + cinfo->jpeg_width = cinfo->image_width; + cinfo->jpeg_height = cinfo->image_height; + cinfo->min_DCT_h_scaled_size = DCTSIZE; + cinfo->min_DCT_v_scaled_size = DCTSIZE; + +#endif /* DCT_SCALING_SUPPORTED */ +} + + +/* + * Master selection of compression modules. + * This is done once at the start of processing an image. We determine + * which modules will be used and give them appropriate initialization calls. + */ + +GLOBAL(void) +jinit_compress_master (j_compress_ptr cinfo) +{ + long samplesperrow; + JDIMENSION jd_samplesperrow; + + /* For now, precision must match compiled-in value... */ + if (cinfo->data_precision != BITS_IN_JSAMPLE) + ERREXIT1(cinfo, JERR_BAD_PRECISION, cinfo->data_precision); + + /* Sanity check on input image dimensions */ + if (cinfo->image_height <= 0 || cinfo->image_width <= 0 || + cinfo->input_components <= 0) + ERREXIT(cinfo, JERR_EMPTY_IMAGE); + + /* Width of an input scanline must be representable as JDIMENSION. */ + samplesperrow = (long) cinfo->image_width * (long) cinfo->input_components; + jd_samplesperrow = (JDIMENSION) samplesperrow; + if ((long) jd_samplesperrow != samplesperrow) + ERREXIT(cinfo, JERR_WIDTH_OVERFLOW); + + /* Compute JPEG image dimensions and related values. */ + jpeg_calc_jpeg_dimensions(cinfo); + + /* Initialize master control (includes parameter checking/processing) */ + jinit_c_master_control(cinfo, FALSE /* full compression */); + + /* Preprocessing */ + if (! cinfo->raw_data_in) { + jinit_color_converter(cinfo); + jinit_downsampler(cinfo); + jinit_c_prep_controller(cinfo, FALSE /* never need full buffer here */); + } + /* Forward DCT */ + jinit_forward_dct(cinfo); + /* Entropy encoding: either Huffman or arithmetic coding. */ + if (cinfo->arith_code) + jinit_arith_encoder(cinfo); + else { + jinit_huff_encoder(cinfo); + } + + /* Need a full-image coefficient buffer in any multi-pass mode. */ + jinit_c_coef_controller(cinfo, + (boolean) (cinfo->num_scans > 1 || cinfo->optimize_coding)); + jinit_c_main_controller(cinfo, FALSE /* never need full buffer here */); + + jinit_marker_writer(cinfo); + + /* We can now tell the memory manager to allocate virtual arrays. */ + (*cinfo->mem->realize_virt_arrays) ((j_common_ptr) cinfo); + + /* Write the datastream header (SOI) immediately. + * Frame and scan headers are postponed till later. + * This lets application insert special markers after the SOI. + */ + (*cinfo->marker->write_file_header) (cinfo); +} diff --git a/cmake/externals/libjpeg/jcmainct.c b/cmake/externals/libjpeg/jcmainct.c new file mode 100644 index 000000000..39b97902e --- /dev/null +++ b/cmake/externals/libjpeg/jcmainct.c @@ -0,0 +1,297 @@ +/* + * jcmainct.c + * + * Copyright (C) 1994-1996, Thomas G. Lane. + * Modified 2003-2012 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains the main buffer controller for compression. + * The main buffer lies between the pre-processor and the JPEG + * compressor proper; it holds downsampled data in the JPEG colorspace. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* Note: currently, there is no operating mode in which a full-image buffer + * is needed at this step. If there were, that mode could not be used with + * "raw data" input, since this module is bypassed in that case. However, + * we've left the code here for possible use in special applications. + */ +#undef FULL_MAIN_BUFFER_SUPPORTED + + +/* Private buffer controller object */ + +typedef struct { + struct jpeg_c_main_controller pub; /* public fields */ + + JDIMENSION cur_iMCU_row; /* number of current iMCU row */ + JDIMENSION rowgroup_ctr; /* counts row groups received in iMCU row */ + boolean suspended; /* remember if we suspended output */ + J_BUF_MODE pass_mode; /* current operating mode */ + + /* If using just a strip buffer, this points to the entire set of buffers + * (we allocate one for each component). In the full-image case, this + * points to the currently accessible strips of the virtual arrays. + */ + JSAMPARRAY buffer[MAX_COMPONENTS]; + +#ifdef FULL_MAIN_BUFFER_SUPPORTED + /* If using full-image storage, this array holds pointers to virtual-array + * control blocks for each component. Unused if not full-image storage. + */ + jvirt_sarray_ptr whole_image[MAX_COMPONENTS]; +#endif +} my_main_controller; + +typedef my_main_controller * my_main_ptr; + + +/* Forward declarations */ +METHODDEF(void) process_data_simple_main + JPP((j_compress_ptr cinfo, JSAMPARRAY input_buf, + JDIMENSION *in_row_ctr, JDIMENSION in_rows_avail)); +#ifdef FULL_MAIN_BUFFER_SUPPORTED +METHODDEF(void) process_data_buffer_main + JPP((j_compress_ptr cinfo, JSAMPARRAY input_buf, + JDIMENSION *in_row_ctr, JDIMENSION in_rows_avail)); +#endif + + +/* + * Initialize for a processing pass. + */ + +METHODDEF(void) +start_pass_main (j_compress_ptr cinfo, J_BUF_MODE pass_mode) +{ + my_main_ptr mainp = (my_main_ptr) cinfo->main; + + /* Do nothing in raw-data mode. */ + if (cinfo->raw_data_in) + return; + + mainp->cur_iMCU_row = 0; /* initialize counters */ + mainp->rowgroup_ctr = 0; + mainp->suspended = FALSE; + mainp->pass_mode = pass_mode; /* save mode for use by process_data */ + + switch (pass_mode) { + case JBUF_PASS_THRU: +#ifdef FULL_MAIN_BUFFER_SUPPORTED + if (mainp->whole_image[0] != NULL) + ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); +#endif + mainp->pub.process_data = process_data_simple_main; + break; +#ifdef FULL_MAIN_BUFFER_SUPPORTED + case JBUF_SAVE_SOURCE: + case JBUF_CRANK_DEST: + case JBUF_SAVE_AND_PASS: + if (mainp->whole_image[0] == NULL) + ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); + mainp->pub.process_data = process_data_buffer_main; + break; +#endif + default: + ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); + break; + } +} + + +/* + * Process some data. + * This routine handles the simple pass-through mode, + * where we have only a strip buffer. + */ + +METHODDEF(void) +process_data_simple_main (j_compress_ptr cinfo, + JSAMPARRAY input_buf, JDIMENSION *in_row_ctr, + JDIMENSION in_rows_avail) +{ + my_main_ptr mainp = (my_main_ptr) cinfo->main; + + while (mainp->cur_iMCU_row < cinfo->total_iMCU_rows) { + /* Read input data if we haven't filled the main buffer yet */ + if (mainp->rowgroup_ctr < (JDIMENSION) cinfo->min_DCT_v_scaled_size) + (*cinfo->prep->pre_process_data) (cinfo, + input_buf, in_row_ctr, in_rows_avail, + mainp->buffer, &mainp->rowgroup_ctr, + (JDIMENSION) cinfo->min_DCT_v_scaled_size); + + /* If we don't have a full iMCU row buffered, return to application for + * more data. Note that preprocessor will always pad to fill the iMCU row + * at the bottom of the image. + */ + if (mainp->rowgroup_ctr != (JDIMENSION) cinfo->min_DCT_v_scaled_size) + return; + + /* Send the completed row to the compressor */ + if (! (*cinfo->coef->compress_data) (cinfo, mainp->buffer)) { + /* If compressor did not consume the whole row, then we must need to + * suspend processing and return to the application. In this situation + * we pretend we didn't yet consume the last input row; otherwise, if + * it happened to be the last row of the image, the application would + * think we were done. + */ + if (! mainp->suspended) { + (*in_row_ctr)--; + mainp->suspended = TRUE; + } + return; + } + /* We did finish the row. Undo our little suspension hack if a previous + * call suspended; then mark the main buffer empty. + */ + if (mainp->suspended) { + (*in_row_ctr)++; + mainp->suspended = FALSE; + } + mainp->rowgroup_ctr = 0; + mainp->cur_iMCU_row++; + } +} + + +#ifdef FULL_MAIN_BUFFER_SUPPORTED + +/* + * Process some data. + * This routine handles all of the modes that use a full-size buffer. + */ + +METHODDEF(void) +process_data_buffer_main (j_compress_ptr cinfo, + JSAMPARRAY input_buf, JDIMENSION *in_row_ctr, + JDIMENSION in_rows_avail) +{ + my_main_ptr mainp = (my_main_ptr) cinfo->main; + int ci; + jpeg_component_info *compptr; + boolean writing = (mainp->pass_mode != JBUF_CRANK_DEST); + + while (mainp->cur_iMCU_row < cinfo->total_iMCU_rows) { + /* Realign the virtual buffers if at the start of an iMCU row. */ + if (mainp->rowgroup_ctr == 0) { + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + mainp->buffer[ci] = (*cinfo->mem->access_virt_sarray) + ((j_common_ptr) cinfo, mainp->whole_image[ci], mainp->cur_iMCU_row * + ((JDIMENSION) (compptr->v_samp_factor * cinfo->min_DCT_v_scaled_size)), + (JDIMENSION) (compptr->v_samp_factor * cinfo->min_DCT_v_scaled_size), + writing); + } + /* In a read pass, pretend we just read some source data. */ + if (! writing) { + *in_row_ctr += (JDIMENSION) + (cinfo->max_v_samp_factor * cinfo->min_DCT_v_scaled_size); + mainp->rowgroup_ctr = (JDIMENSION) cinfo->min_DCT_v_scaled_size; + } + } + + /* If a write pass, read input data until the current iMCU row is full. */ + /* Note: preprocessor will pad if necessary to fill the last iMCU row. */ + if (writing) { + (*cinfo->prep->pre_process_data) (cinfo, + input_buf, in_row_ctr, in_rows_avail, + mainp->buffer, &mainp->rowgroup_ctr, + (JDIMENSION) cinfo->min_DCT_v_scaled_size); + /* Return to application if we need more data to fill the iMCU row. */ + if (mainp->rowgroup_ctr < (JDIMENSION) cinfo->min_DCT_v_scaled_size) + return; + } + + /* Emit data, unless this is a sink-only pass. */ + if (mainp->pass_mode != JBUF_SAVE_SOURCE) { + if (! (*cinfo->coef->compress_data) (cinfo, mainp->buffer)) { + /* If compressor did not consume the whole row, then we must need to + * suspend processing and return to the application. In this situation + * we pretend we didn't yet consume the last input row; otherwise, if + * it happened to be the last row of the image, the application would + * think we were done. + */ + if (! mainp->suspended) { + (*in_row_ctr)--; + mainp->suspended = TRUE; + } + return; + } + /* We did finish the row. Undo our little suspension hack if a previous + * call suspended; then mark the main buffer empty. + */ + if (mainp->suspended) { + (*in_row_ctr)++; + mainp->suspended = FALSE; + } + } + + /* If get here, we are done with this iMCU row. Mark buffer empty. */ + mainp->rowgroup_ctr = 0; + mainp->cur_iMCU_row++; + } +} + +#endif /* FULL_MAIN_BUFFER_SUPPORTED */ + + +/* + * Initialize main buffer controller. + */ + +GLOBAL(void) +jinit_c_main_controller (j_compress_ptr cinfo, boolean need_full_buffer) +{ + my_main_ptr mainp; + int ci; + jpeg_component_info *compptr; + + mainp = (my_main_ptr) + (*cinfo->mem->alloc_small) ((j_common_ptr) cinfo, JPOOL_IMAGE, + SIZEOF(my_main_controller)); + cinfo->main = &mainp->pub; + mainp->pub.start_pass = start_pass_main; + + /* We don't need to create a buffer in raw-data mode. */ + if (cinfo->raw_data_in) + return; + + /* Create the buffer. It holds downsampled data, so each component + * may be of a different size. + */ + if (need_full_buffer) { +#ifdef FULL_MAIN_BUFFER_SUPPORTED + /* Allocate a full-image virtual array for each component */ + /* Note we pad the bottom to a multiple of the iMCU height */ + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + mainp->whole_image[ci] = (*cinfo->mem->request_virt_sarray) + ((j_common_ptr) cinfo, JPOOL_IMAGE, FALSE, + compptr->width_in_blocks * ((JDIMENSION) compptr->DCT_h_scaled_size), + ((JDIMENSION) jround_up((long) compptr->height_in_blocks, + (long) compptr->v_samp_factor)) * + ((JDIMENSION) cinfo->min_DCT_v_scaled_size), + (JDIMENSION) (compptr->v_samp_factor * compptr->DCT_v_scaled_size)); + } +#else + ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); +#endif + } else { +#ifdef FULL_MAIN_BUFFER_SUPPORTED + mainp->whole_image[0] = NULL; /* flag for no virtual arrays */ +#endif + /* Allocate a strip buffer for each component */ + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + mainp->buffer[ci] = (*cinfo->mem->alloc_sarray) + ((j_common_ptr) cinfo, JPOOL_IMAGE, + compptr->width_in_blocks * ((JDIMENSION) compptr->DCT_h_scaled_size), + (JDIMENSION) (compptr->v_samp_factor * compptr->DCT_v_scaled_size)); + } + } +} diff --git a/cmake/externals/libjpeg/jcmarker.c b/cmake/externals/libjpeg/jcmarker.c new file mode 100644 index 000000000..8874cd867 --- /dev/null +++ b/cmake/externals/libjpeg/jcmarker.c @@ -0,0 +1,717 @@ +/* + * jcmarker.c + * + * Copyright (C) 1991-1998, Thomas G. Lane. + * Modified 2003-2019 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains routines to write JPEG datastream markers. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +typedef enum { /* JPEG marker codes */ + M_SOF0 = 0xc0, + M_SOF1 = 0xc1, + M_SOF2 = 0xc2, + M_SOF3 = 0xc3, + + M_SOF5 = 0xc5, + M_SOF6 = 0xc6, + M_SOF7 = 0xc7, + + M_JPG = 0xc8, + M_SOF9 = 0xc9, + M_SOF10 = 0xca, + M_SOF11 = 0xcb, + + M_SOF13 = 0xcd, + M_SOF14 = 0xce, + M_SOF15 = 0xcf, + + M_DHT = 0xc4, + + M_DAC = 0xcc, + + M_RST0 = 0xd0, + M_RST1 = 0xd1, + M_RST2 = 0xd2, + M_RST3 = 0xd3, + M_RST4 = 0xd4, + M_RST5 = 0xd5, + M_RST6 = 0xd6, + M_RST7 = 0xd7, + + M_SOI = 0xd8, + M_EOI = 0xd9, + M_SOS = 0xda, + M_DQT = 0xdb, + M_DNL = 0xdc, + M_DRI = 0xdd, + M_DHP = 0xde, + M_EXP = 0xdf, + + M_APP0 = 0xe0, + M_APP1 = 0xe1, + M_APP2 = 0xe2, + M_APP3 = 0xe3, + M_APP4 = 0xe4, + M_APP5 = 0xe5, + M_APP6 = 0xe6, + M_APP7 = 0xe7, + M_APP8 = 0xe8, + M_APP9 = 0xe9, + M_APP10 = 0xea, + M_APP11 = 0xeb, + M_APP12 = 0xec, + M_APP13 = 0xed, + M_APP14 = 0xee, + M_APP15 = 0xef, + + M_JPG0 = 0xf0, + M_JPG8 = 0xf8, + M_JPG13 = 0xfd, + M_COM = 0xfe, + + M_TEM = 0x01, + + M_ERROR = 0x100 +} JPEG_MARKER; + + +/* Private state */ + +typedef struct { + struct jpeg_marker_writer pub; /* public fields */ + + unsigned int last_restart_interval; /* last DRI value emitted; 0 after SOI */ +} my_marker_writer; + +typedef my_marker_writer * my_marker_ptr; + + +/* + * Basic output routines. + * + * Note that we do not support suspension while writing a marker. + * Therefore, an application using suspension must ensure that there is + * enough buffer space for the initial markers (typ. 600-700 bytes) before + * calling jpeg_start_compress, and enough space to write the trailing EOI + * (a few bytes) before calling jpeg_finish_compress. Multipass compression + * modes are not supported at all with suspension, so those two are the only + * points where markers will be written. + */ + +LOCAL(void) +emit_byte (j_compress_ptr cinfo, int val) +/* Emit a byte */ +{ + struct jpeg_destination_mgr * dest = cinfo->dest; + + *(dest->next_output_byte)++ = (JOCTET) val; + if (--dest->free_in_buffer == 0) { + if (! (*dest->empty_output_buffer) (cinfo)) + ERREXIT(cinfo, JERR_CANT_SUSPEND); + } +} + + +LOCAL(void) +emit_marker (j_compress_ptr cinfo, JPEG_MARKER mark) +/* Emit a marker code */ +{ + emit_byte(cinfo, 0xFF); + emit_byte(cinfo, (int) mark); +} + + +LOCAL(void) +emit_2bytes (j_compress_ptr cinfo, int value) +/* Emit a 2-byte integer; these are always MSB first in JPEG files */ +{ + emit_byte(cinfo, (value >> 8) & 0xFF); + emit_byte(cinfo, value & 0xFF); +} + + +/* + * Routines to write specific marker types. + */ + +LOCAL(int) +emit_dqt (j_compress_ptr cinfo, int index) +/* Emit a DQT marker */ +/* Returns the precision used (0 = 8bits, 1 = 16bits) for baseline checking */ +{ + JQUANT_TBL * qtbl = cinfo->quant_tbl_ptrs[index]; + int prec; + int i; + + if (qtbl == NULL) + ERREXIT1(cinfo, JERR_NO_QUANT_TABLE, index); + + prec = 0; + for (i = 0; i <= cinfo->lim_Se; i++) { + if (qtbl->quantval[cinfo->natural_order[i]] > 255) + prec = 1; + } + + if (! qtbl->sent_table) { + emit_marker(cinfo, M_DQT); + + emit_2bytes(cinfo, + prec ? cinfo->lim_Se * 2 + 2 + 1 + 2 : cinfo->lim_Se + 1 + 1 + 2); + + emit_byte(cinfo, index + (prec<<4)); + + for (i = 0; i <= cinfo->lim_Se; i++) { + /* The table entries must be emitted in zigzag order. */ + unsigned int qval = qtbl->quantval[cinfo->natural_order[i]]; + if (prec) + emit_byte(cinfo, (int) (qval >> 8)); + emit_byte(cinfo, (int) (qval & 0xFF)); + } + + qtbl->sent_table = TRUE; + } + + return prec; +} + + +LOCAL(void) +emit_dht (j_compress_ptr cinfo, int index, boolean is_ac) +/* Emit a DHT marker */ +{ + JHUFF_TBL * htbl; + int length, i; + + if (is_ac) { + htbl = cinfo->ac_huff_tbl_ptrs[index]; + index += 0x10; /* output index has AC bit set */ + } else { + htbl = cinfo->dc_huff_tbl_ptrs[index]; + } + + if (htbl == NULL) + ERREXIT1(cinfo, JERR_NO_HUFF_TABLE, index); + + if (! htbl->sent_table) { + emit_marker(cinfo, M_DHT); + + length = 0; + for (i = 1; i <= 16; i++) + length += htbl->bits[i]; + + emit_2bytes(cinfo, length + 2 + 1 + 16); + emit_byte(cinfo, index); + + for (i = 1; i <= 16; i++) + emit_byte(cinfo, htbl->bits[i]); + + for (i = 0; i < length; i++) + emit_byte(cinfo, htbl->huffval[i]); + + htbl->sent_table = TRUE; + } +} + + +LOCAL(void) +emit_dac (j_compress_ptr cinfo) +/* Emit a DAC marker */ +/* Since the useful info is so small, we want to emit all the tables in */ +/* one DAC marker. Therefore this routine does its own scan of the table. */ +{ +#ifdef C_ARITH_CODING_SUPPORTED + char dc_in_use[NUM_ARITH_TBLS]; + char ac_in_use[NUM_ARITH_TBLS]; + int length, i; + jpeg_component_info *compptr; + + for (i = 0; i < NUM_ARITH_TBLS; i++) + dc_in_use[i] = ac_in_use[i] = 0; + + for (i = 0; i < cinfo->comps_in_scan; i++) { + compptr = cinfo->cur_comp_info[i]; + /* DC needs no table for refinement scan */ + if (cinfo->Ss == 0 && cinfo->Ah == 0) + dc_in_use[compptr->dc_tbl_no] = 1; + /* AC needs no table when not present */ + if (cinfo->Se) + ac_in_use[compptr->ac_tbl_no] = 1; + } + + length = 0; + for (i = 0; i < NUM_ARITH_TBLS; i++) + length += dc_in_use[i] + ac_in_use[i]; + + if (length) { + emit_marker(cinfo, M_DAC); + + emit_2bytes(cinfo, length*2 + 2); + + for (i = 0; i < NUM_ARITH_TBLS; i++) { + if (dc_in_use[i]) { + emit_byte(cinfo, i); + emit_byte(cinfo, cinfo->arith_dc_L[i] + (cinfo->arith_dc_U[i]<<4)); + } + if (ac_in_use[i]) { + emit_byte(cinfo, i + 0x10); + emit_byte(cinfo, cinfo->arith_ac_K[i]); + } + } + } +#endif /* C_ARITH_CODING_SUPPORTED */ +} + + +LOCAL(void) +emit_dri (j_compress_ptr cinfo) +/* Emit a DRI marker */ +{ + emit_marker(cinfo, M_DRI); + + emit_2bytes(cinfo, 4); /* fixed length */ + + emit_2bytes(cinfo, (int) cinfo->restart_interval); +} + + +LOCAL(void) +emit_lse_ict (j_compress_ptr cinfo) +/* Emit an LSE inverse color transform specification marker */ +{ + /* Support only 1 transform */ + if (cinfo->color_transform != JCT_SUBTRACT_GREEN || + cinfo->num_components < 3) + ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); + + emit_marker(cinfo, M_JPG8); + + emit_2bytes(cinfo, 24); /* fixed length */ + + emit_byte(cinfo, 0x0D); /* ID inverse transform specification */ + emit_2bytes(cinfo, MAXJSAMPLE); /* MAXTRANS */ + emit_byte(cinfo, 3); /* Nt=3 */ + emit_byte(cinfo, cinfo->comp_info[1].component_id); + emit_byte(cinfo, cinfo->comp_info[0].component_id); + emit_byte(cinfo, cinfo->comp_info[2].component_id); + emit_byte(cinfo, 0x80); /* F1: CENTER1=1, NORM1=0 */ + emit_2bytes(cinfo, 0); /* A(1,1)=0 */ + emit_2bytes(cinfo, 0); /* A(1,2)=0 */ + emit_byte(cinfo, 0); /* F2: CENTER2=0, NORM2=0 */ + emit_2bytes(cinfo, 1); /* A(2,1)=1 */ + emit_2bytes(cinfo, 0); /* A(2,2)=0 */ + emit_byte(cinfo, 0); /* F3: CENTER3=0, NORM3=0 */ + emit_2bytes(cinfo, 1); /* A(3,1)=1 */ + emit_2bytes(cinfo, 0); /* A(3,2)=0 */ +} + + +LOCAL(void) +emit_sof (j_compress_ptr cinfo, JPEG_MARKER code) +/* Emit a SOF marker */ +{ + int ci; + jpeg_component_info *compptr; + + emit_marker(cinfo, code); + + emit_2bytes(cinfo, 3 * cinfo->num_components + 2 + 5 + 1); /* length */ + + /* Make sure image isn't bigger than SOF field can handle */ + if ((long) cinfo->jpeg_height > 65535L || + (long) cinfo->jpeg_width > 65535L) + ERREXIT1(cinfo, JERR_IMAGE_TOO_BIG, (unsigned int) 65535); + + emit_byte(cinfo, cinfo->data_precision); + emit_2bytes(cinfo, (int) cinfo->jpeg_height); + emit_2bytes(cinfo, (int) cinfo->jpeg_width); + + emit_byte(cinfo, cinfo->num_components); + + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + emit_byte(cinfo, compptr->component_id); + emit_byte(cinfo, (compptr->h_samp_factor << 4) + compptr->v_samp_factor); + emit_byte(cinfo, compptr->quant_tbl_no); + } +} + + +LOCAL(void) +emit_sos (j_compress_ptr cinfo) +/* Emit a SOS marker */ +{ + int i, td, ta; + jpeg_component_info *compptr; + + emit_marker(cinfo, M_SOS); + + emit_2bytes(cinfo, 2 * cinfo->comps_in_scan + 2 + 1 + 3); /* length */ + + emit_byte(cinfo, cinfo->comps_in_scan); + + for (i = 0; i < cinfo->comps_in_scan; i++) { + compptr = cinfo->cur_comp_info[i]; + emit_byte(cinfo, compptr->component_id); + + /* We emit 0 for unused field(s); this is recommended by the P&M text + * but does not seem to be specified in the standard. + */ + + /* DC needs no table for refinement scan */ + td = cinfo->Ss == 0 && cinfo->Ah == 0 ? compptr->dc_tbl_no : 0; + /* AC needs no table when not present */ + ta = cinfo->Se ? compptr->ac_tbl_no : 0; + + emit_byte(cinfo, (td << 4) + ta); + } + + emit_byte(cinfo, cinfo->Ss); + emit_byte(cinfo, cinfo->Se); + emit_byte(cinfo, (cinfo->Ah << 4) + cinfo->Al); +} + + +LOCAL(void) +emit_pseudo_sos (j_compress_ptr cinfo) +/* Emit a pseudo SOS marker */ +{ + emit_marker(cinfo, M_SOS); + + emit_2bytes(cinfo, 2 + 1 + 3); /* length */ + + emit_byte(cinfo, 0); /* Ns */ + + emit_byte(cinfo, 0); /* Ss */ + emit_byte(cinfo, cinfo->block_size * cinfo->block_size - 1); /* Se */ + emit_byte(cinfo, 0); /* Ah/Al */ +} + + +LOCAL(void) +emit_jfif_app0 (j_compress_ptr cinfo) +/* Emit a JFIF-compliant APP0 marker */ +{ + /* + * Length of APP0 block (2 bytes) + * Block ID (4 bytes - ASCII "JFIF") + * Zero byte (1 byte to terminate the ID string) + * Version Major, Minor (2 bytes - major first) + * Units (1 byte - 0x00 = none, 0x01 = inch, 0x02 = cm) + * Xdpu (2 bytes - dots per unit horizontal) + * Ydpu (2 bytes - dots per unit vertical) + * Thumbnail X size (1 byte) + * Thumbnail Y size (1 byte) + */ + + emit_marker(cinfo, M_APP0); + + emit_2bytes(cinfo, 2 + 4 + 1 + 2 + 1 + 2 + 2 + 1 + 1); /* length */ + + emit_byte(cinfo, 0x4A); /* Identifier: ASCII "JFIF" */ + emit_byte(cinfo, 0x46); + emit_byte(cinfo, 0x49); + emit_byte(cinfo, 0x46); + emit_byte(cinfo, 0); + emit_byte(cinfo, cinfo->JFIF_major_version); /* Version fields */ + emit_byte(cinfo, cinfo->JFIF_minor_version); + emit_byte(cinfo, cinfo->density_unit); /* Pixel size information */ + emit_2bytes(cinfo, (int) cinfo->X_density); + emit_2bytes(cinfo, (int) cinfo->Y_density); + emit_byte(cinfo, 0); /* No thumbnail image */ + emit_byte(cinfo, 0); +} + + +LOCAL(void) +emit_adobe_app14 (j_compress_ptr cinfo) +/* Emit an Adobe APP14 marker */ +{ + /* + * Length of APP14 block (2 bytes) + * Block ID (5 bytes - ASCII "Adobe") + * Version Number (2 bytes - currently 100) + * Flags0 (2 bytes - currently 0) + * Flags1 (2 bytes - currently 0) + * Color transform (1 byte) + * + * Although Adobe TN 5116 mentions Version = 101, all the Adobe files + * now in circulation seem to use Version = 100, so that's what we write. + * + * We write the color transform byte as 1 if the JPEG color space is + * YCbCr, 2 if it's YCCK, 0 otherwise. Adobe's definition has to do with + * whether the encoder performed a transformation, which is pretty useless. + */ + + emit_marker(cinfo, M_APP14); + + emit_2bytes(cinfo, 2 + 5 + 2 + 2 + 2 + 1); /* length */ + + emit_byte(cinfo, 0x41); /* Identifier: ASCII "Adobe" */ + emit_byte(cinfo, 0x64); + emit_byte(cinfo, 0x6F); + emit_byte(cinfo, 0x62); + emit_byte(cinfo, 0x65); + emit_2bytes(cinfo, 100); /* Version */ + emit_2bytes(cinfo, 0); /* Flags0 */ + emit_2bytes(cinfo, 0); /* Flags1 */ + switch (cinfo->jpeg_color_space) { + case JCS_YCbCr: + emit_byte(cinfo, 1); /* Color transform = 1 */ + break; + case JCS_YCCK: + emit_byte(cinfo, 2); /* Color transform = 2 */ + break; + default: + emit_byte(cinfo, 0); /* Color transform = 0 */ + } +} + + +/* + * These routines allow writing an arbitrary marker with parameters. + * The only intended use is to emit COM or APPn markers after calling + * write_file_header and before calling write_frame_header. + * Other uses are not guaranteed to produce desirable results. + * Counting the parameter bytes properly is the caller's responsibility. + */ + +METHODDEF(void) +write_marker_header (j_compress_ptr cinfo, int marker, unsigned int datalen) +/* Emit an arbitrary marker header */ +{ + if (datalen > (unsigned int) 65533) /* safety check */ + ERREXIT(cinfo, JERR_BAD_LENGTH); + + emit_marker(cinfo, (JPEG_MARKER) marker); + + emit_2bytes(cinfo, (int) (datalen + 2)); /* total length */ +} + +METHODDEF(void) +write_marker_byte (j_compress_ptr cinfo, int val) +/* Emit one byte of marker parameters following write_marker_header */ +{ + emit_byte(cinfo, val); +} + + +/* + * Write datastream header. + * This consists of an SOI and optional APPn markers. + * We recommend use of the JFIF marker, but not the Adobe marker, + * when using YCbCr or grayscale data. The JFIF marker is also used + * for other standard JPEG colorspaces. The Adobe marker is helpful + * to distinguish RGB, CMYK, and YCCK colorspaces. + * Note that an application can write additional header markers after + * jpeg_start_compress returns. + */ + +METHODDEF(void) +write_file_header (j_compress_ptr cinfo) +{ + my_marker_ptr marker = (my_marker_ptr) cinfo->marker; + + emit_marker(cinfo, M_SOI); /* first the SOI */ + + /* SOI is defined to reset restart interval to 0 */ + marker->last_restart_interval = 0; + + if (cinfo->write_JFIF_header) /* next an optional JFIF APP0 */ + emit_jfif_app0(cinfo); + if (cinfo->write_Adobe_marker) /* next an optional Adobe APP14 */ + emit_adobe_app14(cinfo); +} + + +/* + * Write frame header. + * This consists of DQT and SOFn markers, + * a conditional LSE marker and a conditional pseudo SOS marker. + * Note that we do not emit the SOF until we have emitted the DQT(s). + * This avoids compatibility problems with incorrect implementations that + * try to error-check the quant table numbers as soon as they see the SOF. + */ + +METHODDEF(void) +write_frame_header (j_compress_ptr cinfo) +{ + int ci, prec; + boolean is_baseline; + jpeg_component_info *compptr; + + /* Emit DQT for each quantization table. + * Note that emit_dqt() suppresses any duplicate tables. + */ + prec = 0; + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + prec += emit_dqt(cinfo, compptr->quant_tbl_no); + } + /* now prec is nonzero iff there are any 16-bit quant tables. */ + + /* Check for a non-baseline specification. + * Note we assume that Huffman table numbers won't be changed later. + */ + if (cinfo->arith_code || cinfo->progressive_mode || + cinfo->data_precision != 8 || cinfo->block_size != DCTSIZE) { + is_baseline = FALSE; + } else { + is_baseline = TRUE; + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + if (compptr->dc_tbl_no > 1 || compptr->ac_tbl_no > 1) + is_baseline = FALSE; + } + if (prec && is_baseline) { + is_baseline = FALSE; + /* If it's baseline except for quantizer size, warn the user */ + TRACEMS(cinfo, 0, JTRC_16BIT_TABLES); + } + } + + /* Emit the proper SOF marker */ + if (cinfo->arith_code) { + if (cinfo->progressive_mode) + emit_sof(cinfo, M_SOF10); /* SOF code for progressive arithmetic */ + else + emit_sof(cinfo, M_SOF9); /* SOF code for sequential arithmetic */ + } else { + if (cinfo->progressive_mode) + emit_sof(cinfo, M_SOF2); /* SOF code for progressive Huffman */ + else if (is_baseline) + emit_sof(cinfo, M_SOF0); /* SOF code for baseline implementation */ + else + emit_sof(cinfo, M_SOF1); /* SOF code for non-baseline Huffman file */ + } + + /* Check to emit LSE inverse color transform specification marker */ + if (cinfo->color_transform) + emit_lse_ict(cinfo); + + /* Check to emit pseudo SOS marker */ + if (cinfo->progressive_mode && cinfo->block_size != DCTSIZE) + emit_pseudo_sos(cinfo); +} + + +/* + * Write scan header. + * This consists of DHT or DAC markers, optional DRI, and SOS. + * Compressed data will be written following the SOS. + */ + +METHODDEF(void) +write_scan_header (j_compress_ptr cinfo) +{ + my_marker_ptr marker = (my_marker_ptr) cinfo->marker; + int i; + jpeg_component_info *compptr; + + if (cinfo->arith_code) { + /* Emit arith conditioning info. We may have some duplication + * if the file has multiple scans, but it's so small it's hardly + * worth worrying about. + */ + emit_dac(cinfo); + } else { + /* Emit Huffman tables. + * Note that emit_dht() suppresses any duplicate tables. + */ + for (i = 0; i < cinfo->comps_in_scan; i++) { + compptr = cinfo->cur_comp_info[i]; + /* DC needs no table for refinement scan */ + if (cinfo->Ss == 0 && cinfo->Ah == 0) + emit_dht(cinfo, compptr->dc_tbl_no, FALSE); + /* AC needs no table when not present */ + if (cinfo->Se) + emit_dht(cinfo, compptr->ac_tbl_no, TRUE); + } + } + + /* Emit DRI if required --- note that DRI value could change for each scan. + * We avoid wasting space with unnecessary DRIs, however. + */ + if (cinfo->restart_interval != marker->last_restart_interval) { + emit_dri(cinfo); + marker->last_restart_interval = cinfo->restart_interval; + } + + emit_sos(cinfo); +} + + +/* + * Write datastream trailer. + */ + +METHODDEF(void) +write_file_trailer (j_compress_ptr cinfo) +{ + emit_marker(cinfo, M_EOI); +} + + +/* + * Write an abbreviated table-specification datastream. + * This consists of SOI, DQT and DHT tables, and EOI. + * Any table that is defined and not marked sent_table = TRUE will be + * emitted. Note that all tables will be marked sent_table = TRUE at exit. + */ + +METHODDEF(void) +write_tables_only (j_compress_ptr cinfo) +{ + int i; + + emit_marker(cinfo, M_SOI); + + for (i = 0; i < NUM_QUANT_TBLS; i++) { + if (cinfo->quant_tbl_ptrs[i] != NULL) + (void) emit_dqt(cinfo, i); + } + + if (! cinfo->arith_code) { + for (i = 0; i < NUM_HUFF_TBLS; i++) { + if (cinfo->dc_huff_tbl_ptrs[i] != NULL) + emit_dht(cinfo, i, FALSE); + if (cinfo->ac_huff_tbl_ptrs[i] != NULL) + emit_dht(cinfo, i, TRUE); + } + } + + emit_marker(cinfo, M_EOI); +} + + +/* + * Initialize the marker writer module. + */ + +GLOBAL(void) +jinit_marker_writer (j_compress_ptr cinfo) +{ + my_marker_ptr marker; + + /* Create the subobject */ + marker = (my_marker_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_marker_writer)); + cinfo->marker = &marker->pub; + /* Initialize method pointers */ + marker->pub.write_file_header = write_file_header; + marker->pub.write_frame_header = write_frame_header; + marker->pub.write_scan_header = write_scan_header; + marker->pub.write_file_trailer = write_file_trailer; + marker->pub.write_tables_only = write_tables_only; + marker->pub.write_marker_header = write_marker_header; + marker->pub.write_marker_byte = write_marker_byte; + /* Initialize private state */ + marker->last_restart_interval = 0; +} diff --git a/cmake/externals/libjpeg/jcmaster.c b/cmake/externals/libjpeg/jcmaster.c new file mode 100644 index 000000000..a70af0c02 --- /dev/null +++ b/cmake/externals/libjpeg/jcmaster.c @@ -0,0 +1,675 @@ +/* + * jcmaster.c + * + * Copyright (C) 1991-1997, Thomas G. Lane. + * Modified 2003-2020 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains master control logic for the JPEG compressor. + * These routines are concerned with parameter validation, initial setup, + * and inter-pass control (determining the number of passes and the work + * to be done in each pass). + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* Private state */ + +typedef enum { + main_pass, /* input data, also do first output step */ + huff_opt_pass, /* Huffman code optimization pass */ + output_pass /* data output pass */ +} c_pass_type; + +typedef struct { + struct jpeg_comp_master pub; /* public fields */ + + c_pass_type pass_type; /* the type of the current pass */ + + int pass_number; /* # of passes completed */ + int total_passes; /* total # of passes needed */ + + int scan_number; /* current index in scan_info[] */ +} my_comp_master; + +typedef my_comp_master * my_master_ptr; + + +/* + * Support routines that do various essential calculations. + */ + +LOCAL(void) +initial_setup (j_compress_ptr cinfo) +/* Do computations that are needed before master selection phase */ +{ + int ci, ssize; + jpeg_component_info *compptr; + + /* Sanity check on block_size */ + if (cinfo->block_size < 1 || cinfo->block_size > 16) + ERREXIT2(cinfo, JERR_BAD_DCTSIZE, cinfo->block_size, cinfo->block_size); + + /* Derive natural_order from block_size */ + switch (cinfo->block_size) { + case 2: cinfo->natural_order = jpeg_natural_order2; break; + case 3: cinfo->natural_order = jpeg_natural_order3; break; + case 4: cinfo->natural_order = jpeg_natural_order4; break; + case 5: cinfo->natural_order = jpeg_natural_order5; break; + case 6: cinfo->natural_order = jpeg_natural_order6; break; + case 7: cinfo->natural_order = jpeg_natural_order7; break; + default: cinfo->natural_order = jpeg_natural_order; + } + + /* Derive lim_Se from block_size */ + cinfo->lim_Se = cinfo->block_size < DCTSIZE ? + cinfo->block_size * cinfo->block_size - 1 : DCTSIZE2-1; + + /* Sanity check on image dimensions */ + if (cinfo->jpeg_height <= 0 || cinfo->jpeg_width <= 0 || + cinfo->num_components <= 0) + ERREXIT(cinfo, JERR_EMPTY_IMAGE); + + /* Make sure image isn't bigger than I can handle */ + if ((long) cinfo->jpeg_height > (long) JPEG_MAX_DIMENSION || + (long) cinfo->jpeg_width > (long) JPEG_MAX_DIMENSION) + ERREXIT1(cinfo, JERR_IMAGE_TOO_BIG, (unsigned int) JPEG_MAX_DIMENSION); + + /* Only 8 to 12 bits data precision are supported for DCT based JPEG */ + if (cinfo->data_precision < 8 || cinfo->data_precision > 12) + ERREXIT1(cinfo, JERR_BAD_PRECISION, cinfo->data_precision); + + /* Check that number of components won't exceed internal array sizes */ + if (cinfo->num_components > MAX_COMPONENTS) + ERREXIT2(cinfo, JERR_COMPONENT_COUNT, cinfo->num_components, + MAX_COMPONENTS); + + /* Compute maximum sampling factors; check factor validity */ + cinfo->max_h_samp_factor = 1; + cinfo->max_v_samp_factor = 1; + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + if (compptr->h_samp_factor<=0 || compptr->h_samp_factor>MAX_SAMP_FACTOR || + compptr->v_samp_factor<=0 || compptr->v_samp_factor>MAX_SAMP_FACTOR) + ERREXIT(cinfo, JERR_BAD_SAMPLING); + cinfo->max_h_samp_factor = MAX(cinfo->max_h_samp_factor, + compptr->h_samp_factor); + cinfo->max_v_samp_factor = MAX(cinfo->max_v_samp_factor, + compptr->v_samp_factor); + } + + /* Compute dimensions of components */ + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + /* Fill in the correct component_index value; don't rely on application */ + compptr->component_index = ci; + /* In selecting the actual DCT scaling for each component, we try to + * scale down the chroma components via DCT scaling rather than downsampling. + * This saves time if the downsampler gets to use 1:1 scaling. + * Note this code adapts subsampling ratios which are powers of 2. + */ + ssize = 1; +#ifdef DCT_SCALING_SUPPORTED + if (! cinfo->raw_data_in) + while (cinfo->min_DCT_h_scaled_size * ssize <= + (cinfo->do_fancy_downsampling ? DCTSIZE : DCTSIZE / 2) && + (cinfo->max_h_samp_factor % (compptr->h_samp_factor * ssize * 2)) == + 0) { + ssize = ssize * 2; + } +#endif + compptr->DCT_h_scaled_size = cinfo->min_DCT_h_scaled_size * ssize; + ssize = 1; +#ifdef DCT_SCALING_SUPPORTED + if (! cinfo->raw_data_in) + while (cinfo->min_DCT_v_scaled_size * ssize <= + (cinfo->do_fancy_downsampling ? DCTSIZE : DCTSIZE / 2) && + (cinfo->max_v_samp_factor % (compptr->v_samp_factor * ssize * 2)) == + 0) { + ssize = ssize * 2; + } +#endif + compptr->DCT_v_scaled_size = cinfo->min_DCT_v_scaled_size * ssize; + + /* We don't support DCT ratios larger than 2. */ + if (compptr->DCT_h_scaled_size > compptr->DCT_v_scaled_size * 2) + compptr->DCT_h_scaled_size = compptr->DCT_v_scaled_size * 2; + else if (compptr->DCT_v_scaled_size > compptr->DCT_h_scaled_size * 2) + compptr->DCT_v_scaled_size = compptr->DCT_h_scaled_size * 2; + + /* Size in DCT blocks */ + compptr->width_in_blocks = (JDIMENSION) + jdiv_round_up((long) cinfo->jpeg_width * (long) compptr->h_samp_factor, + (long) (cinfo->max_h_samp_factor * cinfo->block_size)); + compptr->height_in_blocks = (JDIMENSION) + jdiv_round_up((long) cinfo->jpeg_height * (long) compptr->v_samp_factor, + (long) (cinfo->max_v_samp_factor * cinfo->block_size)); + /* Size in samples */ + compptr->downsampled_width = (JDIMENSION) + jdiv_round_up((long) cinfo->jpeg_width * + (long) (compptr->h_samp_factor * compptr->DCT_h_scaled_size), + (long) (cinfo->max_h_samp_factor * cinfo->block_size)); + compptr->downsampled_height = (JDIMENSION) + jdiv_round_up((long) cinfo->jpeg_height * + (long) (compptr->v_samp_factor * compptr->DCT_v_scaled_size), + (long) (cinfo->max_v_samp_factor * cinfo->block_size)); + /* Don't need quantization scale after DCT, + * until color conversion says otherwise. + */ + compptr->component_needed = FALSE; + } + + /* Compute number of fully interleaved MCU rows (number of times that + * main controller will call coefficient controller). + */ + cinfo->total_iMCU_rows = (JDIMENSION) + jdiv_round_up((long) cinfo->jpeg_height, + (long) (cinfo->max_v_samp_factor * cinfo->block_size)); +} + + +#ifdef C_MULTISCAN_FILES_SUPPORTED + +LOCAL(void) +validate_script (j_compress_ptr cinfo) +/* Verify that the scan script in cinfo->scan_info[] is valid; also + * determine whether it uses progressive JPEG, and set cinfo->progressive_mode. + */ +{ + const jpeg_scan_info * scanptr; + int scanno, ncomps, ci, coefi, thisi; + int Ss, Se, Ah, Al; + boolean component_sent[MAX_COMPONENTS]; +#ifdef C_PROGRESSIVE_SUPPORTED + int * last_bitpos_ptr; + int last_bitpos[MAX_COMPONENTS][DCTSIZE2]; + /* -1 until that coefficient has been seen; then last Al for it */ +#endif + + if (cinfo->num_scans <= 0) + ERREXIT1(cinfo, JERR_BAD_SCAN_SCRIPT, 0); + + /* For sequential JPEG, all scans must have Ss=0, Se=DCTSIZE2-1; + * for progressive JPEG, no scan can have this. + */ + scanptr = cinfo->scan_info; + if (scanptr->Ss != 0 || scanptr->Se != DCTSIZE2-1) { +#ifdef C_PROGRESSIVE_SUPPORTED + cinfo->progressive_mode = TRUE; + last_bitpos_ptr = & last_bitpos[0][0]; + for (ci = 0; ci < cinfo->num_components; ci++) + for (coefi = 0; coefi < DCTSIZE2; coefi++) + *last_bitpos_ptr++ = -1; +#else + ERREXIT(cinfo, JERR_NOT_COMPILED); +#endif + } else { + cinfo->progressive_mode = FALSE; + for (ci = 0; ci < cinfo->num_components; ci++) + component_sent[ci] = FALSE; + } + + for (scanno = 1; scanno <= cinfo->num_scans; scanptr++, scanno++) { + /* Validate component indexes */ + ncomps = scanptr->comps_in_scan; + if (ncomps <= 0 || ncomps > MAX_COMPS_IN_SCAN) + ERREXIT2(cinfo, JERR_COMPONENT_COUNT, ncomps, MAX_COMPS_IN_SCAN); + for (ci = 0; ci < ncomps; ci++) { + thisi = scanptr->component_index[ci]; + if (thisi < 0 || thisi >= cinfo->num_components) + ERREXIT1(cinfo, JERR_BAD_SCAN_SCRIPT, scanno); + /* Components must appear in SOF order within each scan */ + if (ci > 0 && thisi <= scanptr->component_index[ci-1]) + ERREXIT1(cinfo, JERR_BAD_SCAN_SCRIPT, scanno); + } + /* Validate progression parameters */ + Ss = scanptr->Ss; + Se = scanptr->Se; + Ah = scanptr->Ah; + Al = scanptr->Al; + if (cinfo->progressive_mode) { +#ifdef C_PROGRESSIVE_SUPPORTED + /* The JPEG spec simply gives the ranges 0..13 for Ah and Al, but that + * seems wrong: the upper bound ought to depend on data precision. + * Perhaps they really meant 0..N+1 for N-bit precision. + * Here we allow 0..10 for 8-bit data; Al larger than 10 results in + * out-of-range reconstructed DC values during the first DC scan, + * which might cause problems for some decoders. + */ + if (Ss < 0 || Ss >= DCTSIZE2 || Se < Ss || Se >= DCTSIZE2 || + Ah < 0 || Ah > (cinfo->data_precision > 8 ? 13 : 10) || + Al < 0 || Al > (cinfo->data_precision > 8 ? 13 : 10)) + ERREXIT1(cinfo, JERR_BAD_PROG_SCRIPT, scanno); + if (Ss == 0) { + if (Se != 0) /* DC and AC together not OK */ + ERREXIT1(cinfo, JERR_BAD_PROG_SCRIPT, scanno); + } else { + if (ncomps != 1) /* AC scans must be for only one component */ + ERREXIT1(cinfo, JERR_BAD_PROG_SCRIPT, scanno); + } + for (ci = 0; ci < ncomps; ci++) { + last_bitpos_ptr = & last_bitpos[scanptr->component_index[ci]][0]; + if (Ss != 0 && last_bitpos_ptr[0] < 0) /* AC without prior DC scan */ + ERREXIT1(cinfo, JERR_BAD_PROG_SCRIPT, scanno); + for (coefi = Ss; coefi <= Se; coefi++) { + if (last_bitpos_ptr[coefi] < 0) { + /* first scan of this coefficient */ + if (Ah != 0) + ERREXIT1(cinfo, JERR_BAD_PROG_SCRIPT, scanno); + } else { + /* not first scan */ + if (Ah != last_bitpos_ptr[coefi] || Al != Ah-1) + ERREXIT1(cinfo, JERR_BAD_PROG_SCRIPT, scanno); + } + last_bitpos_ptr[coefi] = Al; + } + } +#endif + } else { + /* For sequential JPEG, all progression parameters must be these: */ + if (Ss != 0 || Se != DCTSIZE2-1 || Ah != 0 || Al != 0) + ERREXIT1(cinfo, JERR_BAD_PROG_SCRIPT, scanno); + /* Make sure components are not sent twice */ + for (ci = 0; ci < ncomps; ci++) { + thisi = scanptr->component_index[ci]; + if (component_sent[thisi]) + ERREXIT1(cinfo, JERR_BAD_SCAN_SCRIPT, scanno); + component_sent[thisi] = TRUE; + } + } + } + + /* Now verify that everything got sent. */ + if (cinfo->progressive_mode) { +#ifdef C_PROGRESSIVE_SUPPORTED + /* For progressive mode, we only check that at least some DC data + * got sent for each component; the spec does not require that all bits + * of all coefficients be transmitted. Would it be wiser to enforce + * transmission of all coefficient bits?? + */ + for (ci = 0; ci < cinfo->num_components; ci++) { + if (last_bitpos[ci][0] < 0) + ERREXIT(cinfo, JERR_MISSING_DATA); + } +#endif + } else { + for (ci = 0; ci < cinfo->num_components; ci++) { + if (! component_sent[ci]) + ERREXIT(cinfo, JERR_MISSING_DATA); + } + } +} + + +LOCAL(void) +reduce_script (j_compress_ptr cinfo) +/* Adapt scan script for use with reduced block size; + * assume that script has been validated before. + */ +{ + jpeg_scan_info * scanptr; + int idxout, idxin; + + /* Circumvent const declaration for this function */ + scanptr = (jpeg_scan_info *) cinfo->scan_info; + idxout = 0; + + for (idxin = 0; idxin < cinfo->num_scans; idxin++) { + /* After skipping, idxout becomes smaller than idxin */ + if (idxin != idxout) + /* Copy rest of data; + * note we stay in given chunk of allocated memory. + */ + scanptr[idxout] = scanptr[idxin]; + if (scanptr[idxout].Ss > cinfo->lim_Se) + /* Entire scan out of range - skip this entry */ + continue; + if (scanptr[idxout].Se > cinfo->lim_Se) + /* Limit scan to end of block */ + scanptr[idxout].Se = cinfo->lim_Se; + idxout++; + } + + cinfo->num_scans = idxout; +} + +#endif /* C_MULTISCAN_FILES_SUPPORTED */ + + +LOCAL(void) +select_scan_parameters (j_compress_ptr cinfo) +/* Set up the scan parameters for the current scan */ +{ + int ci; + +#ifdef C_MULTISCAN_FILES_SUPPORTED + if (cinfo->scan_info != NULL) { + /* Prepare for current scan --- the script is already validated */ + my_master_ptr master = (my_master_ptr) cinfo->master; + const jpeg_scan_info * scanptr = cinfo->scan_info + master->scan_number; + + cinfo->comps_in_scan = scanptr->comps_in_scan; + for (ci = 0; ci < scanptr->comps_in_scan; ci++) { + cinfo->cur_comp_info[ci] = + &cinfo->comp_info[scanptr->component_index[ci]]; + } + if (cinfo->progressive_mode) { + cinfo->Ss = scanptr->Ss; + cinfo->Se = scanptr->Se; + cinfo->Ah = scanptr->Ah; + cinfo->Al = scanptr->Al; + return; + } + } + else +#endif + { + /* Prepare for single sequential-JPEG scan containing all components */ + if (cinfo->num_components > MAX_COMPS_IN_SCAN) + ERREXIT2(cinfo, JERR_COMPONENT_COUNT, cinfo->num_components, + MAX_COMPS_IN_SCAN); + cinfo->comps_in_scan = cinfo->num_components; + for (ci = 0; ci < cinfo->num_components; ci++) { + cinfo->cur_comp_info[ci] = &cinfo->comp_info[ci]; + } + } + cinfo->Ss = 0; + cinfo->Se = cinfo->block_size * cinfo->block_size - 1; + cinfo->Ah = 0; + cinfo->Al = 0; +} + + +LOCAL(void) +per_scan_setup (j_compress_ptr cinfo) +/* Do computations that are needed before processing a JPEG scan */ +/* cinfo->comps_in_scan and cinfo->cur_comp_info[] are already set */ +{ + int ci, mcublks, tmp; + jpeg_component_info *compptr; + + if (cinfo->comps_in_scan == 1) { + + /* Noninterleaved (single-component) scan */ + compptr = cinfo->cur_comp_info[0]; + + /* Overall image size in MCUs */ + cinfo->MCUs_per_row = compptr->width_in_blocks; + cinfo->MCU_rows_in_scan = compptr->height_in_blocks; + + /* For noninterleaved scan, always one block per MCU */ + compptr->MCU_width = 1; + compptr->MCU_height = 1; + compptr->MCU_blocks = 1; + compptr->MCU_sample_width = compptr->DCT_h_scaled_size; + compptr->last_col_width = 1; + /* For noninterleaved scans, it is convenient to define last_row_height + * as the number of block rows present in the last iMCU row. + */ + tmp = (int) (compptr->height_in_blocks % compptr->v_samp_factor); + if (tmp == 0) tmp = compptr->v_samp_factor; + compptr->last_row_height = tmp; + + /* Prepare array describing MCU composition */ + cinfo->blocks_in_MCU = 1; + cinfo->MCU_membership[0] = 0; + + } else { + + /* Interleaved (multi-component) scan */ + if (cinfo->comps_in_scan <= 0 || cinfo->comps_in_scan > MAX_COMPS_IN_SCAN) + ERREXIT2(cinfo, JERR_COMPONENT_COUNT, cinfo->comps_in_scan, + MAX_COMPS_IN_SCAN); + + /* Overall image size in MCUs */ + cinfo->MCUs_per_row = (JDIMENSION) + jdiv_round_up((long) cinfo->jpeg_width, + (long) (cinfo->max_h_samp_factor * cinfo->block_size)); + cinfo->MCU_rows_in_scan = cinfo->total_iMCU_rows; + + cinfo->blocks_in_MCU = 0; + + for (ci = 0; ci < cinfo->comps_in_scan; ci++) { + compptr = cinfo->cur_comp_info[ci]; + /* Sampling factors give # of blocks of component in each MCU */ + compptr->MCU_width = compptr->h_samp_factor; + compptr->MCU_height = compptr->v_samp_factor; + compptr->MCU_blocks = compptr->MCU_width * compptr->MCU_height; + compptr->MCU_sample_width = compptr->MCU_width * compptr->DCT_h_scaled_size; + /* Figure number of non-dummy blocks in last MCU column & row */ + tmp = (int) (compptr->width_in_blocks % compptr->MCU_width); + if (tmp == 0) tmp = compptr->MCU_width; + compptr->last_col_width = tmp; + tmp = (int) (compptr->height_in_blocks % compptr->MCU_height); + if (tmp == 0) tmp = compptr->MCU_height; + compptr->last_row_height = tmp; + /* Prepare array describing MCU composition */ + mcublks = compptr->MCU_blocks; + if (cinfo->blocks_in_MCU + mcublks > C_MAX_BLOCKS_IN_MCU) + ERREXIT(cinfo, JERR_BAD_MCU_SIZE); + while (mcublks-- > 0) { + cinfo->MCU_membership[cinfo->blocks_in_MCU++] = ci; + } + } + + } + + /* Convert restart specified in rows to actual MCU count. */ + /* Note that count must fit in 16 bits, so we provide limiting. */ + if (cinfo->restart_in_rows > 0) { + long nominal = (long) cinfo->restart_in_rows * (long) cinfo->MCUs_per_row; + cinfo->restart_interval = (unsigned int) MIN(nominal, 65535L); + } +} + + +/* + * Per-pass setup. + * This is called at the beginning of each pass. We determine which modules + * will be active during this pass and give them appropriate start_pass calls. + * We also set is_last_pass to indicate whether any more passes will be + * required. + */ + +METHODDEF(void) +prepare_for_pass (j_compress_ptr cinfo) +{ + my_master_ptr master = (my_master_ptr) cinfo->master; + + switch (master->pass_type) { + case main_pass: + /* Initial pass: will collect input data, and do either Huffman + * optimization or data output for the first scan. + */ + select_scan_parameters(cinfo); + per_scan_setup(cinfo); + if (! cinfo->raw_data_in) { + (*cinfo->cconvert->start_pass) (cinfo); + (*cinfo->downsample->start_pass) (cinfo); + (*cinfo->prep->start_pass) (cinfo, JBUF_PASS_THRU); + } + (*cinfo->fdct->start_pass) (cinfo); + (*cinfo->entropy->start_pass) (cinfo, cinfo->optimize_coding); + (*cinfo->coef->start_pass) (cinfo, + (master->total_passes > 1 ? + JBUF_SAVE_AND_PASS : JBUF_PASS_THRU)); + (*cinfo->main->start_pass) (cinfo, JBUF_PASS_THRU); + if (cinfo->optimize_coding) { + /* No immediate data output; postpone writing frame/scan headers */ + master->pub.call_pass_startup = FALSE; + } else { + /* Will write frame/scan headers at first jpeg_write_scanlines call */ + master->pub.call_pass_startup = TRUE; + } + break; +#ifdef ENTROPY_OPT_SUPPORTED + case huff_opt_pass: + /* Do Huffman optimization for a scan after the first one. */ + select_scan_parameters(cinfo); + per_scan_setup(cinfo); + if (cinfo->Ss != 0 || cinfo->Ah == 0) { + (*cinfo->entropy->start_pass) (cinfo, TRUE); + (*cinfo->coef->start_pass) (cinfo, JBUF_CRANK_DEST); + master->pub.call_pass_startup = FALSE; + break; + } + /* Special case: Huffman DC refinement scans need no Huffman table + * and therefore we can skip the optimization pass for them. + */ + master->pass_type = output_pass; + master->pass_number++; + /*FALLTHROUGH*/ +#endif + case output_pass: + /* Do a data-output pass. */ + /* We need not repeat per-scan setup if prior optimization pass did it. */ + if (! cinfo->optimize_coding) { + select_scan_parameters(cinfo); + per_scan_setup(cinfo); + } + (*cinfo->entropy->start_pass) (cinfo, FALSE); + (*cinfo->coef->start_pass) (cinfo, JBUF_CRANK_DEST); + /* We emit frame/scan headers now */ + if (master->scan_number == 0) + (*cinfo->marker->write_frame_header) (cinfo); + (*cinfo->marker->write_scan_header) (cinfo); + master->pub.call_pass_startup = FALSE; + break; + default: + ERREXIT(cinfo, JERR_NOT_COMPILED); + } + + master->pub.is_last_pass = (master->pass_number == master->total_passes-1); + + /* Set up progress monitor's pass info if present */ + if (cinfo->progress != NULL) { + cinfo->progress->completed_passes = master->pass_number; + cinfo->progress->total_passes = master->total_passes; + } +} + + +/* + * Special start-of-pass hook. + * This is called by jpeg_write_scanlines if call_pass_startup is TRUE. + * In single-pass processing, we need this hook because we don't want to + * write frame/scan headers during jpeg_start_compress; we want to let the + * application write COM markers etc. between jpeg_start_compress and the + * jpeg_write_scanlines loop. + * In multi-pass processing, this routine is not used. + */ + +METHODDEF(void) +pass_startup (j_compress_ptr cinfo) +{ + cinfo->master->call_pass_startup = FALSE; /* reset flag so call only once */ + + (*cinfo->marker->write_frame_header) (cinfo); + (*cinfo->marker->write_scan_header) (cinfo); +} + + +/* + * Finish up at end of pass. + */ + +METHODDEF(void) +finish_pass_master (j_compress_ptr cinfo) +{ + my_master_ptr master = (my_master_ptr) cinfo->master; + + /* The entropy coder always needs an end-of-pass call, + * either to analyze statistics or to flush its output buffer. + */ + (*cinfo->entropy->finish_pass) (cinfo); + + /* Update state for next pass */ + switch (master->pass_type) { + case main_pass: + /* next pass is either output of scan 0 (after optimization) + * or output of scan 1 (if no optimization). + */ + master->pass_type = output_pass; + if (! cinfo->optimize_coding) + master->scan_number++; + break; + case huff_opt_pass: + /* next pass is always output of current scan */ + master->pass_type = output_pass; + break; + case output_pass: + /* next pass is either optimization or output of next scan */ + if (cinfo->optimize_coding) + master->pass_type = huff_opt_pass; + master->scan_number++; + break; + } + + master->pass_number++; +} + + +/* + * Initialize master compression control. + */ + +GLOBAL(void) +jinit_c_master_control (j_compress_ptr cinfo, boolean transcode_only) +{ + my_master_ptr master; + + master = (my_master_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_comp_master)); + cinfo->master = &master->pub; + master->pub.prepare_for_pass = prepare_for_pass; + master->pub.pass_startup = pass_startup; + master->pub.finish_pass = finish_pass_master; + master->pub.is_last_pass = FALSE; + + /* Validate parameters, determine derived values */ + initial_setup(cinfo); + + if (cinfo->scan_info != NULL) { +#ifdef C_MULTISCAN_FILES_SUPPORTED + validate_script(cinfo); + if (cinfo->block_size < DCTSIZE) + reduce_script(cinfo); +#else + ERREXIT(cinfo, JERR_NOT_COMPILED); +#endif + } else { + cinfo->progressive_mode = FALSE; + cinfo->num_scans = 1; + } + + if (cinfo->optimize_coding) + cinfo->arith_code = FALSE; /* disable arithmetic coding */ + else if (! cinfo->arith_code && + (cinfo->progressive_mode || + (cinfo->block_size > 1 && cinfo->block_size < DCTSIZE))) + /* TEMPORARY HACK ??? */ + /* assume default tables no good for progressive or reduced AC mode */ + cinfo->optimize_coding = TRUE; /* force Huffman optimization */ + + /* Initialize my private state */ + if (transcode_only) { + /* no main pass in transcoding */ + if (cinfo->optimize_coding) + master->pass_type = huff_opt_pass; + else + master->pass_type = output_pass; + } else { + /* for normal compression, first pass is always this type: */ + master->pass_type = main_pass; + } + master->scan_number = 0; + master->pass_number = 0; + if (cinfo->optimize_coding) + master->total_passes = cinfo->num_scans * 2; + else + master->total_passes = cinfo->num_scans; +} diff --git a/cmake/externals/libjpeg/jcomapi.c b/cmake/externals/libjpeg/jcomapi.c new file mode 100644 index 000000000..678c5d12d --- /dev/null +++ b/cmake/externals/libjpeg/jcomapi.c @@ -0,0 +1,244 @@ +/* + * jcomapi.c + * + * Copyright (C) 1994-1997, Thomas G. Lane. + * Modified 2019 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains application interface routines that are used for both + * compression and decompression. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* + * Abort processing of a JPEG compression or decompression operation, + * but don't destroy the object itself. + * + * For this, we merely clean up all the nonpermanent memory pools. + * Note that temp files (virtual arrays) are not allowed to belong to + * the permanent pool, so we will be able to close all temp files here. + * Closing a data source or destination, if necessary, is the application's + * responsibility. + */ + +GLOBAL(void) +jpeg_abort (j_common_ptr cinfo) +{ + int pool; + + /* Do nothing if called on a not-initialized or destroyed JPEG object. */ + if (cinfo->mem == NULL) + return; + + /* Releasing pools in reverse order might help avoid fragmentation + * with some (brain-damaged) malloc libraries. + */ + for (pool = JPOOL_NUMPOOLS-1; pool > JPOOL_PERMANENT; pool--) { + (*cinfo->mem->free_pool) (cinfo, pool); + } + + /* Reset overall state for possible reuse of object */ + if (cinfo->is_decompressor) { + cinfo->global_state = DSTATE_START; + /* Try to keep application from accessing now-deleted marker list. + * A bit kludgy to do it here, but this is the most central place. + */ + ((j_decompress_ptr) cinfo)->marker_list = NULL; + } else { + cinfo->global_state = CSTATE_START; + } +} + + +/* + * Destruction of a JPEG object. + * + * Everything gets deallocated except the master jpeg_compress_struct itself + * and the error manager struct. Both of these are supplied by the application + * and must be freed, if necessary, by the application. (Often they are on + * the stack and so don't need to be freed anyway.) + * Closing a data source or destination, if necessary, is the application's + * responsibility. + */ + +GLOBAL(void) +jpeg_destroy (j_common_ptr cinfo) +{ + /* We need only tell the memory manager to release everything. */ + /* NB: mem pointer is NULL if memory mgr failed to initialize. */ + if (cinfo->mem != NULL) + (*cinfo->mem->self_destruct) (cinfo); + cinfo->mem = NULL; /* be safe if jpeg_destroy is called twice */ + cinfo->global_state = 0; /* mark it destroyed */ +} + + +/* + * Convenience routines for allocating quantization and Huffman tables. + * (Would jutils.c be a more reasonable place to put these?) + */ + +GLOBAL(JQUANT_TBL *) +jpeg_alloc_quant_table (j_common_ptr cinfo) +{ + JQUANT_TBL *tbl; + + tbl = (JQUANT_TBL *) + (*cinfo->mem->alloc_small) (cinfo, JPOOL_PERMANENT, SIZEOF(JQUANT_TBL)); + tbl->sent_table = FALSE; /* make sure this is false in any new table */ + return tbl; +} + + +GLOBAL(JHUFF_TBL *) +jpeg_alloc_huff_table (j_common_ptr cinfo) +{ + JHUFF_TBL *tbl; + + tbl = (JHUFF_TBL *) + (*cinfo->mem->alloc_small) (cinfo, JPOOL_PERMANENT, SIZEOF(JHUFF_TBL)); + tbl->sent_table = FALSE; /* make sure this is false in any new table */ + return tbl; +} + + +/* + * Set up the standard Huffman tables (cf. JPEG standard section K.3). + * IMPORTANT: these are only valid for 8-bit data precision! + * (Would jutils.c be a more reasonable place to put this?) + */ + +GLOBAL(JHUFF_TBL *) +jpeg_std_huff_table (j_common_ptr cinfo, boolean isDC, int tblno) +{ + JHUFF_TBL **htblptr, *htbl; + const UINT8 *bits, *val; + int nsymbols, len; + + static const UINT8 bits_dc_luminance[17] = + { /* 0-base */ 0, 0, 1, 5, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0 }; + static const UINT8 val_dc_luminance[] = + { 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 }; + + static const UINT8 bits_dc_chrominance[17] = + { /* 0-base */ 0, 0, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0 }; + static const UINT8 val_dc_chrominance[] = + { 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 }; + + static const UINT8 bits_ac_luminance[17] = + { /* 0-base */ 0, 0, 2, 1, 3, 3, 2, 4, 3, 5, 5, 4, 4, 0, 0, 1, 0x7d }; + static const UINT8 val_ac_luminance[] = + { 0x01, 0x02, 0x03, 0x00, 0x04, 0x11, 0x05, 0x12, + 0x21, 0x31, 0x41, 0x06, 0x13, 0x51, 0x61, 0x07, + 0x22, 0x71, 0x14, 0x32, 0x81, 0x91, 0xa1, 0x08, + 0x23, 0x42, 0xb1, 0xc1, 0x15, 0x52, 0xd1, 0xf0, + 0x24, 0x33, 0x62, 0x72, 0x82, 0x09, 0x0a, 0x16, + 0x17, 0x18, 0x19, 0x1a, 0x25, 0x26, 0x27, 0x28, + 0x29, 0x2a, 0x34, 0x35, 0x36, 0x37, 0x38, 0x39, + 0x3a, 0x43, 0x44, 0x45, 0x46, 0x47, 0x48, 0x49, + 0x4a, 0x53, 0x54, 0x55, 0x56, 0x57, 0x58, 0x59, + 0x5a, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, + 0x6a, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, + 0x7a, 0x83, 0x84, 0x85, 0x86, 0x87, 0x88, 0x89, + 0x8a, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, 0x98, + 0x99, 0x9a, 0xa2, 0xa3, 0xa4, 0xa5, 0xa6, 0xa7, + 0xa8, 0xa9, 0xaa, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, + 0xb7, 0xb8, 0xb9, 0xba, 0xc2, 0xc3, 0xc4, 0xc5, + 0xc6, 0xc7, 0xc8, 0xc9, 0xca, 0xd2, 0xd3, 0xd4, + 0xd5, 0xd6, 0xd7, 0xd8, 0xd9, 0xda, 0xe1, 0xe2, + 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, 0xe8, 0xe9, 0xea, + 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, 0xf8, + 0xf9, 0xfa }; + + static const UINT8 bits_ac_chrominance[17] = + { /* 0-base */ 0, 0, 2, 1, 2, 4, 4, 3, 4, 7, 5, 4, 4, 0, 1, 2, 0x77 }; + static const UINT8 val_ac_chrominance[] = + { 0x00, 0x01, 0x02, 0x03, 0x11, 0x04, 0x05, 0x21, + 0x31, 0x06, 0x12, 0x41, 0x51, 0x07, 0x61, 0x71, + 0x13, 0x22, 0x32, 0x81, 0x08, 0x14, 0x42, 0x91, + 0xa1, 0xb1, 0xc1, 0x09, 0x23, 0x33, 0x52, 0xf0, + 0x15, 0x62, 0x72, 0xd1, 0x0a, 0x16, 0x24, 0x34, + 0xe1, 0x25, 0xf1, 0x17, 0x18, 0x19, 0x1a, 0x26, + 0x27, 0x28, 0x29, 0x2a, 0x35, 0x36, 0x37, 0x38, + 0x39, 0x3a, 0x43, 0x44, 0x45, 0x46, 0x47, 0x48, + 0x49, 0x4a, 0x53, 0x54, 0x55, 0x56, 0x57, 0x58, + 0x59, 0x5a, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, + 0x69, 0x6a, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, + 0x79, 0x7a, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, + 0x88, 0x89, 0x8a, 0x92, 0x93, 0x94, 0x95, 0x96, + 0x97, 0x98, 0x99, 0x9a, 0xa2, 0xa3, 0xa4, 0xa5, + 0xa6, 0xa7, 0xa8, 0xa9, 0xaa, 0xb2, 0xb3, 0xb4, + 0xb5, 0xb6, 0xb7, 0xb8, 0xb9, 0xba, 0xc2, 0xc3, + 0xc4, 0xc5, 0xc6, 0xc7, 0xc8, 0xc9, 0xca, 0xd2, + 0xd3, 0xd4, 0xd5, 0xd6, 0xd7, 0xd8, 0xd9, 0xda, + 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, 0xe8, 0xe9, + 0xea, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, 0xf8, + 0xf9, 0xfa }; + + if (cinfo->is_decompressor) { + if (isDC) + htblptr = ((j_decompress_ptr) cinfo)->dc_huff_tbl_ptrs; + else + htblptr = ((j_decompress_ptr) cinfo)->ac_huff_tbl_ptrs; + } else { + if (isDC) + htblptr = ((j_compress_ptr) cinfo)->dc_huff_tbl_ptrs; + else + htblptr = ((j_compress_ptr) cinfo)->ac_huff_tbl_ptrs; + } + + switch (tblno) { + case 0: + if (isDC) { + bits = bits_dc_luminance; + val = val_dc_luminance; + } else { + bits = bits_ac_luminance; + val = val_ac_luminance; + } + break; + case 1: + if (isDC) { + bits = bits_dc_chrominance; + val = val_dc_chrominance; + } else { + bits = bits_ac_chrominance; + val = val_ac_chrominance; + } + break; + default: + ERREXIT1(cinfo, JERR_NO_HUFF_TABLE, tblno); + return NULL; /* avoid compiler warnings for uninitialized variables */ + } + + if (htblptr[tblno] == NULL) + htblptr[tblno] = jpeg_alloc_huff_table(cinfo); + + htbl = htblptr[tblno]; + + /* Copy the number-of-symbols-of-each-code-length counts */ + MEMCOPY(htbl->bits, bits, SIZEOF(htbl->bits)); + + /* Validate the counts. We do this here mainly so we can copy the right + * number of symbols from the val[] array, without risking marching off + * the end of memory. jxhuff.c will do a more thorough test later. + */ + nsymbols = 0; + for (len = 1; len <= 16; len++) + nsymbols += bits[len]; + if (nsymbols > 256) + ERREXIT(cinfo, JERR_BAD_HUFF_TABLE); + + if (nsymbols > 0) + MEMCOPY(htbl->huffval, val, nsymbols * SIZEOF(UINT8)); + + /* Initialize sent_table FALSE so table will be written to JPEG file. */ + htbl->sent_table = FALSE; + + return htbl; +} diff --git a/cmake/externals/libjpeg/jconfig.h b/cmake/externals/libjpeg/jconfig.h new file mode 100644 index 000000000..8d6c8f5b3 --- /dev/null +++ b/cmake/externals/libjpeg/jconfig.h @@ -0,0 +1,64 @@ +/* jconfig.vc --- jconfig.h for Microsoft Visual C++ on Windows 9x or NT. + * This file also works for Borland/Embarcadero C++ for Win32 or Win64 + * (CLI: bcc32, bcc32c, bcc32x, bcc64; GUI IDE: C++Builder/RAD Studio). + * See jconfig.txt for explanations. + */ + +#define HAVE_PROTOTYPES +#define HAVE_UNSIGNED_CHAR +#define HAVE_UNSIGNED_SHORT +/* #define void char */ +/* #define const */ +#undef CHAR_IS_UNSIGNED +#define HAVE_STDDEF_H +#define HAVE_STDLIB_H +#undef NEED_BSD_STRINGS +#undef NEED_SYS_TYPES_H +#undef NEED_FAR_POINTERS /* we presume a 32-bit flat memory model */ +#undef NEED_SHORT_EXTERNAL_NAMES +#undef INCOMPLETE_TYPES_BROKEN + +/* Define "boolean" as unsigned char, not enum, per Windows custom */ +#ifndef __RPCNDR_H__ /* don't conflict if rpcndr.h already read */ +typedef unsigned char boolean; +#endif +#ifndef FALSE /* in case these macros already exist */ +#define FALSE 0 /* values of boolean */ +#endif +#ifndef TRUE +#define TRUE 1 +#endif +#define HAVE_BOOLEAN /* prevent jmorecfg.h from redefining it */ + +/* Define custom RGB color order, prevent jmorecfg.h from redefinition */ +#undef JPEG_HAVE_RGB_CUSTOM +/* Use Windows custom BGR color order defined in jmorecfg.h */ +#undef JPEG_USE_RGB_CUSTOM + +/* Define custom file I/O functions, prevent jinclude.h from redefinition */ +#undef JPEG_HAVE_FILE_IO_CUSTOM +/* Use Delphi custom file I/O functions defined in jinclude.h */ +#undef JPEG_USE_FILE_IO_CUSTOM + + +#ifdef JPEG_INTERNALS + +#undef RIGHT_SHIFT_IS_UNSIGNED + +#endif /* JPEG_INTERNALS */ + +#ifdef JPEG_CJPEG_DJPEG + +#define BMP_SUPPORTED /* BMP image file format */ +#define GIF_SUPPORTED /* GIF image file format */ +#define PPM_SUPPORTED /* PBMPLUS PPM/PGM image file format */ +#undef RLE_SUPPORTED /* Utah RLE image file format */ +#define TARGA_SUPPORTED /* Targa image file format */ + +#define TWO_FILE_COMMANDLINE /* optional */ +#define USE_SETMODE /* Microsoft/Borland/Embarcadero have setmode() */ +#undef NEED_SIGNAL_CATCHER +#undef DONT_USE_B_MODE +#undef PROGRESS_REPORT /* optional */ + +#endif /* JPEG_CJPEG_DJPEG */ diff --git a/cmake/externals/libjpeg/jcparam.c b/cmake/externals/libjpeg/jcparam.c new file mode 100644 index 000000000..261ae86ca --- /dev/null +++ b/cmake/externals/libjpeg/jcparam.c @@ -0,0 +1,591 @@ +/* + * jcparam.c + * + * Copyright (C) 1991-1998, Thomas G. Lane. + * Modified 2003-2022 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains optional default-setting code for the JPEG compressor. + * Applications do not have to use this file, but those that don't use it + * must know a lot more about the innards of the JPEG code. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* + * Quantization table setup routines + */ + +GLOBAL(void) +jpeg_add_quant_table (j_compress_ptr cinfo, int which_tbl, + const unsigned int *basic_table, + int scale_factor, boolean force_baseline) +/* Define a quantization table equal to the basic_table times + * a scale factor (given as a percentage). + * If force_baseline is TRUE, the computed quantization table entries + * are limited to 1..255 for JPEG baseline compatibility. + */ +{ + JQUANT_TBL ** qtblptr; + int i; + long temp; + + /* Safety check to ensure start_compress not called yet. */ + if (cinfo->global_state != CSTATE_START) + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + + if (which_tbl < 0 || which_tbl >= NUM_QUANT_TBLS) + ERREXIT1(cinfo, JERR_DQT_INDEX, which_tbl); + + qtblptr = & cinfo->quant_tbl_ptrs[which_tbl]; + + if (*qtblptr == NULL) + *qtblptr = jpeg_alloc_quant_table((j_common_ptr) cinfo); + + for (i = 0; i < DCTSIZE2; i++) { + temp = ((long) basic_table[i] * scale_factor + 50L) / 100L; + /* limit the values to the valid range */ + if (temp <= 0L) temp = 1L; + if (temp > 32767L) temp = 32767L; /* max quantizer needed for 12 bits */ + if (force_baseline && temp > 255L) + temp = 255L; /* limit to baseline range if requested */ + (*qtblptr)->quantval[i] = (UINT16) temp; + } + + /* Initialize sent_table FALSE so table will be written to JPEG file. */ + (*qtblptr)->sent_table = FALSE; +} + + +/* These are the sample quantization tables given in JPEG spec section K.1. + * NOTE: chrominance DC value is changed from 17 to 16 for lossless support. + * The spec says that the values given produce "good" quality, + * and when divided by 2, "very good" quality. + */ +static const unsigned int std_luminance_quant_tbl[DCTSIZE2] = { + 16, 11, 10, 16, 24, 40, 51, 61, + 12, 12, 14, 19, 26, 58, 60, 55, + 14, 13, 16, 24, 40, 57, 69, 56, + 14, 17, 22, 29, 51, 87, 80, 62, + 18, 22, 37, 56, 68, 109, 103, 77, + 24, 35, 55, 64, 81, 104, 113, 92, + 49, 64, 78, 87, 103, 121, 120, 101, + 72, 92, 95, 98, 112, 100, 103, 99 +}; +static const unsigned int std_chrominance_quant_tbl[DCTSIZE2] = { + 16, 18, 24, 47, 99, 99, 99, 99, + 18, 21, 26, 66, 99, 99, 99, 99, + 24, 26, 56, 99, 99, 99, 99, 99, + 47, 66, 99, 99, 99, 99, 99, 99, + 99, 99, 99, 99, 99, 99, 99, 99, + 99, 99, 99, 99, 99, 99, 99, 99, + 99, 99, 99, 99, 99, 99, 99, 99, + 99, 99, 99, 99, 99, 99, 99, 99 +}; + + +GLOBAL(void) +jpeg_default_qtables (j_compress_ptr cinfo, boolean force_baseline) +/* Set or change the 'quality' (quantization) setting, using default tables + * and straight percentage-scaling quality scales. + * This entry point allows different scalings for luminance and chrominance. + */ +{ + /* Set up two quantization tables using the specified scaling */ + jpeg_add_quant_table(cinfo, 0, std_luminance_quant_tbl, + cinfo->q_scale_factor[0], force_baseline); + jpeg_add_quant_table(cinfo, 1, std_chrominance_quant_tbl, + cinfo->q_scale_factor[1], force_baseline); +} + + +GLOBAL(void) +jpeg_set_linear_quality (j_compress_ptr cinfo, int scale_factor, + boolean force_baseline) +/* Set or change the 'quality' (quantization) setting, using default tables + * and a straight percentage-scaling quality scale. In most cases it's better + * to use jpeg_set_quality (below); this entry point is provided for + * applications that insist on a linear percentage scaling. + */ +{ + /* Set up two quantization tables using the specified scaling */ + jpeg_add_quant_table(cinfo, 0, std_luminance_quant_tbl, + scale_factor, force_baseline); + jpeg_add_quant_table(cinfo, 1, std_chrominance_quant_tbl, + scale_factor, force_baseline); +} + + +GLOBAL(int) +jpeg_quality_scaling (int quality) +/* Convert a user-specified quality rating to a percentage scaling factor + * for an underlying quantization table, using our recommended scaling curve. + * The input 'quality' factor should be 0 (terrible) to 100 (very good). + */ +{ + /* Safety limit on quality factor. Convert 0 to 1 to avoid zero divide. */ + if (quality <= 0) quality = 1; + if (quality > 100) quality = 100; + + /* The basic table is used as-is (scaling 100) for a quality of 50. + * Qualities 50..100 are converted to scaling percentage 200 - 2*Q; + * note that at Q=100 the scaling is 0, which will cause jpeg_add_quant_table + * to make all the table entries 1 (hence, minimum quantization loss). + * Qualities 1..50 are converted to scaling percentage 5000/Q. + */ + if (quality < 50) + quality = 5000 / quality; + else + quality = 200 - quality*2; + + return quality; +} + + +GLOBAL(void) +jpeg_set_quality (j_compress_ptr cinfo, int quality, boolean force_baseline) +/* Set or change the 'quality' (quantization) setting, using default tables. + * This is the standard quality-adjusting entry point for typical user + * interfaces; only those who want detailed control over quantization tables + * would use the preceding routines directly. + */ +{ + /* Convert user 0-100 rating to percentage scaling */ + quality = jpeg_quality_scaling(quality); + + /* Set up standard quality tables */ + jpeg_set_linear_quality(cinfo, quality, force_baseline); +} + + +/* + * Reset standard Huffman tables + */ + +LOCAL(void) +std_huff_tables (j_compress_ptr cinfo) +{ + if (cinfo->dc_huff_tbl_ptrs[0] != NULL) + (void) jpeg_std_huff_table((j_common_ptr) cinfo, TRUE, 0); + + if (cinfo->ac_huff_tbl_ptrs[0] != NULL) + (void) jpeg_std_huff_table((j_common_ptr) cinfo, FALSE, 0); + + if (cinfo->dc_huff_tbl_ptrs[1] != NULL) + (void) jpeg_std_huff_table((j_common_ptr) cinfo, TRUE, 1); + + if (cinfo->ac_huff_tbl_ptrs[1] != NULL) + (void) jpeg_std_huff_table((j_common_ptr) cinfo, FALSE, 1); +} + + +/* + * Default parameter setup for compression. + * + * Applications that don't choose to use this routine must do their + * own setup of all these parameters. Alternately, you can call this + * to establish defaults and then alter parameters selectively. This + * is the recommended approach since, if we add any new parameters, + * your code will still work (they'll be set to reasonable defaults). + */ + +GLOBAL(void) +jpeg_set_defaults (j_compress_ptr cinfo) +{ + int i; + + /* Safety check to ensure start_compress not called yet. */ + if (cinfo->global_state != CSTATE_START) + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + + /* Allocate comp_info array large enough for maximum component count. + * Array is made permanent in case application wants to compress + * multiple images at same param settings. + */ + if (cinfo->comp_info == NULL) + cinfo->comp_info = (jpeg_component_info *) + (*cinfo->mem->alloc_small) ((j_common_ptr) cinfo, JPOOL_PERMANENT, + MAX_COMPONENTS * SIZEOF(jpeg_component_info)); + + /* Initialize everything not dependent on the color space */ + + cinfo->scale_num = 1; /* 1:1 scaling */ + cinfo->scale_denom = 1; + cinfo->data_precision = BITS_IN_JSAMPLE; + /* Set up two quantization tables using default quality of 75 */ + jpeg_set_quality(cinfo, 75, TRUE); + /* Reset standard Huffman tables */ + std_huff_tables(cinfo); + + /* Initialize default arithmetic coding conditioning */ + for (i = 0; i < NUM_ARITH_TBLS; i++) { + cinfo->arith_dc_L[i] = 0; + cinfo->arith_dc_U[i] = 1; + cinfo->arith_ac_K[i] = 5; + } + + /* Default is no multiple-scan output */ + cinfo->scan_info = NULL; + cinfo->num_scans = 0; + + /* Expect normal source image, not raw downsampled data */ + cinfo->raw_data_in = FALSE; + + /* The standard Huffman tables are only valid for 8-bit data precision. + * If the precision is higher, use arithmetic coding. + * (Alternatively, using Huffman coding would be possible with forcing + * optimization on so that usable tables will be computed, or by + * supplying default tables that are valid for the desired precision.) + * Otherwise, use Huffman coding by default. + */ + cinfo->arith_code = cinfo->data_precision > 8 ? TRUE : FALSE; + + /* By default, don't do extra passes to optimize entropy coding */ + cinfo->optimize_coding = FALSE; + + /* By default, use the simpler non-cosited sampling alignment */ + cinfo->CCIR601_sampling = FALSE; + + /* By default, apply fancy downsampling */ + cinfo->do_fancy_downsampling = TRUE; + + /* No input smoothing */ + cinfo->smoothing_factor = 0; + + /* DCT algorithm preference */ + cinfo->dct_method = JDCT_DEFAULT; + + /* No restart markers */ + cinfo->restart_interval = 0; + cinfo->restart_in_rows = 0; + + /* Fill in default JFIF marker parameters. Note that whether the marker + * will actually be written is determined by jpeg_set_colorspace. + * + * By default, the library emits JFIF version code 1.01. + * An application that wants to emit JFIF 1.02 extension markers should set + * JFIF_minor_version to 2. We could probably get away with just defaulting + * to 1.02, but there may still be some decoders in use that will complain + * about that; saying 1.01 should minimize compatibility problems. + * + * For wide gamut colorspaces (BG_RGB and BG_YCC), the major version will be + * overridden by jpeg_set_colorspace and set to 2. + */ + cinfo->JFIF_major_version = 1; /* Default JFIF version = 1.01 */ + cinfo->JFIF_minor_version = 1; + cinfo->density_unit = 0; /* Pixel size is unknown by default */ + cinfo->X_density = 1; /* Pixel aspect ratio is square by default */ + cinfo->Y_density = 1; + + /* No color transform */ + cinfo->color_transform = JCT_NONE; + + /* Choose JPEG colorspace based on input space, set defaults accordingly */ + + jpeg_default_colorspace(cinfo); +} + + +/* + * Select an appropriate JPEG colorspace for in_color_space. + */ + +GLOBAL(void) +jpeg_default_colorspace (j_compress_ptr cinfo) +{ + switch (cinfo->in_color_space) { + case JCS_UNKNOWN: + jpeg_set_colorspace(cinfo, JCS_UNKNOWN); + break; + case JCS_GRAYSCALE: + jpeg_set_colorspace(cinfo, JCS_GRAYSCALE); + break; + case JCS_RGB: + jpeg_set_colorspace(cinfo, JCS_YCbCr); + break; + case JCS_YCbCr: + jpeg_set_colorspace(cinfo, JCS_YCbCr); + break; + case JCS_CMYK: + jpeg_set_colorspace(cinfo, JCS_CMYK); /* By default, no translation */ + break; + case JCS_YCCK: + jpeg_set_colorspace(cinfo, JCS_YCCK); + break; + case JCS_BG_RGB: + /* No translation for now -- conversion to BG_YCC not yet supportet */ + jpeg_set_colorspace(cinfo, JCS_BG_RGB); + break; + case JCS_BG_YCC: + jpeg_set_colorspace(cinfo, JCS_BG_YCC); + break; + default: + ERREXIT(cinfo, JERR_BAD_IN_COLORSPACE); + } +} + + +/* + * Set the JPEG colorspace, and choose colorspace-dependent default values. + */ + +GLOBAL(void) +jpeg_set_colorspace (j_compress_ptr cinfo, J_COLOR_SPACE colorspace) +{ + jpeg_component_info * compptr; + int ci; + +#define SET_COMP(index,id,hsamp,vsamp,quant,dctbl,actbl) \ + (compptr = &cinfo->comp_info[index], \ + compptr->component_id = (id), \ + compptr->h_samp_factor = (hsamp), \ + compptr->v_samp_factor = (vsamp), \ + compptr->quant_tbl_no = (quant), \ + compptr->dc_tbl_no = (dctbl), \ + compptr->ac_tbl_no = (actbl) ) + + /* Safety check to ensure start_compress not called yet. */ + if (cinfo->global_state != CSTATE_START) + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + + /* For all colorspaces, we use Q and Huff tables 0 for luminance components, + * tables 1 for chrominance components. + */ + + cinfo->jpeg_color_space = colorspace; + + cinfo->write_JFIF_header = FALSE; /* No marker for non-JFIF colorspaces */ + cinfo->write_Adobe_marker = FALSE; /* write no Adobe marker by default */ + + switch (colorspace) { + case JCS_UNKNOWN: + cinfo->num_components = cinfo->input_components; + if (cinfo->num_components < 1 || cinfo->num_components > MAX_COMPONENTS) + ERREXIT2(cinfo, JERR_COMPONENT_COUNT, cinfo->num_components, + MAX_COMPONENTS); + for (ci = 0; ci < cinfo->num_components; ci++) { + SET_COMP(ci, ci, 1,1, 0, 0,0); + } + break; + case JCS_GRAYSCALE: + cinfo->write_JFIF_header = TRUE; /* Write a JFIF marker */ + cinfo->num_components = 1; + /* JFIF specifies component ID 1 */ + SET_COMP(0, 0x01, 1,1, 0, 0,0); + break; + case JCS_RGB: + cinfo->write_Adobe_marker = TRUE; /* write Adobe marker to flag RGB */ + cinfo->num_components = 3; + SET_COMP(0, 0x52 /* 'R' */, 1,1, + cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0, + cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0, + cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0); + SET_COMP(1, 0x47 /* 'G' */, 1,1, 0, 0,0); + SET_COMP(2, 0x42 /* 'B' */, 1,1, + cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0, + cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0, + cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0); + break; + case JCS_YCbCr: + cinfo->write_JFIF_header = TRUE; /* Write a JFIF marker */ + cinfo->num_components = 3; + /* JFIF specifies component IDs 1,2,3 */ + /* We default to 2x2 subsamples of chrominance */ + SET_COMP(0, 0x01, 2,2, 0, 0,0); + SET_COMP(1, 0x02, 1,1, 1, 1,1); + SET_COMP(2, 0x03, 1,1, 1, 1,1); + break; + case JCS_CMYK: + cinfo->write_Adobe_marker = TRUE; /* write Adobe marker to flag CMYK */ + cinfo->num_components = 4; + SET_COMP(0, 0x43 /* 'C' */, 1,1, 0, 0,0); + SET_COMP(1, 0x4D /* 'M' */, 1,1, 0, 0,0); + SET_COMP(2, 0x59 /* 'Y' */, 1,1, 0, 0,0); + SET_COMP(3, 0x4B /* 'K' */, 1,1, 0, 0,0); + break; + case JCS_YCCK: + cinfo->write_Adobe_marker = TRUE; /* write Adobe marker to flag YCCK */ + cinfo->num_components = 4; + SET_COMP(0, 0x01, 2,2, 0, 0,0); + SET_COMP(1, 0x02, 1,1, 1, 1,1); + SET_COMP(2, 0x03, 1,1, 1, 1,1); + SET_COMP(3, 0x04, 2,2, 0, 0,0); + break; + case JCS_BG_RGB: + cinfo->write_JFIF_header = TRUE; /* Write a JFIF marker */ + cinfo->JFIF_major_version = 2; /* Set JFIF major version = 2 */ + cinfo->num_components = 3; + /* Add offset 0x20 to the normal R/G/B component IDs */ + SET_COMP(0, 0x72 /* 'r' */, 1,1, + cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0, + cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0, + cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0); + SET_COMP(1, 0x67 /* 'g' */, 1,1, 0, 0,0); + SET_COMP(2, 0x62 /* 'b' */, 1,1, + cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0, + cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0, + cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0); + break; + case JCS_BG_YCC: + cinfo->write_JFIF_header = TRUE; /* Write a JFIF marker */ + cinfo->JFIF_major_version = 2; /* Set JFIF major version = 2 */ + cinfo->num_components = 3; + /* Add offset 0x20 to the normal Cb/Cr component IDs */ + /* We default to 2x2 subsamples of chrominance */ + SET_COMP(0, 0x01, 2,2, 0, 0,0); + SET_COMP(1, 0x22, 1,1, 1, 1,1); + SET_COMP(2, 0x23, 1,1, 1, 1,1); + break; + default: + ERREXIT(cinfo, JERR_BAD_J_COLORSPACE); + } +} + + +#ifdef C_PROGRESSIVE_SUPPORTED + +LOCAL(jpeg_scan_info *) +fill_a_scan (jpeg_scan_info * scanptr, int ci, + int Ss, int Se, int Ah, int Al) +/* Support routine: generate one scan for specified component */ +{ + scanptr->comps_in_scan = 1; + scanptr->component_index[0] = ci; + scanptr->Ss = Ss; + scanptr->Se = Se; + scanptr->Ah = Ah; + scanptr->Al = Al; + scanptr++; + return scanptr; +} + +LOCAL(jpeg_scan_info *) +fill_scans (jpeg_scan_info * scanptr, int ncomps, + int Ss, int Se, int Ah, int Al) +/* Support routine: generate one scan for each component */ +{ + int ci; + + for (ci = 0; ci < ncomps; ci++) { + scanptr->comps_in_scan = 1; + scanptr->component_index[0] = ci; + scanptr->Ss = Ss; + scanptr->Se = Se; + scanptr->Ah = Ah; + scanptr->Al = Al; + scanptr++; + } + return scanptr; +} + +LOCAL(jpeg_scan_info *) +fill_dc_scans (jpeg_scan_info * scanptr, int ncomps, int Ah, int Al) +/* Support routine: generate interleaved DC scan if possible, else N scans */ +{ + int ci; + + if (ncomps <= MAX_COMPS_IN_SCAN) { + /* Single interleaved DC scan */ + scanptr->comps_in_scan = ncomps; + for (ci = 0; ci < ncomps; ci++) + scanptr->component_index[ci] = ci; + scanptr->Ss = scanptr->Se = 0; + scanptr->Ah = Ah; + scanptr->Al = Al; + scanptr++; + } else { + /* Noninterleaved DC scan for each component */ + scanptr = fill_scans(scanptr, ncomps, 0, 0, Ah, Al); + } + return scanptr; +} + + +/* + * Create a recommended progressive-JPEG script. + * cinfo->num_components and cinfo->jpeg_color_space must be correct. + */ + +GLOBAL(void) +jpeg_simple_progression (j_compress_ptr cinfo) +{ + int ncomps = cinfo->num_components; + int nscans; + jpeg_scan_info * scanptr; + + /* Safety check to ensure start_compress not called yet. */ + if (cinfo->global_state != CSTATE_START) + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + + /* Figure space needed for script. Calculation must match code below! */ + if (ncomps == 3 && + (cinfo->jpeg_color_space == JCS_YCbCr || + cinfo->jpeg_color_space == JCS_BG_YCC)) { + /* Custom script for YCC color images. */ + nscans = 10; + } else { + /* All-purpose script for other color spaces. */ + if (ncomps > MAX_COMPS_IN_SCAN) + nscans = 6 * ncomps; /* 2 DC + 4 AC scans per component */ + else + nscans = 2 + 4 * ncomps; /* 2 DC scans; 4 AC scans per component */ + } + + /* Allocate space for script. + * We need to put it in the permanent pool in case the application performs + * multiple compressions without changing the settings. To avoid a memory + * leak if jpeg_simple_progression is called repeatedly for the same JPEG + * object, we try to re-use previously allocated space, and we allocate + * enough space to handle YCC even if initially asked for grayscale. + */ + if (cinfo->script_space == NULL || cinfo->script_space_size < nscans) { + cinfo->script_space_size = MAX(nscans, 10); + cinfo->script_space = (jpeg_scan_info *) + (*cinfo->mem->alloc_small) ((j_common_ptr) cinfo, JPOOL_PERMANENT, + cinfo->script_space_size * SIZEOF(jpeg_scan_info)); + } + scanptr = cinfo->script_space; + cinfo->scan_info = scanptr; + cinfo->num_scans = nscans; + + if (ncomps == 3 && + (cinfo->jpeg_color_space == JCS_YCbCr || + cinfo->jpeg_color_space == JCS_BG_YCC)) { + /* Custom script for YCC color images. */ + /* Initial DC scan */ + scanptr = fill_dc_scans(scanptr, ncomps, 0, 1); + /* Initial AC scan: get some luma data out in a hurry */ + scanptr = fill_a_scan(scanptr, 0, 1, 5, 0, 2); + /* Chroma data is too small to be worth expending many scans on */ + scanptr = fill_a_scan(scanptr, 2, 1, 63, 0, 1); + scanptr = fill_a_scan(scanptr, 1, 1, 63, 0, 1); + /* Complete spectral selection for luma AC */ + scanptr = fill_a_scan(scanptr, 0, 6, 63, 0, 2); + /* Refine next bit of luma AC */ + scanptr = fill_a_scan(scanptr, 0, 1, 63, 2, 1); + /* Finish DC successive approximation */ + scanptr = fill_dc_scans(scanptr, ncomps, 1, 0); + /* Finish AC successive approximation */ + scanptr = fill_a_scan(scanptr, 2, 1, 63, 1, 0); + scanptr = fill_a_scan(scanptr, 1, 1, 63, 1, 0); + /* Luma bottom bit comes last since it's usually largest scan */ + scanptr = fill_a_scan(scanptr, 0, 1, 63, 1, 0); + } else { + /* All-purpose script for other color spaces. */ + /* Successive approximation first pass */ + scanptr = fill_dc_scans(scanptr, ncomps, 0, 1); + scanptr = fill_scans(scanptr, ncomps, 1, 5, 0, 2); + scanptr = fill_scans(scanptr, ncomps, 6, 63, 0, 2); + /* Successive approximation second pass */ + scanptr = fill_scans(scanptr, ncomps, 1, 63, 2, 1); + /* Successive approximation final pass */ + scanptr = fill_dc_scans(scanptr, ncomps, 1, 0); + scanptr = fill_scans(scanptr, ncomps, 1, 63, 1, 0); + } +} + +#endif /* C_PROGRESSIVE_SUPPORTED */ diff --git a/cmake/externals/libjpeg/jcprepct.c b/cmake/externals/libjpeg/jcprepct.c new file mode 100644 index 000000000..586964bd4 --- /dev/null +++ b/cmake/externals/libjpeg/jcprepct.c @@ -0,0 +1,358 @@ +/* + * jcprepct.c + * + * Copyright (C) 1994-1996, Thomas G. Lane. + * Modified 2003-2020 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains the compression preprocessing controller. + * This controller manages the color conversion, downsampling, + * and edge expansion steps. + * + * Most of the complexity here is associated with buffering input rows + * as required by the downsampler. See the comments at the head of + * jcsample.c for the downsampler's needs. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* At present, jcsample.c can request context rows only for smoothing. + * In the future, we might also need context rows for CCIR601 sampling + * or other more-complex downsampling procedures. The code to support + * context rows should be compiled only if needed. + */ +#ifdef INPUT_SMOOTHING_SUPPORTED +#define CONTEXT_ROWS_SUPPORTED +#endif + + +/* + * For the simple (no-context-row) case, we just need to buffer one + * row group's worth of pixels for the downsampling step. At the bottom of + * the image, we pad to a full row group by replicating the last pixel row. + * The downsampler's last output row is then replicated if needed to pad + * out to a full iMCU row. + * + * When providing context rows, we must buffer three row groups' worth of + * pixels. Three row groups are physically allocated, but the row pointer + * arrays are made five row groups high, with the extra pointers above and + * below "wrapping around" to point to the last and first real row groups. + * This allows the downsampler to access the proper context rows. + * At the top and bottom of the image, we create dummy context rows by + * copying the first or last real pixel row. This copying could be avoided + * by pointer hacking as is done in jdmainct.c, but it doesn't seem worth the + * trouble on the compression side. + */ + + +/* Private buffer controller object */ + +typedef struct { + struct jpeg_c_prep_controller pub; /* public fields */ + + /* Downsampling input buffer. This buffer holds color-converted data + * until we have enough to do a downsample step. + */ + JSAMPARRAY color_buf[MAX_COMPONENTS]; + + JDIMENSION rows_to_go; /* counts rows remaining in source image */ + int next_buf_row; /* index of next row to store in color_buf */ + +#ifdef CONTEXT_ROWS_SUPPORTED /* only needed for context case */ + int this_row_group; /* starting row index of group to process */ + int next_buf_stop; /* downsample when we reach this index */ +#endif +} my_prep_controller; + +typedef my_prep_controller * my_prep_ptr; + + +/* + * Initialize for a processing pass. + */ + +METHODDEF(void) +start_pass_prep (j_compress_ptr cinfo, J_BUF_MODE pass_mode) +{ + my_prep_ptr prep = (my_prep_ptr) cinfo->prep; + + if (pass_mode != JBUF_PASS_THRU) + ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); + + /* Initialize total-height counter for detecting bottom of image */ + prep->rows_to_go = cinfo->image_height; + /* Mark the conversion buffer empty */ + prep->next_buf_row = 0; +#ifdef CONTEXT_ROWS_SUPPORTED + /* Preset additional state variables for context mode. + * These aren't used in non-context mode, so we needn't test which mode. + */ + prep->this_row_group = 0; + /* Set next_buf_stop to stop after two row groups have been read in. */ + prep->next_buf_stop = 2 * cinfo->max_v_samp_factor; +#endif +} + + +/* + * Expand an image vertically from height input_rows to height output_rows, + * by duplicating the bottom row. + */ + +LOCAL(void) +expand_bottom_edge (JSAMPARRAY image_data, JDIMENSION num_cols, + int input_rows, int output_rows) +{ + register int row; + + for (row = input_rows; row < output_rows; row++) { + jcopy_sample_rows(image_data + input_rows - 1, + image_data + row, + 1, num_cols); + } +} + + +/* + * Process some data in the simple no-context case. + * + * Preprocessor output data is counted in "row groups". A row group + * is defined to be v_samp_factor sample rows of each component. + * Downsampling will produce this much data from each max_v_samp_factor + * input rows. + */ + +METHODDEF(void) +pre_process_data (j_compress_ptr cinfo, + JSAMPARRAY input_buf, JDIMENSION *in_row_ctr, + JDIMENSION in_rows_avail, + JSAMPIMAGE output_buf, JDIMENSION *out_row_group_ctr, + JDIMENSION out_row_groups_avail) +{ + my_prep_ptr prep = (my_prep_ptr) cinfo->prep; + int numrows, ci; + JDIMENSION inrows; + jpeg_component_info * compptr; + + while (*in_row_ctr < in_rows_avail && + *out_row_group_ctr < out_row_groups_avail) { + /* Do color conversion to fill the conversion buffer. */ + inrows = in_rows_avail - *in_row_ctr; + numrows = cinfo->max_v_samp_factor - prep->next_buf_row; + numrows = (int) MIN((JDIMENSION) numrows, inrows); + (*cinfo->cconvert->color_convert) (cinfo, input_buf + *in_row_ctr, + prep->color_buf, + (JDIMENSION) prep->next_buf_row, + numrows); + *in_row_ctr += numrows; + prep->next_buf_row += numrows; + prep->rows_to_go -= numrows; + /* If at bottom of image, pad to fill the conversion buffer. */ + if (prep->rows_to_go == 0 && + prep->next_buf_row < cinfo->max_v_samp_factor) { + for (ci = 0; ci < cinfo->num_components; ci++) { + expand_bottom_edge(prep->color_buf[ci], cinfo->image_width, + prep->next_buf_row, cinfo->max_v_samp_factor); + } + prep->next_buf_row = cinfo->max_v_samp_factor; + } + /* If we've filled the conversion buffer, empty it. */ + if (prep->next_buf_row == cinfo->max_v_samp_factor) { + (*cinfo->downsample->downsample) (cinfo, + prep->color_buf, (JDIMENSION) 0, + output_buf, *out_row_group_ctr); + prep->next_buf_row = 0; + (*out_row_group_ctr)++; + } + /* If at bottom of image, pad the output to a full iMCU height. + * Note we assume the caller is providing a one-iMCU-height output buffer! + */ + if (prep->rows_to_go == 0 && + *out_row_group_ctr < out_row_groups_avail) { + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + numrows = (compptr->v_samp_factor * compptr->DCT_v_scaled_size) / + cinfo->min_DCT_v_scaled_size; + expand_bottom_edge(output_buf[ci], + compptr->width_in_blocks * compptr->DCT_h_scaled_size, + (int) (*out_row_group_ctr * numrows), + (int) (out_row_groups_avail * numrows)); + } + *out_row_group_ctr = out_row_groups_avail; + break; /* can exit outer loop without test */ + } + } +} + + +#ifdef CONTEXT_ROWS_SUPPORTED + +/* + * Process some data in the context case. + */ + +METHODDEF(void) +pre_process_context (j_compress_ptr cinfo, + JSAMPARRAY input_buf, JDIMENSION *in_row_ctr, + JDIMENSION in_rows_avail, + JSAMPIMAGE output_buf, JDIMENSION *out_row_group_ctr, + JDIMENSION out_row_groups_avail) +{ + my_prep_ptr prep = (my_prep_ptr) cinfo->prep; + int numrows, ci; + int buf_height = cinfo->max_v_samp_factor * 3; + JDIMENSION inrows; + + while (*out_row_group_ctr < out_row_groups_avail) { + if (*in_row_ctr < in_rows_avail) { + /* Do color conversion to fill the conversion buffer. */ + inrows = in_rows_avail - *in_row_ctr; + numrows = prep->next_buf_stop - prep->next_buf_row; + numrows = (int) MIN((JDIMENSION) numrows, inrows); + (*cinfo->cconvert->color_convert) (cinfo, input_buf + *in_row_ctr, + prep->color_buf, + (JDIMENSION) prep->next_buf_row, + numrows); + /* Pad at top of image, if first time through */ + if (prep->rows_to_go == cinfo->image_height) { + for (ci = 0; ci < cinfo->num_components; ci++) { + int row; + for (row = 1; row <= cinfo->max_v_samp_factor; row++) { + jcopy_sample_rows(prep->color_buf[ci], + prep->color_buf[ci] - row, + 1, cinfo->image_width); + } + } + } + *in_row_ctr += numrows; + prep->next_buf_row += numrows; + prep->rows_to_go -= numrows; + } else { + /* Return for more data, unless we are at the bottom of the image. */ + if (prep->rows_to_go != 0) + break; + /* When at bottom of image, pad to fill the conversion buffer. */ + if (prep->next_buf_row < prep->next_buf_stop) { + for (ci = 0; ci < cinfo->num_components; ci++) { + expand_bottom_edge(prep->color_buf[ci], cinfo->image_width, + prep->next_buf_row, prep->next_buf_stop); + } + prep->next_buf_row = prep->next_buf_stop; + } + } + /* If we've gotten enough data, downsample a row group. */ + if (prep->next_buf_row == prep->next_buf_stop) { + (*cinfo->downsample->downsample) (cinfo, + prep->color_buf, + (JDIMENSION) prep->this_row_group, + output_buf, *out_row_group_ctr); + (*out_row_group_ctr)++; + /* Advance pointers with wraparound as necessary. */ + prep->this_row_group += cinfo->max_v_samp_factor; + if (prep->this_row_group >= buf_height) + prep->this_row_group = 0; + if (prep->next_buf_row >= buf_height) + prep->next_buf_row = 0; + prep->next_buf_stop = prep->next_buf_row + cinfo->max_v_samp_factor; + } + } +} + + +/* + * Create the wrapped-around downsampling input buffer needed for context mode. + */ + +LOCAL(void) +create_context_buffer (j_compress_ptr cinfo) +{ + my_prep_ptr prep = (my_prep_ptr) cinfo->prep; + int rgroup_height = cinfo->max_v_samp_factor; + int ci, i; + jpeg_component_info * compptr; + JSAMPARRAY true_buffer, fake_buffer; + + /* Grab enough space for fake row pointers for all the components; + * we need five row groups' worth of pointers for each component. + */ + fake_buffer = (JSAMPARRAY) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, + (cinfo->num_components * 5 * rgroup_height) * SIZEOF(JSAMPROW)); + + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + /* Allocate the actual buffer space (3 row groups) for this component. + * We make the buffer wide enough to allow the downsampler to edge-expand + * horizontally within the buffer, if it so chooses. + */ + true_buffer = (*cinfo->mem->alloc_sarray) + ((j_common_ptr) cinfo, JPOOL_IMAGE, + (JDIMENSION) (((long) compptr->width_in_blocks * + cinfo->min_DCT_h_scaled_size * + cinfo->max_h_samp_factor) / compptr->h_samp_factor), + (JDIMENSION) (3 * rgroup_height)); + /* Copy true buffer row pointers into the middle of the fake row array */ + MEMCOPY(fake_buffer + rgroup_height, true_buffer, + 3 * rgroup_height * SIZEOF(JSAMPROW)); + /* Fill in the above and below wraparound pointers */ + for (i = 0; i < rgroup_height; i++) { + fake_buffer[i] = true_buffer[2 * rgroup_height + i]; + fake_buffer[4 * rgroup_height + i] = true_buffer[i]; + } + prep->color_buf[ci] = fake_buffer + rgroup_height; + fake_buffer += 5 * rgroup_height; /* point to space for next component */ + } +} + +#endif /* CONTEXT_ROWS_SUPPORTED */ + + +/* + * Initialize preprocessing controller. + */ + +GLOBAL(void) +jinit_c_prep_controller (j_compress_ptr cinfo, boolean need_full_buffer) +{ + my_prep_ptr prep; + int ci; + jpeg_component_info * compptr; + + if (need_full_buffer) /* safety check */ + ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); + + prep = (my_prep_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_prep_controller)); + cinfo->prep = &prep->pub; + prep->pub.start_pass = start_pass_prep; + + /* Allocate the color conversion buffer. + * We make the buffer wide enough to allow the downsampler to edge-expand + * horizontally within the buffer, if it so chooses. + */ + if (cinfo->downsample->need_context_rows) { + /* Set up to provide context rows */ +#ifdef CONTEXT_ROWS_SUPPORTED + prep->pub.pre_process_data = pre_process_context; + create_context_buffer(cinfo); +#else + ERREXIT(cinfo, JERR_NOT_COMPILED); +#endif + } else { + /* No context, just make it tall enough for one row group */ + prep->pub.pre_process_data = pre_process_data; + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + prep->color_buf[ci] = (*cinfo->mem->alloc_sarray) + ((j_common_ptr) cinfo, JPOOL_IMAGE, + (JDIMENSION) (((long) compptr->width_in_blocks * + cinfo->min_DCT_h_scaled_size * + cinfo->max_h_samp_factor) / compptr->h_samp_factor), + (JDIMENSION) cinfo->max_v_samp_factor); + } + } +} diff --git a/cmake/externals/libjpeg/jcsample.c b/cmake/externals/libjpeg/jcsample.c new file mode 100644 index 000000000..2372c4173 --- /dev/null +++ b/cmake/externals/libjpeg/jcsample.c @@ -0,0 +1,545 @@ +/* + * jcsample.c + * + * Copyright (C) 1991-1996, Thomas G. Lane. + * Modified 2003-2020 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains downsampling routines. + * + * Downsampling input data is counted in "row groups". A row group + * is defined to be max_v_samp_factor pixel rows of each component, + * from which the downsampler produces v_samp_factor sample rows. + * A single row group is processed in each call to the downsampler module. + * + * The downsampler is responsible for edge-expansion of its output data + * to fill an integral number of DCT blocks horizontally. The source buffer + * may be modified if it is helpful for this purpose (the source buffer is + * allocated wide enough to correspond to the desired output width). + * The caller (the prep controller) is responsible for vertical padding. + * + * The downsampler may request "context rows" by setting need_context_rows + * during startup. In this case, the input arrays will contain at least + * one row group's worth of pixels above and below the passed-in data; + * the caller will create dummy rows at image top and bottom by replicating + * the first or last real pixel row. + * + * An excellent reference for image resampling is + * Digital Image Warping, George Wolberg, 1990. + * Pub. by IEEE Computer Society Press, Los Alamitos, CA. ISBN 0-8186-8944-7. + * + * The downsampling algorithm used here is a simple average of the source + * pixels covered by the output pixel. The hi-falutin sampling literature + * refers to this as a "box filter". In general the characteristics of a box + * filter are not very good, but for the specific cases we normally use (1:1 + * and 2:1 ratios) the box is equivalent to a "triangle filter" which is not + * nearly so bad. If you intend to use other sampling ratios, you'd be well + * advised to improve this code. + * + * A simple input-smoothing capability is provided. This is mainly intended + * for cleaning up color-dithered GIF input files (if you find it inadequate, + * we suggest using an external filtering program such as pnmconvol). When + * enabled, each input pixel P is replaced by a weighted sum of itself and its + * eight neighbors. P's weight is 1-8*SF and each neighbor's weight is SF, + * where SF = (smoothing_factor / 1024). + * Currently, smoothing is only supported for 2h2v sampling factors. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* Pointer to routine to downsample a single component */ +typedef JMETHOD(void, downsample1_ptr, + (j_compress_ptr cinfo, jpeg_component_info * compptr, + JSAMPARRAY input_data, JSAMPARRAY output_data)); + +/* Private subobject */ + +typedef struct { + struct jpeg_downsampler pub; /* public fields */ + + /* Downsampling method pointers, one per component */ + downsample1_ptr methods[MAX_COMPONENTS]; + + /* Height of an output row group for each component. */ + int rowgroup_height[MAX_COMPONENTS]; + + /* These arrays save pixel expansion factors so that int_downsample need not + * recompute them each time. They are unused for other downsampling methods. + */ + UINT8 h_expand[MAX_COMPONENTS]; + UINT8 v_expand[MAX_COMPONENTS]; +} my_downsampler; + +typedef my_downsampler * my_downsample_ptr; + + +/* + * Initialize for a downsampling pass. + */ + +METHODDEF(void) +start_pass_downsample (j_compress_ptr cinfo) +{ + /* no work for now */ +} + + +/* + * Expand a component horizontally from width input_cols to width output_cols, + * by duplicating the rightmost samples. + */ + +LOCAL(void) +expand_right_edge (JSAMPARRAY image_data, int num_rows, + JDIMENSION input_cols, JDIMENSION output_cols) +{ + register JSAMPROW ptr; + register JSAMPLE pixval; + register int count; + int row; + int numcols = (int) (output_cols - input_cols); + + if (numcols > 0) { + for (row = 0; row < num_rows; row++) { + ptr = image_data[row] + input_cols; + pixval = ptr[-1]; /* don't need GETJSAMPLE() here */ + for (count = numcols; count > 0; count--) + *ptr++ = pixval; + } + } +} + + +/* + * Do downsampling for a whole row group (all components). + * + * In this version we simply downsample each component independently. + */ + +METHODDEF(void) +sep_downsample (j_compress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION in_row_index, + JSAMPIMAGE output_buf, JDIMENSION out_row_group_index) +{ + my_downsample_ptr downsample = (my_downsample_ptr) cinfo->downsample; + int ci; + jpeg_component_info * compptr; + JSAMPARRAY in_ptr, out_ptr; + + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + in_ptr = input_buf[ci] + in_row_index; + out_ptr = output_buf[ci] + + (out_row_group_index * downsample->rowgroup_height[ci]); + (*downsample->methods[ci]) (cinfo, compptr, in_ptr, out_ptr); + } +} + + +/* + * Downsample pixel values of a single component. + * One row group is processed per call. + * This version handles arbitrary integral sampling ratios, without smoothing. + * Note that this version is not actually used for customary sampling ratios. + */ + +METHODDEF(void) +int_downsample (j_compress_ptr cinfo, jpeg_component_info * compptr, + JSAMPARRAY input_data, JSAMPARRAY output_data) +{ + my_downsample_ptr downsample = (my_downsample_ptr) cinfo->downsample; + int inrow, outrow, h_expand, v_expand, numpix, numpix2, h, v; + JDIMENSION outcol, outcol_h; /* outcol_h == outcol*h_expand */ + JDIMENSION output_cols = compptr->width_in_blocks * compptr->DCT_h_scaled_size; + JSAMPROW inptr, outptr; + INT32 outvalue; + + h_expand = downsample->h_expand[compptr->component_index]; + v_expand = downsample->v_expand[compptr->component_index]; + numpix = h_expand * v_expand; + numpix2 = numpix/2; + + /* Expand input data enough to let all the output samples be generated + * by the standard loop. Special-casing padded output would be more + * efficient. + */ + expand_right_edge(input_data, cinfo->max_v_samp_factor, + cinfo->image_width, output_cols * h_expand); + + inrow = outrow = 0; + while (inrow < cinfo->max_v_samp_factor) { + outptr = output_data[outrow]; + for (outcol = 0, outcol_h = 0; outcol < output_cols; + outcol++, outcol_h += h_expand) { + outvalue = 0; + for (v = 0; v < v_expand; v++) { + inptr = input_data[inrow+v] + outcol_h; + for (h = 0; h < h_expand; h++) { + outvalue += (INT32) GETJSAMPLE(*inptr++); + } + } + *outptr++ = (JSAMPLE) ((outvalue + numpix2) / numpix); + } + inrow += v_expand; + outrow++; + } +} + + +/* + * Downsample pixel values of a single component. + * This version handles the special case of a full-size component, + * without smoothing. + */ + +METHODDEF(void) +fullsize_downsample (j_compress_ptr cinfo, jpeg_component_info * compptr, + JSAMPARRAY input_data, JSAMPARRAY output_data) +{ + /* Copy the data */ + jcopy_sample_rows(input_data, output_data, + cinfo->max_v_samp_factor, cinfo->image_width); + /* Edge-expand */ + expand_right_edge(output_data, cinfo->max_v_samp_factor, cinfo->image_width, + compptr->width_in_blocks * compptr->DCT_h_scaled_size); +} + + +/* + * Downsample pixel values of a single component. + * This version handles the common case of 2:1 horizontal and 1:1 vertical, + * without smoothing. + * + * A note about the "bias" calculations: when rounding fractional values to + * integer, we do not want to always round 0.5 up to the next integer. + * If we did that, we'd introduce a noticeable bias towards larger values. + * Instead, this code is arranged so that 0.5 will be rounded up or down at + * alternate pixel locations (a simple ordered dither pattern). + */ + +METHODDEF(void) +h2v1_downsample (j_compress_ptr cinfo, jpeg_component_info * compptr, + JSAMPARRAY input_data, JSAMPARRAY output_data) +{ + int inrow; + JDIMENSION outcol; + JDIMENSION output_cols = compptr->width_in_blocks * compptr->DCT_h_scaled_size; + register JSAMPROW inptr, outptr; + register int bias; + + /* Expand input data enough to let all the output samples be generated + * by the standard loop. Special-casing padded output would be more + * efficient. + */ + expand_right_edge(input_data, cinfo->max_v_samp_factor, + cinfo->image_width, output_cols * 2); + + for (inrow = 0; inrow < cinfo->max_v_samp_factor; inrow++) { + outptr = output_data[inrow]; + inptr = input_data[inrow]; + bias = 0; /* bias = 0,1,0,1,... for successive samples */ + for (outcol = 0; outcol < output_cols; outcol++) { + *outptr++ = (JSAMPLE) ((GETJSAMPLE(*inptr) + GETJSAMPLE(inptr[1]) + + bias) >> 1); + bias ^= 1; /* 0=>1, 1=>0 */ + inptr += 2; + } + } +} + + +/* + * Downsample pixel values of a single component. + * This version handles the standard case of 2:1 horizontal and 2:1 vertical, + * without smoothing. + */ + +METHODDEF(void) +h2v2_downsample (j_compress_ptr cinfo, jpeg_component_info * compptr, + JSAMPARRAY input_data, JSAMPARRAY output_data) +{ + int inrow, outrow; + JDIMENSION outcol; + JDIMENSION output_cols = compptr->width_in_blocks * compptr->DCT_h_scaled_size; + register JSAMPROW inptr0, inptr1, outptr; + register int bias; + + /* Expand input data enough to let all the output samples be generated + * by the standard loop. Special-casing padded output would be more + * efficient. + */ + expand_right_edge(input_data, cinfo->max_v_samp_factor, + cinfo->image_width, output_cols * 2); + + inrow = outrow = 0; + while (inrow < cinfo->max_v_samp_factor) { + outptr = output_data[outrow]; + inptr0 = input_data[inrow]; + inptr1 = input_data[inrow+1]; + bias = 1; /* bias = 1,2,1,2,... for successive samples */ + for (outcol = 0; outcol < output_cols; outcol++) { + *outptr++ = (JSAMPLE) ((GETJSAMPLE(*inptr0) + GETJSAMPLE(inptr0[1]) + + GETJSAMPLE(*inptr1) + GETJSAMPLE(inptr1[1]) + + bias) >> 2); + bias ^= 3; /* 1=>2, 2=>1 */ + inptr0 += 2; inptr1 += 2; + } + inrow += 2; + outrow++; + } +} + + +#ifdef INPUT_SMOOTHING_SUPPORTED + +/* + * Downsample pixel values of a single component. + * This version handles the standard case of 2:1 horizontal and 2:1 vertical, + * with smoothing. One row of context is required. + */ + +METHODDEF(void) +h2v2_smooth_downsample (j_compress_ptr cinfo, jpeg_component_info * compptr, + JSAMPARRAY input_data, JSAMPARRAY output_data) +{ + int inrow, outrow; + JDIMENSION colctr; + JDIMENSION output_cols = compptr->width_in_blocks * compptr->DCT_h_scaled_size; + register JSAMPROW inptr0, inptr1, above_ptr, below_ptr, outptr; + INT32 membersum, neighsum, memberscale, neighscale; + + /* Expand input data enough to let all the output samples be generated + * by the standard loop. Special-casing padded output would be more + * efficient. + */ + expand_right_edge(input_data - 1, cinfo->max_v_samp_factor + 2, + cinfo->image_width, output_cols * 2); + + /* We don't bother to form the individual "smoothed" input pixel values; + * we can directly compute the output which is the average of the four + * smoothed values. Each of the four member pixels contributes a fraction + * (1-8*SF) to its own smoothed image and a fraction SF to each of the three + * other smoothed pixels, therefore a total fraction (1-5*SF)/4 to the final + * output. The four corner-adjacent neighbor pixels contribute a fraction + * SF to just one smoothed pixel, or SF/4 to the final output; while the + * eight edge-adjacent neighbors contribute SF to each of two smoothed + * pixels, or SF/2 overall. In order to use integer arithmetic, these + * factors are scaled by 2^16 = 65536. + * Also recall that SF = smoothing_factor / 1024. + */ + + memberscale = 16384 - cinfo->smoothing_factor * 80; /* scaled (1-5*SF)/4 */ + neighscale = cinfo->smoothing_factor * 16; /* scaled SF/4 */ + + inrow = outrow = 0; + while (inrow < cinfo->max_v_samp_factor) { + outptr = output_data[outrow]; + inptr0 = input_data[inrow]; + inptr1 = input_data[inrow+1]; + above_ptr = input_data[inrow-1]; + below_ptr = input_data[inrow+2]; + + /* Special case for first column: pretend column -1 is same as column 0 */ + membersum = GETJSAMPLE(*inptr0) + GETJSAMPLE(inptr0[1]) + + GETJSAMPLE(*inptr1) + GETJSAMPLE(inptr1[1]); + neighsum = GETJSAMPLE(*above_ptr) + GETJSAMPLE(above_ptr[1]) + + GETJSAMPLE(*below_ptr) + GETJSAMPLE(below_ptr[1]) + + GETJSAMPLE(*inptr0) + GETJSAMPLE(inptr0[2]) + + GETJSAMPLE(*inptr1) + GETJSAMPLE(inptr1[2]); + neighsum += neighsum; + neighsum += GETJSAMPLE(*above_ptr) + GETJSAMPLE(above_ptr[2]) + + GETJSAMPLE(*below_ptr) + GETJSAMPLE(below_ptr[2]); + membersum = membersum * memberscale + neighsum * neighscale; + *outptr++ = (JSAMPLE) ((membersum + 32768) >> 16); + inptr0 += 2; inptr1 += 2; above_ptr += 2; below_ptr += 2; + + for (colctr = output_cols - 2; colctr > 0; colctr--) { + /* sum of pixels directly mapped to this output element */ + membersum = GETJSAMPLE(*inptr0) + GETJSAMPLE(inptr0[1]) + + GETJSAMPLE(*inptr1) + GETJSAMPLE(inptr1[1]); + /* sum of edge-neighbor pixels */ + neighsum = GETJSAMPLE(*above_ptr) + GETJSAMPLE(above_ptr[1]) + + GETJSAMPLE(*below_ptr) + GETJSAMPLE(below_ptr[1]) + + GETJSAMPLE(inptr0[-1]) + GETJSAMPLE(inptr0[2]) + + GETJSAMPLE(inptr1[-1]) + GETJSAMPLE(inptr1[2]); + /* The edge-neighbors count twice as much as corner-neighbors */ + neighsum += neighsum; + /* Add in the corner-neighbors */ + neighsum += GETJSAMPLE(above_ptr[-1]) + GETJSAMPLE(above_ptr[2]) + + GETJSAMPLE(below_ptr[-1]) + GETJSAMPLE(below_ptr[2]); + /* form final output scaled up by 2^16 */ + membersum = membersum * memberscale + neighsum * neighscale; + /* round, descale and output it */ + *outptr++ = (JSAMPLE) ((membersum + 32768) >> 16); + inptr0 += 2; inptr1 += 2; above_ptr += 2; below_ptr += 2; + } + + /* Special case for last column */ + membersum = GETJSAMPLE(*inptr0) + GETJSAMPLE(inptr0[1]) + + GETJSAMPLE(*inptr1) + GETJSAMPLE(inptr1[1]); + neighsum = GETJSAMPLE(*above_ptr) + GETJSAMPLE(above_ptr[1]) + + GETJSAMPLE(*below_ptr) + GETJSAMPLE(below_ptr[1]) + + GETJSAMPLE(inptr0[-1]) + GETJSAMPLE(inptr0[1]) + + GETJSAMPLE(inptr1[-1]) + GETJSAMPLE(inptr1[1]); + neighsum += neighsum; + neighsum += GETJSAMPLE(above_ptr[-1]) + GETJSAMPLE(above_ptr[1]) + + GETJSAMPLE(below_ptr[-1]) + GETJSAMPLE(below_ptr[1]); + membersum = membersum * memberscale + neighsum * neighscale; + *outptr = (JSAMPLE) ((membersum + 32768) >> 16); + + inrow += 2; + outrow++; + } +} + + +/* + * Downsample pixel values of a single component. + * This version handles the special case of a full-size component, + * with smoothing. One row of context is required. + */ + +METHODDEF(void) +fullsize_smooth_downsample (j_compress_ptr cinfo, jpeg_component_info *compptr, + JSAMPARRAY input_data, JSAMPARRAY output_data) +{ + int inrow; + JDIMENSION colctr; + JDIMENSION output_cols = compptr->width_in_blocks * compptr->DCT_h_scaled_size; + register JSAMPROW inptr, above_ptr, below_ptr, outptr; + INT32 membersum, neighsum, memberscale, neighscale; + int colsum, lastcolsum, nextcolsum; + + /* Expand input data enough to let all the output samples be generated + * by the standard loop. Special-casing padded output would be more + * efficient. + */ + expand_right_edge(input_data - 1, cinfo->max_v_samp_factor + 2, + cinfo->image_width, output_cols); + + /* Each of the eight neighbor pixels contributes a fraction SF to the + * smoothed pixel, while the main pixel contributes (1-8*SF). In order + * to use integer arithmetic, these factors are multiplied by 2^16 = 65536. + * Also recall that SF = smoothing_factor / 1024. + */ + + memberscale = 65536L - cinfo->smoothing_factor * 512L; /* scaled 1-8*SF */ + neighscale = cinfo->smoothing_factor * 64; /* scaled SF */ + + for (inrow = 0; inrow < cinfo->max_v_samp_factor; inrow++) { + outptr = output_data[inrow]; + inptr = input_data[inrow]; + above_ptr = input_data[inrow-1]; + below_ptr = input_data[inrow+1]; + + /* Special case for first column */ + colsum = GETJSAMPLE(*above_ptr++) + GETJSAMPLE(*below_ptr++) + + GETJSAMPLE(*inptr); + membersum = GETJSAMPLE(*inptr++); + nextcolsum = GETJSAMPLE(*above_ptr) + GETJSAMPLE(*below_ptr) + + GETJSAMPLE(*inptr); + neighsum = colsum + (colsum - membersum) + nextcolsum; + membersum = membersum * memberscale + neighsum * neighscale; + *outptr++ = (JSAMPLE) ((membersum + 32768) >> 16); + lastcolsum = colsum; colsum = nextcolsum; + + for (colctr = output_cols - 2; colctr > 0; colctr--) { + membersum = GETJSAMPLE(*inptr++); + above_ptr++; below_ptr++; + nextcolsum = GETJSAMPLE(*above_ptr) + GETJSAMPLE(*below_ptr) + + GETJSAMPLE(*inptr); + neighsum = lastcolsum + (colsum - membersum) + nextcolsum; + membersum = membersum * memberscale + neighsum * neighscale; + *outptr++ = (JSAMPLE) ((membersum + 32768) >> 16); + lastcolsum = colsum; colsum = nextcolsum; + } + + /* Special case for last column */ + membersum = GETJSAMPLE(*inptr); + neighsum = lastcolsum + (colsum - membersum) + colsum; + membersum = membersum * memberscale + neighsum * neighscale; + *outptr = (JSAMPLE) ((membersum + 32768) >> 16); + + } +} + +#endif /* INPUT_SMOOTHING_SUPPORTED */ + + +/* + * Module initialization routine for downsampling. + * Note that we must select a routine for each component. + */ + +GLOBAL(void) +jinit_downsampler (j_compress_ptr cinfo) +{ + my_downsample_ptr downsample; + int ci; + jpeg_component_info * compptr; + boolean smoothok = TRUE; + int h_in_group, v_in_group, h_out_group, v_out_group; + + downsample = (my_downsample_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_downsampler)); + cinfo->downsample = &downsample->pub; + downsample->pub.start_pass = start_pass_downsample; + downsample->pub.downsample = sep_downsample; + downsample->pub.need_context_rows = FALSE; + + if (cinfo->CCIR601_sampling) + ERREXIT(cinfo, JERR_CCIR601_NOTIMPL); + + /* Verify we can handle the sampling factors, and set up method pointers */ + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + /* Compute size of an "output group" for DCT scaling. This many samples + * are to be converted from max_h_samp_factor * max_v_samp_factor pixels. + */ + h_out_group = (compptr->h_samp_factor * compptr->DCT_h_scaled_size) / + cinfo->min_DCT_h_scaled_size; + v_out_group = (compptr->v_samp_factor * compptr->DCT_v_scaled_size) / + cinfo->min_DCT_v_scaled_size; + h_in_group = cinfo->max_h_samp_factor; + v_in_group = cinfo->max_v_samp_factor; + downsample->rowgroup_height[ci] = v_out_group; /* save for use later */ + if (h_in_group == h_out_group && v_in_group == v_out_group) { +#ifdef INPUT_SMOOTHING_SUPPORTED + if (cinfo->smoothing_factor) { + downsample->methods[ci] = fullsize_smooth_downsample; + downsample->pub.need_context_rows = TRUE; + } else +#endif + downsample->methods[ci] = fullsize_downsample; + } else if (h_in_group == h_out_group * 2 && + v_in_group == v_out_group) { + smoothok = FALSE; + downsample->methods[ci] = h2v1_downsample; + } else if (h_in_group == h_out_group * 2 && + v_in_group == v_out_group * 2) { +#ifdef INPUT_SMOOTHING_SUPPORTED + if (cinfo->smoothing_factor) { + downsample->methods[ci] = h2v2_smooth_downsample; + downsample->pub.need_context_rows = TRUE; + } else +#endif + downsample->methods[ci] = h2v2_downsample; + } else if ((h_in_group % h_out_group) == 0 && + (v_in_group % v_out_group) == 0) { + smoothok = FALSE; + downsample->methods[ci] = int_downsample; + downsample->h_expand[ci] = (UINT8) (h_in_group / h_out_group); + downsample->v_expand[ci] = (UINT8) (v_in_group / v_out_group); + } else + ERREXIT(cinfo, JERR_FRACT_SAMPLE_NOTIMPL); + } + +#ifdef INPUT_SMOOTHING_SUPPORTED + if (cinfo->smoothing_factor && !smoothok) + TRACEMS(cinfo, 0, JTRC_SMOOTH_NOTIMPL); +#endif +} diff --git a/cmake/externals/libjpeg/jctrans.c b/cmake/externals/libjpeg/jctrans.c new file mode 100644 index 000000000..261dd2996 --- /dev/null +++ b/cmake/externals/libjpeg/jctrans.c @@ -0,0 +1,399 @@ +/* + * jctrans.c + * + * Copyright (C) 1995-1998, Thomas G. Lane. + * Modified 2000-2020 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains library routines for transcoding compression, + * that is, writing raw DCT coefficient arrays to an output JPEG file. + * The routines in jcapimin.c will also be needed by a transcoder. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* Forward declarations */ +LOCAL(void) transencode_master_selection + JPP((j_compress_ptr cinfo, jvirt_barray_ptr * coef_arrays)); +LOCAL(void) transencode_coef_controller + JPP((j_compress_ptr cinfo, jvirt_barray_ptr * coef_arrays)); + + +/* + * Compression initialization for writing raw-coefficient data. + * Before calling this, all parameters and a data destination must be set up. + * Call jpeg_finish_compress() to actually write the data. + * + * The number of passed virtual arrays must match cinfo->num_components. + * Note that the virtual arrays need not be filled or even realized at + * the time write_coefficients is called; indeed, if the virtual arrays + * were requested from this compression object's memory manager, they + * typically will be realized during this routine and filled afterwards. + */ + +GLOBAL(void) +jpeg_write_coefficients (j_compress_ptr cinfo, jvirt_barray_ptr * coef_arrays) +{ + if (cinfo->global_state != CSTATE_START) + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + /* Mark all tables to be written */ + jpeg_suppress_tables(cinfo, FALSE); + /* (Re)initialize error mgr and destination modules */ + (*cinfo->err->reset_error_mgr) ((j_common_ptr) cinfo); + (*cinfo->dest->init_destination) (cinfo); + /* Perform master selection of active modules */ + transencode_master_selection(cinfo, coef_arrays); + /* Wait for jpeg_finish_compress() call */ + cinfo->next_scanline = 0; /* so jpeg_write_marker works */ + cinfo->global_state = CSTATE_WRCOEFS; +} + + +/* + * Initialize the compression object with default parameters, + * then copy from the source object all parameters needed for lossless + * transcoding. Parameters that can be varied without loss (such as + * scan script and Huffman optimization) are left in their default states. + */ + +GLOBAL(void) +jpeg_copy_critical_parameters (j_decompress_ptr srcinfo, + j_compress_ptr dstinfo) +{ + JQUANT_TBL ** qtblptr; + jpeg_component_info *incomp, *outcomp; + JQUANT_TBL *c_quant, *slot_quant; + int tblno, ci, coefi; + + /* Safety check to ensure start_compress not called yet. */ + if (dstinfo->global_state != CSTATE_START) + ERREXIT1(dstinfo, JERR_BAD_STATE, dstinfo->global_state); + /* Copy fundamental image dimensions */ + dstinfo->image_width = srcinfo->image_width; + dstinfo->image_height = srcinfo->image_height; + dstinfo->input_components = srcinfo->num_components; + dstinfo->in_color_space = srcinfo->jpeg_color_space; + dstinfo->jpeg_width = srcinfo->output_width; + dstinfo->jpeg_height = srcinfo->output_height; + dstinfo->min_DCT_h_scaled_size = srcinfo->min_DCT_h_scaled_size; + dstinfo->min_DCT_v_scaled_size = srcinfo->min_DCT_v_scaled_size; + /* Initialize all parameters to default values */ + jpeg_set_defaults(dstinfo); + /* jpeg_set_defaults may choose wrong colorspace, eg YCbCr if input is RGB. + * Fix it to get the right header markers for the image colorspace. + * Note: Entropy table assignment in jpeg_set_colorspace + * depends on color_transform. + * Adaption is also required for setting the appropriate + * entropy coding mode dependent on image data precision. + */ + dstinfo->color_transform = srcinfo->color_transform; + jpeg_set_colorspace(dstinfo, srcinfo->jpeg_color_space); + dstinfo->data_precision = srcinfo->data_precision; + dstinfo->arith_code = srcinfo->data_precision > 8 ? TRUE : FALSE; + dstinfo->CCIR601_sampling = srcinfo->CCIR601_sampling; + /* Copy the source's quantization tables. */ + for (tblno = 0; tblno < NUM_QUANT_TBLS; tblno++) { + if (srcinfo->quant_tbl_ptrs[tblno] != NULL) { + qtblptr = & dstinfo->quant_tbl_ptrs[tblno]; + if (*qtblptr == NULL) + *qtblptr = jpeg_alloc_quant_table((j_common_ptr) dstinfo); + MEMCOPY((*qtblptr)->quantval, + srcinfo->quant_tbl_ptrs[tblno]->quantval, + SIZEOF((*qtblptr)->quantval)); + (*qtblptr)->sent_table = FALSE; + } + } + /* Copy the source's per-component info. + * Note we assume jpeg_set_defaults has allocated the dest comp_info array. + */ + dstinfo->num_components = srcinfo->num_components; + if (dstinfo->num_components < 1 || dstinfo->num_components > MAX_COMPONENTS) + ERREXIT2(dstinfo, JERR_COMPONENT_COUNT, dstinfo->num_components, + MAX_COMPONENTS); + for (ci = 0, incomp = srcinfo->comp_info, outcomp = dstinfo->comp_info; + ci < dstinfo->num_components; ci++, incomp++, outcomp++) { + outcomp->component_id = incomp->component_id; + outcomp->h_samp_factor = incomp->h_samp_factor; + outcomp->v_samp_factor = incomp->v_samp_factor; + outcomp->quant_tbl_no = incomp->quant_tbl_no; + /* Make sure saved quantization table for component matches the qtable + * slot. If not, the input file re-used this qtable slot. + * IJG encoder currently cannot duplicate this. + */ + tblno = outcomp->quant_tbl_no; + if (tblno < 0 || tblno >= NUM_QUANT_TBLS || + srcinfo->quant_tbl_ptrs[tblno] == NULL) + ERREXIT1(dstinfo, JERR_NO_QUANT_TABLE, tblno); + slot_quant = srcinfo->quant_tbl_ptrs[tblno]; + c_quant = incomp->quant_table; + if (c_quant != NULL) { + for (coefi = 0; coefi < DCTSIZE2; coefi++) { + if (c_quant->quantval[coefi] != slot_quant->quantval[coefi]) + ERREXIT1(dstinfo, JERR_MISMATCHED_QUANT_TABLE, tblno); + } + } + /* Note: we do not copy the source's entropy table assignments; + * instead we rely on jpeg_set_colorspace to have made a suitable choice. + */ + } + /* Also copy JFIF version and resolution information, if available. + * Strictly speaking this isn't "critical" info, but it's nearly + * always appropriate to copy it if available. In particular, + * if the application chooses to copy JFIF 1.02 extension markers from + * the source file, we need to copy the version to make sure we don't + * emit a file that has 1.02 extensions but a claimed version of 1.01. + */ + if (srcinfo->saw_JFIF_marker) { + if (srcinfo->JFIF_major_version == 1 || + srcinfo->JFIF_major_version == 2) { + dstinfo->JFIF_major_version = srcinfo->JFIF_major_version; + dstinfo->JFIF_minor_version = srcinfo->JFIF_minor_version; + } + dstinfo->density_unit = srcinfo->density_unit; + dstinfo->X_density = srcinfo->X_density; + dstinfo->Y_density = srcinfo->Y_density; + } +} + + +LOCAL(void) +jpeg_calc_trans_dimensions (j_compress_ptr cinfo) +/* Do computations that are needed before master selection phase */ +{ + if (cinfo->min_DCT_h_scaled_size != cinfo->min_DCT_v_scaled_size) + ERREXIT2(cinfo, JERR_BAD_DCTSIZE, + cinfo->min_DCT_h_scaled_size, cinfo->min_DCT_v_scaled_size); + + cinfo->block_size = cinfo->min_DCT_h_scaled_size; +} + + +/* + * Master selection of compression modules for transcoding. + * This substitutes for jcinit.c's initialization of the full compressor. + */ + +LOCAL(void) +transencode_master_selection (j_compress_ptr cinfo, + jvirt_barray_ptr * coef_arrays) +{ + /* Do computations that are needed before master selection phase */ + jpeg_calc_trans_dimensions(cinfo); + + /* Initialize master control (includes parameter checking/processing) */ + jinit_c_master_control(cinfo, TRUE /* transcode only */); + + /* Entropy encoding: either Huffman or arithmetic coding. */ + if (cinfo->arith_code) + jinit_arith_encoder(cinfo); + else { + jinit_huff_encoder(cinfo); + } + + /* We need a special coefficient buffer controller. */ + transencode_coef_controller(cinfo, coef_arrays); + + jinit_marker_writer(cinfo); + + /* We can now tell the memory manager to allocate virtual arrays. */ + (*cinfo->mem->realize_virt_arrays) ((j_common_ptr) cinfo); + + /* Write the datastream header (SOI, JFIF) immediately. + * Frame and scan headers are postponed till later. + * This lets application insert special markers after the SOI. + */ + (*cinfo->marker->write_file_header) (cinfo); +} + + +/* + * The rest of this file is a special implementation of the coefficient + * buffer controller. This is similar to jccoefct.c, but it handles only + * output from presupplied virtual arrays. Furthermore, we generate any + * dummy padding blocks on-the-fly rather than expecting them to be present + * in the arrays. + */ + +/* Private buffer controller object */ + +typedef struct { + struct jpeg_c_coef_controller pub; /* public fields */ + + JDIMENSION iMCU_row_num; /* iMCU row # within image */ + JDIMENSION MCU_ctr; /* counts MCUs processed in current row */ + int MCU_vert_offset; /* counts MCU rows within iMCU row */ + int MCU_rows_per_iMCU_row; /* number of such rows needed */ + + /* Virtual block array for each component. */ + jvirt_barray_ptr * whole_image; + + /* Workspace for constructing dummy blocks at right/bottom edges. */ + JBLOCK dummy_buffer[C_MAX_BLOCKS_IN_MCU]; +} my_coef_controller; + +typedef my_coef_controller * my_coef_ptr; + + +LOCAL(void) +start_iMCU_row (j_compress_ptr cinfo) +/* Reset within-iMCU-row counters for a new row */ +{ + my_coef_ptr coef = (my_coef_ptr) cinfo->coef; + + /* In an interleaved scan, an MCU row is the same as an iMCU row. + * In a noninterleaved scan, an iMCU row has v_samp_factor MCU rows. + * But at the bottom of the image, process only what's left. + */ + if (cinfo->comps_in_scan > 1) { + coef->MCU_rows_per_iMCU_row = 1; + } else { + if (coef->iMCU_row_num < (cinfo->total_iMCU_rows-1)) + coef->MCU_rows_per_iMCU_row = cinfo->cur_comp_info[0]->v_samp_factor; + else + coef->MCU_rows_per_iMCU_row = cinfo->cur_comp_info[0]->last_row_height; + } + + coef->MCU_ctr = 0; + coef->MCU_vert_offset = 0; +} + + +/* + * Initialize for a processing pass. + */ + +METHODDEF(void) +start_pass_coef (j_compress_ptr cinfo, J_BUF_MODE pass_mode) +{ + my_coef_ptr coef = (my_coef_ptr) cinfo->coef; + + if (pass_mode != JBUF_CRANK_DEST) + ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); + + coef->iMCU_row_num = 0; + start_iMCU_row(cinfo); +} + + +/* + * Process some data. + * We process the equivalent of one fully interleaved MCU row ("iMCU" row) + * per call, ie, v_samp_factor block rows for each component in the scan. + * The data is obtained from the virtual arrays and fed to the entropy coder. + * Returns TRUE if the iMCU row is completed, FALSE if suspended. + * + * NB: input_buf is ignored; it is likely to be a NULL pointer. + */ + +METHODDEF(boolean) +compress_output (j_compress_ptr cinfo, JSAMPIMAGE input_buf) +{ + my_coef_ptr coef = (my_coef_ptr) cinfo->coef; + JDIMENSION MCU_col_num; /* index of current MCU within row */ + JDIMENSION last_MCU_col = cinfo->MCUs_per_row - 1; + JDIMENSION last_iMCU_row = cinfo->total_iMCU_rows - 1; + int blkn, ci, xindex, yindex, yoffset, blockcnt; + JDIMENSION start_col; + JBLOCKARRAY buffer[MAX_COMPS_IN_SCAN]; + JBLOCKROW MCU_buffer[C_MAX_BLOCKS_IN_MCU]; + JBLOCKROW buffer_ptr; + jpeg_component_info *compptr; + + /* Align the virtual buffers for the components used in this scan. */ + for (ci = 0; ci < cinfo->comps_in_scan; ci++) { + compptr = cinfo->cur_comp_info[ci]; + buffer[ci] = (*cinfo->mem->access_virt_barray) + ((j_common_ptr) cinfo, coef->whole_image[compptr->component_index], + coef->iMCU_row_num * compptr->v_samp_factor, + (JDIMENSION) compptr->v_samp_factor, FALSE); + } + + /* Loop to process one whole iMCU row */ + for (yoffset = coef->MCU_vert_offset; yoffset < coef->MCU_rows_per_iMCU_row; + yoffset++) { + for (MCU_col_num = coef->MCU_ctr; MCU_col_num <= last_MCU_col; + MCU_col_num++) { + /* Construct list of pointers to DCT blocks belonging to this MCU */ + blkn = 0; /* index of current DCT block within MCU */ + for (ci = 0; ci < cinfo->comps_in_scan; ci++) { + compptr = cinfo->cur_comp_info[ci]; + blockcnt = (MCU_col_num < last_MCU_col) ? compptr->MCU_width + : compptr->last_col_width; + start_col = MCU_col_num * compptr->MCU_width; + for (yindex = 0; yindex < compptr->MCU_height; yindex++) { + if (coef->iMCU_row_num < last_iMCU_row || + yoffset + yindex < compptr->last_row_height) { + /* Fill in pointers to real blocks in this row */ + buffer_ptr = buffer[ci][yoffset + yindex] + start_col; + xindex = blockcnt; + do { + MCU_buffer[blkn++] = buffer_ptr++; + } while (--xindex); + /* Dummy blocks at right edge */ + if ((xindex = compptr->MCU_width - blockcnt) == 0) + continue; + } else { + /* At bottom of image, need a whole row of dummy blocks */ + xindex = compptr->MCU_width; + } + /* Fill in any dummy blocks needed in this row. + * Dummy blocks are filled in the same way as in jccoefct.c: + * all zeroes in the AC entries, DC entries equal to previous + * block's DC value. The init routine has already zeroed the + * AC entries, so we need only set the DC entries correctly. + */ + buffer_ptr = coef->dummy_buffer + blkn; + do { + buffer_ptr[0][0] = MCU_buffer[blkn-1][0][0]; + MCU_buffer[blkn++] = buffer_ptr++; + } while (--xindex); + } + } + /* Try to write the MCU. */ + if (! (*cinfo->entropy->encode_mcu) (cinfo, MCU_buffer)) { + /* Suspension forced; update state counters and exit */ + coef->MCU_vert_offset = yoffset; + coef->MCU_ctr = MCU_col_num; + return FALSE; + } + } + /* Completed an MCU row, but perhaps not an iMCU row */ + coef->MCU_ctr = 0; + } + /* Completed the iMCU row, advance counters for next one */ + coef->iMCU_row_num++; + start_iMCU_row(cinfo); + return TRUE; +} + + +/* + * Initialize coefficient buffer controller. + * + * Each passed coefficient array must be the right size for that + * coefficient: width_in_blocks wide and height_in_blocks high, + * with unitheight at least v_samp_factor. + */ + +LOCAL(void) +transencode_coef_controller (j_compress_ptr cinfo, + jvirt_barray_ptr * coef_arrays) +{ + my_coef_ptr coef; + + coef = (my_coef_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_coef_controller)); + cinfo->coef = &coef->pub; + coef->pub.start_pass = start_pass_coef; + coef->pub.compress_data = compress_output; + + /* Save pointer to virtual arrays */ + coef->whole_image = coef_arrays; + + /* Pre-zero space for dummy DCT blocks */ + MEMZERO(coef->dummy_buffer, SIZEOF(coef->dummy_buffer)); +} diff --git a/cmake/externals/libjpeg/jdapimin.c b/cmake/externals/libjpeg/jdapimin.c new file mode 100644 index 000000000..785e52722 --- /dev/null +++ b/cmake/externals/libjpeg/jdapimin.c @@ -0,0 +1,412 @@ +/* + * jdapimin.c + * + * Copyright (C) 1994-1998, Thomas G. Lane. + * Modified 2009-2020 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains application interface code for the decompression half + * of the JPEG library. These are the "minimum" API routines that may be + * needed in either the normal full-decompression case or the + * transcoding-only case. + * + * Most of the routines intended to be called directly by an application + * are in this file or in jdapistd.c. But also see jcomapi.c for routines + * shared by compression and decompression, and jdtrans.c for the transcoding + * case. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* + * Initialization of a JPEG decompression object. + * The error manager must already be set up (in case memory manager fails). + */ + +GLOBAL(void) +jpeg_CreateDecompress (j_decompress_ptr cinfo, int version, size_t structsize) +{ + int i; + + /* Guard against version mismatches between library and caller. */ + cinfo->mem = NULL; /* so jpeg_destroy knows mem mgr not called */ + if (version != JPEG_LIB_VERSION) + ERREXIT2(cinfo, JERR_BAD_LIB_VERSION, JPEG_LIB_VERSION, version); + if (structsize != SIZEOF(struct jpeg_decompress_struct)) + ERREXIT2(cinfo, JERR_BAD_STRUCT_SIZE, + (int) SIZEOF(struct jpeg_decompress_struct), (int) structsize); + + /* For debugging purposes, we zero the whole master structure. + * But the application has already set the err pointer, and may have set + * client_data, so we have to save and restore those fields. + * Note: if application hasn't set client_data, tools like Purify may + * complain here. + */ + { + struct jpeg_error_mgr * err = cinfo->err; + void * client_data = cinfo->client_data; /* ignore Purify complaint here */ + MEMZERO(cinfo, SIZEOF(struct jpeg_decompress_struct)); + cinfo->err = err; + cinfo->client_data = client_data; + } + cinfo->is_decompressor = TRUE; + + /* Initialize a memory manager instance for this object */ + jinit_memory_mgr((j_common_ptr) cinfo); + + /* Zero out pointers to permanent structures. */ + cinfo->progress = NULL; + cinfo->src = NULL; + + for (i = 0; i < NUM_QUANT_TBLS; i++) + cinfo->quant_tbl_ptrs[i] = NULL; + + for (i = 0; i < NUM_HUFF_TBLS; i++) { + cinfo->dc_huff_tbl_ptrs[i] = NULL; + cinfo->ac_huff_tbl_ptrs[i] = NULL; + } + + /* Initialize marker processor so application can override methods + * for COM, APPn markers before calling jpeg_read_header. + */ + cinfo->marker_list = NULL; + jinit_marker_reader(cinfo); + + /* And initialize the overall input controller. */ + jinit_input_controller(cinfo); + + /* OK, I'm ready */ + cinfo->global_state = DSTATE_START; +} + + +/* + * Destruction of a JPEG decompression object + */ + +GLOBAL(void) +jpeg_destroy_decompress (j_decompress_ptr cinfo) +{ + jpeg_destroy((j_common_ptr) cinfo); /* use common routine */ +} + + +/* + * Abort processing of a JPEG decompression operation, + * but don't destroy the object itself. + */ + +GLOBAL(void) +jpeg_abort_decompress (j_decompress_ptr cinfo) +{ + jpeg_abort((j_common_ptr) cinfo); /* use common routine */ +} + + +/* + * Set default decompression parameters. + */ + +LOCAL(void) +default_decompress_parms (j_decompress_ptr cinfo) +{ + int cid0, cid1, cid2, cid3; + + /* Guess the input colorspace, and set output colorspace accordingly. */ + /* Note application may override our guesses. */ + switch (cinfo->num_components) { + case 1: + cinfo->jpeg_color_space = JCS_GRAYSCALE; + cinfo->out_color_space = JCS_GRAYSCALE; + break; + + case 3: + cid0 = cinfo->comp_info[0].component_id; + cid1 = cinfo->comp_info[1].component_id; + cid2 = cinfo->comp_info[2].component_id; + + /* For robust detection of standard colorspaces + * regardless of the presence of special markers, + * check component IDs from SOF marker first. + */ + if (cid0 == 0x01 && cid1 == 0x02 && cid2 == 0x03) + cinfo->jpeg_color_space = JCS_YCbCr; + else if (cid0 == 0x01 && cid1 == 0x22 && cid2 == 0x23) + cinfo->jpeg_color_space = JCS_BG_YCC; + else if (cid0 == 0x52 && cid1 == 0x47 && cid2 == 0x42) + cinfo->jpeg_color_space = JCS_RGB; /* ASCII 'R', 'G', 'B' */ + else if (cid0 == 0x72 && cid1 == 0x67 && cid2 == 0x62) + cinfo->jpeg_color_space = JCS_BG_RGB; /* ASCII 'r', 'g', 'b' */ + else if (cinfo->saw_JFIF_marker) + cinfo->jpeg_color_space = JCS_YCbCr; /* assume it's YCbCr */ + else if (cinfo->saw_Adobe_marker) { + switch (cinfo->Adobe_transform) { + case 0: + cinfo->jpeg_color_space = JCS_RGB; + break; + case 1: + cinfo->jpeg_color_space = JCS_YCbCr; + break; + default: + WARNMS1(cinfo, JWRN_ADOBE_XFORM, cinfo->Adobe_transform); + cinfo->jpeg_color_space = JCS_YCbCr; /* assume it's YCbCr */ + } + } else { + TRACEMS3(cinfo, 1, JTRC_UNKNOWN_IDS, cid0, cid1, cid2); + cinfo->jpeg_color_space = JCS_YCbCr; /* assume it's YCbCr */ + } + /* Always guess RGB is proper output colorspace. */ + cinfo->out_color_space = JCS_RGB; + break; + + case 4: + cid0 = cinfo->comp_info[0].component_id; + cid1 = cinfo->comp_info[1].component_id; + cid2 = cinfo->comp_info[2].component_id; + cid3 = cinfo->comp_info[3].component_id; + + /* For robust detection of standard colorspaces + * regardless of the presence of special markers, + * check component IDs from SOF marker first. + */ + if (cid0 == 0x01 && cid1 == 0x02 && cid2 == 0x03 && cid3 == 0x04) + cinfo->jpeg_color_space = JCS_YCCK; + else if (cid0 == 0x43 && cid1 == 0x4D && cid2 == 0x59 && cid3 == 0x4B) + cinfo->jpeg_color_space = JCS_CMYK; /* ASCII 'C', 'M', 'Y', 'K' */ + else if (cinfo->saw_Adobe_marker) { + switch (cinfo->Adobe_transform) { + case 0: + cinfo->jpeg_color_space = JCS_CMYK; + break; + case 2: + cinfo->jpeg_color_space = JCS_YCCK; + break; + default: + WARNMS1(cinfo, JWRN_ADOBE_XFORM, cinfo->Adobe_transform); + cinfo->jpeg_color_space = JCS_YCCK; /* assume it's YCCK */ + } + } else { + /* Unknown IDs and no special markers, assume straight CMYK. */ + cinfo->jpeg_color_space = JCS_CMYK; + } + cinfo->out_color_space = JCS_CMYK; + break; + + default: + cinfo->jpeg_color_space = JCS_UNKNOWN; + cinfo->out_color_space = JCS_UNKNOWN; + } + + /* Set defaults for other decompression parameters. */ + cinfo->scale_num = cinfo->block_size; /* 1:1 scaling */ + cinfo->scale_denom = cinfo->block_size; + cinfo->output_gamma = 1.0; + cinfo->buffered_image = FALSE; + cinfo->raw_data_out = FALSE; + cinfo->dct_method = JDCT_DEFAULT; + cinfo->do_fancy_upsampling = TRUE; + cinfo->do_block_smoothing = TRUE; + cinfo->quantize_colors = FALSE; + /* We set these in case application only sets quantize_colors. */ + cinfo->dither_mode = JDITHER_FS; +#ifdef QUANT_2PASS_SUPPORTED + cinfo->two_pass_quantize = TRUE; +#else + cinfo->two_pass_quantize = FALSE; +#endif + cinfo->desired_number_of_colors = 256; + cinfo->colormap = NULL; + /* Initialize for no mode change in buffered-image mode. */ + cinfo->enable_1pass_quant = FALSE; + cinfo->enable_external_quant = FALSE; + cinfo->enable_2pass_quant = FALSE; +} + + +/* + * Decompression startup: read start of JPEG datastream to see what's there. + * Need only initialize JPEG object and supply a data source before calling. + * + * This routine will read as far as the first SOS marker (ie, actual start of + * compressed data), and will save all tables and parameters in the JPEG + * object. It will also initialize the decompression parameters to default + * values, and finally return JPEG_HEADER_OK. On return, the application may + * adjust the decompression parameters and then call jpeg_start_decompress. + * (Or, if the application only wanted to determine the image parameters, + * the data need not be decompressed. In that case, call jpeg_abort or + * jpeg_destroy to release any temporary space.) + * If an abbreviated (tables only) datastream is presented, the routine will + * return JPEG_HEADER_TABLES_ONLY upon reaching EOI. The application may then + * re-use the JPEG object to read the abbreviated image datastream(s). + * It is unnecessary (but OK) to call jpeg_abort in this case. + * The JPEG_SUSPENDED return code only occurs if the data source module + * requests suspension of the decompressor. In this case the application + * should load more source data and then re-call jpeg_read_header to resume + * processing. + * If a non-suspending data source is used and require_image is TRUE, then the + * return code need not be inspected since only JPEG_HEADER_OK is possible. + * + * This routine is now just a front end to jpeg_consume_input, with some + * extra error checking. + */ + +GLOBAL(int) +jpeg_read_header (j_decompress_ptr cinfo, boolean require_image) +{ + int retcode; + + if (cinfo->global_state != DSTATE_START && + cinfo->global_state != DSTATE_INHEADER) + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + + retcode = jpeg_consume_input(cinfo); + + switch (retcode) { + case JPEG_REACHED_SOS: + retcode = JPEG_HEADER_OK; + break; + case JPEG_REACHED_EOI: + if (require_image) /* Complain if application wanted an image */ + ERREXIT(cinfo, JERR_NO_IMAGE); + /* Reset to start state; it would be safer to require the application to + * call jpeg_abort, but we can't change it now for compatibility reasons. + * A side effect is to free any temporary memory (there shouldn't be any). + */ + jpeg_abort((j_common_ptr) cinfo); /* sets state = DSTATE_START */ + retcode = JPEG_HEADER_TABLES_ONLY; + break; + case JPEG_SUSPENDED: + /* no work */ + break; + } + + return retcode; +} + + +/* + * Consume data in advance of what the decompressor requires. + * This can be called at any time once the decompressor object has + * been created and a data source has been set up. + * + * This routine is essentially a state machine that handles a couple + * of critical state-transition actions, namely initial setup and + * transition from header scanning to ready-for-start_decompress. + * All the actual input is done via the input controller's consume_input + * method. + */ + +GLOBAL(int) +jpeg_consume_input (j_decompress_ptr cinfo) +{ + int retcode = JPEG_SUSPENDED; + + /* NB: every possible DSTATE value should be listed in this switch */ + switch (cinfo->global_state) { + case DSTATE_START: + /* Start-of-datastream actions: reset appropriate modules */ + (*cinfo->inputctl->reset_input_controller) (cinfo); + /* Initialize application's data source module */ + (*cinfo->src->init_source) (cinfo); + cinfo->global_state = DSTATE_INHEADER; + /*FALLTHROUGH*/ + case DSTATE_INHEADER: + retcode = (*cinfo->inputctl->consume_input) (cinfo); + if (retcode == JPEG_REACHED_SOS) { /* Found SOS, prepare to decompress */ + /* Set up default parameters based on header data */ + default_decompress_parms(cinfo); + /* Set global state: ready for start_decompress */ + cinfo->global_state = DSTATE_READY; + } + break; + case DSTATE_READY: + /* Can't advance past first SOS until start_decompress is called */ + retcode = JPEG_REACHED_SOS; + break; + case DSTATE_PRELOAD: + case DSTATE_PRESCAN: + case DSTATE_SCANNING: + case DSTATE_RAW_OK: + case DSTATE_BUFIMAGE: + case DSTATE_BUFPOST: + case DSTATE_STOPPING: + retcode = (*cinfo->inputctl->consume_input) (cinfo); + break; + default: + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + } + return retcode; +} + + +/* + * Have we finished reading the input file? + */ + +GLOBAL(boolean) +jpeg_input_complete (j_decompress_ptr cinfo) +{ + /* Check for valid jpeg object */ + if (cinfo->global_state < DSTATE_START || + cinfo->global_state > DSTATE_STOPPING) + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + return cinfo->inputctl->eoi_reached; +} + + +/* + * Is there more than one scan? + */ + +GLOBAL(boolean) +jpeg_has_multiple_scans (j_decompress_ptr cinfo) +{ + /* Only valid after jpeg_read_header completes */ + if (cinfo->global_state < DSTATE_READY || + cinfo->global_state > DSTATE_STOPPING) + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + return cinfo->inputctl->has_multiple_scans; +} + + +/* + * Finish JPEG decompression. + * + * This will normally just verify the file trailer and release temp storage. + * + * Returns FALSE if suspended. The return value need be inspected only if + * a suspending data source is used. + */ + +GLOBAL(boolean) +jpeg_finish_decompress (j_decompress_ptr cinfo) +{ + if ((cinfo->global_state == DSTATE_SCANNING || + cinfo->global_state == DSTATE_RAW_OK) && ! cinfo->buffered_image) { + /* Terminate final pass of non-buffered mode */ + if (cinfo->output_scanline < cinfo->output_height) + ERREXIT(cinfo, JERR_TOO_LITTLE_DATA); + (*cinfo->master->finish_output_pass) (cinfo); + cinfo->global_state = DSTATE_STOPPING; + } else if (cinfo->global_state == DSTATE_BUFIMAGE) { + /* Finishing after a buffered-image operation */ + cinfo->global_state = DSTATE_STOPPING; + } else if (cinfo->global_state != DSTATE_STOPPING) { + /* STOPPING = repeat call after a suspension, anything else is error */ + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + } + /* Read until EOI */ + while (! cinfo->inputctl->eoi_reached) { + if ((*cinfo->inputctl->consume_input) (cinfo) == JPEG_SUSPENDED) + return FALSE; /* Suspend, come back later */ + } + /* Do final cleanup */ + (*cinfo->src->term_source) (cinfo); + /* We can use jpeg_abort to release memory and reset global_state */ + jpeg_abort((j_common_ptr) cinfo); + return TRUE; +} diff --git a/cmake/externals/libjpeg/jdapistd.c b/cmake/externals/libjpeg/jdapistd.c new file mode 100644 index 000000000..7f3a78b25 --- /dev/null +++ b/cmake/externals/libjpeg/jdapistd.c @@ -0,0 +1,276 @@ +/* + * jdapistd.c + * + * Copyright (C) 1994-1996, Thomas G. Lane. + * Modified 2002-2013 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains application interface code for the decompression half + * of the JPEG library. These are the "standard" API routines that are + * used in the normal full-decompression case. They are not used by a + * transcoding-only application. Note that if an application links in + * jpeg_start_decompress, it will end up linking in the entire decompressor. + * We thus must separate this file from jdapimin.c to avoid linking the + * whole decompression library into a transcoder. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* Forward declarations */ +LOCAL(boolean) output_pass_setup JPP((j_decompress_ptr cinfo)); + + +/* + * Decompression initialization. + * jpeg_read_header must be completed before calling this. + * + * If a multipass operating mode was selected, this will do all but the + * last pass, and thus may take a great deal of time. + * + * Returns FALSE if suspended. The return value need be inspected only if + * a suspending data source is used. + */ + +GLOBAL(boolean) +jpeg_start_decompress (j_decompress_ptr cinfo) +{ + if (cinfo->global_state == DSTATE_READY) { + /* First call: initialize master control, select active modules */ + jinit_master_decompress(cinfo); + if (cinfo->buffered_image) { + /* No more work here; expecting jpeg_start_output next */ + cinfo->global_state = DSTATE_BUFIMAGE; + return TRUE; + } + cinfo->global_state = DSTATE_PRELOAD; + } + if (cinfo->global_state == DSTATE_PRELOAD) { + /* If file has multiple scans, absorb them all into the coef buffer */ + if (cinfo->inputctl->has_multiple_scans) { +#ifdef D_MULTISCAN_FILES_SUPPORTED + for (;;) { + int retcode; + /* Call progress monitor hook if present */ + if (cinfo->progress != NULL) + (*cinfo->progress->progress_monitor) ((j_common_ptr) cinfo); + /* Absorb some more input */ + retcode = (*cinfo->inputctl->consume_input) (cinfo); + if (retcode == JPEG_SUSPENDED) + return FALSE; + if (retcode == JPEG_REACHED_EOI) + break; + /* Advance progress counter if appropriate */ + if (cinfo->progress != NULL && + (retcode == JPEG_ROW_COMPLETED || retcode == JPEG_REACHED_SOS)) { + if (++cinfo->progress->pass_counter >= cinfo->progress->pass_limit) { + /* jdmaster underestimated number of scans; ratchet up one scan */ + cinfo->progress->pass_limit += (long) cinfo->total_iMCU_rows; + } + } + } +#else + ERREXIT(cinfo, JERR_NOT_COMPILED); +#endif /* D_MULTISCAN_FILES_SUPPORTED */ + } + cinfo->output_scan_number = cinfo->input_scan_number; + } else if (cinfo->global_state != DSTATE_PRESCAN) + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + /* Perform any dummy output passes, and set up for the final pass */ + return output_pass_setup(cinfo); +} + + +/* + * Set up for an output pass, and perform any dummy pass(es) needed. + * Common subroutine for jpeg_start_decompress and jpeg_start_output. + * Entry: global_state = DSTATE_PRESCAN only if previously suspended. + * Exit: If done, returns TRUE and sets global_state for proper output mode. + * If suspended, returns FALSE and sets global_state = DSTATE_PRESCAN. + */ + +LOCAL(boolean) +output_pass_setup (j_decompress_ptr cinfo) +{ + if (cinfo->global_state != DSTATE_PRESCAN) { + /* First call: do pass setup */ + (*cinfo->master->prepare_for_output_pass) (cinfo); + cinfo->output_scanline = 0; + cinfo->global_state = DSTATE_PRESCAN; + } + /* Loop over any required dummy passes */ + while (cinfo->master->is_dummy_pass) { +#ifdef QUANT_2PASS_SUPPORTED + /* Crank through the dummy pass */ + while (cinfo->output_scanline < cinfo->output_height) { + JDIMENSION last_scanline; + /* Call progress monitor hook if present */ + if (cinfo->progress != NULL) { + cinfo->progress->pass_counter = (long) cinfo->output_scanline; + cinfo->progress->pass_limit = (long) cinfo->output_height; + (*cinfo->progress->progress_monitor) ((j_common_ptr) cinfo); + } + /* Process some data */ + last_scanline = cinfo->output_scanline; + (*cinfo->main->process_data) (cinfo, (JSAMPARRAY) NULL, + &cinfo->output_scanline, (JDIMENSION) 0); + if (cinfo->output_scanline == last_scanline) + return FALSE; /* No progress made, must suspend */ + } + /* Finish up dummy pass, and set up for another one */ + (*cinfo->master->finish_output_pass) (cinfo); + (*cinfo->master->prepare_for_output_pass) (cinfo); + cinfo->output_scanline = 0; +#else + ERREXIT(cinfo, JERR_NOT_COMPILED); +#endif /* QUANT_2PASS_SUPPORTED */ + } + /* Ready for application to drive output pass through + * jpeg_read_scanlines or jpeg_read_raw_data. + */ + cinfo->global_state = cinfo->raw_data_out ? DSTATE_RAW_OK : DSTATE_SCANNING; + return TRUE; +} + + +/* + * Read some scanlines of data from the JPEG decompressor. + * + * The return value will be the number of lines actually read. + * This may be less than the number requested in several cases, + * including bottom of image, data source suspension, and operating + * modes that emit multiple scanlines at a time. + * + * Note: we warn about excess calls to jpeg_read_scanlines() since + * this likely signals an application programmer error. However, + * an oversize buffer (max_lines > scanlines remaining) is not an error. + */ + +GLOBAL(JDIMENSION) +jpeg_read_scanlines (j_decompress_ptr cinfo, JSAMPARRAY scanlines, + JDIMENSION max_lines) +{ + JDIMENSION row_ctr; + + if (cinfo->global_state != DSTATE_SCANNING) + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + if (cinfo->output_scanline >= cinfo->output_height) { + WARNMS(cinfo, JWRN_TOO_MUCH_DATA); + return 0; + } + + /* Call progress monitor hook if present */ + if (cinfo->progress != NULL) { + cinfo->progress->pass_counter = (long) cinfo->output_scanline; + cinfo->progress->pass_limit = (long) cinfo->output_height; + (*cinfo->progress->progress_monitor) ((j_common_ptr) cinfo); + } + + /* Process some data */ + row_ctr = 0; + (*cinfo->main->process_data) (cinfo, scanlines, &row_ctr, max_lines); + cinfo->output_scanline += row_ctr; + return row_ctr; +} + + +/* + * Alternate entry point to read raw data. + * Processes exactly one iMCU row per call, unless suspended. + */ + +GLOBAL(JDIMENSION) +jpeg_read_raw_data (j_decompress_ptr cinfo, JSAMPIMAGE data, + JDIMENSION max_lines) +{ + JDIMENSION lines_per_iMCU_row; + + if (cinfo->global_state != DSTATE_RAW_OK) + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + if (cinfo->output_scanline >= cinfo->output_height) { + WARNMS(cinfo, JWRN_TOO_MUCH_DATA); + return 0; + } + + /* Call progress monitor hook if present */ + if (cinfo->progress != NULL) { + cinfo->progress->pass_counter = (long) cinfo->output_scanline; + cinfo->progress->pass_limit = (long) cinfo->output_height; + (*cinfo->progress->progress_monitor) ((j_common_ptr) cinfo); + } + + /* Verify that at least one iMCU row can be returned. */ + lines_per_iMCU_row = cinfo->max_v_samp_factor * cinfo->min_DCT_v_scaled_size; + if (max_lines < lines_per_iMCU_row) + ERREXIT(cinfo, JERR_BUFFER_SIZE); + + /* Decompress directly into user's buffer. */ + if (! (*cinfo->coef->decompress_data) (cinfo, data)) + return 0; /* suspension forced, can do nothing more */ + + /* OK, we processed one iMCU row. */ + cinfo->output_scanline += lines_per_iMCU_row; + return lines_per_iMCU_row; +} + + +/* Additional entry points for buffered-image mode. */ + +#ifdef D_MULTISCAN_FILES_SUPPORTED + +/* + * Initialize for an output pass in buffered-image mode. + */ + +GLOBAL(boolean) +jpeg_start_output (j_decompress_ptr cinfo, int scan_number) +{ + if (cinfo->global_state != DSTATE_BUFIMAGE && + cinfo->global_state != DSTATE_PRESCAN) + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + /* Limit scan number to valid range */ + if (scan_number <= 0) + scan_number = 1; + if (cinfo->inputctl->eoi_reached && + scan_number > cinfo->input_scan_number) + scan_number = cinfo->input_scan_number; + cinfo->output_scan_number = scan_number; + /* Perform any dummy output passes, and set up for the real pass */ + return output_pass_setup(cinfo); +} + + +/* + * Finish up after an output pass in buffered-image mode. + * + * Returns FALSE if suspended. The return value need be inspected only if + * a suspending data source is used. + */ + +GLOBAL(boolean) +jpeg_finish_output (j_decompress_ptr cinfo) +{ + if ((cinfo->global_state == DSTATE_SCANNING || + cinfo->global_state == DSTATE_RAW_OK) && cinfo->buffered_image) { + /* Terminate this pass. */ + /* We do not require the whole pass to have been completed. */ + (*cinfo->master->finish_output_pass) (cinfo); + cinfo->global_state = DSTATE_BUFPOST; + } else if (cinfo->global_state != DSTATE_BUFPOST) { + /* BUFPOST = repeat call after a suspension, anything else is error */ + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + } + /* Read markers looking for SOS or EOI */ + while (cinfo->input_scan_number <= cinfo->output_scan_number && + ! cinfo->inputctl->eoi_reached) { + if ((*cinfo->inputctl->consume_input) (cinfo) == JPEG_SUSPENDED) + return FALSE; /* Suspend, come back later */ + } + cinfo->global_state = DSTATE_BUFIMAGE; + return TRUE; +} + +#endif /* D_MULTISCAN_FILES_SUPPORTED */ diff --git a/cmake/externals/libjpeg/jdarith.c b/cmake/externals/libjpeg/jdarith.c new file mode 100644 index 000000000..2c9abe23f --- /dev/null +++ b/cmake/externals/libjpeg/jdarith.c @@ -0,0 +1,796 @@ +/* + * jdarith.c + * + * Developed 1997-2020 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains portable arithmetic entropy decoding routines for JPEG + * (implementing the ISO/IEC IS 10918-1 and CCITT Recommendation ITU-T T.81). + * + * Both sequential and progressive modes are supported in this single module. + * + * Suspension is not currently supported in this module. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* Expanded entropy decoder object for arithmetic decoding. */ + +typedef struct { + struct jpeg_entropy_decoder pub; /* public fields */ + + INT32 c; /* C register, base of coding interval + input bit buffer */ + INT32 a; /* A register, normalized size of coding interval */ + int ct; /* bit shift counter, # of bits left in bit buffer part of C */ + /* init: ct = -16 */ + /* run: ct = 0..7 */ + /* error: ct = -1 */ + int last_dc_val[MAX_COMPS_IN_SCAN]; /* last DC coef for each component */ + int dc_context[MAX_COMPS_IN_SCAN]; /* context index for DC conditioning */ + + unsigned int restarts_to_go; /* MCUs left in this restart interval */ + + /* Pointers to statistics areas (these workspaces have image lifespan) */ + unsigned char * dc_stats[NUM_ARITH_TBLS]; + unsigned char * ac_stats[NUM_ARITH_TBLS]; + + /* Statistics bin for coding with fixed probability 0.5 */ + unsigned char fixed_bin[4]; +} arith_entropy_decoder; + +typedef arith_entropy_decoder * arith_entropy_ptr; + +/* The following two definitions specify the allocation chunk size + * for the statistics area. + * According to sections F.1.4.4.1.3 and F.1.4.4.2, we need at least + * 49 statistics bins for DC, and 245 statistics bins for AC coding. + * + * We use a compact representation with 1 byte per statistics bin, + * thus the numbers directly represent byte sizes. + * This 1 byte per statistics bin contains the meaning of the MPS + * (more probable symbol) in the highest bit (mask 0x80), and the + * index into the probability estimation state machine table + * in the lower bits (mask 0x7F). + */ + +#define DC_STAT_BINS 64 +#define AC_STAT_BINS 256 + + +LOCAL(int) +get_byte (j_decompress_ptr cinfo) +/* Read next input byte; we do not support suspension in this module. */ +{ + struct jpeg_source_mgr * src = cinfo->src; + + if (src->bytes_in_buffer == 0) + if (! (*src->fill_input_buffer) (cinfo)) + ERREXIT(cinfo, JERR_CANT_SUSPEND); + src->bytes_in_buffer--; + return GETJOCTET(*src->next_input_byte++); +} + + +/* + * The core arithmetic decoding routine (common in JPEG and JBIG). + * This needs to go as fast as possible. + * Machine-dependent optimization facilities + * are not utilized in this portable implementation. + * However, this code should be fairly efficient and + * may be a good base for further optimizations anyway. + * + * Return value is 0 or 1 (binary decision). + * + * Note: I've changed the handling of the code base & bit + * buffer register C compared to other implementations + * based on the standards layout & procedures. + * While it also contains both the actual base of the + * coding interval (16 bits) and the next-bits buffer, + * the cut-point between these two parts is floating + * (instead of fixed) with the bit shift counter CT. + * Thus, we also need only one (variable instead of + * fixed size) shift for the LPS/MPS decision, and + * we can do away with any renormalization update + * of C (except for new data insertion, of course). + * + * I've also introduced a new scheme for accessing + * the probability estimation state machine table, + * derived from Markus Kuhn's JBIG implementation. + */ + +LOCAL(int) +arith_decode (j_decompress_ptr cinfo, unsigned char *st) +{ + register arith_entropy_ptr e = (arith_entropy_ptr) cinfo->entropy; + register unsigned char nl, nm; + register INT32 qe, temp; + register int sv, data; + + /* Renormalization & data input per section D.2.6 */ + while (e->a < 0x8000L) { + if (--e->ct < 0) { + /* Need to fetch next data byte */ + if (cinfo->unread_marker) + data = 0; /* stuff zero data */ + else { + data = get_byte(cinfo); /* read next input byte */ + if (data == 0xFF) { /* zero stuff or marker code */ + do data = get_byte(cinfo); + while (data == 0xFF); /* swallow extra 0xFF bytes */ + if (data == 0) + data = 0xFF; /* discard stuffed zero byte */ + else { + /* Note: Different from the Huffman decoder, hitting + * a marker while processing the compressed data + * segment is legal in arithmetic coding. + * The convention is to supply zero data + * then until decoding is complete. + */ + cinfo->unread_marker = data; + data = 0; + } + } + } + e->c = (e->c << 8) | data; /* insert data into C register */ + if ((e->ct += 8) < 0) /* update bit shift counter */ + /* Need more initial bytes */ + if (++e->ct == 0) + /* Got 2 initial bytes -> re-init A and exit loop */ + e->a = 0x8000L; /* => e->a = 0x10000L after loop exit */ + } + e->a <<= 1; + } + + /* Fetch values from our compact representation of Table D.3(D.2): + * Qe values and probability estimation state machine + */ + sv = *st; + qe = jpeg_aritab[sv & 0x7F]; /* => Qe_Value */ + nl = qe & 0xFF; qe >>= 8; /* Next_Index_LPS + Switch_MPS */ + nm = qe & 0xFF; qe >>= 8; /* Next_Index_MPS */ + + /* Decode & estimation procedures per sections D.2.4 & D.2.5 */ + temp = e->a - qe; + e->a = temp; + temp <<= e->ct; + if (e->c >= temp) { + e->c -= temp; + /* Conditional LPS (less probable symbol) exchange */ + if (e->a < qe) { + e->a = qe; + *st = (sv & 0x80) ^ nm; /* Estimate_after_MPS */ + } else { + e->a = qe; + *st = (sv & 0x80) ^ nl; /* Estimate_after_LPS */ + sv ^= 0x80; /* Exchange LPS/MPS */ + } + } else if (e->a < 0x8000L) { + /* Conditional MPS (more probable symbol) exchange */ + if (e->a < qe) { + *st = (sv & 0x80) ^ nl; /* Estimate_after_LPS */ + sv ^= 0x80; /* Exchange LPS/MPS */ + } else { + *st = (sv & 0x80) ^ nm; /* Estimate_after_MPS */ + } + } + + return sv >> 7; +} + + +/* + * Check for a restart marker & resynchronize decoder. + */ + +LOCAL(void) +process_restart (j_decompress_ptr cinfo) +{ + arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; + int ci; + jpeg_component_info * compptr; + + /* Advance past the RSTn marker */ + if (! (*cinfo->marker->read_restart_marker) (cinfo)) + ERREXIT(cinfo, JERR_CANT_SUSPEND); + + /* Re-initialize statistics areas */ + for (ci = 0; ci < cinfo->comps_in_scan; ci++) { + compptr = cinfo->cur_comp_info[ci]; + if (! cinfo->progressive_mode || (cinfo->Ss == 0 && cinfo->Ah == 0)) { + MEMZERO(entropy->dc_stats[compptr->dc_tbl_no], DC_STAT_BINS); + /* Reset DC predictions to 0 */ + entropy->last_dc_val[ci] = 0; + entropy->dc_context[ci] = 0; + } + if ((! cinfo->progressive_mode && cinfo->lim_Se) || + (cinfo->progressive_mode && cinfo->Ss)) { + MEMZERO(entropy->ac_stats[compptr->ac_tbl_no], AC_STAT_BINS); + } + } + + /* Reset arithmetic decoding variables */ + entropy->c = 0; + entropy->a = 0; + entropy->ct = -16; /* force reading 2 initial bytes to fill C */ + + /* Reset restart counter */ + entropy->restarts_to_go = cinfo->restart_interval; +} + + +/* + * Arithmetic MCU decoding. + * Each of these routines decodes and returns one MCU's worth of + * arithmetic-compressed coefficients. + * The coefficients are reordered from zigzag order into natural array order, + * but are not dequantized. + * + * The i'th block of the MCU is stored into the block pointed to by + * MCU_data[i]. WE ASSUME THIS AREA IS INITIALLY ZEROED BY THE CALLER. + */ + +/* + * MCU decoding for DC initial scan (either spectral selection, + * or first pass of successive approximation). + */ + +METHODDEF(boolean) +decode_mcu_DC_first (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data) +{ + arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; + JBLOCKROW block; + unsigned char *st; + int blkn, ci, tbl, sign; + int v, m; + + /* Process restart marker if needed */ + if (cinfo->restart_interval) { + if (entropy->restarts_to_go == 0) + process_restart(cinfo); + entropy->restarts_to_go--; + } + + if (entropy->ct == -1) return TRUE; /* if error do nothing */ + + /* Outer loop handles each block in the MCU */ + + for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { + block = MCU_data[blkn]; + ci = cinfo->MCU_membership[blkn]; + tbl = cinfo->cur_comp_info[ci]->dc_tbl_no; + + /* Sections F.2.4.1 & F.1.4.4.1: Decoding of DC coefficients */ + + /* Table F.4: Point to statistics bin S0 for DC coefficient coding */ + st = entropy->dc_stats[tbl] + entropy->dc_context[ci]; + + /* Figure F.19: Decode_DC_DIFF */ + if (arith_decode(cinfo, st) == 0) + entropy->dc_context[ci] = 0; + else { + /* Figure F.21: Decoding nonzero value v */ + /* Figure F.22: Decoding the sign of v */ + sign = arith_decode(cinfo, st + 1); + st += 2; st += sign; + /* Figure F.23: Decoding the magnitude category of v */ + if ((m = arith_decode(cinfo, st)) != 0) { + st = entropy->dc_stats[tbl] + 20; /* Table F.4: X1 = 20 */ + while (arith_decode(cinfo, st)) { + if ((m <<= 1) == (int) 0x8000U) { + WARNMS(cinfo, JWRN_ARITH_BAD_CODE); + entropy->ct = -1; /* magnitude overflow */ + return TRUE; + } + st += 1; + } + } + /* Section F.1.4.4.1.2: Establish dc_context conditioning category */ + if (m < (int) ((1L << cinfo->arith_dc_L[tbl]) >> 1)) + entropy->dc_context[ci] = 0; /* zero diff category */ + else if (m > (int) ((1L << cinfo->arith_dc_U[tbl]) >> 1)) + entropy->dc_context[ci] = 12 + (sign * 4); /* large diff category */ + else + entropy->dc_context[ci] = 4 + (sign * 4); /* small diff category */ + v = m; + /* Figure F.24: Decoding the magnitude bit pattern of v */ + st += 14; + while (m >>= 1) + if (arith_decode(cinfo, st)) v |= m; + v += 1; if (sign) v = -v; + entropy->last_dc_val[ci] += v; + } + + /* Scale and output the DC coefficient (assumes jpeg_natural_order[0]=0) */ + (*block)[0] = (JCOEF) (entropy->last_dc_val[ci] << cinfo->Al); + } + + return TRUE; +} + + +/* + * MCU decoding for AC initial scan (either spectral selection, + * or first pass of successive approximation). + */ + +METHODDEF(boolean) +decode_mcu_AC_first (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data) +{ + arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; + JBLOCKROW block; + unsigned char *st; + int tbl, sign, k; + int v, m; + const int * natural_order; + + /* Process restart marker if needed */ + if (cinfo->restart_interval) { + if (entropy->restarts_to_go == 0) + process_restart(cinfo); + entropy->restarts_to_go--; + } + + if (entropy->ct == -1) return TRUE; /* if error do nothing */ + + natural_order = cinfo->natural_order; + + /* There is always only one block per MCU */ + block = MCU_data[0]; + tbl = cinfo->cur_comp_info[0]->ac_tbl_no; + + /* Sections F.2.4.2 & F.1.4.4.2: Decoding of AC coefficients */ + + /* Figure F.20: Decode_AC_coefficients */ + k = cinfo->Ss - 1; + do { + st = entropy->ac_stats[tbl] + 3 * k; + if (arith_decode(cinfo, st)) break; /* EOB flag */ + for (;;) { + k++; + if (arith_decode(cinfo, st + 1)) break; + st += 3; + if (k >= cinfo->Se) { + WARNMS(cinfo, JWRN_ARITH_BAD_CODE); + entropy->ct = -1; /* spectral overflow */ + return TRUE; + } + } + /* Figure F.21: Decoding nonzero value v */ + /* Figure F.22: Decoding the sign of v */ + sign = arith_decode(cinfo, entropy->fixed_bin); + st += 2; + /* Figure F.23: Decoding the magnitude category of v */ + if ((m = arith_decode(cinfo, st)) != 0) { + if (arith_decode(cinfo, st)) { + m <<= 1; + st = entropy->ac_stats[tbl] + + (k <= cinfo->arith_ac_K[tbl] ? 189 : 217); + while (arith_decode(cinfo, st)) { + if ((m <<= 1) == (int) 0x8000U) { + WARNMS(cinfo, JWRN_ARITH_BAD_CODE); + entropy->ct = -1; /* magnitude overflow */ + return TRUE; + } + st += 1; + } + } + } + v = m; + /* Figure F.24: Decoding the magnitude bit pattern of v */ + st += 14; + while (m >>= 1) + if (arith_decode(cinfo, st)) v |= m; + v += 1; if (sign) v = -v; + /* Scale and output coefficient in natural (dezigzagged) order */ + (*block)[natural_order[k]] = (JCOEF) (v << cinfo->Al); + } while (k < cinfo->Se); + + return TRUE; +} + + +/* + * MCU decoding for DC successive approximation refinement scan. + * Note: we assume such scans can be multi-component, + * although the spec is not very clear on the point. + */ + +METHODDEF(boolean) +decode_mcu_DC_refine (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data) +{ + arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; + unsigned char *st; + JCOEF p1; + int blkn; + + /* Process restart marker if needed */ + if (cinfo->restart_interval) { + if (entropy->restarts_to_go == 0) + process_restart(cinfo); + entropy->restarts_to_go--; + } + + st = entropy->fixed_bin; /* use fixed probability estimation */ + p1 = 1 << cinfo->Al; /* 1 in the bit position being coded */ + + /* Outer loop handles each block in the MCU */ + + for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { + /* Encoded data is simply the next bit of the two's-complement DC value */ + if (arith_decode(cinfo, st)) + MCU_data[blkn][0][0] |= p1; + } + + return TRUE; +} + + +/* + * MCU decoding for AC successive approximation refinement scan. + */ + +METHODDEF(boolean) +decode_mcu_AC_refine (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data) +{ + arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; + JBLOCKROW block; + JCOEFPTR thiscoef; + unsigned char *st; + int tbl, k, kex; + JCOEF p1, m1; + const int * natural_order; + + /* Process restart marker if needed */ + if (cinfo->restart_interval) { + if (entropy->restarts_to_go == 0) + process_restart(cinfo); + entropy->restarts_to_go--; + } + + if (entropy->ct == -1) return TRUE; /* if error do nothing */ + + natural_order = cinfo->natural_order; + + /* There is always only one block per MCU */ + block = MCU_data[0]; + tbl = cinfo->cur_comp_info[0]->ac_tbl_no; + + p1 = 1 << cinfo->Al; /* 1 in the bit position being coded */ + m1 = -p1; /* -1 in the bit position being coded */ + + /* Establish EOBx (previous stage end-of-block) index */ + kex = cinfo->Se; + do { + if ((*block)[natural_order[kex]]) break; + } while (--kex); + + k = cinfo->Ss - 1; + do { + st = entropy->ac_stats[tbl] + 3 * k; + if (k >= kex) + if (arith_decode(cinfo, st)) break; /* EOB flag */ + for (;;) { + thiscoef = *block + natural_order[++k]; + if (*thiscoef) { /* previously nonzero coef */ + if (arith_decode(cinfo, st + 2)) { + if (*thiscoef < 0) + *thiscoef += m1; + else + *thiscoef += p1; + } + break; + } + if (arith_decode(cinfo, st + 1)) { /* newly nonzero coef */ + if (arith_decode(cinfo, entropy->fixed_bin)) + *thiscoef = m1; + else + *thiscoef = p1; + break; + } + st += 3; + if (k >= cinfo->Se) { + WARNMS(cinfo, JWRN_ARITH_BAD_CODE); + entropy->ct = -1; /* spectral overflow */ + return TRUE; + } + } + } while (k < cinfo->Se); + + return TRUE; +} + + +/* + * Decode one MCU's worth of arithmetic-compressed coefficients. + */ + +METHODDEF(boolean) +decode_mcu (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data) +{ + arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; + jpeg_component_info * compptr; + JBLOCKROW block; + unsigned char *st; + int blkn, ci, tbl, sign, k; + int v, m; + const int * natural_order; + + /* Process restart marker if needed */ + if (cinfo->restart_interval) { + if (entropy->restarts_to_go == 0) + process_restart(cinfo); + entropy->restarts_to_go--; + } + + if (entropy->ct == -1) return TRUE; /* if error do nothing */ + + natural_order = cinfo->natural_order; + + /* Outer loop handles each block in the MCU */ + + for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { + block = MCU_data[blkn]; + ci = cinfo->MCU_membership[blkn]; + compptr = cinfo->cur_comp_info[ci]; + + /* Sections F.2.4.1 & F.1.4.4.1: Decoding of DC coefficients */ + + tbl = compptr->dc_tbl_no; + + /* Table F.4: Point to statistics bin S0 for DC coefficient coding */ + st = entropy->dc_stats[tbl] + entropy->dc_context[ci]; + + /* Figure F.19: Decode_DC_DIFF */ + if (arith_decode(cinfo, st) == 0) + entropy->dc_context[ci] = 0; + else { + /* Figure F.21: Decoding nonzero value v */ + /* Figure F.22: Decoding the sign of v */ + sign = arith_decode(cinfo, st + 1); + st += 2; st += sign; + /* Figure F.23: Decoding the magnitude category of v */ + if ((m = arith_decode(cinfo, st)) != 0) { + st = entropy->dc_stats[tbl] + 20; /* Table F.4: X1 = 20 */ + while (arith_decode(cinfo, st)) { + if ((m <<= 1) == (int) 0x8000U) { + WARNMS(cinfo, JWRN_ARITH_BAD_CODE); + entropy->ct = -1; /* magnitude overflow */ + return TRUE; + } + st += 1; + } + } + /* Section F.1.4.4.1.2: Establish dc_context conditioning category */ + if (m < (int) ((1L << cinfo->arith_dc_L[tbl]) >> 1)) + entropy->dc_context[ci] = 0; /* zero diff category */ + else if (m > (int) ((1L << cinfo->arith_dc_U[tbl]) >> 1)) + entropy->dc_context[ci] = 12 + (sign * 4); /* large diff category */ + else + entropy->dc_context[ci] = 4 + (sign * 4); /* small diff category */ + v = m; + /* Figure F.24: Decoding the magnitude bit pattern of v */ + st += 14; + while (m >>= 1) + if (arith_decode(cinfo, st)) v |= m; + v += 1; if (sign) v = -v; + entropy->last_dc_val[ci] += v; + } + + (*block)[0] = (JCOEF) entropy->last_dc_val[ci]; + + /* Sections F.2.4.2 & F.1.4.4.2: Decoding of AC coefficients */ + + if (cinfo->lim_Se == 0) continue; + tbl = compptr->ac_tbl_no; + k = 0; + + /* Figure F.20: Decode_AC_coefficients */ + do { + st = entropy->ac_stats[tbl] + 3 * k; + if (arith_decode(cinfo, st)) break; /* EOB flag */ + for (;;) { + k++; + if (arith_decode(cinfo, st + 1)) break; + st += 3; + if (k >= cinfo->lim_Se) { + WARNMS(cinfo, JWRN_ARITH_BAD_CODE); + entropy->ct = -1; /* spectral overflow */ + return TRUE; + } + } + /* Figure F.21: Decoding nonzero value v */ + /* Figure F.22: Decoding the sign of v */ + sign = arith_decode(cinfo, entropy->fixed_bin); + st += 2; + /* Figure F.23: Decoding the magnitude category of v */ + if ((m = arith_decode(cinfo, st)) != 0) { + if (arith_decode(cinfo, st)) { + m <<= 1; + st = entropy->ac_stats[tbl] + + (k <= cinfo->arith_ac_K[tbl] ? 189 : 217); + while (arith_decode(cinfo, st)) { + if ((m <<= 1) == (int) 0x8000U) { + WARNMS(cinfo, JWRN_ARITH_BAD_CODE); + entropy->ct = -1; /* magnitude overflow */ + return TRUE; + } + st += 1; + } + } + } + v = m; + /* Figure F.24: Decoding the magnitude bit pattern of v */ + st += 14; + while (m >>= 1) + if (arith_decode(cinfo, st)) v |= m; + v += 1; if (sign) v = -v; + (*block)[natural_order[k]] = (JCOEF) v; + } while (k < cinfo->lim_Se); + } + + return TRUE; +} + + +/* + * Initialize for an arithmetic-compressed scan. + */ + +METHODDEF(void) +start_pass (j_decompress_ptr cinfo) +{ + arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; + int ci, tbl; + jpeg_component_info * compptr; + + if (cinfo->progressive_mode) { + /* Validate progressive scan parameters */ + if (cinfo->Ss == 0) { + if (cinfo->Se != 0) + goto bad; + } else { + /* need not check Ss/Se < 0 since they came from unsigned bytes */ + if (cinfo->Se < cinfo->Ss || cinfo->Se > cinfo->lim_Se) + goto bad; + /* AC scans may have only one component */ + if (cinfo->comps_in_scan != 1) + goto bad; + } + if (cinfo->Ah != 0) { + /* Successive approximation refinement scan: must have Al = Ah-1. */ + if (cinfo->Ah-1 != cinfo->Al) + goto bad; + } + if (cinfo->Al > 13) { /* need not check for < 0 */ + bad: + ERREXIT4(cinfo, JERR_BAD_PROGRESSION, + cinfo->Ss, cinfo->Se, cinfo->Ah, cinfo->Al); + } + /* Update progression status, and verify that scan order is legal. + * Note that inter-scan inconsistencies are treated as warnings + * not fatal errors ... not clear if this is right way to behave. + */ + for (ci = 0; ci < cinfo->comps_in_scan; ci++) { + int coefi, cindex = cinfo->cur_comp_info[ci]->component_index; + int *coef_bit_ptr = & cinfo->coef_bits[cindex][0]; + if (cinfo->Ss && coef_bit_ptr[0] < 0) /* AC without prior DC scan */ + WARNMS2(cinfo, JWRN_BOGUS_PROGRESSION, cindex, 0); + for (coefi = cinfo->Ss; coefi <= cinfo->Se; coefi++) { + int expected = (coef_bit_ptr[coefi] < 0) ? 0 : coef_bit_ptr[coefi]; + if (cinfo->Ah != expected) + WARNMS2(cinfo, JWRN_BOGUS_PROGRESSION, cindex, coefi); + coef_bit_ptr[coefi] = cinfo->Al; + } + } + /* Select MCU decoding routine */ + if (cinfo->Ah == 0) { + if (cinfo->Ss == 0) + entropy->pub.decode_mcu = decode_mcu_DC_first; + else + entropy->pub.decode_mcu = decode_mcu_AC_first; + } else { + if (cinfo->Ss == 0) + entropy->pub.decode_mcu = decode_mcu_DC_refine; + else + entropy->pub.decode_mcu = decode_mcu_AC_refine; + } + } else { + /* Check that the scan parameters Ss, Se, Ah/Al are OK for sequential JPEG. + * This ought to be an error condition, but we make it a warning. + */ + if (cinfo->Ss != 0 || cinfo->Ah != 0 || cinfo->Al != 0 || + (cinfo->Se < DCTSIZE2 && cinfo->Se != cinfo->lim_Se)) + WARNMS(cinfo, JWRN_NOT_SEQUENTIAL); + /* Select MCU decoding routine */ + entropy->pub.decode_mcu = decode_mcu; + } + + /* Allocate & initialize requested statistics areas */ + for (ci = 0; ci < cinfo->comps_in_scan; ci++) { + compptr = cinfo->cur_comp_info[ci]; + if (! cinfo->progressive_mode || (cinfo->Ss == 0 && cinfo->Ah == 0)) { + tbl = compptr->dc_tbl_no; + if (tbl < 0 || tbl >= NUM_ARITH_TBLS) + ERREXIT1(cinfo, JERR_NO_ARITH_TABLE, tbl); + if (entropy->dc_stats[tbl] == NULL) + entropy->dc_stats[tbl] = (unsigned char *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, DC_STAT_BINS); + MEMZERO(entropy->dc_stats[tbl], DC_STAT_BINS); + /* Initialize DC predictions to 0 */ + entropy->last_dc_val[ci] = 0; + entropy->dc_context[ci] = 0; + } + if ((! cinfo->progressive_mode && cinfo->lim_Se) || + (cinfo->progressive_mode && cinfo->Ss)) { + tbl = compptr->ac_tbl_no; + if (tbl < 0 || tbl >= NUM_ARITH_TBLS) + ERREXIT1(cinfo, JERR_NO_ARITH_TABLE, tbl); + if (entropy->ac_stats[tbl] == NULL) + entropy->ac_stats[tbl] = (unsigned char *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, AC_STAT_BINS); + MEMZERO(entropy->ac_stats[tbl], AC_STAT_BINS); + } + } + + /* Initialize arithmetic decoding variables */ + entropy->c = 0; + entropy->a = 0; + entropy->ct = -16; /* force reading 2 initial bytes to fill C */ + + /* Initialize restart counter */ + entropy->restarts_to_go = cinfo->restart_interval; +} + + +/* + * Finish up at the end of an arithmetic-compressed scan. + */ + +METHODDEF(void) +finish_pass (j_decompress_ptr cinfo) +{ + /* no work necessary here */ +} + + +/* + * Module initialization routine for arithmetic entropy decoding. + */ + +GLOBAL(void) +jinit_arith_decoder (j_decompress_ptr cinfo) +{ + arith_entropy_ptr entropy; + int i; + + entropy = (arith_entropy_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(arith_entropy_decoder)); + cinfo->entropy = &entropy->pub; + entropy->pub.start_pass = start_pass; + entropy->pub.finish_pass = finish_pass; + + /* Mark tables unallocated */ + for (i = 0; i < NUM_ARITH_TBLS; i++) { + entropy->dc_stats[i] = NULL; + entropy->ac_stats[i] = NULL; + } + + /* Initialize index for fixed probability estimation */ + entropy->fixed_bin[0] = 113; + + if (cinfo->progressive_mode) { + /* Create progression status table */ + int *coef_bit_ptr, ci; + cinfo->coef_bits = (int (*)[DCTSIZE2]) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, + cinfo->num_components * DCTSIZE2 * SIZEOF(int)); + coef_bit_ptr = & cinfo->coef_bits[0][0]; + for (ci = 0; ci < cinfo->num_components; ci++) + for (i = 0; i < DCTSIZE2; i++) + *coef_bit_ptr++ = -1; + } +} diff --git a/cmake/externals/libjpeg/jdatadst.c b/cmake/externals/libjpeg/jdatadst.c new file mode 100644 index 000000000..b3b4798ea --- /dev/null +++ b/cmake/externals/libjpeg/jdatadst.c @@ -0,0 +1,263 @@ +/* + * jdatadst.c + * + * Copyright (C) 1994-1996, Thomas G. Lane. + * Modified 2009-2022 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains compression data destination routines for the case of + * emitting JPEG data to memory or to a file (or any stdio stream). + * While these routines are sufficient for most applications, + * some will want to use a different destination manager. + * IMPORTANT: we assume that fwrite() will correctly transcribe an array of + * JOCTETs into 8-bit-wide elements on external storage. If char is wider + * than 8 bits on your machine, you may need to do some tweaking. + */ + +/* this is not a core library module, so it doesn't define JPEG_INTERNALS */ +#include "jinclude.h" +#include "jpeglib.h" +#include "jerror.h" + +#ifndef HAVE_STDLIB_H /* should declare malloc(),free() */ +extern void * malloc JPP((size_t size)); +extern void free JPP((void *ptr)); +#endif + + +/* Expanded data destination object for stdio output */ + +#define OUTPUT_BUF_SIZE 4096 /* choose an efficiently fwrite'able size */ + +typedef struct { + struct jpeg_destination_mgr pub; /* public fields */ + + FILE * outfile; /* target stream */ + JOCTET buffer[OUTPUT_BUF_SIZE]; /* output buffer */ +} my_destination_mgr; + +typedef my_destination_mgr * my_dest_ptr; + + +/* Expanded data destination object for memory output */ + +typedef struct { + struct jpeg_destination_mgr pub; /* public fields */ + + unsigned char ** outbuffer; /* target buffer */ + size_t * outsize; + unsigned char * newbuffer; /* newly allocated buffer */ + JOCTET * buffer; /* start of buffer */ + size_t bufsize; +} my_mem_destination_mgr; + +typedef my_mem_destination_mgr * my_mem_dest_ptr; + + +/* + * Initialize destination --- called by jpeg_start_compress + * before any data is actually written. + */ + +METHODDEF(void) +init_destination (j_compress_ptr cinfo) +{ + my_dest_ptr dest = (my_dest_ptr) cinfo->dest; + + dest->pub.next_output_byte = dest->buffer; + dest->pub.free_in_buffer = OUTPUT_BUF_SIZE; +} + +METHODDEF(void) +init_mem_destination (j_compress_ptr cinfo) +{ + /* no work necessary here */ +} + + +/* + * Empty the output buffer --- called whenever buffer fills up. + * + * In typical applications, this should write the entire output buffer + * (ignoring the current state of next_output_byte & free_in_buffer), + * reset the pointer & count to the start of the buffer, and return TRUE + * indicating that the buffer has been dumped. + * + * In applications that need to be able to suspend compression due to output + * overrun, a FALSE return indicates that the buffer cannot be emptied now. + * In this situation, the compressor will return to its caller (possibly with + * an indication that it has not accepted all the supplied scanlines). The + * application should resume compression after it has made more room in the + * output buffer. Note that there are substantial restrictions on the use of + * suspension --- see the documentation. + * + * When suspending, the compressor will back up to a convenient restart point + * (typically the start of the current MCU). next_output_byte & free_in_buffer + * indicate where the restart point will be if the current call returns FALSE. + * Data beyond this point will be regenerated after resumption, so do not + * write it out when emptying the buffer externally. + */ + +METHODDEF(boolean) +empty_output_buffer (j_compress_ptr cinfo) +{ + my_dest_ptr dest = (my_dest_ptr) cinfo->dest; + + if (JFWRITE(dest->outfile, dest->buffer, OUTPUT_BUF_SIZE) != + (size_t) OUTPUT_BUF_SIZE) + ERREXIT(cinfo, JERR_FILE_WRITE); + + dest->pub.next_output_byte = dest->buffer; + dest->pub.free_in_buffer = OUTPUT_BUF_SIZE; + + return TRUE; +} + +METHODDEF(boolean) +empty_mem_output_buffer (j_compress_ptr cinfo) +{ + size_t nextsize; + JOCTET * nextbuffer; + my_mem_dest_ptr dest = (my_mem_dest_ptr) cinfo->dest; + + /* Try to allocate new buffer with double size */ + nextsize = dest->bufsize * 2; + nextbuffer = (JOCTET *) malloc(nextsize); + + if (nextbuffer == NULL) + ERREXIT1(cinfo, JERR_OUT_OF_MEMORY, 11); + + MEMCOPY(nextbuffer, dest->buffer, dest->bufsize); + + if (dest->newbuffer != NULL) + free(dest->newbuffer); + + dest->newbuffer = nextbuffer; + + dest->pub.next_output_byte = nextbuffer + dest->bufsize; + dest->pub.free_in_buffer = dest->bufsize; + + dest->buffer = nextbuffer; + dest->bufsize = nextsize; + + return TRUE; +} + + +/* + * Terminate destination --- called by jpeg_finish_compress + * after all data has been written. Usually needs to flush buffer. + * + * NB: *not* called by jpeg_abort or jpeg_destroy; surrounding + * application must deal with any cleanup that should happen even + * for error exit. + */ + +METHODDEF(void) +term_destination (j_compress_ptr cinfo) +{ + my_dest_ptr dest = (my_dest_ptr) cinfo->dest; + size_t datacount = OUTPUT_BUF_SIZE - dest->pub.free_in_buffer; + + /* Write any data remaining in the buffer */ + if (datacount > 0) { + if (JFWRITE(dest->outfile, dest->buffer, datacount) != datacount) + ERREXIT(cinfo, JERR_FILE_WRITE); + } + JFFLUSH(dest->outfile); + /* Make sure we wrote the output file OK */ + if (JFERROR(dest->outfile)) + ERREXIT(cinfo, JERR_FILE_WRITE); +} + +METHODDEF(void) +term_mem_destination (j_compress_ptr cinfo) +{ + my_mem_dest_ptr dest = (my_mem_dest_ptr) cinfo->dest; + + *dest->outbuffer = dest->buffer; + *dest->outsize = dest->bufsize - dest->pub.free_in_buffer; +} + + +/* + * Prepare for output to a stdio stream. + * The caller must have already opened the stream, + * and is responsible for closing it after finishing compression. + */ + +GLOBAL(void) +jpeg_stdio_dest (j_compress_ptr cinfo, FILE * outfile) +{ + my_dest_ptr dest; + + /* The destination object is made permanent so that multiple JPEG images + * can be written to the same file without re-executing jpeg_stdio_dest. + * This makes it dangerous to use this manager and a different destination + * manager serially with the same JPEG object, because their private object + * sizes may be different. Caveat programmer. + */ + if (cinfo->dest == NULL) { /* first time for this JPEG object? */ + cinfo->dest = (struct jpeg_destination_mgr *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_PERMANENT, SIZEOF(my_destination_mgr)); + } + + dest = (my_dest_ptr) cinfo->dest; + dest->pub.init_destination = init_destination; + dest->pub.empty_output_buffer = empty_output_buffer; + dest->pub.term_destination = term_destination; + dest->outfile = outfile; +} + + +/* + * Prepare for output to a memory buffer. + * The caller may supply an own initial buffer with appropriate size. + * Otherwise, or when the actual data output exceeds the given size, + * the library adapts the buffer size as necessary. + * The standard library functions malloc/free are used for allocating + * larger memory, so the buffer is available to the application after + * finishing compression, and then the application is responsible for + * freeing the requested memory. + * Note: An initial buffer supplied by the caller is expected to be + * managed by the application. The library does not free such buffer + * when allocating a larger buffer. + */ + +GLOBAL(void) +jpeg_mem_dest (j_compress_ptr cinfo, + unsigned char ** outbuffer, size_t * outsize) +{ + my_mem_dest_ptr dest; + + if (outbuffer == NULL || outsize == NULL) /* sanity check */ + ERREXIT(cinfo, JERR_BUFFER_SIZE); + + /* The destination object is made permanent so that multiple JPEG images + * can be written to the same buffer without re-executing jpeg_mem_dest. + */ + if (cinfo->dest == NULL) { /* first time for this JPEG object? */ + cinfo->dest = (struct jpeg_destination_mgr *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_PERMANENT, SIZEOF(my_mem_destination_mgr)); + } + + dest = (my_mem_dest_ptr) cinfo->dest; + dest->pub.init_destination = init_mem_destination; + dest->pub.empty_output_buffer = empty_mem_output_buffer; + dest->pub.term_destination = term_mem_destination; + dest->outbuffer = outbuffer; + dest->outsize = outsize; + dest->newbuffer = NULL; + + if (*outbuffer == NULL || *outsize == 0) { + /* Allocate initial buffer */ + dest->newbuffer = *outbuffer = (unsigned char *) malloc(OUTPUT_BUF_SIZE); + if (dest->newbuffer == NULL) + ERREXIT1(cinfo, JERR_OUT_OF_MEMORY, 10); + *outsize = OUTPUT_BUF_SIZE; + } + + dest->pub.next_output_byte = dest->buffer = *outbuffer; + dest->pub.free_in_buffer = dest->bufsize = *outsize; +} diff --git a/cmake/externals/libjpeg/jdatasrc.c b/cmake/externals/libjpeg/jdatasrc.c new file mode 100644 index 000000000..fd7a1a594 --- /dev/null +++ b/cmake/externals/libjpeg/jdatasrc.c @@ -0,0 +1,271 @@ +/* + * jdatasrc.c + * + * Copyright (C) 1994-1996, Thomas G. Lane. + * Modified 2009-2022 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains decompression data source routines for the case of + * reading JPEG data from memory or from a file (or any stdio stream). + * While these routines are sufficient for most applications, + * some will want to use a different source manager. + * IMPORTANT: we assume that fread() will correctly transcribe an array of + * JOCTETs from 8-bit-wide elements on external storage. If char is wider + * than 8 bits on your machine, you may need to do some tweaking. + */ + +/* this is not a core library module, so it doesn't define JPEG_INTERNALS */ +#include "jinclude.h" +#include "jpeglib.h" +#include "jerror.h" + + +/* Expanded data source object for stdio input */ + +#define INPUT_BUF_SIZE 4096 /* choose an efficiently fread'able size */ + +typedef struct { + struct jpeg_source_mgr pub; /* public fields */ + + FILE * infile; /* source stream */ + JOCTET buffer[INPUT_BUF_SIZE]; /* input buffer */ + boolean start_of_file; /* have we gotten any data yet? */ +} my_source_mgr; + +typedef my_source_mgr * my_src_ptr; + + +/* + * Initialize source --- called by jpeg_read_header + * before any data is actually read. + */ + +METHODDEF(void) +init_source (j_decompress_ptr cinfo) +{ + my_src_ptr src = (my_src_ptr) cinfo->src; + + /* We reset the empty-input-file flag for each image, + * but we don't clear the input buffer. + * This is correct behavior for reading a series of images from one source. + */ + src->start_of_file = TRUE; +} + +METHODDEF(void) +init_mem_source (j_decompress_ptr cinfo) +{ + /* no work necessary here */ +} + + +/* + * Fill the input buffer --- called whenever buffer is emptied. + * + * In typical applications, this should read fresh data into the buffer + * (ignoring the current state of next_input_byte & bytes_in_buffer), + * reset the pointer & count to the start of the buffer, and return TRUE + * indicating that the buffer has been reloaded. It is not necessary to + * fill the buffer entirely, only to obtain at least one more byte. + * + * There is no such thing as an EOF return. If the end of the file has been + * reached, the routine has a choice of ERREXIT() or inserting fake data into + * the buffer. In most cases, generating a warning message and inserting a + * fake EOI marker is the best course of action --- this will allow the + * decompressor to output however much of the image is there. However, + * the resulting error message is misleading if the real problem is an empty + * input file, so we handle that case specially. + * + * In applications that need to be able to suspend compression due to input + * not being available yet, a FALSE return indicates that no more data can be + * obtained right now, but more may be forthcoming later. In this situation, + * the decompressor will return to its caller (with an indication of the + * number of scanlines it has read, if any). The application should resume + * decompression after it has loaded more data into the input buffer. Note + * that there are substantial restrictions on the use of suspension --- see + * the documentation. + * + * When suspending, the decompressor will back up to a convenient restart point + * (typically the start of the current MCU). next_input_byte & bytes_in_buffer + * indicate where the restart point will be if the current call returns FALSE. + * Data beyond this point must be rescanned after resumption, so move it to + * the front of the buffer rather than discarding it. + */ + +METHODDEF(boolean) +fill_input_buffer (j_decompress_ptr cinfo) +{ + my_src_ptr src = (my_src_ptr) cinfo->src; + size_t nbytes; + + nbytes = JFREAD(src->infile, src->buffer, INPUT_BUF_SIZE); + + if (nbytes <= 0) { + if (src->start_of_file) /* Treat empty input file as fatal error */ + ERREXIT(cinfo, JERR_INPUT_EMPTY); + WARNMS(cinfo, JWRN_JPEG_EOF); + /* Insert a fake EOI marker */ + src->buffer[0] = (JOCTET) 0xFF; + src->buffer[1] = (JOCTET) JPEG_EOI; + nbytes = 2; + } + + src->pub.next_input_byte = src->buffer; + src->pub.bytes_in_buffer = nbytes; + src->start_of_file = FALSE; + + return TRUE; +} + +METHODDEF(boolean) +fill_mem_input_buffer (j_decompress_ptr cinfo) +{ + static const JOCTET mybuffer[4] = { + (JOCTET) 0xFF, (JOCTET) JPEG_EOI, 0, 0 + }; + + /* The whole JPEG data is expected to reside in the supplied memory + * buffer, so any request for more data beyond the given buffer size + * is treated as an error. + */ + WARNMS(cinfo, JWRN_JPEG_EOF); + + /* Insert a fake EOI marker */ + + cinfo->src->next_input_byte = mybuffer; + cinfo->src->bytes_in_buffer = 2; + + return TRUE; +} + + +/* + * Skip data --- used to skip over a potentially large amount of + * uninteresting data (such as an APPn marker). + * + * Writers of suspendable-input applications must note that skip_input_data + * is not granted the right to give a suspension return. If the skip extends + * beyond the data currently in the buffer, the buffer can be marked empty so + * that the next read will cause a fill_input_buffer call that can suspend. + * Arranging for additional bytes to be discarded before reloading the input + * buffer is the application writer's problem. + */ + +METHODDEF(void) +skip_input_data (j_decompress_ptr cinfo, long num_bytes) +{ + struct jpeg_source_mgr * src = cinfo->src; + size_t nbytes; + + /* Just a dumb implementation for now. Could use fseek() except + * it doesn't work on pipes. Not clear that being smart is worth + * any trouble anyway --- large skips are infrequent. + */ + if (num_bytes > 0) { + nbytes = (size_t) num_bytes; + while (nbytes > src->bytes_in_buffer) { + nbytes -= src->bytes_in_buffer; + (void) (*src->fill_input_buffer) (cinfo); + /* note we assume that fill_input_buffer will never return FALSE, + * so suspension need not be handled. + */ + } + src->next_input_byte += nbytes; + src->bytes_in_buffer -= nbytes; + } +} + + +/* + * An additional method that can be provided by data source modules is the + * resync_to_restart method for error recovery in the presence of RST markers. + * For the moment, this source module just uses the default resync method + * provided by the JPEG library. That method assumes that no backtracking + * is possible. + */ + + +/* + * Terminate source --- called by jpeg_finish_decompress + * after all data has been read. Often a no-op. + * + * NB: *not* called by jpeg_abort or jpeg_destroy; surrounding + * application must deal with any cleanup that should happen even + * for error exit. + */ + +METHODDEF(void) +term_source (j_decompress_ptr cinfo) +{ + /* no work necessary here */ +} + + +/* + * Prepare for input from a stdio stream. + * The caller must have already opened the stream, + * and is responsible for closing it after finishing decompression. + */ + +GLOBAL(void) +jpeg_stdio_src (j_decompress_ptr cinfo, FILE * infile) +{ + my_src_ptr src; + + /* The source object including the input buffer is made permanent so that + * a series of JPEG images can be read from the same file by calling + * jpeg_stdio_src only before the first one. (If we discarded the buffer + * at the end of one image, we'd likely lose the start of the next one.) + * This makes it unsafe to use this manager and a different source + * manager serially with the same JPEG object. Caveat programmer. + */ + if (cinfo->src == NULL) { /* first time for this JPEG object? */ + cinfo->src = (struct jpeg_source_mgr *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_PERMANENT, SIZEOF(my_source_mgr)); + } + + src = (my_src_ptr) cinfo->src; + src->pub.init_source = init_source; + src->pub.fill_input_buffer = fill_input_buffer; + src->pub.skip_input_data = skip_input_data; + src->pub.resync_to_restart = jpeg_resync_to_restart; /* use default method */ + src->pub.term_source = term_source; + src->infile = infile; + src->pub.bytes_in_buffer = 0; /* forces fill_input_buffer on first read */ + src->pub.next_input_byte = NULL; /* until buffer loaded */ +} + + +/* + * Prepare for input from a supplied memory buffer. + * The buffer must contain the whole JPEG data. + */ + +GLOBAL(void) +jpeg_mem_src (j_decompress_ptr cinfo, + const unsigned char * inbuffer, size_t insize) +{ + struct jpeg_source_mgr * src; + + if (inbuffer == NULL || insize == 0) /* Treat empty input as fatal error */ + ERREXIT(cinfo, JERR_INPUT_EMPTY); + + /* The source object is made permanent so that a series of JPEG images + * can be read from the same buffer by calling jpeg_mem_src only before + * the first one. + */ + if (cinfo->src == NULL) { /* first time for this JPEG object? */ + cinfo->src = (struct jpeg_source_mgr *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_PERMANENT, SIZEOF(struct jpeg_source_mgr)); + } + + src = cinfo->src; + src->init_source = init_mem_source; + src->fill_input_buffer = fill_mem_input_buffer; + src->skip_input_data = skip_input_data; + src->resync_to_restart = jpeg_resync_to_restart; /* use default method */ + src->term_source = term_source; + src->bytes_in_buffer = insize; + src->next_input_byte = (const JOCTET *) inbuffer; +} diff --git a/cmake/externals/libjpeg/jdcoefct.c b/cmake/externals/libjpeg/jdcoefct.c new file mode 100644 index 000000000..79ba42014 --- /dev/null +++ b/cmake/externals/libjpeg/jdcoefct.c @@ -0,0 +1,744 @@ +/* + * jdcoefct.c + * + * Copyright (C) 1994-1997, Thomas G. Lane. + * Modified 2002-2020 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains the coefficient buffer controller for decompression. + * This controller is the top level of the JPEG decompressor proper. + * The coefficient buffer lies between entropy decoding and inverse-DCT steps. + * + * In buffered-image mode, this controller is the interface between + * input-oriented processing and output-oriented processing. + * Also, the input side (only) is used when reading a file for transcoding. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* Block smoothing is only applicable for progressive JPEG, so: */ +#ifndef D_PROGRESSIVE_SUPPORTED +#undef BLOCK_SMOOTHING_SUPPORTED +#endif + + +/* Private buffer controller object */ + +typedef struct { + struct jpeg_d_coef_controller pub; /* public fields */ + + /* These variables keep track of the current location of the input side. */ + /* cinfo->input_iMCU_row is also used for this. */ + JDIMENSION MCU_ctr; /* counts MCUs processed in current row */ + int MCU_vert_offset; /* counts MCU rows within iMCU row */ + int MCU_rows_per_iMCU_row; /* number of such rows needed */ + + /* The output side's location is represented by cinfo->output_iMCU_row. */ + + /* In single-pass modes, it's sufficient to buffer just one MCU. + * We append a workspace of D_MAX_BLOCKS_IN_MCU coefficient blocks, + * and let the entropy decoder write into that workspace each time. + * In multi-pass modes, this array points to the current MCU's blocks + * within the virtual arrays; it is used only by the input side. + */ + JBLOCKROW MCU_buffer[D_MAX_BLOCKS_IN_MCU]; + +#ifdef D_MULTISCAN_FILES_SUPPORTED + /* In multi-pass modes, we need a virtual block array for each component. */ + jvirt_barray_ptr whole_image[MAX_COMPONENTS]; +#endif + +#ifdef BLOCK_SMOOTHING_SUPPORTED + /* When doing block smoothing, we latch coefficient Al values here */ + int * coef_bits_latch; +#define SAVED_COEFS 6 /* we save coef_bits[0..5] */ +#endif + + /* Workspace for single-pass modes (omitted otherwise). */ + JBLOCK blk_buffer[D_MAX_BLOCKS_IN_MCU]; +} my_coef_controller; + +typedef my_coef_controller * my_coef_ptr; + + +/* Forward declarations */ +METHODDEF(int) decompress_onepass + JPP((j_decompress_ptr cinfo, JSAMPIMAGE output_buf)); +#ifdef D_MULTISCAN_FILES_SUPPORTED +METHODDEF(int) decompress_data + JPP((j_decompress_ptr cinfo, JSAMPIMAGE output_buf)); +#endif +#ifdef BLOCK_SMOOTHING_SUPPORTED +LOCAL(boolean) smoothing_ok JPP((j_decompress_ptr cinfo)); +METHODDEF(int) decompress_smooth_data + JPP((j_decompress_ptr cinfo, JSAMPIMAGE output_buf)); +#endif + + +LOCAL(void) +start_iMCU_row (j_decompress_ptr cinfo) +/* Reset within-iMCU-row counters for a new row (input side) */ +{ + my_coef_ptr coef = (my_coef_ptr) cinfo->coef; + + /* In an interleaved scan, an MCU row is the same as an iMCU row. + * In a noninterleaved scan, an iMCU row has v_samp_factor MCU rows. + * But at the bottom of the image, process only what's left. + */ + if (cinfo->comps_in_scan > 1) { + coef->MCU_rows_per_iMCU_row = 1; + } else { + if (cinfo->input_iMCU_row < (cinfo->total_iMCU_rows-1)) + coef->MCU_rows_per_iMCU_row = cinfo->cur_comp_info[0]->v_samp_factor; + else + coef->MCU_rows_per_iMCU_row = cinfo->cur_comp_info[0]->last_row_height; + } + + coef->MCU_ctr = 0; + coef->MCU_vert_offset = 0; +} + + +/* + * Initialize for an input processing pass. + */ + +METHODDEF(void) +start_input_pass (j_decompress_ptr cinfo) +{ + cinfo->input_iMCU_row = 0; + start_iMCU_row(cinfo); +} + + +/* + * Initialize for an output processing pass. + */ + +METHODDEF(void) +start_output_pass (j_decompress_ptr cinfo) +{ +#ifdef BLOCK_SMOOTHING_SUPPORTED + my_coef_ptr coef = (my_coef_ptr) cinfo->coef; + + /* If multipass, check to see whether to use block smoothing on this pass */ + if (coef->pub.coef_arrays != NULL) { + if (cinfo->do_block_smoothing && smoothing_ok(cinfo)) + coef->pub.decompress_data = decompress_smooth_data; + else + coef->pub.decompress_data = decompress_data; + } +#endif + cinfo->output_iMCU_row = 0; +} + + +/* + * Decompress and return some data in the single-pass case. + * Always attempts to emit one fully interleaved MCU row ("iMCU" row). + * Input and output must run in lockstep since we have only a one-MCU buffer. + * Return value is JPEG_ROW_COMPLETED, JPEG_SCAN_COMPLETED, or JPEG_SUSPENDED. + * + * NB: output_buf contains a plane for each component in image, + * which we index according to the component's SOF position. + */ + +METHODDEF(int) +decompress_onepass (j_decompress_ptr cinfo, JSAMPIMAGE output_buf) +{ + my_coef_ptr coef = (my_coef_ptr) cinfo->coef; + JDIMENSION MCU_col_num; /* index of current MCU within row */ + JDIMENSION last_MCU_col = cinfo->MCUs_per_row - 1; + JDIMENSION last_iMCU_row = cinfo->total_iMCU_rows - 1; + int ci, xindex, yindex, yoffset, useful_width; + JBLOCKROW blkp; + JSAMPARRAY output_ptr; + JDIMENSION start_col, output_col; + jpeg_component_info *compptr; + inverse_DCT_method_ptr inverse_DCT; + + /* Loop to process as much as one whole iMCU row */ + for (yoffset = coef->MCU_vert_offset; yoffset < coef->MCU_rows_per_iMCU_row; + yoffset++) { + for (MCU_col_num = coef->MCU_ctr; MCU_col_num <= last_MCU_col; + MCU_col_num++) { + blkp = coef->blk_buffer; /* pointer to current DCT block within MCU */ + /* Try to fetch an MCU. Entropy decoder expects buffer to be zeroed. */ + if (cinfo->lim_Se) /* can bypass in DC only case */ + MEMZERO(blkp, cinfo->blocks_in_MCU * SIZEOF(JBLOCK)); + if (! (*cinfo->entropy->decode_mcu) (cinfo, coef->MCU_buffer)) { + /* Suspension forced; update state counters and exit */ + coef->MCU_vert_offset = yoffset; + coef->MCU_ctr = MCU_col_num; + return JPEG_SUSPENDED; + } + /* Determine where data should go in output_buf and do the IDCT thing. + * We skip dummy blocks at the right and bottom edges (but blkp gets + * incremented past them!). + */ + for (ci = 0; ci < cinfo->comps_in_scan; ci++) { + compptr = cinfo->cur_comp_info[ci]; + /* Don't bother to IDCT an uninteresting component. */ + if (! compptr->component_needed) { + blkp += compptr->MCU_blocks; + continue; + } + inverse_DCT = cinfo->idct->inverse_DCT[compptr->component_index]; + output_ptr = output_buf[compptr->component_index] + + yoffset * compptr->DCT_v_scaled_size; + useful_width = (MCU_col_num < last_MCU_col) ? compptr->MCU_width + : compptr->last_col_width; + start_col = MCU_col_num * compptr->MCU_sample_width; + for (yindex = 0; yindex < compptr->MCU_height; yindex++) { + if (cinfo->input_iMCU_row < last_iMCU_row || + yoffset + yindex < compptr->last_row_height) { + output_col = start_col; + for (xindex = 0; xindex < useful_width; xindex++) { + (*inverse_DCT) (cinfo, compptr, (JCOEFPTR) (blkp + xindex), + output_ptr, output_col); + output_col += compptr->DCT_h_scaled_size; + } + output_ptr += compptr->DCT_v_scaled_size; + } + blkp += compptr->MCU_width; + } + } + } + /* Completed an MCU row, but perhaps not an iMCU row */ + coef->MCU_ctr = 0; + } + /* Completed the iMCU row, advance counters for next one */ + cinfo->output_iMCU_row++; + if (++(cinfo->input_iMCU_row) <= last_iMCU_row) { + start_iMCU_row(cinfo); + return JPEG_ROW_COMPLETED; + } + /* Completed the scan */ + (*cinfo->inputctl->finish_input_pass) (cinfo); + return JPEG_SCAN_COMPLETED; +} + + +/* + * Dummy consume-input routine for single-pass operation. + */ + +METHODDEF(int) +dummy_consume_data (j_decompress_ptr cinfo) +{ + return JPEG_SUSPENDED; /* Always indicate nothing was done */ +} + + +#ifdef D_MULTISCAN_FILES_SUPPORTED + +/* + * Consume input data and store it in the full-image coefficient buffer. + * We read as much as one fully interleaved MCU row ("iMCU" row) per call, + * ie, v_samp_factor block rows for each component in the scan. + * Return value is JPEG_ROW_COMPLETED, JPEG_SCAN_COMPLETED, or JPEG_SUSPENDED. + */ + +METHODDEF(int) +consume_data (j_decompress_ptr cinfo) +{ + my_coef_ptr coef = (my_coef_ptr) cinfo->coef; + JDIMENSION MCU_col_num; /* index of current MCU within row */ + int ci, xindex, yindex, yoffset; + JDIMENSION start_col; + JBLOCKARRAY blkp; + JBLOCKARRAY buffer[MAX_COMPS_IN_SCAN]; + JBLOCKROW buffer_ptr; + jpeg_component_info *compptr; + + /* Align the virtual buffers for the components used in this scan. */ + for (ci = 0; ci < cinfo->comps_in_scan; ci++) { + compptr = cinfo->cur_comp_info[ci]; + buffer[ci] = (*cinfo->mem->access_virt_barray) + ((j_common_ptr) cinfo, coef->whole_image[compptr->component_index], + cinfo->input_iMCU_row * compptr->v_samp_factor, + (JDIMENSION) compptr->v_samp_factor, TRUE); + /* Note: entropy decoder expects buffer to be zeroed, + * but this is handled automatically by the memory manager + * because we requested a pre-zeroed array. + */ + } + + /* Loop to process one whole iMCU row */ + for (yoffset = coef->MCU_vert_offset; yoffset < coef->MCU_rows_per_iMCU_row; + yoffset++) { + for (MCU_col_num = coef->MCU_ctr; MCU_col_num < cinfo->MCUs_per_row; + MCU_col_num++) { + /* Construct list of pointers to DCT blocks belonging to this MCU */ + blkp = coef->MCU_buffer; /* pointer to current DCT block within MCU */ + for (ci = 0; ci < cinfo->comps_in_scan; ci++) { + compptr = cinfo->cur_comp_info[ci]; + start_col = MCU_col_num * compptr->MCU_width; + for (yindex = 0; yindex < compptr->MCU_height; yindex++) { + buffer_ptr = buffer[ci][yoffset + yindex] + start_col; + xindex = compptr->MCU_width; + do { + *blkp++ = buffer_ptr++; + } while (--xindex); + } + } + /* Try to fetch the MCU. */ + if (! (*cinfo->entropy->decode_mcu) (cinfo, coef->MCU_buffer)) { + /* Suspension forced; update state counters and exit */ + coef->MCU_vert_offset = yoffset; + coef->MCU_ctr = MCU_col_num; + return JPEG_SUSPENDED; + } + } + /* Completed an MCU row, but perhaps not an iMCU row */ + coef->MCU_ctr = 0; + } + /* Completed the iMCU row, advance counters for next one */ + if (++(cinfo->input_iMCU_row) < cinfo->total_iMCU_rows) { + start_iMCU_row(cinfo); + return JPEG_ROW_COMPLETED; + } + /* Completed the scan */ + (*cinfo->inputctl->finish_input_pass) (cinfo); + return JPEG_SCAN_COMPLETED; +} + + +/* + * Decompress and return some data in the multi-pass case. + * Always attempts to emit one fully interleaved MCU row ("iMCU" row). + * Return value is JPEG_ROW_COMPLETED, JPEG_SCAN_COMPLETED, or JPEG_SUSPENDED. + * + * NB: output_buf contains a plane for each component in image. + */ + +METHODDEF(int) +decompress_data (j_decompress_ptr cinfo, JSAMPIMAGE output_buf) +{ + my_coef_ptr coef = (my_coef_ptr) cinfo->coef; + JDIMENSION last_iMCU_row = cinfo->total_iMCU_rows - 1; + JDIMENSION block_num; + int ci, block_row, block_rows; + JBLOCKARRAY buffer; + JBLOCKROW buffer_ptr; + JSAMPARRAY output_ptr; + JDIMENSION output_col; + jpeg_component_info *compptr; + inverse_DCT_method_ptr inverse_DCT; + + /* Force some input to be done if we are getting ahead of the input. */ + while (cinfo->input_scan_number < cinfo->output_scan_number || + (cinfo->input_scan_number == cinfo->output_scan_number && + cinfo->input_iMCU_row <= cinfo->output_iMCU_row)) { + if ((*cinfo->inputctl->consume_input)(cinfo) == JPEG_SUSPENDED) + return JPEG_SUSPENDED; + } + + /* OK, output from the virtual arrays. */ + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + /* Don't bother to IDCT an uninteresting component. */ + if (! compptr->component_needed) + continue; + /* Align the virtual buffer for this component. */ + buffer = (*cinfo->mem->access_virt_barray) + ((j_common_ptr) cinfo, coef->whole_image[ci], + cinfo->output_iMCU_row * compptr->v_samp_factor, + (JDIMENSION) compptr->v_samp_factor, FALSE); + /* Count non-dummy DCT block rows in this iMCU row. */ + if (cinfo->output_iMCU_row < last_iMCU_row) + block_rows = compptr->v_samp_factor; + else { + /* NB: can't use last_row_height here; it is input-side-dependent! */ + block_rows = (int) (compptr->height_in_blocks % compptr->v_samp_factor); + if (block_rows == 0) block_rows = compptr->v_samp_factor; + } + inverse_DCT = cinfo->idct->inverse_DCT[ci]; + output_ptr = output_buf[ci]; + /* Loop over all DCT blocks to be processed. */ + for (block_row = 0; block_row < block_rows; block_row++) { + buffer_ptr = buffer[block_row]; + output_col = 0; + for (block_num = 0; block_num < compptr->width_in_blocks; block_num++) { + (*inverse_DCT) (cinfo, compptr, (JCOEFPTR) buffer_ptr, + output_ptr, output_col); + buffer_ptr++; + output_col += compptr->DCT_h_scaled_size; + } + output_ptr += compptr->DCT_v_scaled_size; + } + } + + if (++(cinfo->output_iMCU_row) <= last_iMCU_row) + return JPEG_ROW_COMPLETED; + return JPEG_SCAN_COMPLETED; +} + +#endif /* D_MULTISCAN_FILES_SUPPORTED */ + + +#ifdef BLOCK_SMOOTHING_SUPPORTED + +/* + * This code applies interblock smoothing as described by section K.8 + * of the JPEG standard: the first 5 AC coefficients are estimated from + * the DC values of a DCT block and its 8 neighboring blocks. + * We apply smoothing only for progressive JPEG decoding, and only if + * the coefficients it can estimate are not yet known to full precision. + */ + +/* Natural-order array positions of the first 5 zigzag-order coefficients */ +#define Q01_POS 1 +#define Q10_POS 8 +#define Q20_POS 16 +#define Q11_POS 9 +#define Q02_POS 2 + +/* + * Determine whether block smoothing is applicable and safe. + * We also latch the current states of the coef_bits[] entries for the + * AC coefficients; otherwise, if the input side of the decompressor + * advances into a new scan, we might think the coefficients are known + * more accurately than they really are. + */ + +LOCAL(boolean) +smoothing_ok (j_decompress_ptr cinfo) +{ + my_coef_ptr coef = (my_coef_ptr) cinfo->coef; + boolean smoothing_useful = FALSE; + int ci, coefi; + jpeg_component_info *compptr; + JQUANT_TBL * qtable; + int * coef_bits; + int * coef_bits_latch; + + if (! cinfo->progressive_mode || cinfo->coef_bits == NULL) + return FALSE; + + /* Allocate latch area if not already done */ + if (coef->coef_bits_latch == NULL) + coef->coef_bits_latch = (int *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, + cinfo->num_components * (SAVED_COEFS * SIZEOF(int))); + coef_bits_latch = coef->coef_bits_latch; + + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + /* All components' quantization values must already be latched. */ + if ((qtable = compptr->quant_table) == NULL) + return FALSE; + /* Verify DC & first 5 AC quantizers are nonzero to avoid zero-divide. */ + if (qtable->quantval[0] == 0 || + qtable->quantval[Q01_POS] == 0 || + qtable->quantval[Q10_POS] == 0 || + qtable->quantval[Q20_POS] == 0 || + qtable->quantval[Q11_POS] == 0 || + qtable->quantval[Q02_POS] == 0) + return FALSE; + /* DC values must be at least partly known for all components. */ + coef_bits = cinfo->coef_bits[ci]; + if (coef_bits[0] < 0) + return FALSE; + /* Block smoothing is helpful if some AC coefficients remain inaccurate. */ + for (coefi = 1; coefi <= 5; coefi++) { + coef_bits_latch[coefi] = coef_bits[coefi]; + if (coef_bits[coefi] != 0) + smoothing_useful = TRUE; + } + coef_bits_latch += SAVED_COEFS; + } + + return smoothing_useful; +} + + +/* + * Variant of decompress_data for use when doing block smoothing. + */ + +METHODDEF(int) +decompress_smooth_data (j_decompress_ptr cinfo, JSAMPIMAGE output_buf) +{ + my_coef_ptr coef = (my_coef_ptr) cinfo->coef; + JDIMENSION last_iMCU_row = cinfo->total_iMCU_rows - 1; + JDIMENSION block_num, last_block_column; + int ci, block_row, block_rows, access_rows; + JBLOCKARRAY buffer; + JBLOCKROW buffer_ptr, prev_block_row, next_block_row; + JSAMPARRAY output_ptr; + JDIMENSION output_col; + jpeg_component_info *compptr; + inverse_DCT_method_ptr inverse_DCT; + boolean first_row, last_row; + JBLOCK workspace; + int *coef_bits; + JQUANT_TBL *quanttbl; + INT32 Q00,Q01,Q02,Q10,Q11,Q20, num; + int DC1,DC2,DC3,DC4,DC5,DC6,DC7,DC8,DC9; + int Al, pred; + + /* Force some input to be done if we are getting ahead of the input. */ + while (cinfo->input_scan_number <= cinfo->output_scan_number && + ! cinfo->inputctl->eoi_reached) { + if (cinfo->input_scan_number == cinfo->output_scan_number) { + /* If input is working on current scan, we ordinarily want it to + * have completed the current row. But if input scan is DC, + * we want it to keep one row ahead so that next block row's DC + * values are up to date. + */ + JDIMENSION delta = (cinfo->Ss == 0) ? 1 : 0; + if (cinfo->input_iMCU_row > cinfo->output_iMCU_row+delta) + break; + } + if ((*cinfo->inputctl->consume_input)(cinfo) == JPEG_SUSPENDED) + return JPEG_SUSPENDED; + } + + /* OK, output from the virtual arrays. */ + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + /* Don't bother to IDCT an uninteresting component. */ + if (! compptr->component_needed) + continue; + /* Count non-dummy DCT block rows in this iMCU row. */ + if (cinfo->output_iMCU_row < last_iMCU_row) { + block_rows = compptr->v_samp_factor; + access_rows = block_rows * 2; /* this and next iMCU row */ + last_row = FALSE; + } else { + /* NB: can't use last_row_height here; it is input-side-dependent! */ + block_rows = (int) (compptr->height_in_blocks % compptr->v_samp_factor); + if (block_rows == 0) block_rows = compptr->v_samp_factor; + access_rows = block_rows; /* this iMCU row only */ + last_row = TRUE; + } + /* Align the virtual buffer for this component. */ + if (cinfo->output_iMCU_row > 0) { + access_rows += compptr->v_samp_factor; /* prior iMCU row too */ + buffer = (*cinfo->mem->access_virt_barray) + ((j_common_ptr) cinfo, coef->whole_image[ci], + (cinfo->output_iMCU_row - 1) * compptr->v_samp_factor, + (JDIMENSION) access_rows, FALSE); + buffer += compptr->v_samp_factor; /* point to current iMCU row */ + first_row = FALSE; + } else { + buffer = (*cinfo->mem->access_virt_barray) + ((j_common_ptr) cinfo, coef->whole_image[ci], + (JDIMENSION) 0, (JDIMENSION) access_rows, FALSE); + first_row = TRUE; + } + /* Fetch component-dependent info */ + coef_bits = coef->coef_bits_latch + (ci * SAVED_COEFS); + quanttbl = compptr->quant_table; + Q00 = quanttbl->quantval[0]; + Q01 = quanttbl->quantval[Q01_POS]; + Q10 = quanttbl->quantval[Q10_POS]; + Q20 = quanttbl->quantval[Q20_POS]; + Q11 = quanttbl->quantval[Q11_POS]; + Q02 = quanttbl->quantval[Q02_POS]; + inverse_DCT = cinfo->idct->inverse_DCT[ci]; + output_ptr = output_buf[ci]; + /* Loop over all DCT blocks to be processed. */ + for (block_row = 0; block_row < block_rows; block_row++) { + buffer_ptr = buffer[block_row]; + if (first_row && block_row == 0) + prev_block_row = buffer_ptr; + else + prev_block_row = buffer[block_row-1]; + if (last_row && block_row == block_rows-1) + next_block_row = buffer_ptr; + else + next_block_row = buffer[block_row+1]; + /* We fetch the surrounding DC values using a sliding-register approach. + * Initialize all nine here so as to do the right thing on narrow pics. + */ + DC1 = DC2 = DC3 = (int) prev_block_row[0][0]; + DC4 = DC5 = DC6 = (int) buffer_ptr[0][0]; + DC7 = DC8 = DC9 = (int) next_block_row[0][0]; + output_col = 0; + last_block_column = compptr->width_in_blocks - 1; + for (block_num = 0; block_num <= last_block_column; block_num++) { + /* Fetch current DCT block into workspace so we can modify it. */ + jcopy_block_row(buffer_ptr, (JBLOCKROW) workspace, (JDIMENSION) 1); + /* Update DC values */ + if (block_num < last_block_column) { + DC3 = (int) prev_block_row[1][0]; + DC6 = (int) buffer_ptr[1][0]; + DC9 = (int) next_block_row[1][0]; + } + /* Compute coefficient estimates per K.8. + * An estimate is applied only if coefficient is still zero, + * and is not known to be fully accurate. + */ + /* AC01 */ + if ((Al=coef_bits[1]) != 0 && workspace[1] == 0) { + num = 36 * Q00 * (DC4 - DC6); + if (num >= 0) { + pred = (int) (((Q01<<7) + num) / (Q01<<8)); + if (Al > 0 && pred >= (1< 0 && pred >= (1<= 0) { + pred = (int) (((Q10<<7) + num) / (Q10<<8)); + if (Al > 0 && pred >= (1< 0 && pred >= (1<= 0) { + pred = (int) (((Q20<<7) + num) / (Q20<<8)); + if (Al > 0 && pred >= (1< 0 && pred >= (1<= 0) { + pred = (int) (((Q11<<7) + num) / (Q11<<8)); + if (Al > 0 && pred >= (1< 0 && pred >= (1<= 0) { + pred = (int) (((Q02<<7) + num) / (Q02<<8)); + if (Al > 0 && pred >= (1< 0 && pred >= (1<DCT_h_scaled_size; + } + output_ptr += compptr->DCT_v_scaled_size; + } + } + + if (++(cinfo->output_iMCU_row) <= last_iMCU_row) + return JPEG_ROW_COMPLETED; + return JPEG_SCAN_COMPLETED; +} + +#endif /* BLOCK_SMOOTHING_SUPPORTED */ + + +/* + * Initialize coefficient buffer controller. + */ + +GLOBAL(void) +jinit_d_coef_controller (j_decompress_ptr cinfo, boolean need_full_buffer) +{ + my_coef_ptr coef; + + if (need_full_buffer) { +#ifdef D_MULTISCAN_FILES_SUPPORTED + /* Allocate a full-image virtual array for each component, */ + /* padded to a multiple of samp_factor DCT blocks in each direction. */ + /* Note we ask for a pre-zeroed array. */ + int ci, access_rows; + jpeg_component_info *compptr; + + coef = (my_coef_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, + SIZEOF(my_coef_controller) - SIZEOF(coef->blk_buffer)); + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + access_rows = compptr->v_samp_factor; +#ifdef BLOCK_SMOOTHING_SUPPORTED + /* If block smoothing could be used, need a bigger window */ + if (cinfo->progressive_mode) + access_rows *= 3; +#endif + coef->whole_image[ci] = (*cinfo->mem->request_virt_barray) + ((j_common_ptr) cinfo, JPOOL_IMAGE, TRUE, + (JDIMENSION) jround_up((long) compptr->width_in_blocks, + (long) compptr->h_samp_factor), + (JDIMENSION) jround_up((long) compptr->height_in_blocks, + (long) compptr->v_samp_factor), + (JDIMENSION) access_rows); + } + coef->pub.consume_data = consume_data; + coef->pub.decompress_data = decompress_data; + coef->pub.coef_arrays = coef->whole_image; /* link to virtual arrays */ +#else + ERREXIT(cinfo, JERR_NOT_COMPILED); +#endif + } else { + /* We only need a single-MCU buffer. */ + JBLOCKARRAY blkp; + JBLOCKROW buffer_ptr; + int bi; + + coef = (my_coef_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_coef_controller)); + buffer_ptr = coef->blk_buffer; + if (cinfo->lim_Se == 0) /* DC only case: want to bypass later */ + MEMZERO(buffer_ptr, SIZEOF(coef->blk_buffer)); + blkp = coef->MCU_buffer; + bi = D_MAX_BLOCKS_IN_MCU; + do { + *blkp++ = buffer_ptr++; + } while (--bi); + coef->pub.consume_data = dummy_consume_data; + coef->pub.decompress_data = decompress_onepass; + coef->pub.coef_arrays = NULL; /* flag for no virtual arrays */ + } + + coef->pub.start_input_pass = start_input_pass; + coef->pub.start_output_pass = start_output_pass; +#ifdef BLOCK_SMOOTHING_SUPPORTED + coef->coef_bits_latch = NULL; +#endif + cinfo->coef = &coef->pub; +} diff --git a/cmake/externals/libjpeg/jdcolor.c b/cmake/externals/libjpeg/jdcolor.c new file mode 100644 index 000000000..6b40fb534 --- /dev/null +++ b/cmake/externals/libjpeg/jdcolor.c @@ -0,0 +1,769 @@ +/* + * jdcolor.c + * + * Copyright (C) 1991-1997, Thomas G. Lane. + * Modified 2011-2023 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains output colorspace conversion routines. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +#if RANGE_BITS < 2 + /* Deliberate syntax err */ + Sorry, this code requires 2 or more range extension bits. +#endif + + +/* Private subobject */ + +typedef struct { + struct jpeg_color_deconverter pub; /* public fields */ + + /* Private state for YCbCr->RGB and BG_YCC->RGB conversion */ + int * Cr_r_tab; /* => table for Cr to R conversion */ + int * Cb_b_tab; /* => table for Cb to B conversion */ + INT32 * Cr_g_tab; /* => table for Cr to G conversion */ + INT32 * Cb_g_tab; /* => table for Cb to G conversion */ + + /* Private state for RGB->Y conversion */ + INT32 * R_y_tab; /* => table for R to Y conversion */ + INT32 * G_y_tab; /* => table for G to Y conversion */ + INT32 * B_y_tab; /* => table for B to Y conversion */ +} my_color_deconverter; + +typedef my_color_deconverter * my_cconvert_ptr; + + +/*************** YCbCr -> RGB conversion: most common case **************/ +/*************** BG_YCC -> RGB conversion: less common case **************/ +/*************** RGB -> Y conversion: less common case **************/ + +/* + * YCbCr is defined per Recommendation ITU-R BT.601-7 (03/2011), + * previously known as Recommendation CCIR 601-1, except that Cb and Cr + * are normalized to the range 0..MAXJSAMPLE rather than -0.5 .. 0.5. + * sRGB (standard RGB color space) is defined per IEC 61966-2-1:1999. + * sYCC (standard luma-chroma-chroma color space with extended gamut) + * is defined per IEC 61966-2-1:1999 Amendment A1:2003 Annex F. + * bg-sRGB and bg-sYCC (big gamut standard color spaces) + * are defined per IEC 61966-2-1:1999 Amendment A1:2003 Annex G. + * Note that the derived conversion coefficients given in some of these + * documents are imprecise. The general conversion equations are + * + * R = Y + K * (1 - Kr) * Cr + * G = Y - K * (Kb * (1 - Kb) * Cb + Kr * (1 - Kr) * Cr) / (1 - Kr - Kb) + * B = Y + K * (1 - Kb) * Cb + * + * Y = Kr * R + (1 - Kr - Kb) * G + Kb * B + * + * With Kr = 0.299 and Kb = 0.114 (derived according to SMPTE RP 177-1993 + * from the 1953 FCC NTSC primaries and CIE Illuminant C), K = 2 for sYCC, + * the conversion equations to be implemented are therefore + * + * R = Y + 1.402 * Cr + * G = Y - 0.344136286 * Cb - 0.714136286 * Cr + * B = Y + 1.772 * Cb + * + * Y = 0.299 * R + 0.587 * G + 0.114 * B + * + * where Cb and Cr represent the incoming values less CENTERJSAMPLE. + * For bg-sYCC, with K = 4, the equations are + * + * R = Y + 2.804 * Cr + * G = Y - 0.688272572 * Cb - 1.428272572 * Cr + * B = Y + 3.544 * Cb + * + * To avoid floating-point arithmetic, we represent the fractional constants + * as integers scaled up by 2^16 (about 4 digits precision); we have to divide + * the products by 2^16, with appropriate rounding, to get the correct answer. + * Notice that Y, being an integral input, does not contribute any fraction + * so it need not participate in the rounding. + * + * For even more speed, we avoid doing any multiplications in the inner loop + * by precalculating the constants times Cb and Cr for all possible values. + * For 8-bit JSAMPLEs this is very reasonable (only 256 entries per table); + * for 9-bit to 12-bit samples it is still acceptable. It's not very + * reasonable for 16-bit samples, but if you want lossless storage + * you shouldn't be changing colorspace anyway. + * The Cr=>R and Cb=>B values can be rounded to integers in advance; + * the values for the G calculation are left scaled up, + * since we must add them together before rounding. + */ + +#define SCALEBITS 16 /* speediest right-shift on some machines */ +#define ONE_HALF ((INT32) 1 << (SCALEBITS-1)) +#define FIX(x) ((INT32) ((x) * (1L<RGB and BG_YCC->RGB colorspace conversion. + */ + +LOCAL(void) +build_ycc_rgb_table (j_decompress_ptr cinfo) +/* Normal case, sYCC */ +{ + my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; + int i; + INT32 x; + SHIFT_TEMPS + + cconvert->Cr_r_tab = (int *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(int)); + cconvert->Cb_b_tab = (int *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(int)); + cconvert->Cr_g_tab = (INT32 *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(INT32)); + cconvert->Cb_g_tab = (INT32 *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(INT32)); + + for (i = 0, x = -CENTERJSAMPLE; i <= MAXJSAMPLE; i++, x++) { + /* i is the actual input pixel value, in the range 0..MAXJSAMPLE */ + /* The Cb or Cr value we are thinking of is x = i - CENTERJSAMPLE */ + /* Cr=>R value is nearest int to 1.402 * x */ + cconvert->Cr_r_tab[i] = (int) DESCALE(FIX(1.402) * x, SCALEBITS); + /* Cb=>B value is nearest int to 1.772 * x */ + cconvert->Cb_b_tab[i] = (int) DESCALE(FIX(1.772) * x, SCALEBITS); + /* Cr=>G value is scaled-up -0.714136286 * x */ + cconvert->Cr_g_tab[i] = (- FIX(0.714136286)) * x; + /* Cb=>G value is scaled-up -0.344136286 * x */ + /* We also add in ONE_HALF so that need not do it in inner loop */ + cconvert->Cb_g_tab[i] = (- FIX(0.344136286)) * x + ONE_HALF; + } +} + + +LOCAL(void) +build_bg_ycc_rgb_table (j_decompress_ptr cinfo) +/* Wide gamut case, bg-sYCC */ +{ + my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; + int i; + INT32 x; + SHIFT_TEMPS + + cconvert->Cr_r_tab = (int *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(int)); + cconvert->Cb_b_tab = (int *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(int)); + cconvert->Cr_g_tab = (INT32 *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(INT32)); + cconvert->Cb_g_tab = (INT32 *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(INT32)); + + for (i = 0, x = -CENTERJSAMPLE; i <= MAXJSAMPLE; i++, x++) { + /* i is the actual input pixel value, in the range 0..MAXJSAMPLE */ + /* The Cb or Cr value we are thinking of is x = i - CENTERJSAMPLE */ + /* Cr=>R value is nearest int to 2.804 * x */ + cconvert->Cr_r_tab[i] = (int) DESCALE(FIX(2.804) * x, SCALEBITS); + /* Cb=>B value is nearest int to 3.544 * x */ + cconvert->Cb_b_tab[i] = (int) DESCALE(FIX(3.544) * x, SCALEBITS); + /* Cr=>G value is scaled-up -1.428272572 * x */ + cconvert->Cr_g_tab[i] = (- FIX(1.428272572)) * x; + /* Cb=>G value is scaled-up -0.688272572 * x */ + /* We also add in ONE_HALF so that need not do it in inner loop */ + cconvert->Cb_g_tab[i] = (- FIX(0.688272572)) * x + ONE_HALF; + } +} + + +/* + * Convert some rows of samples to the output colorspace. + * + * Note that we change from noninterleaved, one-plane-per-component format + * to interleaved-pixel format. The output buffer is therefore three times + * as wide as the input buffer. + * + * A starting row offset is provided only for the input buffer. The caller + * can easily adjust the passed output_buf value to accommodate any row + * offset required on that side. + */ + +METHODDEF(void) +ycc_rgb_convert (j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION input_row, + JSAMPARRAY output_buf, int num_rows) +{ + my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; + register int y, cb, cr; + register JSAMPROW outptr; + register JSAMPROW inptr0, inptr1, inptr2; + register JDIMENSION col; + JDIMENSION num_cols = cinfo->output_width; + /* copy these pointers into registers if possible */ + register JSAMPLE * range_limit = cinfo->sample_range_limit; + register int * Crrtab = cconvert->Cr_r_tab; + register int * Cbbtab = cconvert->Cb_b_tab; + register INT32 * Crgtab = cconvert->Cr_g_tab; + register INT32 * Cbgtab = cconvert->Cb_g_tab; + SHIFT_TEMPS + + while (--num_rows >= 0) { + inptr0 = input_buf[0][input_row]; + inptr1 = input_buf[1][input_row]; + inptr2 = input_buf[2][input_row]; + input_row++; + outptr = *output_buf++; + for (col = 0; col < num_cols; col++) { + y = GETJSAMPLE(inptr0[col]); + cb = GETJSAMPLE(inptr1[col]); + cr = GETJSAMPLE(inptr2[col]); + /* Range-limiting is essential due to noise introduced by DCT losses, + * for extended gamut (sYCC) and wide gamut (bg-sYCC) encodings. + */ + outptr[RGB_RED] = range_limit[y + Crrtab[cr]]; + outptr[RGB_GREEN] = range_limit[y + + ((int) RIGHT_SHIFT(Cbgtab[cb] + Crgtab[cr], + SCALEBITS))]; + outptr[RGB_BLUE] = range_limit[y + Cbbtab[cb]]; + outptr += RGB_PIXELSIZE; + } + } +} + + +/**************** Cases other than YCC -> RGB ****************/ + + +/* + * Initialize for RGB->grayscale colorspace conversion. + */ + +LOCAL(void) +build_rgb_y_table (j_decompress_ptr cinfo) +{ + my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; + INT32 i; + + cconvert->R_y_tab = (INT32 *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(INT32)); + cconvert->G_y_tab = (INT32 *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(INT32)); + cconvert->B_y_tab = (INT32 *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(INT32)); + + for (i = 0; i <= MAXJSAMPLE; i++) { + cconvert->R_y_tab[i] = FIX(0.299) * i; + cconvert->G_y_tab[i] = FIX(0.587) * i; + cconvert->B_y_tab[i] = FIX(0.114) * i + ONE_HALF; + } +} + + +/* + * Convert RGB to grayscale. + */ + +METHODDEF(void) +rgb_gray_convert (j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION input_row, + JSAMPARRAY output_buf, int num_rows) +{ + my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; + register INT32 y; + register INT32 * Rytab = cconvert->R_y_tab; + register INT32 * Gytab = cconvert->G_y_tab; + register INT32 * Bytab = cconvert->B_y_tab; + register JSAMPROW outptr; + register JSAMPROW inptr0, inptr1, inptr2; + register JDIMENSION col; + JDIMENSION num_cols = cinfo->output_width; + + while (--num_rows >= 0) { + inptr0 = input_buf[0][input_row]; + inptr1 = input_buf[1][input_row]; + inptr2 = input_buf[2][input_row]; + input_row++; + outptr = *output_buf++; + for (col = 0; col < num_cols; col++) { + y = Rytab[GETJSAMPLE(inptr0[col])]; + y += Gytab[GETJSAMPLE(inptr1[col])]; + y += Bytab[GETJSAMPLE(inptr2[col])]; + outptr[col] = (JSAMPLE) (y >> SCALEBITS); + } + } +} + + +/* + * Convert some rows of samples to the output colorspace. + * [R-G,G,B-G] to [R,G,B] conversion with modulo calculation + * (inverse color transform). + * This can be seen as an adaption of the general YCbCr->RGB + * conversion equation with Kr = Kb = 0, while replacing the + * normalization by modulo calculation. + */ + +METHODDEF(void) +rgb1_rgb_convert (j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION input_row, + JSAMPARRAY output_buf, int num_rows) +{ + register int r, g, b; + register JSAMPROW outptr; + register JSAMPROW inptr0, inptr1, inptr2; + register JDIMENSION col; + JDIMENSION num_cols = cinfo->output_width; + + while (--num_rows >= 0) { + inptr0 = input_buf[0][input_row]; + inptr1 = input_buf[1][input_row]; + inptr2 = input_buf[2][input_row]; + input_row++; + outptr = *output_buf++; + for (col = 0; col < num_cols; col++) { + r = GETJSAMPLE(inptr0[col]); + g = GETJSAMPLE(inptr1[col]); + b = GETJSAMPLE(inptr2[col]); + /* Assume that MAXJSAMPLE+1 is a power of 2, so that the MOD + * (modulo) operator is equivalent to the bitmask operator AND. + */ + outptr[RGB_RED] = (JSAMPLE) ((r + g - CENTERJSAMPLE) & MAXJSAMPLE); + outptr[RGB_GREEN] = (JSAMPLE) g; + outptr[RGB_BLUE] = (JSAMPLE) ((b + g - CENTERJSAMPLE) & MAXJSAMPLE); + outptr += RGB_PIXELSIZE; + } + } +} + + +/* + * [R-G,G,B-G] to grayscale conversion with modulo calculation + * (inverse color transform). + */ + +METHODDEF(void) +rgb1_gray_convert (j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION input_row, + JSAMPARRAY output_buf, int num_rows) +{ + my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; + register int r, g, b; + register INT32 y; + register INT32 * Rytab = cconvert->R_y_tab; + register INT32 * Gytab = cconvert->G_y_tab; + register INT32 * Bytab = cconvert->B_y_tab; + register JSAMPROW outptr; + register JSAMPROW inptr0, inptr1, inptr2; + register JDIMENSION col; + JDIMENSION num_cols = cinfo->output_width; + + while (--num_rows >= 0) { + inptr0 = input_buf[0][input_row]; + inptr1 = input_buf[1][input_row]; + inptr2 = input_buf[2][input_row]; + input_row++; + outptr = *output_buf++; + for (col = 0; col < num_cols; col++) { + r = GETJSAMPLE(inptr0[col]); + g = GETJSAMPLE(inptr1[col]); + b = GETJSAMPLE(inptr2[col]); + /* Assume that MAXJSAMPLE+1 is a power of 2, so that the MOD + * (modulo) operator is equivalent to the bitmask operator AND. + */ + y = Rytab[(r + g - CENTERJSAMPLE) & MAXJSAMPLE]; + y += Gytab[g]; + y += Bytab[(b + g - CENTERJSAMPLE) & MAXJSAMPLE]; + outptr[col] = (JSAMPLE) (y >> SCALEBITS); + } + } +} + + +/* + * Convert some rows of samples to the output colorspace. + * No colorspace change, but conversion from separate-planes + * to interleaved representation. + */ + +METHODDEF(void) +rgb_convert (j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION input_row, + JSAMPARRAY output_buf, int num_rows) +{ + register JSAMPROW outptr; + register JSAMPROW inptr0, inptr1, inptr2; + register JDIMENSION col; + JDIMENSION num_cols = cinfo->output_width; + + while (--num_rows >= 0) { + inptr0 = input_buf[0][input_row]; + inptr1 = input_buf[1][input_row]; + inptr2 = input_buf[2][input_row]; + input_row++; + outptr = *output_buf++; + for (col = 0; col < num_cols; col++) { + /* We can dispense with GETJSAMPLE() here */ + outptr[RGB_RED] = inptr0[col]; + outptr[RGB_GREEN] = inptr1[col]; + outptr[RGB_BLUE] = inptr2[col]; + outptr += RGB_PIXELSIZE; + } + } +} + + +/* + * Color conversion for no colorspace change: just copy the data, + * converting from separate-planes to interleaved representation. + * Note: Omit uninteresting components in output buffer. + */ + +METHODDEF(void) +null_convert (j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION input_row, + JSAMPARRAY output_buf, int num_rows) +{ + register JSAMPROW outptr; + register JSAMPROW inptr; + register JDIMENSION count; + register int out_comps = cinfo->out_color_components; + JDIMENSION num_cols = cinfo->output_width; + JSAMPROW startptr; + int ci; + jpeg_component_info *compptr; + + while (--num_rows >= 0) { + /* It seems fastest to make a separate pass for each component. */ + startptr = *output_buf++; + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + if (! compptr->component_needed) + continue; /* skip uninteresting component */ + inptr = input_buf[ci][input_row]; + outptr = startptr++; + for (count = num_cols; count > 0; count--) { + *outptr = *inptr++; /* don't need GETJSAMPLE() here */ + outptr += out_comps; + } + } + input_row++; + } +} + + +/* + * Color conversion for grayscale: just copy the data. + * This also works for YCC -> grayscale conversion, in which + * we just copy the Y (luminance) component and ignore chrominance. + */ + +METHODDEF(void) +grayscale_convert (j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION input_row, + JSAMPARRAY output_buf, int num_rows) +{ + jcopy_sample_rows(input_buf[0] + input_row, output_buf, + num_rows, cinfo->output_width); +} + + +/* + * Convert grayscale to RGB: just duplicate the graylevel three times. + * This is provided to support applications that don't want to cope + * with grayscale as a separate case. + */ + +METHODDEF(void) +gray_rgb_convert (j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION input_row, + JSAMPARRAY output_buf, int num_rows) +{ + register JSAMPROW outptr; + register JSAMPROW inptr; + register JDIMENSION col; + JDIMENSION num_cols = cinfo->output_width; + + while (--num_rows >= 0) { + inptr = input_buf[0][input_row++]; + outptr = *output_buf++; + for (col = 0; col < num_cols; col++) { + /* We can dispense with GETJSAMPLE() here */ + outptr[RGB_RED] = outptr[RGB_GREEN] = outptr[RGB_BLUE] = inptr[col]; + outptr += RGB_PIXELSIZE; + } + } +} + + +/* + * Convert some rows of samples to the output colorspace. + * This version handles Adobe-style YCCK->CMYK conversion, + * where we convert YCbCr to R=1-C, G=1-M, and B=1-Y using the + * same conversion as above, while passing K (black) unchanged. + * We assume build_ycc_rgb_table has been called. + */ + +METHODDEF(void) +ycck_cmyk_convert (j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION input_row, + JSAMPARRAY output_buf, int num_rows) +{ + my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; + register int y, cb, cr; + register JSAMPROW outptr; + register JSAMPROW inptr0, inptr1, inptr2, inptr3; + register JDIMENSION col; + JDIMENSION num_cols = cinfo->output_width; + /* copy these pointers into registers if possible */ + register JSAMPLE * range_limit = cinfo->sample_range_limit; + register int * Crrtab = cconvert->Cr_r_tab; + register int * Cbbtab = cconvert->Cb_b_tab; + register INT32 * Crgtab = cconvert->Cr_g_tab; + register INT32 * Cbgtab = cconvert->Cb_g_tab; + SHIFT_TEMPS + + while (--num_rows >= 0) { + inptr0 = input_buf[0][input_row]; + inptr1 = input_buf[1][input_row]; + inptr2 = input_buf[2][input_row]; + inptr3 = input_buf[3][input_row]; + input_row++; + outptr = *output_buf++; + for (col = 0; col < num_cols; col++) { + y = GETJSAMPLE(inptr0[col]); + cb = GETJSAMPLE(inptr1[col]); + cr = GETJSAMPLE(inptr2[col]); + /* Range-limiting is essential due to noise introduced by DCT losses, + * and for extended gamut encodings (sYCC). + */ + outptr[0] = range_limit[MAXJSAMPLE - (y + Crrtab[cr])]; /* red */ + outptr[1] = range_limit[MAXJSAMPLE - (y + /* green */ + ((int) RIGHT_SHIFT(Cbgtab[cb] + Crgtab[cr], + SCALEBITS)))]; + outptr[2] = range_limit[MAXJSAMPLE - (y + Cbbtab[cb])]; /* blue */ + /* K passes through unchanged */ + outptr[3] = inptr3[col]; /* don't need GETJSAMPLE here */ + outptr += 4; + } + } +} + + +/* + * Convert CMYK to YK part of YCCK for colorless output. + * We assume build_rgb_y_table has been called. + */ + +METHODDEF(void) +cmyk_yk_convert (j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION input_row, + JSAMPARRAY output_buf, int num_rows) +{ + my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; + register INT32 y; + register INT32 * Rytab = cconvert->R_y_tab; + register INT32 * Gytab = cconvert->G_y_tab; + register INT32 * Bytab = cconvert->B_y_tab; + register JSAMPROW outptr; + register JSAMPROW inptr0, inptr1, inptr2, inptr3; + register JDIMENSION col; + JDIMENSION num_cols = cinfo->output_width; + + while (--num_rows >= 0) { + inptr0 = input_buf[0][input_row]; + inptr1 = input_buf[1][input_row]; + inptr2 = input_buf[2][input_row]; + inptr3 = input_buf[3][input_row]; + input_row++; + outptr = *output_buf++; + for (col = 0; col < num_cols; col++) { + y = Rytab[MAXJSAMPLE - GETJSAMPLE(inptr0[col])]; + y += Gytab[MAXJSAMPLE - GETJSAMPLE(inptr1[col])]; + y += Bytab[MAXJSAMPLE - GETJSAMPLE(inptr2[col])]; + outptr[0] = (JSAMPLE) (y >> SCALEBITS); + /* K passes through unchanged */ + outptr[1] = inptr3[col]; /* don't need GETJSAMPLE here */ + outptr += 2; + } + } +} + + +/* + * Empty method for start_pass. + */ + +METHODDEF(void) +start_pass_dcolor (j_decompress_ptr cinfo) +{ + /* no work needed */ +} + + +/* + * Module initialization routine for output colorspace conversion. + */ + +GLOBAL(void) +jinit_color_deconverter (j_decompress_ptr cinfo) +{ + my_cconvert_ptr cconvert; + int ci, i; + + cconvert = (my_cconvert_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_color_deconverter)); + cinfo->cconvert = &cconvert->pub; + cconvert->pub.start_pass = start_pass_dcolor; + + /* Make sure num_components agrees with jpeg_color_space */ + switch (cinfo->jpeg_color_space) { + case JCS_GRAYSCALE: + if (cinfo->num_components != 1) + ERREXIT(cinfo, JERR_BAD_J_COLORSPACE); + break; + + case JCS_RGB: + case JCS_YCbCr: + case JCS_BG_RGB: + case JCS_BG_YCC: + if (cinfo->num_components != 3) + ERREXIT(cinfo, JERR_BAD_J_COLORSPACE); + break; + + case JCS_CMYK: + case JCS_YCCK: + if (cinfo->num_components != 4) + ERREXIT(cinfo, JERR_BAD_J_COLORSPACE); + break; + + default: /* JCS_UNKNOWN can be anything */ + if (cinfo->num_components < 1) + ERREXIT(cinfo, JERR_BAD_J_COLORSPACE); + } + + /* Support color transform only for RGB colorspaces */ + if (cinfo->color_transform && + cinfo->jpeg_color_space != JCS_RGB && + cinfo->jpeg_color_space != JCS_BG_RGB) + ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); + + /* Set out_color_components and conversion method based on requested space. + * Also adjust the component_needed flags for any unused components, + * so that earlier pipeline stages can avoid useless computation. + */ + + switch (cinfo->out_color_space) { + case JCS_GRAYSCALE: + cinfo->out_color_components = 1; + switch (cinfo->jpeg_color_space) { + case JCS_GRAYSCALE: + case JCS_YCbCr: + case JCS_BG_YCC: + cconvert->pub.color_convert = grayscale_convert; + /* For color->grayscale conversion, only the Y (0) component is needed */ + for (ci = 1; ci < cinfo->num_components; ci++) + cinfo->comp_info[ci].component_needed = FALSE; + break; + case JCS_RGB: + switch (cinfo->color_transform) { + case JCT_NONE: + cconvert->pub.color_convert = rgb_gray_convert; + break; + case JCT_SUBTRACT_GREEN: + cconvert->pub.color_convert = rgb1_gray_convert; + break; + default: + ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); + } + build_rgb_y_table(cinfo); + break; + default: + ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); + } + break; + + case JCS_RGB: + cinfo->out_color_components = RGB_PIXELSIZE; + switch (cinfo->jpeg_color_space) { + case JCS_GRAYSCALE: + cconvert->pub.color_convert = gray_rgb_convert; + break; + case JCS_YCbCr: + cconvert->pub.color_convert = ycc_rgb_convert; + build_ycc_rgb_table(cinfo); + break; + case JCS_BG_YCC: + cconvert->pub.color_convert = ycc_rgb_convert; + build_bg_ycc_rgb_table(cinfo); + break; + case JCS_RGB: + switch (cinfo->color_transform) { + case JCT_NONE: + cconvert->pub.color_convert = rgb_convert; + break; + case JCT_SUBTRACT_GREEN: + cconvert->pub.color_convert = rgb1_rgb_convert; + break; + default: + ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); + } + break; + default: + ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); + } + break; + + case JCS_BG_RGB: + if (cinfo->jpeg_color_space != JCS_BG_RGB) + ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); + cinfo->out_color_components = RGB_PIXELSIZE; + switch (cinfo->color_transform) { + case JCT_NONE: + cconvert->pub.color_convert = rgb_convert; + break; + case JCT_SUBTRACT_GREEN: + cconvert->pub.color_convert = rgb1_rgb_convert; + break; + default: + ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); + } + break; + + case JCS_CMYK: + if (cinfo->jpeg_color_space != JCS_YCCK) + goto def_label; + cinfo->out_color_components = 4; + cconvert->pub.color_convert = ycck_cmyk_convert; + build_ycc_rgb_table(cinfo); + break; + + case JCS_YCCK: + if (cinfo->jpeg_color_space != JCS_CMYK || + /* Support only YK part of YCCK for colorless output */ + ! cinfo->comp_info[0].component_needed || + cinfo->comp_info[1].component_needed || + cinfo->comp_info[2].component_needed || + ! cinfo->comp_info[3].component_needed) + goto def_label; + cinfo->out_color_components = 2; + /* Need all components on input side */ + cinfo->comp_info[1].component_needed = TRUE; + cinfo->comp_info[2].component_needed = TRUE; + cconvert->pub.color_convert = cmyk_yk_convert; + build_rgb_y_table(cinfo); + break; + + default: def_label: /* permit null conversion to same output space */ + if (cinfo->out_color_space != cinfo->jpeg_color_space) + /* unsupported non-null conversion */ + ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); + i = 0; + for (ci = 0; ci < cinfo->num_components; ci++) + if (cinfo->comp_info[ci].component_needed) + i++; /* count output color components */ + cinfo->out_color_components = i; + cconvert->pub.color_convert = null_convert; + } + + if (cinfo->quantize_colors) + cinfo->output_components = 1; /* single colormapped output component */ + else + cinfo->output_components = cinfo->out_color_components; +} diff --git a/cmake/externals/libjpeg/jdct.h b/cmake/externals/libjpeg/jdct.h new file mode 100644 index 000000000..0f251590c --- /dev/null +++ b/cmake/externals/libjpeg/jdct.h @@ -0,0 +1,409 @@ +/* + * jdct.h + * + * Copyright (C) 1994-1996, Thomas G. Lane. + * Modified 2002-2023 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This include file contains common declarations for the forward and + * inverse DCT modules. These declarations are private to the DCT managers + * (jcdctmgr.c, jddctmgr.c) and the individual DCT algorithms. + * The individual DCT algorithms are kept in separate files to ease + * machine-dependent tuning (e.g., assembly coding). + */ + + +/* + * A forward DCT routine is given a pointer to an input sample array and + * a pointer to a work area of type DCTELEM[]; the DCT is to be performed + * in-place in that buffer. Type DCTELEM is int for 8-bit samples, INT32 + * for 12-bit samples. (NOTE: Floating-point DCT implementations use an + * array of type FAST_FLOAT, instead.) + * The input data is to be fetched from the sample array starting at a + * specified column. (Any row offset needed will be applied to the array + * pointer before it is passed to the FDCT code.) + * Note that the number of samples fetched by the FDCT routine is + * DCT_h_scaled_size * DCT_v_scaled_size. + * The DCT outputs are returned scaled up by a factor of 8; they therefore + * have a range of +-8K for 8-bit data, +-128K for 12-bit data. This + * convention improves accuracy in integer implementations and saves some + * work in floating-point ones. + * Quantization of the output coefficients is done by jcdctmgr.c. + */ + +#if BITS_IN_JSAMPLE == 8 +typedef int DCTELEM; /* 16 or 32 bits is fine */ +#else +typedef INT32 DCTELEM; /* must have 32 bits */ +#endif + +typedef JMETHOD(void, forward_DCT_method_ptr, (DCTELEM * data, + JSAMPARRAY sample_data, + JDIMENSION start_col)); +typedef JMETHOD(void, float_DCT_method_ptr, (FAST_FLOAT * data, + JSAMPARRAY sample_data, + JDIMENSION start_col)); + + +/* + * An inverse DCT routine is given a pointer to the input JBLOCK and a pointer + * to an output sample array. The routine must dequantize the input data as + * well as perform the IDCT; for dequantization, it uses the multiplier table + * pointed to by compptr->dct_table. The output data is to be placed into the + * sample array starting at a specified column. (Any row offset needed will + * be applied to the array pointer before it is passed to the IDCT code.) + * Note that the number of samples emitted by the IDCT routine is + * DCT_h_scaled_size * DCT_v_scaled_size. + */ + +/* typedef inverse_DCT_method_ptr is declared in jpegint.h */ + +/* + * Each IDCT routine has its own ideas about the best dct_table element type. + */ + +typedef MULTIPLIER ISLOW_MULT_TYPE; /* short or int, whichever is faster */ +#if BITS_IN_JSAMPLE == 8 +typedef MULTIPLIER IFAST_MULT_TYPE; /* 16 bits is OK, use short if faster */ +#define IFAST_SCALE_BITS 2 /* fractional bits in scale factors */ +#else +typedef INT32 IFAST_MULT_TYPE; /* need 32 bits for scaled quantizers */ +#define IFAST_SCALE_BITS 13 /* fractional bits in scale factors */ +#endif +typedef FAST_FLOAT FLOAT_MULT_TYPE; /* preferred floating type */ + + +/* + * Each IDCT routine is responsible for range-limiting its results and + * converting them to unsigned form (0..MAXJSAMPLE). The raw outputs could + * be quite far out of range if the input data is corrupt, so a bulletproof + * range-limiting step is required. We use a mask-and-table-lookup method + * to do the combined operations quickly, assuming that RANGE_CENTER + * (defined in jpegint.h) is a power of 2. See the comments with + * prepare_range_limit_table (in jdmaster.c) for more info. + */ + +#define RANGE_MASK (RANGE_CENTER * 2 - 1) +#define RANGE_SUBSET (RANGE_CENTER - CENTERJSAMPLE) + +#define IDCT_range_limit(cinfo) ((cinfo)->sample_range_limit - RANGE_SUBSET) + + +/* Short forms of external names for systems with brain-damaged linkers. */ + +#ifdef NEED_SHORT_EXTERNAL_NAMES +#define jpeg_fdct_islow jFDislow +#define jpeg_fdct_ifast jFDifast +#define jpeg_fdct_float jFDfloat +#define jpeg_fdct_7x7 jFD7x7 +#define jpeg_fdct_6x6 jFD6x6 +#define jpeg_fdct_5x5 jFD5x5 +#define jpeg_fdct_4x4 jFD4x4 +#define jpeg_fdct_3x3 jFD3x3 +#define jpeg_fdct_2x2 jFD2x2 +#define jpeg_fdct_1x1 jFD1x1 +#define jpeg_fdct_9x9 jFD9x9 +#define jpeg_fdct_10x10 jFD10x10 +#define jpeg_fdct_11x11 jFD11x11 +#define jpeg_fdct_12x12 jFD12x12 +#define jpeg_fdct_13x13 jFD13x13 +#define jpeg_fdct_14x14 jFD14x14 +#define jpeg_fdct_15x15 jFD15x15 +#define jpeg_fdct_16x16 jFD16x16 +#define jpeg_fdct_16x8 jFD16x8 +#define jpeg_fdct_14x7 jFD14x7 +#define jpeg_fdct_12x6 jFD12x6 +#define jpeg_fdct_10x5 jFD10x5 +#define jpeg_fdct_8x4 jFD8x4 +#define jpeg_fdct_6x3 jFD6x3 +#define jpeg_fdct_4x2 jFD4x2 +#define jpeg_fdct_2x1 jFD2x1 +#define jpeg_fdct_8x16 jFD8x16 +#define jpeg_fdct_7x14 jFD7x14 +#define jpeg_fdct_6x12 jFD6x12 +#define jpeg_fdct_5x10 jFD5x10 +#define jpeg_fdct_4x8 jFD4x8 +#define jpeg_fdct_3x6 jFD3x6 +#define jpeg_fdct_2x4 jFD2x4 +#define jpeg_fdct_1x2 jFD1x2 +#define jpeg_idct_islow jRDislow +#define jpeg_idct_ifast jRDifast +#define jpeg_idct_float jRDfloat +#define jpeg_idct_7x7 jRD7x7 +#define jpeg_idct_6x6 jRD6x6 +#define jpeg_idct_5x5 jRD5x5 +#define jpeg_idct_4x4 jRD4x4 +#define jpeg_idct_3x3 jRD3x3 +#define jpeg_idct_2x2 jRD2x2 +#define jpeg_idct_1x1 jRD1x1 +#define jpeg_idct_9x9 jRD9x9 +#define jpeg_idct_10x10 jRD10x10 +#define jpeg_idct_11x11 jRD11x11 +#define jpeg_idct_12x12 jRD12x12 +#define jpeg_idct_13x13 jRD13x13 +#define jpeg_idct_14x14 jRD14x14 +#define jpeg_idct_15x15 jRD15x15 +#define jpeg_idct_16x16 jRD16x16 +#define jpeg_idct_16x8 jRD16x8 +#define jpeg_idct_14x7 jRD14x7 +#define jpeg_idct_12x6 jRD12x6 +#define jpeg_idct_10x5 jRD10x5 +#define jpeg_idct_8x4 jRD8x4 +#define jpeg_idct_6x3 jRD6x3 +#define jpeg_idct_4x2 jRD4x2 +#define jpeg_idct_2x1 jRD2x1 +#define jpeg_idct_8x16 jRD8x16 +#define jpeg_idct_7x14 jRD7x14 +#define jpeg_idct_6x12 jRD6x12 +#define jpeg_idct_5x10 jRD5x10 +#define jpeg_idct_4x8 jRD4x8 +#define jpeg_idct_3x6 jRD3x6 +#define jpeg_idct_2x4 jRD2x4 +#define jpeg_idct_1x2 jRD1x2 +#endif /* NEED_SHORT_EXTERNAL_NAMES */ + +/* Extern declarations for the forward and inverse DCT routines. */ + +EXTERN(void) jpeg_fdct_islow + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_ifast + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_float + JPP((FAST_FLOAT * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_7x7 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_6x6 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_5x5 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_4x4 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_3x3 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_2x2 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_1x1 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_9x9 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_10x10 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_11x11 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_12x12 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_13x13 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_14x14 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_15x15 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_16x16 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_16x8 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_14x7 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_12x6 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_10x5 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_8x4 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_6x3 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_4x2 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_2x1 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_8x16 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_7x14 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_6x12 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_5x10 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_4x8 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_3x6 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_2x4 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); +EXTERN(void) jpeg_fdct_1x2 + JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); + +EXTERN(void) jpeg_idct_islow + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_ifast + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_float + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_7x7 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_6x6 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_5x5 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_4x4 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_3x3 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_2x2 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_1x1 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_9x9 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_10x10 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_11x11 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_12x12 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_13x13 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_14x14 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_15x15 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_16x16 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_16x8 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_14x7 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_12x6 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_10x5 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_8x4 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_6x3 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_4x2 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_2x1 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_8x16 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_7x14 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_6x12 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_5x10 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_4x8 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_3x6 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_2x4 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); +EXTERN(void) jpeg_idct_1x2 + JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); + + +/* + * Macros for handling fixed-point arithmetic; these are used by many + * but not all of the DCT/IDCT modules. + * + * All values are expected to be of type INT32. + * Fractional constants are scaled left by CONST_BITS bits. + * CONST_BITS is defined within each module using these macros, + * and may differ from one module to the next. + */ + +#define ONE ((INT32) 1) +#define CONST_SCALE (ONE << CONST_BITS) + +/* Convert a positive real constant to an integer scaled by CONST_SCALE. + * Caution: some C compilers fail to reduce "FIX(constant)" at compile time, + * thus causing a lot of useless floating-point operations at run time. + */ + +#define FIX(x) ((INT32) ((x) * CONST_SCALE + 0.5)) + +/* Multiply an INT32 variable by an INT32 constant to yield an INT32 result. + * This macro is used only when the two inputs will actually be no more than + * 16 bits wide, so that a 16x16->32 bit multiply can be used instead of a + * full 32x32 multiply. This provides a useful speedup on many machines. + * Unfortunately there is no way to specify a 16x16->32 multiply portably + * in C, but some C compilers will do the right thing if you provide the + * correct combination of casts. + */ + +#ifdef SHORTxSHORT_32 /* may work if 'int' is 32 bits */ +#define MULTIPLY16C16(var,const) (((INT16) (var)) * ((INT16) (const))) +#endif +#ifdef SHORTxLCONST_32 /* known to work with Microsoft C 6.0 */ +#define MULTIPLY16C16(var,const) (((INT16) (var)) * ((INT32) (const))) +#endif + +#ifndef MULTIPLY16C16 /* default definition */ +#define MULTIPLY16C16(var,const) ((var) * (const)) +#endif + +/* Same except both inputs are variables. */ + +#ifdef SHORTxSHORT_32 /* may work if 'int' is 32 bits */ +#define MULTIPLY16V16(var1,var2) (((INT16) (var1)) * ((INT16) (var2))) +#endif + +#ifndef MULTIPLY16V16 /* default definition */ +#define MULTIPLY16V16(var1,var2) ((var1) * (var2)) +#endif + +/* Like RIGHT_SHIFT, but applies to a DCTELEM. + * We assume that int right shift is unsigned if INT32 right shift is. + */ + +#ifdef RIGHT_SHIFT_IS_UNSIGNED +#define ISHIFT_TEMPS DCTELEM ishift_temp; +#if BITS_IN_JSAMPLE == 8 +#define DCTELEMBITS 16 /* DCTELEM may be 16 or 32 bits */ +#else +#define DCTELEMBITS 32 /* DCTELEM must be 32 bits */ +#endif +#define IRIGHT_SHIFT(x,shft) \ + ((ishift_temp = (x)) < 0 ? \ + (ishift_temp >> (shft)) | ((~((DCTELEM) 0)) << (DCTELEMBITS-(shft))) : \ + (ishift_temp >> (shft))) +#else +#define ISHIFT_TEMPS +#define IRIGHT_SHIFT(x,shft) ((x) >> (shft)) +#endif diff --git a/cmake/externals/libjpeg/jddctmgr.c b/cmake/externals/libjpeg/jddctmgr.c new file mode 100644 index 000000000..9ecfbb510 --- /dev/null +++ b/cmake/externals/libjpeg/jddctmgr.c @@ -0,0 +1,384 @@ +/* + * jddctmgr.c + * + * Copyright (C) 1994-1996, Thomas G. Lane. + * Modified 2002-2013 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains the inverse-DCT management logic. + * This code selects a particular IDCT implementation to be used, + * and it performs related housekeeping chores. No code in this file + * is executed per IDCT step, only during output pass setup. + * + * Note that the IDCT routines are responsible for performing coefficient + * dequantization as well as the IDCT proper. This module sets up the + * dequantization multiplier table needed by the IDCT routine. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" +#include "jdct.h" /* Private declarations for DCT subsystem */ + + +/* + * The decompressor input side (jdinput.c) saves away the appropriate + * quantization table for each component at the start of the first scan + * involving that component. (This is necessary in order to correctly + * decode files that reuse Q-table slots.) + * When we are ready to make an output pass, the saved Q-table is converted + * to a multiplier table that will actually be used by the IDCT routine. + * The multiplier table contents are IDCT-method-dependent. To support + * application changes in IDCT method between scans, we can remake the + * multiplier tables if necessary. + * In buffered-image mode, the first output pass may occur before any data + * has been seen for some components, and thus before their Q-tables have + * been saved away. To handle this case, multiplier tables are preset + * to zeroes; the result of the IDCT will be a neutral gray level. + */ + + +/* Private subobject for this module */ + +typedef struct { + struct jpeg_inverse_dct pub; /* public fields */ + + /* This array contains the IDCT method code that each multiplier table + * is currently set up for, or -1 if it's not yet set up. + * The actual multiplier tables are pointed to by dct_table in the + * per-component comp_info structures. + */ + int cur_method[MAX_COMPONENTS]; +} my_idct_controller; + +typedef my_idct_controller * my_idct_ptr; + + +/* Allocated multiplier tables: big enough for any supported variant */ + +typedef union { + ISLOW_MULT_TYPE islow_array[DCTSIZE2]; +#ifdef DCT_IFAST_SUPPORTED + IFAST_MULT_TYPE ifast_array[DCTSIZE2]; +#endif +#ifdef DCT_FLOAT_SUPPORTED + FLOAT_MULT_TYPE float_array[DCTSIZE2]; +#endif +} multiplier_table; + + +/* The current scaled-IDCT routines require ISLOW-style multiplier tables, + * so be sure to compile that code if either ISLOW or SCALING is requested. + */ +#ifdef DCT_ISLOW_SUPPORTED +#define PROVIDE_ISLOW_TABLES +#else +#ifdef IDCT_SCALING_SUPPORTED +#define PROVIDE_ISLOW_TABLES +#endif +#endif + + +/* + * Prepare for an output pass. + * Here we select the proper IDCT routine for each component and build + * a matching multiplier table. + */ + +METHODDEF(void) +start_pass (j_decompress_ptr cinfo) +{ + my_idct_ptr idct = (my_idct_ptr) cinfo->idct; + int ci, i; + jpeg_component_info *compptr; + int method = 0; + inverse_DCT_method_ptr method_ptr = NULL; + JQUANT_TBL * qtbl; + + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + /* Select the proper IDCT routine for this component's scaling */ + switch ((compptr->DCT_h_scaled_size << 8) + compptr->DCT_v_scaled_size) { +#ifdef IDCT_SCALING_SUPPORTED + case ((1 << 8) + 1): + method_ptr = jpeg_idct_1x1; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((2 << 8) + 2): + method_ptr = jpeg_idct_2x2; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((3 << 8) + 3): + method_ptr = jpeg_idct_3x3; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((4 << 8) + 4): + method_ptr = jpeg_idct_4x4; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((5 << 8) + 5): + method_ptr = jpeg_idct_5x5; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((6 << 8) + 6): + method_ptr = jpeg_idct_6x6; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((7 << 8) + 7): + method_ptr = jpeg_idct_7x7; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((9 << 8) + 9): + method_ptr = jpeg_idct_9x9; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((10 << 8) + 10): + method_ptr = jpeg_idct_10x10; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((11 << 8) + 11): + method_ptr = jpeg_idct_11x11; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((12 << 8) + 12): + method_ptr = jpeg_idct_12x12; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((13 << 8) + 13): + method_ptr = jpeg_idct_13x13; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((14 << 8) + 14): + method_ptr = jpeg_idct_14x14; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((15 << 8) + 15): + method_ptr = jpeg_idct_15x15; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((16 << 8) + 16): + method_ptr = jpeg_idct_16x16; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((16 << 8) + 8): + method_ptr = jpeg_idct_16x8; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((14 << 8) + 7): + method_ptr = jpeg_idct_14x7; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((12 << 8) + 6): + method_ptr = jpeg_idct_12x6; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((10 << 8) + 5): + method_ptr = jpeg_idct_10x5; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((8 << 8) + 4): + method_ptr = jpeg_idct_8x4; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((6 << 8) + 3): + method_ptr = jpeg_idct_6x3; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((4 << 8) + 2): + method_ptr = jpeg_idct_4x2; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((2 << 8) + 1): + method_ptr = jpeg_idct_2x1; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((8 << 8) + 16): + method_ptr = jpeg_idct_8x16; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((7 << 8) + 14): + method_ptr = jpeg_idct_7x14; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((6 << 8) + 12): + method_ptr = jpeg_idct_6x12; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((5 << 8) + 10): + method_ptr = jpeg_idct_5x10; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((4 << 8) + 8): + method_ptr = jpeg_idct_4x8; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((3 << 8) + 6): + method_ptr = jpeg_idct_3x6; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((2 << 8) + 4): + method_ptr = jpeg_idct_2x4; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; + case ((1 << 8) + 2): + method_ptr = jpeg_idct_1x2; + method = JDCT_ISLOW; /* jidctint uses islow-style table */ + break; +#endif + case ((DCTSIZE << 8) + DCTSIZE): + switch (cinfo->dct_method) { +#ifdef DCT_ISLOW_SUPPORTED + case JDCT_ISLOW: + method_ptr = jpeg_idct_islow; + method = JDCT_ISLOW; + break; +#endif +#ifdef DCT_IFAST_SUPPORTED + case JDCT_IFAST: + method_ptr = jpeg_idct_ifast; + method = JDCT_IFAST; + break; +#endif +#ifdef DCT_FLOAT_SUPPORTED + case JDCT_FLOAT: + method_ptr = jpeg_idct_float; + method = JDCT_FLOAT; + break; +#endif + default: + ERREXIT(cinfo, JERR_NOT_COMPILED); + break; + } + break; + default: + ERREXIT2(cinfo, JERR_BAD_DCTSIZE, + compptr->DCT_h_scaled_size, compptr->DCT_v_scaled_size); + break; + } + idct->pub.inverse_DCT[ci] = method_ptr; + /* Create multiplier table from quant table. + * However, we can skip this if the component is uninteresting + * or if we already built the table. Also, if no quant table + * has yet been saved for the component, we leave the + * multiplier table all-zero; we'll be reading zeroes from the + * coefficient controller's buffer anyway. + */ + if (! compptr->component_needed || idct->cur_method[ci] == method) + continue; + qtbl = compptr->quant_table; + if (qtbl == NULL) /* happens if no data yet for component */ + continue; + idct->cur_method[ci] = method; + switch (method) { +#ifdef PROVIDE_ISLOW_TABLES + case JDCT_ISLOW: + { + /* For LL&M IDCT method, multipliers are equal to raw quantization + * coefficients, but are stored as ints to ensure access efficiency. + */ + ISLOW_MULT_TYPE * ismtbl = (ISLOW_MULT_TYPE *) compptr->dct_table; + for (i = 0; i < DCTSIZE2; i++) { + ismtbl[i] = (ISLOW_MULT_TYPE) qtbl->quantval[i]; + } + } + break; +#endif +#ifdef DCT_IFAST_SUPPORTED + case JDCT_IFAST: + { + /* For AA&N IDCT method, multipliers are equal to quantization + * coefficients scaled by scalefactor[row]*scalefactor[col], where + * scalefactor[0] = 1 + * scalefactor[k] = cos(k*PI/16) * sqrt(2) for k=1..7 + * For integer operation, the multiplier table is to be scaled by + * IFAST_SCALE_BITS. + */ + IFAST_MULT_TYPE * ifmtbl = (IFAST_MULT_TYPE *) compptr->dct_table; +#define CONST_BITS 14 + static const INT16 aanscales[DCTSIZE2] = { + /* precomputed values scaled up by 14 bits */ + 16384, 22725, 21407, 19266, 16384, 12873, 8867, 4520, + 22725, 31521, 29692, 26722, 22725, 17855, 12299, 6270, + 21407, 29692, 27969, 25172, 21407, 16819, 11585, 5906, + 19266, 26722, 25172, 22654, 19266, 15137, 10426, 5315, + 16384, 22725, 21407, 19266, 16384, 12873, 8867, 4520, + 12873, 17855, 16819, 15137, 12873, 10114, 6967, 3552, + 8867, 12299, 11585, 10426, 8867, 6967, 4799, 2446, + 4520, 6270, 5906, 5315, 4520, 3552, 2446, 1247 + }; + SHIFT_TEMPS + + for (i = 0; i < DCTSIZE2; i++) { + ifmtbl[i] = (IFAST_MULT_TYPE) + DESCALE(MULTIPLY16V16((INT32) qtbl->quantval[i], + (INT32) aanscales[i]), + CONST_BITS-IFAST_SCALE_BITS); + } + } + break; +#endif +#ifdef DCT_FLOAT_SUPPORTED + case JDCT_FLOAT: + { + /* For float AA&N IDCT method, multipliers are equal to quantization + * coefficients scaled by scalefactor[row]*scalefactor[col], where + * scalefactor[0] = 1 + * scalefactor[k] = cos(k*PI/16) * sqrt(2) for k=1..7 + * We apply a further scale factor of 1/8. + */ + FLOAT_MULT_TYPE * fmtbl = (FLOAT_MULT_TYPE *) compptr->dct_table; + int row, col; + static const double aanscalefactor[DCTSIZE] = { + 1.0, 1.387039845, 1.306562965, 1.175875602, + 1.0, 0.785694958, 0.541196100, 0.275899379 + }; + + i = 0; + for (row = 0; row < DCTSIZE; row++) { + for (col = 0; col < DCTSIZE; col++) { + fmtbl[i] = (FLOAT_MULT_TYPE) + ((double) qtbl->quantval[i] * + aanscalefactor[row] * aanscalefactor[col] * 0.125); + i++; + } + } + } + break; +#endif + default: + ERREXIT(cinfo, JERR_NOT_COMPILED); + break; + } + } +} + + +/* + * Initialize IDCT manager. + */ + +GLOBAL(void) +jinit_inverse_dct (j_decompress_ptr cinfo) +{ + my_idct_ptr idct; + int ci; + jpeg_component_info *compptr; + + idct = (my_idct_ptr) + (*cinfo->mem->alloc_small) ((j_common_ptr) cinfo, JPOOL_IMAGE, + SIZEOF(my_idct_controller)); + cinfo->idct = &idct->pub; + idct->pub.start_pass = start_pass; + + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + /* Allocate and pre-zero a multiplier table for each component */ + compptr->dct_table = + (*cinfo->mem->alloc_small) ((j_common_ptr) cinfo, JPOOL_IMAGE, + SIZEOF(multiplier_table)); + MEMZERO(compptr->dct_table, SIZEOF(multiplier_table)); + /* Mark multiplier table not yet set up for any method */ + idct->cur_method[ci] = -1; + } +} diff --git a/cmake/externals/libjpeg/jdhuff.c b/cmake/externals/libjpeg/jdhuff.c new file mode 100644 index 000000000..f175f0c32 --- /dev/null +++ b/cmake/externals/libjpeg/jdhuff.c @@ -0,0 +1,1559 @@ +/* + * jdhuff.c + * + * Copyright (C) 1991-1997, Thomas G. Lane. + * Modified 2006-2020 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains Huffman entropy decoding routines. + * Both sequential and progressive modes are supported in this single module. + * + * Much of the complexity here has to do with supporting input suspension. + * If the data source module demands suspension, we want to be able to back + * up to the start of the current MCU. To do this, we copy state variables + * into local working storage, and update them back to the permanent + * storage only upon successful completion of an MCU. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* Derived data constructed for each Huffman table */ + +#define HUFF_LOOKAHEAD 8 /* # of bits of lookahead */ + +typedef struct { + /* Basic tables: (element [0] of each array is unused) */ + INT32 maxcode[18]; /* largest code of length k (-1 if none) */ + /* (maxcode[17] is a sentinel to ensure jpeg_huff_decode terminates) */ + INT32 valoffset[17]; /* huffval[] offset for codes of length k */ + /* valoffset[k] = huffval[] index of 1st symbol of code length k, less + * the smallest code of length k; so given a code of length k, the + * corresponding symbol is huffval[code + valoffset[k]] + */ + + /* Link to public Huffman table (needed only in jpeg_huff_decode) */ + JHUFF_TBL *pub; + + /* Lookahead tables: indexed by the next HUFF_LOOKAHEAD bits of + * the input data stream. If the next Huffman code is no more + * than HUFF_LOOKAHEAD bits long, we can obtain its length and + * the corresponding symbol directly from these tables. + */ + int look_nbits[1< 32 bits on your machine, and shifting/masking longs is + * reasonably fast, making bit_buf_type be long and setting BIT_BUF_SIZE + * appropriately should be a win. Unfortunately we can't define the size + * with something like #define BIT_BUF_SIZE (sizeof(bit_buf_type)*8) + * because not all machines measure sizeof in 8-bit bytes. + */ + +typedef struct { /* Bitreading state saved across MCUs */ + bit_buf_type get_buffer; /* current bit-extraction buffer */ + int bits_left; /* # of unused bits in it */ +} bitread_perm_state; + +typedef struct { /* Bitreading working state within an MCU */ + /* Current data source location */ + /* We need a copy, rather than munging the original, in case of suspension */ + const JOCTET * next_input_byte; /* => next byte to read from source */ + size_t bytes_in_buffer; /* # of bytes remaining in source buffer */ + /* Bit input buffer --- note these values are kept in register variables, + * not in this struct, inside the inner loops. + */ + bit_buf_type get_buffer; /* current bit-extraction buffer */ + int bits_left; /* # of unused bits in it */ + /* Pointer needed by jpeg_fill_bit_buffer. */ + j_decompress_ptr cinfo; /* back link to decompress master record */ +} bitread_working_state; + +/* Macros to declare and load/save bitread local variables. */ +#define BITREAD_STATE_VARS \ + register bit_buf_type get_buffer; \ + register int bits_left; \ + bitread_working_state br_state + +#define BITREAD_LOAD_STATE(cinfop,permstate) \ + br_state.cinfo = cinfop; \ + br_state.next_input_byte = cinfop->src->next_input_byte; \ + br_state.bytes_in_buffer = cinfop->src->bytes_in_buffer; \ + get_buffer = permstate.get_buffer; \ + bits_left = permstate.bits_left; + +#define BITREAD_SAVE_STATE(cinfop,permstate) \ + cinfop->src->next_input_byte = br_state.next_input_byte; \ + cinfop->src->bytes_in_buffer = br_state.bytes_in_buffer; \ + permstate.get_buffer = get_buffer; \ + permstate.bits_left = bits_left + +/* + * These macros provide the in-line portion of bit fetching. + * Use CHECK_BIT_BUFFER to ensure there are N bits in get_buffer + * before using GET_BITS, PEEK_BITS, or DROP_BITS. + * The variables get_buffer and bits_left are assumed to be locals, + * but the state struct might not be (jpeg_huff_decode needs this). + * CHECK_BIT_BUFFER(state,n,action); + * Ensure there are N bits in get_buffer; if suspend, take action. + * val = GET_BITS(n); + * Fetch next N bits. + * val = PEEK_BITS(n); + * Fetch next N bits without removing them from the buffer. + * DROP_BITS(n); + * Discard next N bits. + * The value N should be a simple variable, not an expression, because it + * is evaluated multiple times. + */ + +#define CHECK_BIT_BUFFER(state,nbits,action) \ + { if (bits_left < (nbits)) { \ + if (! jpeg_fill_bit_buffer(&(state),get_buffer,bits_left,nbits)) \ + { action; } \ + get_buffer = (state).get_buffer; bits_left = (state).bits_left; } } + +#define GET_BITS(nbits) \ + (((int) (get_buffer >> (bits_left -= (nbits)))) & BIT_MASK(nbits)) + +#define PEEK_BITS(nbits) \ + (((int) (get_buffer >> (bits_left - (nbits)))) & BIT_MASK(nbits)) + +#define DROP_BITS(nbits) \ + (bits_left -= (nbits)) + + +/* + * Code for extracting next Huffman-coded symbol from input bit stream. + * Again, this is time-critical and we make the main paths be macros. + * + * We use a lookahead table to process codes of up to HUFF_LOOKAHEAD bits + * without looping. Usually, more than 95% of the Huffman codes will be 8 + * or fewer bits long. The few overlength codes are handled with a loop, + * which need not be inline code. + * + * Notes about the HUFF_DECODE macro: + * 1. Near the end of the data segment, we may fail to get enough bits + * for a lookahead. In that case, we do it the hard way. + * 2. If the lookahead table contains no entry, the next code must be + * more than HUFF_LOOKAHEAD bits long. + * 3. jpeg_huff_decode returns -1 if forced to suspend. + */ + +#define HUFF_DECODE(result,state,htbl,failaction,slowlabel) \ +{ register int nb, look; \ + if (bits_left < HUFF_LOOKAHEAD) { \ + if (! jpeg_fill_bit_buffer(&state,get_buffer,bits_left, 0)) {failaction;} \ + get_buffer = state.get_buffer; bits_left = state.bits_left; \ + if (bits_left < HUFF_LOOKAHEAD) { \ + nb = 1; goto slowlabel; \ + } \ + } \ + look = PEEK_BITS(HUFF_LOOKAHEAD); \ + if ((nb = htbl->look_nbits[look]) != 0) { \ + DROP_BITS(nb); \ + result = htbl->look_sym[look]; \ + } else { \ + nb = HUFF_LOOKAHEAD+1; \ +slowlabel: \ + if ((result=jpeg_huff_decode(&state,get_buffer,bits_left,htbl,nb)) < 0) \ + { failaction; } \ + get_buffer = state.get_buffer; bits_left = state.bits_left; \ + } \ +} + + +/* + * Expanded entropy decoder object for Huffman decoding. + * + * The savable_state subrecord contains fields that change within an MCU, + * but must not be updated permanently until we complete the MCU. + */ + +typedef struct { + unsigned int EOBRUN; /* remaining EOBs in EOBRUN */ + int last_dc_val[MAX_COMPS_IN_SCAN]; /* last DC coef for each component */ +} savable_state; + +/* This macro is to work around compilers with missing or broken + * structure assignment. You'll need to fix this code if you have + * such a compiler and you change MAX_COMPS_IN_SCAN. + */ + +#ifndef NO_STRUCT_ASSIGN +#define ASSIGN_STATE(dest,src) ((dest) = (src)) +#else +#if MAX_COMPS_IN_SCAN == 4 +#define ASSIGN_STATE(dest,src) \ + ((dest).EOBRUN = (src).EOBRUN, \ + (dest).last_dc_val[0] = (src).last_dc_val[0], \ + (dest).last_dc_val[1] = (src).last_dc_val[1], \ + (dest).last_dc_val[2] = (src).last_dc_val[2], \ + (dest).last_dc_val[3] = (src).last_dc_val[3]) +#endif +#endif + + +typedef struct { + struct jpeg_entropy_decoder pub; /* public fields */ + + /* These fields are loaded into local variables at start of each MCU. + * In case of suspension, we exit WITHOUT updating them. + */ + bitread_perm_state bitstate; /* Bit buffer at start of MCU */ + savable_state saved; /* Other state at start of MCU */ + + /* These fields are NOT loaded into local working state. */ + boolean insufficient_data; /* set TRUE after emitting warning */ + unsigned int restarts_to_go; /* MCUs left in this restart interval */ + + /* Following two fields used only in progressive mode */ + + /* Pointers to derived tables (these workspaces have image lifespan) */ + d_derived_tbl * derived_tbls[NUM_HUFF_TBLS]; + + d_derived_tbl * ac_derived_tbl; /* active table during an AC scan */ + + /* Following fields used only in sequential mode */ + + /* Pointers to derived tables (these workspaces have image lifespan) */ + d_derived_tbl * dc_derived_tbls[NUM_HUFF_TBLS]; + d_derived_tbl * ac_derived_tbls[NUM_HUFF_TBLS]; + + /* Precalculated info set up by start_pass for use in decode_mcu: */ + + /* Pointers to derived tables to be used for each block within an MCU */ + d_derived_tbl * dc_cur_tbls[D_MAX_BLOCKS_IN_MCU]; + d_derived_tbl * ac_cur_tbls[D_MAX_BLOCKS_IN_MCU]; + /* Whether we care about the DC and AC coefficient values for each block */ + int coef_limit[D_MAX_BLOCKS_IN_MCU]; +} huff_entropy_decoder; + +typedef huff_entropy_decoder * huff_entropy_ptr; + + +static const int jpeg_zigzag_order[8][8] = { + { 0, 1, 5, 6, 14, 15, 27, 28 }, + { 2, 4, 7, 13, 16, 26, 29, 42 }, + { 3, 8, 12, 17, 25, 30, 41, 43 }, + { 9, 11, 18, 24, 31, 40, 44, 53 }, + { 10, 19, 23, 32, 39, 45, 52, 54 }, + { 20, 22, 33, 38, 46, 51, 55, 60 }, + { 21, 34, 37, 47, 50, 56, 59, 61 }, + { 35, 36, 48, 49, 57, 58, 62, 63 } +}; + +static const int jpeg_zigzag_order7[7][7] = { + { 0, 1, 5, 6, 14, 15, 27 }, + { 2, 4, 7, 13, 16, 26, 28 }, + { 3, 8, 12, 17, 25, 29, 38 }, + { 9, 11, 18, 24, 30, 37, 39 }, + { 10, 19, 23, 31, 36, 40, 45 }, + { 20, 22, 32, 35, 41, 44, 46 }, + { 21, 33, 34, 42, 43, 47, 48 } +}; + +static const int jpeg_zigzag_order6[6][6] = { + { 0, 1, 5, 6, 14, 15 }, + { 2, 4, 7, 13, 16, 25 }, + { 3, 8, 12, 17, 24, 26 }, + { 9, 11, 18, 23, 27, 32 }, + { 10, 19, 22, 28, 31, 33 }, + { 20, 21, 29, 30, 34, 35 } +}; + +static const int jpeg_zigzag_order5[5][5] = { + { 0, 1, 5, 6, 14 }, + { 2, 4, 7, 13, 15 }, + { 3, 8, 12, 16, 21 }, + { 9, 11, 17, 20, 22 }, + { 10, 18, 19, 23, 24 } +}; + +static const int jpeg_zigzag_order4[4][4] = { + { 0, 1, 5, 6 }, + { 2, 4, 7, 12 }, + { 3, 8, 11, 13 }, + { 9, 10, 14, 15 } +}; + +static const int jpeg_zigzag_order3[3][3] = { + { 0, 1, 5 }, + { 2, 4, 6 }, + { 3, 7, 8 } +}; + +static const int jpeg_zigzag_order2[2][2] = { + { 0, 1 }, + { 2, 3 } +}; + + +/* + * Compute the derived values for a Huffman table. + * This routine also performs some validation checks on the table. + */ + +LOCAL(void) +jpeg_make_d_derived_tbl (j_decompress_ptr cinfo, boolean isDC, int tblno, + d_derived_tbl ** pdtbl) +{ + JHUFF_TBL *htbl; + d_derived_tbl *dtbl; + int p, i, l, si, numsymbols; + int lookbits, ctr; + char huffsize[257]; + unsigned int huffcode[257]; + unsigned int code; + + /* Note that huffsize[] and huffcode[] are filled in code-length order, + * paralleling the order of the symbols themselves in htbl->huffval[]. + */ + + /* Find the input Huffman table */ + if (tblno < 0 || tblno >= NUM_HUFF_TBLS) + ERREXIT1(cinfo, JERR_NO_HUFF_TABLE, tblno); + htbl = + isDC ? cinfo->dc_huff_tbl_ptrs[tblno] : cinfo->ac_huff_tbl_ptrs[tblno]; + if (htbl == NULL) + htbl = jpeg_std_huff_table((j_common_ptr) cinfo, isDC, tblno); + + /* Allocate a workspace if we haven't already done so. */ + if (*pdtbl == NULL) + *pdtbl = (d_derived_tbl *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(d_derived_tbl)); + dtbl = *pdtbl; + dtbl->pub = htbl; /* fill in back link */ + + /* Figure C.1: make table of Huffman code length for each symbol */ + + p = 0; + for (l = 1; l <= 16; l++) { + i = (int) htbl->bits[l]; + if (i < 0 || p + i > 256) /* protect against table overrun */ + ERREXIT(cinfo, JERR_BAD_HUFF_TABLE); + while (i--) + huffsize[p++] = (char) l; + } + huffsize[p] = 0; + numsymbols = p; + + /* Figure C.2: generate the codes themselves */ + /* We also validate that the counts represent a legal Huffman code tree. */ + + code = 0; + si = huffsize[0]; + p = 0; + while (huffsize[p]) { + while (((int) huffsize[p]) == si) { + huffcode[p++] = code; + code++; + } + /* code is now 1 more than the last code used for codelength si; but + * it must still fit in si bits, since no code is allowed to be all ones. + */ + if (((INT32) code) >= (((INT32) 1) << si)) + ERREXIT(cinfo, JERR_BAD_HUFF_TABLE); + code <<= 1; + si++; + } + + /* Figure F.15: generate decoding tables for bit-sequential decoding */ + + p = 0; + for (l = 1; l <= 16; l++) { + if (htbl->bits[l]) { + /* valoffset[l] = huffval[] index of 1st symbol of code length l, + * minus the minimum code of length l + */ + dtbl->valoffset[l] = (INT32) p - (INT32) huffcode[p]; + p += htbl->bits[l]; + dtbl->maxcode[l] = huffcode[p-1]; /* maximum code of length l */ + } else { + dtbl->maxcode[l] = -1; /* -1 if no codes of this length */ + } + } + dtbl->maxcode[17] = 0xFFFFFL; /* ensures jpeg_huff_decode terminates */ + + /* Compute lookahead tables to speed up decoding. + * First we set all the table entries to 0, indicating "too long"; + * then we iterate through the Huffman codes that are short enough and + * fill in all the entries that correspond to bit sequences starting + * with that code. + */ + + MEMZERO(dtbl->look_nbits, SIZEOF(dtbl->look_nbits)); + + p = 0; + for (l = 1; l <= HUFF_LOOKAHEAD; l++) { + for (i = 1; i <= (int) htbl->bits[l]; i++, p++) { + /* l = current code's length, p = its index in huffcode[] & huffval[]. */ + /* Generate left-justified code followed by all possible bit sequences */ + lookbits = huffcode[p] << (HUFF_LOOKAHEAD-l); + for (ctr = 1 << (HUFF_LOOKAHEAD-l); ctr > 0; ctr--) { + dtbl->look_nbits[lookbits] = l; + dtbl->look_sym[lookbits] = htbl->huffval[p]; + lookbits++; + } + } + } + + /* Validate symbols as being reasonable. + * For AC tables, we make no check, but accept all byte values 0..255. + * For DC tables, we require the symbols to be in range 0..15. + * (Tighter bounds could be applied depending on the data depth and mode, + * but this is sufficient to ensure safe decoding.) + */ + if (isDC) { + for (i = 0; i < numsymbols; i++) { + int sym = htbl->huffval[i]; + if (sym < 0 || sym > 15) + ERREXIT(cinfo, JERR_BAD_HUFF_TABLE); + } + } +} + + +/* + * Out-of-line code for bit fetching. + * Note: current values of get_buffer and bits_left are passed as parameters, + * but are returned in the corresponding fields of the state struct. + * + * On most machines MIN_GET_BITS should be 25 to allow the full 32-bit width + * of get_buffer to be used. (On machines with wider words, an even larger + * buffer could be used.) However, on some machines 32-bit shifts are + * quite slow and take time proportional to the number of places shifted. + * (This is true with most PC compilers, for instance.) In this case it may + * be a win to set MIN_GET_BITS to the minimum value of 15. This reduces the + * average shift distance at the cost of more calls to jpeg_fill_bit_buffer. + */ + +#ifdef SLOW_SHIFT_32 +#define MIN_GET_BITS 15 /* minimum allowable value */ +#else +#define MIN_GET_BITS (BIT_BUF_SIZE-7) +#endif + + +LOCAL(boolean) +jpeg_fill_bit_buffer (bitread_working_state * state, + register bit_buf_type get_buffer, register int bits_left, + int nbits) +/* Load up the bit buffer to a depth of at least nbits */ +{ + /* Copy heavily used state fields into locals (hopefully registers) */ + register const JOCTET * next_input_byte = state->next_input_byte; + register size_t bytes_in_buffer = state->bytes_in_buffer; + j_decompress_ptr cinfo = state->cinfo; + + /* Attempt to load at least MIN_GET_BITS bits into get_buffer. */ + /* (It is assumed that no request will be for more than that many bits.) */ + /* We fail to do so only if we hit a marker or are forced to suspend. */ + + if (cinfo->unread_marker == 0) { /* cannot advance past a marker */ + while (bits_left < MIN_GET_BITS) { + register int c; + + /* Attempt to read a byte */ + if (bytes_in_buffer == 0) { + if (! (*cinfo->src->fill_input_buffer) (cinfo)) + return FALSE; + next_input_byte = cinfo->src->next_input_byte; + bytes_in_buffer = cinfo->src->bytes_in_buffer; + } + bytes_in_buffer--; + c = GETJOCTET(*next_input_byte++); + + /* If it's 0xFF, check and discard stuffed zero byte */ + if (c == 0xFF) { + /* Loop here to discard any padding FF's on terminating marker, + * so that we can save a valid unread_marker value. NOTE: we will + * accept multiple FF's followed by a 0 as meaning a single FF data + * byte. This data pattern is not valid according to the standard. + */ + do { + if (bytes_in_buffer == 0) { + if (! (*cinfo->src->fill_input_buffer) (cinfo)) + return FALSE; + next_input_byte = cinfo->src->next_input_byte; + bytes_in_buffer = cinfo->src->bytes_in_buffer; + } + bytes_in_buffer--; + c = GETJOCTET(*next_input_byte++); + } while (c == 0xFF); + + if (c == 0) { + /* Found FF/00, which represents an FF data byte */ + c = 0xFF; + } else { + /* Oops, it's actually a marker indicating end of compressed data. + * Save the marker code for later use. + * Fine point: it might appear that we should save the marker into + * bitread working state, not straight into permanent state. But + * once we have hit a marker, we cannot need to suspend within the + * current MCU, because we will read no more bytes from the data + * source. So it is OK to update permanent state right away. + */ + cinfo->unread_marker = c; + /* See if we need to insert some fake zero bits. */ + goto no_more_bytes; + } + } + + /* OK, load c into get_buffer */ + get_buffer = (get_buffer << 8) | c; + bits_left += 8; + } /* end while */ + } else { + no_more_bytes: + /* We get here if we've read the marker that terminates the compressed + * data segment. There should be enough bits in the buffer register + * to satisfy the request; if so, no problem. + */ + if (nbits > bits_left) { + /* Uh-oh. Report corrupted data to user and stuff zeroes into + * the data stream, so that we can produce some kind of image. + * We use a nonvolatile flag to ensure that only one warning message + * appears per data segment. + */ + if (! ((huff_entropy_ptr) cinfo->entropy)->insufficient_data) { + WARNMS(cinfo, JWRN_HIT_MARKER); + ((huff_entropy_ptr) cinfo->entropy)->insufficient_data = TRUE; + } + /* Fill the buffer with zero bits */ + get_buffer <<= MIN_GET_BITS - bits_left; + bits_left = MIN_GET_BITS; + } + } + + /* Unload the local registers */ + state->next_input_byte = next_input_byte; + state->bytes_in_buffer = bytes_in_buffer; + state->get_buffer = get_buffer; + state->bits_left = bits_left; + + return TRUE; +} + + +/* + * Figure F.12: extend sign bit. + * On some machines, a shift and sub will be faster than a table lookup. + */ + +#ifdef AVOID_TABLES + +#define BIT_MASK(nbits) ((1<<(nbits))-1) +#define HUFF_EXTEND(x,s) ((x) < (1<<((s)-1)) ? (x) - ((1<<(s))-1) : (x)) + +#else + +#define BIT_MASK(nbits) bmask[nbits] +#define HUFF_EXTEND(x,s) ((x) <= bmask[(s) - 1] ? (x) - bmask[s] : (x)) + +static const int bmask[16] = /* bmask[n] is mask for n rightmost bits */ + { 0, 0x0001, 0x0003, 0x0007, 0x000F, 0x001F, 0x003F, 0x007F, 0x00FF, + 0x01FF, 0x03FF, 0x07FF, 0x0FFF, 0x1FFF, 0x3FFF, 0x7FFF }; + +#endif /* AVOID_TABLES */ + + +/* + * Out-of-line code for Huffman code decoding. + */ + +LOCAL(int) +jpeg_huff_decode (bitread_working_state * state, + register bit_buf_type get_buffer, register int bits_left, + d_derived_tbl * htbl, int min_bits) +{ + register int l = min_bits; + register INT32 code; + + /* HUFF_DECODE has determined that the code is at least min_bits */ + /* bits long, so fetch that many bits in one swoop. */ + + CHECK_BIT_BUFFER(*state, l, return -1); + code = GET_BITS(l); + + /* Collect the rest of the Huffman code one bit at a time. */ + /* This is per Figure F.16 in the JPEG spec. */ + + while (code > htbl->maxcode[l]) { + code <<= 1; + CHECK_BIT_BUFFER(*state, 1, return -1); + code |= GET_BITS(1); + l++; + } + + /* Unload the local registers */ + state->get_buffer = get_buffer; + state->bits_left = bits_left; + + /* With garbage input we may reach the sentinel value l = 17. */ + + if (l > 16) { + WARNMS(state->cinfo, JWRN_HUFF_BAD_CODE); + return 0; /* fake a zero as the safest result */ + } + + return htbl->pub->huffval[ (int) (code + htbl->valoffset[l]) ]; +} + + +/* + * Finish up at the end of a Huffman-compressed scan. + */ + +METHODDEF(void) +finish_pass_huff (j_decompress_ptr cinfo) +{ + huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; + + /* Throw away any unused bits remaining in bit buffer; */ + /* include any full bytes in next_marker's count of discarded bytes */ + cinfo->marker->discarded_bytes += entropy->bitstate.bits_left / 8; + entropy->bitstate.bits_left = 0; +} + + +/* + * Check for a restart marker & resynchronize decoder. + * Returns FALSE if must suspend. + */ + +LOCAL(boolean) +process_restart (j_decompress_ptr cinfo) +{ + huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; + int ci; + + finish_pass_huff(cinfo); + + /* Advance past the RSTn marker */ + if (! (*cinfo->marker->read_restart_marker) (cinfo)) + return FALSE; + + /* Re-initialize DC predictions to 0 */ + for (ci = 0; ci < cinfo->comps_in_scan; ci++) + entropy->saved.last_dc_val[ci] = 0; + /* Re-init EOB run count, too */ + entropy->saved.EOBRUN = 0; + + /* Reset restart counter */ + entropy->restarts_to_go = cinfo->restart_interval; + + /* Reset out-of-data flag, unless read_restart_marker left us smack up + * against a marker. In that case we will end up treating the next data + * segment as empty, and we can avoid producing bogus output pixels by + * leaving the flag set. + */ + if (cinfo->unread_marker == 0) + entropy->insufficient_data = FALSE; + + return TRUE; +} + + +/* + * Huffman MCU decoding. + * Each of these routines decodes and returns one MCU's worth of + * Huffman-compressed coefficients. + * The coefficients are reordered from zigzag order into natural array order, + * but are not dequantized. + * + * The i'th block of the MCU is stored into the block pointed to by + * MCU_data[i]. WE ASSUME THIS AREA IS INITIALLY ZEROED BY THE CALLER. + * (Wholesale zeroing is usually a little faster than retail...) + * + * We return FALSE if data source requested suspension. In that case no + * changes have been made to permanent state. (Exception: some output + * coefficients may already have been assigned. This is harmless for + * spectral selection, since we'll just re-assign them on the next call. + * Successive approximation AC refinement has to be more careful, however.) + */ + +/* + * MCU decoding for DC initial scan (either spectral selection, + * or first pass of successive approximation). + */ + +METHODDEF(boolean) +decode_mcu_DC_first (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data) +{ + huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; + int Al = cinfo->Al; + register int s, r; + int blkn, ci; + JBLOCKROW block; + BITREAD_STATE_VARS; + savable_state state; + d_derived_tbl * tbl; + jpeg_component_info * compptr; + + /* Process restart marker if needed; may have to suspend */ + if (cinfo->restart_interval) { + if (entropy->restarts_to_go == 0) + if (! process_restart(cinfo)) + return FALSE; + } + + /* If we've run out of data, just leave the MCU set to zeroes. + * This way, we return uniform gray for the remainder of the segment. + */ + if (! entropy->insufficient_data) { + + /* Load up working state */ + BITREAD_LOAD_STATE(cinfo, entropy->bitstate); + ASSIGN_STATE(state, entropy->saved); + + /* Outer loop handles each block in the MCU */ + + for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { + block = MCU_data[blkn]; + ci = cinfo->MCU_membership[blkn]; + compptr = cinfo->cur_comp_info[ci]; + tbl = entropy->derived_tbls[compptr->dc_tbl_no]; + + /* Decode a single block's worth of coefficients */ + + /* Section F.2.2.1: decode the DC coefficient difference */ + HUFF_DECODE(s, br_state, tbl, return FALSE, label1); + if (s) { + CHECK_BIT_BUFFER(br_state, s, return FALSE); + r = GET_BITS(s); + s = HUFF_EXTEND(r, s); + } + + /* Convert DC difference to actual value, update last_dc_val */ + s += state.last_dc_val[ci]; + state.last_dc_val[ci] = s; + /* Scale and output the coefficient (assumes jpeg_natural_order[0]=0) */ + (*block)[0] = (JCOEF) (s << Al); + } + + /* Completed MCU, so update state */ + BITREAD_SAVE_STATE(cinfo, entropy->bitstate); + ASSIGN_STATE(entropy->saved, state); + } + + /* Account for restart interval if using restarts */ + if (cinfo->restart_interval) + entropy->restarts_to_go--; + + return TRUE; +} + + +/* + * MCU decoding for AC initial scan (either spectral selection, + * or first pass of successive approximation). + */ + +METHODDEF(boolean) +decode_mcu_AC_first (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data) +{ + huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; + register int s, k, r; + unsigned int EOBRUN; + int Se, Al; + const int * natural_order; + JBLOCKROW block; + BITREAD_STATE_VARS; + d_derived_tbl * tbl; + + /* Process restart marker if needed; may have to suspend */ + if (cinfo->restart_interval) { + if (entropy->restarts_to_go == 0) + if (! process_restart(cinfo)) + return FALSE; + } + + /* If we've run out of data, just leave the MCU set to zeroes. + * This way, we return uniform gray for the remainder of the segment. + */ + if (! entropy->insufficient_data) { + + /* Load up working state. + * We can avoid loading/saving bitread state if in an EOB run. + */ + EOBRUN = entropy->saved.EOBRUN; /* only part of saved state we need */ + + /* There is always only one block per MCU */ + + if (EOBRUN) /* if it's a band of zeroes... */ + EOBRUN--; /* ...process it now (we do nothing) */ + else { + BITREAD_LOAD_STATE(cinfo, entropy->bitstate); + Se = cinfo->Se; + Al = cinfo->Al; + natural_order = cinfo->natural_order; + block = MCU_data[0]; + tbl = entropy->ac_derived_tbl; + + for (k = cinfo->Ss; k <= Se; k++) { + HUFF_DECODE(s, br_state, tbl, return FALSE, label2); + r = s >> 4; + s &= 15; + if (s) { + k += r; + CHECK_BIT_BUFFER(br_state, s, return FALSE); + r = GET_BITS(s); + s = HUFF_EXTEND(r, s); + /* Scale and output coefficient in natural (dezigzagged) order */ + (*block)[natural_order[k]] = (JCOEF) (s << Al); + } else { + if (r != 15) { /* EOBr, run length is 2^r + appended bits */ + if (r) { /* EOBr, r > 0 */ + EOBRUN = 1 << r; + CHECK_BIT_BUFFER(br_state, r, return FALSE); + r = GET_BITS(r); + EOBRUN += r; + EOBRUN--; /* this band is processed at this moment */ + } + break; /* force end-of-band */ + } + k += 15; /* ZRL: skip 15 zeroes in band */ + } + } + + BITREAD_SAVE_STATE(cinfo, entropy->bitstate); + } + + /* Completed MCU, so update state */ + entropy->saved.EOBRUN = EOBRUN; /* only part of saved state we need */ + } + + /* Account for restart interval if using restarts */ + if (cinfo->restart_interval) + entropy->restarts_to_go--; + + return TRUE; +} + + +/* + * MCU decoding for DC successive approximation refinement scan. + * Note: we assume such scans can be multi-component, + * although the spec is not very clear on the point. + */ + +METHODDEF(boolean) +decode_mcu_DC_refine (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data) +{ + huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; + JCOEF p1; + int blkn; + BITREAD_STATE_VARS; + + /* Process restart marker if needed; may have to suspend */ + if (cinfo->restart_interval) { + if (entropy->restarts_to_go == 0) + if (! process_restart(cinfo)) + return FALSE; + } + + /* Not worth the cycles to check insufficient_data here, + * since we will not change the data anyway if we read zeroes. + */ + + /* Load up working state */ + BITREAD_LOAD_STATE(cinfo, entropy->bitstate); + + p1 = 1 << cinfo->Al; /* 1 in the bit position being coded */ + + /* Outer loop handles each block in the MCU */ + + for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { + /* Encoded data is simply the next bit of the two's-complement DC value */ + CHECK_BIT_BUFFER(br_state, 1, return FALSE); + if (GET_BITS(1)) + MCU_data[blkn][0][0] |= p1; + /* Note: since we use |=, repeating the assignment later is safe */ + } + + /* Completed MCU, so update state */ + BITREAD_SAVE_STATE(cinfo, entropy->bitstate); + + /* Account for restart interval if using restarts */ + if (cinfo->restart_interval) + entropy->restarts_to_go--; + + return TRUE; +} + + +/* + * MCU decoding for AC successive approximation refinement scan. + */ + +METHODDEF(boolean) +decode_mcu_AC_refine (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data) +{ + huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; + register int s, k, r; + unsigned int EOBRUN; + int Se; + JCOEF p1, m1; + const int * natural_order; + JBLOCKROW block; + JCOEFPTR thiscoef; + BITREAD_STATE_VARS; + d_derived_tbl * tbl; + int num_newnz; + int newnz_pos[DCTSIZE2]; + + /* Process restart marker if needed; may have to suspend */ + if (cinfo->restart_interval) { + if (entropy->restarts_to_go == 0) + if (! process_restart(cinfo)) + return FALSE; + } + + /* If we've run out of data, don't modify the MCU. + */ + if (! entropy->insufficient_data) { + + Se = cinfo->Se; + p1 = 1 << cinfo->Al; /* 1 in the bit position being coded */ + m1 = -p1; /* -1 in the bit position being coded */ + natural_order = cinfo->natural_order; + + /* Load up working state */ + BITREAD_LOAD_STATE(cinfo, entropy->bitstate); + EOBRUN = entropy->saved.EOBRUN; /* only part of saved state we need */ + + /* There is always only one block per MCU */ + block = MCU_data[0]; + tbl = entropy->ac_derived_tbl; + + /* If we are forced to suspend, we must undo the assignments to any newly + * nonzero coefficients in the block, because otherwise we'd get confused + * next time about which coefficients were already nonzero. + * But we need not undo addition of bits to already-nonzero coefficients; + * instead, we can test the current bit to see if we already did it. + */ + num_newnz = 0; + + /* initialize coefficient loop counter to start of band */ + k = cinfo->Ss; + + if (EOBRUN == 0) { + do { + HUFF_DECODE(s, br_state, tbl, goto undoit, label3); + r = s >> 4; + s &= 15; + if (s) { + if (s != 1) /* size of new coef should always be 1 */ + WARNMS(cinfo, JWRN_HUFF_BAD_CODE); + CHECK_BIT_BUFFER(br_state, 1, goto undoit); + if (GET_BITS(1)) + s = p1; /* newly nonzero coef is positive */ + else + s = m1; /* newly nonzero coef is negative */ + } else { + if (r != 15) { + EOBRUN = 1 << r; /* EOBr, run length is 2^r + appended bits */ + if (r) { + CHECK_BIT_BUFFER(br_state, r, goto undoit); + r = GET_BITS(r); + EOBRUN += r; + } + break; /* rest of block is handled by EOB logic */ + } + /* note s = 0 for processing ZRL */ + } + /* Advance over already-nonzero coefs and r still-zero coefs, + * appending correction bits to the nonzeroes. A correction bit is 1 + * if the absolute value of the coefficient must be increased. + */ + do { + thiscoef = *block + natural_order[k]; + if (*thiscoef) { + CHECK_BIT_BUFFER(br_state, 1, goto undoit); + if (GET_BITS(1)) { + if ((*thiscoef & p1) == 0) { /* do nothing if already set it */ + if (*thiscoef >= 0) + *thiscoef += p1; + else + *thiscoef += m1; + } + } + } else { + if (--r < 0) + break; /* reached target zero coefficient */ + } + k++; + } while (k <= Se); + if (s) { + int pos = natural_order[k]; + /* Output newly nonzero coefficient */ + (*block)[pos] = (JCOEF) s; + /* Remember its position in case we have to suspend */ + newnz_pos[num_newnz++] = pos; + } + k++; + } while (k <= Se); + } + + if (EOBRUN) { + /* Scan any remaining coefficient positions after the end-of-band + * (the last newly nonzero coefficient, if any). Append a correction + * bit to each already-nonzero coefficient. A correction bit is 1 + * if the absolute value of the coefficient must be increased. + */ + do { + thiscoef = *block + natural_order[k]; + if (*thiscoef) { + CHECK_BIT_BUFFER(br_state, 1, goto undoit); + if (GET_BITS(1)) { + if ((*thiscoef & p1) == 0) { /* do nothing if already changed it */ + if (*thiscoef >= 0) + *thiscoef += p1; + else + *thiscoef += m1; + } + } + } + k++; + } while (k <= Se); + /* Count one block completed in EOB run */ + EOBRUN--; + } + + /* Completed MCU, so update state */ + BITREAD_SAVE_STATE(cinfo, entropy->bitstate); + entropy->saved.EOBRUN = EOBRUN; /* only part of saved state we need */ + } + + /* Account for restart interval if using restarts */ + if (cinfo->restart_interval) + entropy->restarts_to_go--; + + return TRUE; + +undoit: + /* Re-zero any output coefficients that we made newly nonzero */ + while (num_newnz) + (*block)[newnz_pos[--num_newnz]] = 0; + + return FALSE; +} + + +/* + * Decode one MCU's worth of Huffman-compressed coefficients, + * partial blocks. + */ + +METHODDEF(boolean) +decode_mcu_sub (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data) +{ + huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; + const int * natural_order; + int Se, blkn; + BITREAD_STATE_VARS; + savable_state state; + + /* Process restart marker if needed; may have to suspend */ + if (cinfo->restart_interval) { + if (entropy->restarts_to_go == 0) + if (! process_restart(cinfo)) + return FALSE; + } + + /* If we've run out of data, just leave the MCU set to zeroes. + * This way, we return uniform gray for the remainder of the segment. + */ + if (! entropy->insufficient_data) { + + natural_order = cinfo->natural_order; + Se = cinfo->lim_Se; + + /* Load up working state */ + BITREAD_LOAD_STATE(cinfo, entropy->bitstate); + ASSIGN_STATE(state, entropy->saved); + + /* Outer loop handles each block in the MCU */ + + for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { + JBLOCKROW block = MCU_data[blkn]; + d_derived_tbl * htbl; + register int s, k, r; + int coef_limit, ci; + + /* Decode a single block's worth of coefficients */ + + /* Section F.2.2.1: decode the DC coefficient difference */ + htbl = entropy->dc_cur_tbls[blkn]; + HUFF_DECODE(s, br_state, htbl, return FALSE, label1); + + htbl = entropy->ac_cur_tbls[blkn]; + k = 1; + coef_limit = entropy->coef_limit[blkn]; + if (coef_limit) { + /* Convert DC difference to actual value, update last_dc_val */ + if (s) { + CHECK_BIT_BUFFER(br_state, s, return FALSE); + r = GET_BITS(s); + s = HUFF_EXTEND(r, s); + } + ci = cinfo->MCU_membership[blkn]; + s += state.last_dc_val[ci]; + state.last_dc_val[ci] = s; + /* Output the DC coefficient */ + (*block)[0] = (JCOEF) s; + + /* Section F.2.2.2: decode the AC coefficients */ + /* Since zeroes are skipped, output area must be cleared beforehand */ + for (; k < coef_limit; k++) { + HUFF_DECODE(s, br_state, htbl, return FALSE, label2); + + r = s >> 4; + s &= 15; + + if (s) { + k += r; + CHECK_BIT_BUFFER(br_state, s, return FALSE); + r = GET_BITS(s); + s = HUFF_EXTEND(r, s); + /* Output coefficient in natural (dezigzagged) order. + * Note: the extra entries in natural_order[] will save us + * if k > Se, which could happen if the data is corrupted. + */ + (*block)[natural_order[k]] = (JCOEF) s; + } else { + if (r != 15) + goto EndOfBlock; + k += 15; + } + } + } else { + if (s) { + CHECK_BIT_BUFFER(br_state, s, return FALSE); + DROP_BITS(s); + } + } + + /* Section F.2.2.2: decode the AC coefficients */ + /* In this path we just discard the values */ + for (; k <= Se; k++) { + HUFF_DECODE(s, br_state, htbl, return FALSE, label3); + + r = s >> 4; + s &= 15; + + if (s) { + k += r; + CHECK_BIT_BUFFER(br_state, s, return FALSE); + DROP_BITS(s); + } else { + if (r != 15) + break; + k += 15; + } + } + + EndOfBlock: ; + } + + /* Completed MCU, so update state */ + BITREAD_SAVE_STATE(cinfo, entropy->bitstate); + ASSIGN_STATE(entropy->saved, state); + } + + /* Account for restart interval if using restarts */ + if (cinfo->restart_interval) + entropy->restarts_to_go--; + + return TRUE; +} + + +/* + * Decode one MCU's worth of Huffman-compressed coefficients, + * full-size blocks. + */ + +METHODDEF(boolean) +decode_mcu (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data) +{ + huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; + int blkn; + BITREAD_STATE_VARS; + savable_state state; + + /* Process restart marker if needed; may have to suspend */ + if (cinfo->restart_interval) { + if (entropy->restarts_to_go == 0) + if (! process_restart(cinfo)) + return FALSE; + } + + /* If we've run out of data, just leave the MCU set to zeroes. + * This way, we return uniform gray for the remainder of the segment. + */ + if (! entropy->insufficient_data) { + + /* Load up working state */ + BITREAD_LOAD_STATE(cinfo, entropy->bitstate); + ASSIGN_STATE(state, entropy->saved); + + /* Outer loop handles each block in the MCU */ + + for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { + JBLOCKROW block = MCU_data[blkn]; + d_derived_tbl * htbl; + register int s, k, r; + int coef_limit, ci; + + /* Decode a single block's worth of coefficients */ + + /* Section F.2.2.1: decode the DC coefficient difference */ + htbl = entropy->dc_cur_tbls[blkn]; + HUFF_DECODE(s, br_state, htbl, return FALSE, label1); + + htbl = entropy->ac_cur_tbls[blkn]; + k = 1; + coef_limit = entropy->coef_limit[blkn]; + if (coef_limit) { + /* Convert DC difference to actual value, update last_dc_val */ + if (s) { + CHECK_BIT_BUFFER(br_state, s, return FALSE); + r = GET_BITS(s); + s = HUFF_EXTEND(r, s); + } + ci = cinfo->MCU_membership[blkn]; + s += state.last_dc_val[ci]; + state.last_dc_val[ci] = s; + /* Output the DC coefficient */ + (*block)[0] = (JCOEF) s; + + /* Section F.2.2.2: decode the AC coefficients */ + /* Since zeroes are skipped, output area must be cleared beforehand */ + for (; k < coef_limit; k++) { + HUFF_DECODE(s, br_state, htbl, return FALSE, label2); + + r = s >> 4; + s &= 15; + + if (s) { + k += r; + CHECK_BIT_BUFFER(br_state, s, return FALSE); + r = GET_BITS(s); + s = HUFF_EXTEND(r, s); + /* Output coefficient in natural (dezigzagged) order. + * Note: the extra entries in jpeg_natural_order[] will save us + * if k >= DCTSIZE2, which could happen if the data is corrupted. + */ + (*block)[jpeg_natural_order[k]] = (JCOEF) s; + } else { + if (r != 15) + goto EndOfBlock; + k += 15; + } + } + } else { + if (s) { + CHECK_BIT_BUFFER(br_state, s, return FALSE); + DROP_BITS(s); + } + } + + /* Section F.2.2.2: decode the AC coefficients */ + /* In this path we just discard the values */ + for (; k < DCTSIZE2; k++) { + HUFF_DECODE(s, br_state, htbl, return FALSE, label3); + + r = s >> 4; + s &= 15; + + if (s) { + k += r; + CHECK_BIT_BUFFER(br_state, s, return FALSE); + DROP_BITS(s); + } else { + if (r != 15) + break; + k += 15; + } + } + + EndOfBlock: ; + } + + /* Completed MCU, so update state */ + BITREAD_SAVE_STATE(cinfo, entropy->bitstate); + ASSIGN_STATE(entropy->saved, state); + } + + /* Account for restart interval if using restarts */ + if (cinfo->restart_interval) + entropy->restarts_to_go--; + + return TRUE; +} + + +/* + * Initialize for a Huffman-compressed scan. + */ + +METHODDEF(void) +start_pass_huff_decoder (j_decompress_ptr cinfo) +{ + huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; + int ci, blkn, tbl, i; + jpeg_component_info * compptr; + + if (cinfo->progressive_mode) { + /* Validate progressive scan parameters */ + if (cinfo->Ss == 0) { + if (cinfo->Se != 0) + goto bad; + } else { + /* need not check Ss/Se < 0 since they came from unsigned bytes */ + if (cinfo->Se < cinfo->Ss || cinfo->Se > cinfo->lim_Se) + goto bad; + /* AC scans may have only one component */ + if (cinfo->comps_in_scan != 1) + goto bad; + } + if (cinfo->Ah != 0) { + /* Successive approximation refinement scan: must have Al = Ah-1. */ + if (cinfo->Ah-1 != cinfo->Al) + goto bad; + } + if (cinfo->Al > 13) { /* need not check for < 0 */ + /* Arguably the maximum Al value should be less than 13 for 8-bit + * precision, but the spec doesn't say so, and we try to be liberal + * about what we accept. Note: large Al values could result in + * out-of-range DC coefficients during early scans, leading to bizarre + * displays due to overflows in the IDCT math. But we won't crash. + */ + bad: + ERREXIT4(cinfo, JERR_BAD_PROGRESSION, + cinfo->Ss, cinfo->Se, cinfo->Ah, cinfo->Al); + } + /* Update progression status, and verify that scan order is legal. + * Note that inter-scan inconsistencies are treated as warnings + * not fatal errors ... not clear if this is right way to behave. + */ + for (ci = 0; ci < cinfo->comps_in_scan; ci++) { + int coefi, cindex = cinfo->cur_comp_info[ci]->component_index; + int *coef_bit_ptr = & cinfo->coef_bits[cindex][0]; + if (cinfo->Ss && coef_bit_ptr[0] < 0) /* AC without prior DC scan */ + WARNMS2(cinfo, JWRN_BOGUS_PROGRESSION, cindex, 0); + for (coefi = cinfo->Ss; coefi <= cinfo->Se; coefi++) { + int expected = (coef_bit_ptr[coefi] < 0) ? 0 : coef_bit_ptr[coefi]; + if (cinfo->Ah != expected) + WARNMS2(cinfo, JWRN_BOGUS_PROGRESSION, cindex, coefi); + coef_bit_ptr[coefi] = cinfo->Al; + } + } + + /* Select MCU decoding routine */ + if (cinfo->Ah == 0) { + if (cinfo->Ss == 0) + entropy->pub.decode_mcu = decode_mcu_DC_first; + else + entropy->pub.decode_mcu = decode_mcu_AC_first; + } else { + if (cinfo->Ss == 0) + entropy->pub.decode_mcu = decode_mcu_DC_refine; + else + entropy->pub.decode_mcu = decode_mcu_AC_refine; + } + + for (ci = 0; ci < cinfo->comps_in_scan; ci++) { + compptr = cinfo->cur_comp_info[ci]; + /* Make sure requested tables are present, and compute derived tables. + * We may build same derived table more than once, but it's not expensive. + */ + if (cinfo->Ss == 0) { + if (cinfo->Ah == 0) { /* DC refinement needs no table */ + tbl = compptr->dc_tbl_no; + jpeg_make_d_derived_tbl(cinfo, TRUE, tbl, + & entropy->derived_tbls[tbl]); + } + } else { + tbl = compptr->ac_tbl_no; + jpeg_make_d_derived_tbl(cinfo, FALSE, tbl, + & entropy->derived_tbls[tbl]); + /* remember the single active table */ + entropy->ac_derived_tbl = entropy->derived_tbls[tbl]; + } + /* Initialize DC predictions to 0 */ + entropy->saved.last_dc_val[ci] = 0; + } + + /* Initialize private state variables */ + entropy->saved.EOBRUN = 0; + } else { + /* Check that the scan parameters Ss, Se, Ah/Al are OK for sequential JPEG. + * This ought to be an error condition, but we make it a warning because + * there are some baseline files out there with all zeroes in these bytes. + */ + if (cinfo->Ss != 0 || cinfo->Ah != 0 || cinfo->Al != 0 || + ((cinfo->is_baseline || cinfo->Se < DCTSIZE2) && + cinfo->Se != cinfo->lim_Se)) + WARNMS(cinfo, JWRN_NOT_SEQUENTIAL); + + /* Select MCU decoding routine */ + /* We retain the hard-coded case for full-size blocks. + * This is not necessary, but it appears that this version is slightly + * more performant in the given implementation. + * With an improved implementation we would prefer a single optimized + * function. + */ + if (cinfo->lim_Se != DCTSIZE2-1) + entropy->pub.decode_mcu = decode_mcu_sub; + else + entropy->pub.decode_mcu = decode_mcu; + + for (ci = 0; ci < cinfo->comps_in_scan; ci++) { + compptr = cinfo->cur_comp_info[ci]; + /* Compute derived values for Huffman tables */ + /* We may do this more than once for a table, but it's not expensive */ + tbl = compptr->dc_tbl_no; + jpeg_make_d_derived_tbl(cinfo, TRUE, tbl, + & entropy->dc_derived_tbls[tbl]); + if (cinfo->lim_Se) { /* AC needs no table when not present */ + tbl = compptr->ac_tbl_no; + jpeg_make_d_derived_tbl(cinfo, FALSE, tbl, + & entropy->ac_derived_tbls[tbl]); + } + /* Initialize DC predictions to 0 */ + entropy->saved.last_dc_val[ci] = 0; + } + + /* Precalculate decoding info for each block in an MCU of this scan */ + for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { + ci = cinfo->MCU_membership[blkn]; + compptr = cinfo->cur_comp_info[ci]; + /* Precalculate which table to use for each block */ + entropy->dc_cur_tbls[blkn] = entropy->dc_derived_tbls[compptr->dc_tbl_no]; + entropy->ac_cur_tbls[blkn] = /* AC needs no table when not present */ + cinfo->lim_Se ? entropy->ac_derived_tbls[compptr->ac_tbl_no] : NULL; + /* Decide whether we really care about the coefficient values */ + if (compptr->component_needed) { + ci = compptr->DCT_v_scaled_size; + i = compptr->DCT_h_scaled_size; + switch (cinfo->lim_Se) { + case (1*1-1): + entropy->coef_limit[blkn] = 1; + break; + case (2*2-1): + if (ci <= 0 || ci > 2) ci = 2; + if (i <= 0 || i > 2) i = 2; + entropy->coef_limit[blkn] = 1 + jpeg_zigzag_order2[ci - 1][i - 1]; + break; + case (3*3-1): + if (ci <= 0 || ci > 3) ci = 3; + if (i <= 0 || i > 3) i = 3; + entropy->coef_limit[blkn] = 1 + jpeg_zigzag_order3[ci - 1][i - 1]; + break; + case (4*4-1): + if (ci <= 0 || ci > 4) ci = 4; + if (i <= 0 || i > 4) i = 4; + entropy->coef_limit[blkn] = 1 + jpeg_zigzag_order4[ci - 1][i - 1]; + break; + case (5*5-1): + if (ci <= 0 || ci > 5) ci = 5; + if (i <= 0 || i > 5) i = 5; + entropy->coef_limit[blkn] = 1 + jpeg_zigzag_order5[ci - 1][i - 1]; + break; + case (6*6-1): + if (ci <= 0 || ci > 6) ci = 6; + if (i <= 0 || i > 6) i = 6; + entropy->coef_limit[blkn] = 1 + jpeg_zigzag_order6[ci - 1][i - 1]; + break; + case (7*7-1): + if (ci <= 0 || ci > 7) ci = 7; + if (i <= 0 || i > 7) i = 7; + entropy->coef_limit[blkn] = 1 + jpeg_zigzag_order7[ci - 1][i - 1]; + break; + default: + if (ci <= 0 || ci > 8) ci = 8; + if (i <= 0 || i > 8) i = 8; + entropy->coef_limit[blkn] = 1 + jpeg_zigzag_order[ci - 1][i - 1]; + } + } else { + entropy->coef_limit[blkn] = 0; + } + } + } + + /* Initialize bitread state variables */ + entropy->bitstate.bits_left = 0; + entropy->bitstate.get_buffer = 0; /* unnecessary, but keeps Purify quiet */ + entropy->insufficient_data = FALSE; + + /* Initialize restart counter */ + entropy->restarts_to_go = cinfo->restart_interval; +} + + +/* + * Module initialization routine for Huffman entropy decoding. + */ + +GLOBAL(void) +jinit_huff_decoder (j_decompress_ptr cinfo) +{ + huff_entropy_ptr entropy; + int i; + + entropy = (huff_entropy_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(huff_entropy_decoder)); + cinfo->entropy = &entropy->pub; + entropy->pub.start_pass = start_pass_huff_decoder; + entropy->pub.finish_pass = finish_pass_huff; + + if (cinfo->progressive_mode) { + /* Create progression status table */ + int *coef_bit_ptr, ci; + cinfo->coef_bits = (int (*)[DCTSIZE2]) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, + cinfo->num_components * DCTSIZE2 * SIZEOF(int)); + coef_bit_ptr = & cinfo->coef_bits[0][0]; + for (ci = 0; ci < cinfo->num_components; ci++) + for (i = 0; i < DCTSIZE2; i++) + *coef_bit_ptr++ = -1; + + /* Mark derived tables unallocated */ + for (i = 0; i < NUM_HUFF_TBLS; i++) { + entropy->derived_tbls[i] = NULL; + } + } else { + /* Mark derived tables unallocated */ + for (i = 0; i < NUM_HUFF_TBLS; i++) { + entropy->dc_derived_tbls[i] = entropy->ac_derived_tbls[i] = NULL; + } + } +} diff --git a/cmake/externals/libjpeg/jdinput.c b/cmake/externals/libjpeg/jdinput.c new file mode 100644 index 000000000..29fbef90b --- /dev/null +++ b/cmake/externals/libjpeg/jdinput.c @@ -0,0 +1,657 @@ +/* + * jdinput.c + * + * Copyright (C) 1991-1997, Thomas G. Lane. + * Modified 2002-2020 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains input control logic for the JPEG decompressor. + * These routines are concerned with controlling the decompressor's input + * processing (marker reading and coefficient decoding). The actual input + * reading is done in jdmarker.c, jdhuff.c, and jdarith.c. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* Private state */ + +typedef struct { + struct jpeg_input_controller pub; /* public fields */ + + int inheaders; /* Nonzero until first SOS is reached */ +} my_input_controller; + +typedef my_input_controller * my_inputctl_ptr; + + +/* Forward declarations */ +METHODDEF(int) consume_markers JPP((j_decompress_ptr cinfo)); + + +/* + * Routines to calculate various quantities related to the size of the image. + */ + + +/* + * Compute output image dimensions and related values. + * NOTE: this is exported for possible use by application. + * Hence it mustn't do anything that can't be done twice. + */ + +GLOBAL(void) +jpeg_core_output_dimensions (j_decompress_ptr cinfo) +/* Do computations that are needed before master selection phase. + * This function is used for transcoding and full decompression. + */ +{ +#ifdef IDCT_SCALING_SUPPORTED + int ci; + jpeg_component_info *compptr; + + /* Compute actual output image dimensions and DCT scaling choices. */ + if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom) { + /* Provide 1/block_size scaling */ + cinfo->output_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width, (long) cinfo->block_size); + cinfo->output_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height, (long) cinfo->block_size); + cinfo->min_DCT_h_scaled_size = 1; + cinfo->min_DCT_v_scaled_size = 1; + } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 2) { + /* Provide 2/block_size scaling */ + cinfo->output_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * 2L, (long) cinfo->block_size); + cinfo->output_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * 2L, (long) cinfo->block_size); + cinfo->min_DCT_h_scaled_size = 2; + cinfo->min_DCT_v_scaled_size = 2; + } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 3) { + /* Provide 3/block_size scaling */ + cinfo->output_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * 3L, (long) cinfo->block_size); + cinfo->output_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * 3L, (long) cinfo->block_size); + cinfo->min_DCT_h_scaled_size = 3; + cinfo->min_DCT_v_scaled_size = 3; + } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 4) { + /* Provide 4/block_size scaling */ + cinfo->output_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * 4L, (long) cinfo->block_size); + cinfo->output_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * 4L, (long) cinfo->block_size); + cinfo->min_DCT_h_scaled_size = 4; + cinfo->min_DCT_v_scaled_size = 4; + } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 5) { + /* Provide 5/block_size scaling */ + cinfo->output_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * 5L, (long) cinfo->block_size); + cinfo->output_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * 5L, (long) cinfo->block_size); + cinfo->min_DCT_h_scaled_size = 5; + cinfo->min_DCT_v_scaled_size = 5; + } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 6) { + /* Provide 6/block_size scaling */ + cinfo->output_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * 6L, (long) cinfo->block_size); + cinfo->output_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * 6L, (long) cinfo->block_size); + cinfo->min_DCT_h_scaled_size = 6; + cinfo->min_DCT_v_scaled_size = 6; + } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 7) { + /* Provide 7/block_size scaling */ + cinfo->output_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * 7L, (long) cinfo->block_size); + cinfo->output_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * 7L, (long) cinfo->block_size); + cinfo->min_DCT_h_scaled_size = 7; + cinfo->min_DCT_v_scaled_size = 7; + } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 8) { + /* Provide 8/block_size scaling */ + cinfo->output_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * 8L, (long) cinfo->block_size); + cinfo->output_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * 8L, (long) cinfo->block_size); + cinfo->min_DCT_h_scaled_size = 8; + cinfo->min_DCT_v_scaled_size = 8; + } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 9) { + /* Provide 9/block_size scaling */ + cinfo->output_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * 9L, (long) cinfo->block_size); + cinfo->output_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * 9L, (long) cinfo->block_size); + cinfo->min_DCT_h_scaled_size = 9; + cinfo->min_DCT_v_scaled_size = 9; + } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 10) { + /* Provide 10/block_size scaling */ + cinfo->output_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * 10L, (long) cinfo->block_size); + cinfo->output_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * 10L, (long) cinfo->block_size); + cinfo->min_DCT_h_scaled_size = 10; + cinfo->min_DCT_v_scaled_size = 10; + } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 11) { + /* Provide 11/block_size scaling */ + cinfo->output_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * 11L, (long) cinfo->block_size); + cinfo->output_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * 11L, (long) cinfo->block_size); + cinfo->min_DCT_h_scaled_size = 11; + cinfo->min_DCT_v_scaled_size = 11; + } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 12) { + /* Provide 12/block_size scaling */ + cinfo->output_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * 12L, (long) cinfo->block_size); + cinfo->output_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * 12L, (long) cinfo->block_size); + cinfo->min_DCT_h_scaled_size = 12; + cinfo->min_DCT_v_scaled_size = 12; + } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 13) { + /* Provide 13/block_size scaling */ + cinfo->output_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * 13L, (long) cinfo->block_size); + cinfo->output_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * 13L, (long) cinfo->block_size); + cinfo->min_DCT_h_scaled_size = 13; + cinfo->min_DCT_v_scaled_size = 13; + } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 14) { + /* Provide 14/block_size scaling */ + cinfo->output_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * 14L, (long) cinfo->block_size); + cinfo->output_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * 14L, (long) cinfo->block_size); + cinfo->min_DCT_h_scaled_size = 14; + cinfo->min_DCT_v_scaled_size = 14; + } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 15) { + /* Provide 15/block_size scaling */ + cinfo->output_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * 15L, (long) cinfo->block_size); + cinfo->output_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * 15L, (long) cinfo->block_size); + cinfo->min_DCT_h_scaled_size = 15; + cinfo->min_DCT_v_scaled_size = 15; + } else { + /* Provide 16/block_size scaling */ + cinfo->output_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * 16L, (long) cinfo->block_size); + cinfo->output_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * 16L, (long) cinfo->block_size); + cinfo->min_DCT_h_scaled_size = 16; + cinfo->min_DCT_v_scaled_size = 16; + } + + /* Recompute dimensions of components */ + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + compptr->DCT_h_scaled_size = cinfo->min_DCT_h_scaled_size; + compptr->DCT_v_scaled_size = cinfo->min_DCT_v_scaled_size; + } + +#else /* !IDCT_SCALING_SUPPORTED */ + + /* Hardwire it to "no scaling" */ + cinfo->output_width = cinfo->image_width; + cinfo->output_height = cinfo->image_height; + /* initial_setup has already initialized DCT_scaled_size, + * and has computed unscaled downsampled_width and downsampled_height. + */ + +#endif /* IDCT_SCALING_SUPPORTED */ +} + + +LOCAL(void) +initial_setup (j_decompress_ptr cinfo) +/* Called once, when first SOS marker is reached */ +{ + int ci; + jpeg_component_info *compptr; + + /* Make sure image isn't bigger than I can handle */ + if ((long) cinfo->image_height > (long) JPEG_MAX_DIMENSION || + (long) cinfo->image_width > (long) JPEG_MAX_DIMENSION) + ERREXIT1(cinfo, JERR_IMAGE_TOO_BIG, (unsigned int) JPEG_MAX_DIMENSION); + + /* Only 8 to 12 bits data precision are supported for DCT based JPEG */ + if (cinfo->data_precision < 8 || cinfo->data_precision > 12) + ERREXIT1(cinfo, JERR_BAD_PRECISION, cinfo->data_precision); + + /* Check that number of components won't exceed internal array sizes */ + if (cinfo->num_components > MAX_COMPONENTS) + ERREXIT2(cinfo, JERR_COMPONENT_COUNT, cinfo->num_components, + MAX_COMPONENTS); + + /* Compute maximum sampling factors; check factor validity */ + cinfo->max_h_samp_factor = 1; + cinfo->max_v_samp_factor = 1; + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + if (compptr->h_samp_factor<=0 || compptr->h_samp_factor>MAX_SAMP_FACTOR || + compptr->v_samp_factor<=0 || compptr->v_samp_factor>MAX_SAMP_FACTOR) + ERREXIT(cinfo, JERR_BAD_SAMPLING); + cinfo->max_h_samp_factor = MAX(cinfo->max_h_samp_factor, + compptr->h_samp_factor); + cinfo->max_v_samp_factor = MAX(cinfo->max_v_samp_factor, + compptr->v_samp_factor); + } + + /* Derive block_size, natural_order, and lim_Se */ + if (cinfo->is_baseline || (cinfo->progressive_mode && + cinfo->comps_in_scan)) { /* no pseudo SOS marker */ + cinfo->block_size = DCTSIZE; + cinfo->natural_order = jpeg_natural_order; + cinfo->lim_Se = DCTSIZE2-1; + } else + switch (cinfo->Se) { + case (1*1-1): + cinfo->block_size = 1; + cinfo->natural_order = jpeg_natural_order; /* not needed */ + cinfo->lim_Se = cinfo->Se; + break; + case (2*2-1): + cinfo->block_size = 2; + cinfo->natural_order = jpeg_natural_order2; + cinfo->lim_Se = cinfo->Se; + break; + case (3*3-1): + cinfo->block_size = 3; + cinfo->natural_order = jpeg_natural_order3; + cinfo->lim_Se = cinfo->Se; + break; + case (4*4-1): + cinfo->block_size = 4; + cinfo->natural_order = jpeg_natural_order4; + cinfo->lim_Se = cinfo->Se; + break; + case (5*5-1): + cinfo->block_size = 5; + cinfo->natural_order = jpeg_natural_order5; + cinfo->lim_Se = cinfo->Se; + break; + case (6*6-1): + cinfo->block_size = 6; + cinfo->natural_order = jpeg_natural_order6; + cinfo->lim_Se = cinfo->Se; + break; + case (7*7-1): + cinfo->block_size = 7; + cinfo->natural_order = jpeg_natural_order7; + cinfo->lim_Se = cinfo->Se; + break; + case (8*8-1): + cinfo->block_size = 8; + cinfo->natural_order = jpeg_natural_order; + cinfo->lim_Se = DCTSIZE2-1; + break; + case (9*9-1): + cinfo->block_size = 9; + cinfo->natural_order = jpeg_natural_order; + cinfo->lim_Se = DCTSIZE2-1; + break; + case (10*10-1): + cinfo->block_size = 10; + cinfo->natural_order = jpeg_natural_order; + cinfo->lim_Se = DCTSIZE2-1; + break; + case (11*11-1): + cinfo->block_size = 11; + cinfo->natural_order = jpeg_natural_order; + cinfo->lim_Se = DCTSIZE2-1; + break; + case (12*12-1): + cinfo->block_size = 12; + cinfo->natural_order = jpeg_natural_order; + cinfo->lim_Se = DCTSIZE2-1; + break; + case (13*13-1): + cinfo->block_size = 13; + cinfo->natural_order = jpeg_natural_order; + cinfo->lim_Se = DCTSIZE2-1; + break; + case (14*14-1): + cinfo->block_size = 14; + cinfo->natural_order = jpeg_natural_order; + cinfo->lim_Se = DCTSIZE2-1; + break; + case (15*15-1): + cinfo->block_size = 15; + cinfo->natural_order = jpeg_natural_order; + cinfo->lim_Se = DCTSIZE2-1; + break; + case (16*16-1): + cinfo->block_size = 16; + cinfo->natural_order = jpeg_natural_order; + cinfo->lim_Se = DCTSIZE2-1; + break; + default: + ERREXIT4(cinfo, JERR_BAD_PROGRESSION, + cinfo->Ss, cinfo->Se, cinfo->Ah, cinfo->Al); + } + + /* We initialize DCT_scaled_size and min_DCT_scaled_size to block_size. + * In the full decompressor, + * this will be overridden by jpeg_calc_output_dimensions in jdmaster.c; + * but in the transcoder, + * jpeg_calc_output_dimensions is not used, so we must do it here. + */ + cinfo->min_DCT_h_scaled_size = cinfo->block_size; + cinfo->min_DCT_v_scaled_size = cinfo->block_size; + + /* Compute dimensions of components */ + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + compptr->DCT_h_scaled_size = cinfo->block_size; + compptr->DCT_v_scaled_size = cinfo->block_size; + /* Size in DCT blocks */ + compptr->width_in_blocks = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * (long) compptr->h_samp_factor, + (long) (cinfo->max_h_samp_factor * cinfo->block_size)); + compptr->height_in_blocks = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * (long) compptr->v_samp_factor, + (long) (cinfo->max_v_samp_factor * cinfo->block_size)); + /* downsampled_width and downsampled_height will also be overridden by + * jdmaster.c if we are doing full decompression. The transcoder library + * doesn't use these values, but the calling application might. + */ + /* Size in samples */ + compptr->downsampled_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * (long) compptr->h_samp_factor, + (long) cinfo->max_h_samp_factor); + compptr->downsampled_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * (long) compptr->v_samp_factor, + (long) cinfo->max_v_samp_factor); + /* Mark component needed, until color conversion says otherwise */ + compptr->component_needed = TRUE; + /* Mark no quantization table yet saved for component */ + compptr->quant_table = NULL; + } + + /* Compute number of fully interleaved MCU rows. */ + cinfo->total_iMCU_rows = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height, + (long) (cinfo->max_v_samp_factor * cinfo->block_size)); + + /* Decide whether file contains multiple scans */ + if (cinfo->comps_in_scan < cinfo->num_components || cinfo->progressive_mode) + cinfo->inputctl->has_multiple_scans = TRUE; + else + cinfo->inputctl->has_multiple_scans = FALSE; +} + + +LOCAL(void) +per_scan_setup (j_decompress_ptr cinfo) +/* Do computations that are needed before processing a JPEG scan */ +/* cinfo->comps_in_scan and cinfo->cur_comp_info[] were set from SOS marker */ +{ + int ci, mcublks, tmp; + jpeg_component_info *compptr; + + if (cinfo->comps_in_scan == 1) { + + /* Noninterleaved (single-component) scan */ + compptr = cinfo->cur_comp_info[0]; + + /* Overall image size in MCUs */ + cinfo->MCUs_per_row = compptr->width_in_blocks; + cinfo->MCU_rows_in_scan = compptr->height_in_blocks; + + /* For noninterleaved scan, always one block per MCU */ + compptr->MCU_width = 1; + compptr->MCU_height = 1; + compptr->MCU_blocks = 1; + compptr->MCU_sample_width = compptr->DCT_h_scaled_size; + compptr->last_col_width = 1; + /* For noninterleaved scans, it is convenient to define last_row_height + * as the number of block rows present in the last iMCU row. + */ + tmp = (int) (compptr->height_in_blocks % compptr->v_samp_factor); + if (tmp == 0) tmp = compptr->v_samp_factor; + compptr->last_row_height = tmp; + + /* Prepare array describing MCU composition */ + cinfo->blocks_in_MCU = 1; + cinfo->MCU_membership[0] = 0; + + } else { + + /* Interleaved (multi-component) scan */ + if (cinfo->comps_in_scan <= 0 || cinfo->comps_in_scan > MAX_COMPS_IN_SCAN) + ERREXIT2(cinfo, JERR_COMPONENT_COUNT, cinfo->comps_in_scan, + MAX_COMPS_IN_SCAN); + + /* Overall image size in MCUs */ + cinfo->MCUs_per_row = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width, + (long) (cinfo->max_h_samp_factor * cinfo->block_size)); + cinfo->MCU_rows_in_scan = cinfo->total_iMCU_rows; + + cinfo->blocks_in_MCU = 0; + + for (ci = 0; ci < cinfo->comps_in_scan; ci++) { + compptr = cinfo->cur_comp_info[ci]; + /* Sampling factors give # of blocks of component in each MCU */ + compptr->MCU_width = compptr->h_samp_factor; + compptr->MCU_height = compptr->v_samp_factor; + compptr->MCU_blocks = compptr->MCU_width * compptr->MCU_height; + compptr->MCU_sample_width = compptr->MCU_width * compptr->DCT_h_scaled_size; + /* Figure number of non-dummy blocks in last MCU column & row */ + tmp = (int) (compptr->width_in_blocks % compptr->MCU_width); + if (tmp == 0) tmp = compptr->MCU_width; + compptr->last_col_width = tmp; + tmp = (int) (compptr->height_in_blocks % compptr->MCU_height); + if (tmp == 0) tmp = compptr->MCU_height; + compptr->last_row_height = tmp; + /* Prepare array describing MCU composition */ + mcublks = compptr->MCU_blocks; + if (cinfo->blocks_in_MCU + mcublks > D_MAX_BLOCKS_IN_MCU) + ERREXIT(cinfo, JERR_BAD_MCU_SIZE); + while (mcublks-- > 0) { + cinfo->MCU_membership[cinfo->blocks_in_MCU++] = ci; + } + } + + } +} + + +/* + * Save away a copy of the Q-table referenced by each component present + * in the current scan, unless already saved during a prior scan. + * + * In a multiple-scan JPEG file, the encoder could assign different components + * the same Q-table slot number, but change table definitions between scans + * so that each component uses a different Q-table. (The IJG encoder is not + * currently capable of doing this, but other encoders might.) Since we want + * to be able to dequantize all the components at the end of the file, this + * means that we have to save away the table actually used for each component. + * We do this by copying the table at the start of the first scan containing + * the component. + * The JPEG spec prohibits the encoder from changing the contents of a Q-table + * slot between scans of a component using that slot. If the encoder does so + * anyway, this decoder will simply use the Q-table values that were current + * at the start of the first scan for the component. + * + * The decompressor output side looks only at the saved quant tables, + * not at the current Q-table slots. + */ + +LOCAL(void) +latch_quant_tables (j_decompress_ptr cinfo) +{ + int ci, qtblno; + jpeg_component_info *compptr; + JQUANT_TBL * qtbl; + + for (ci = 0; ci < cinfo->comps_in_scan; ci++) { + compptr = cinfo->cur_comp_info[ci]; + /* No work if we already saved Q-table for this component */ + if (compptr->quant_table != NULL) + continue; + /* Make sure specified quantization table is present */ + qtblno = compptr->quant_tbl_no; + if (qtblno < 0 || qtblno >= NUM_QUANT_TBLS || + cinfo->quant_tbl_ptrs[qtblno] == NULL) + ERREXIT1(cinfo, JERR_NO_QUANT_TABLE, qtblno); + /* OK, save away the quantization table */ + qtbl = (JQUANT_TBL *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(JQUANT_TBL)); + MEMCOPY(qtbl, cinfo->quant_tbl_ptrs[qtblno], SIZEOF(JQUANT_TBL)); + compptr->quant_table = qtbl; + } +} + + +/* + * Initialize the input modules to read a scan of compressed data. + * The first call to this is done by jdmaster.c after initializing + * the entire decompressor (during jpeg_start_decompress). + * Subsequent calls come from consume_markers, below. + */ + +METHODDEF(void) +start_input_pass (j_decompress_ptr cinfo) +{ + per_scan_setup(cinfo); + latch_quant_tables(cinfo); + (*cinfo->entropy->start_pass) (cinfo); + (*cinfo->coef->start_input_pass) (cinfo); + cinfo->inputctl->consume_input = cinfo->coef->consume_data; +} + + +/* + * Finish up after inputting a compressed-data scan. + * This is called by the coefficient controller after it's read all + * the expected data of the scan. + */ + +METHODDEF(void) +finish_input_pass (j_decompress_ptr cinfo) +{ + (*cinfo->entropy->finish_pass) (cinfo); + cinfo->inputctl->consume_input = consume_markers; +} + + +/* + * Read JPEG markers before, between, or after compressed-data scans. + * Change state as necessary when a new scan is reached. + * Return value is JPEG_SUSPENDED, JPEG_REACHED_SOS, or JPEG_REACHED_EOI. + * + * The consume_input method pointer points either here or to the + * coefficient controller's consume_data routine, depending on whether + * we are reading a compressed data segment or inter-segment markers. + * + * Note: This function should NOT return a pseudo SOS marker (with zero + * component number) to the caller. A pseudo marker received by + * read_markers is processed and then skipped for other markers. + */ + +METHODDEF(int) +consume_markers (j_decompress_ptr cinfo) +{ + my_inputctl_ptr inputctl = (my_inputctl_ptr) cinfo->inputctl; + int val; + + if (inputctl->pub.eoi_reached) /* After hitting EOI, read no further */ + return JPEG_REACHED_EOI; + + for (;;) { /* Loop to pass pseudo SOS marker */ + val = (*cinfo->marker->read_markers) (cinfo); + + switch (val) { + case JPEG_REACHED_SOS: /* Found SOS */ + if (inputctl->inheaders) { /* 1st SOS */ + if (inputctl->inheaders == 1) + initial_setup(cinfo); + if (cinfo->comps_in_scan == 0) { /* pseudo SOS marker */ + inputctl->inheaders = 2; + break; + } + inputctl->inheaders = 0; + /* Note: start_input_pass must be called by jdmaster.c + * before any more input can be consumed. jdapimin.c is + * responsible for enforcing this sequencing. + */ + } else { /* 2nd or later SOS marker */ + if (! inputctl->pub.has_multiple_scans) + ERREXIT(cinfo, JERR_EOI_EXPECTED); /* Oops, I wasn't expecting this! */ + if (cinfo->comps_in_scan == 0) /* unexpected pseudo SOS marker */ + break; + start_input_pass(cinfo); + } + return val; + case JPEG_REACHED_EOI: /* Found EOI */ + inputctl->pub.eoi_reached = TRUE; + if (inputctl->inheaders) { /* Tables-only datastream, apparently */ + if (cinfo->marker->saw_SOF) + ERREXIT(cinfo, JERR_SOF_NO_SOS); + } else { + /* Prevent infinite loop in coef ctlr's decompress_data routine + * if user set output_scan_number larger than number of scans. + */ + if (cinfo->output_scan_number > cinfo->input_scan_number) + cinfo->output_scan_number = cinfo->input_scan_number; + } + return val; + case JPEG_SUSPENDED: + return val; + default: + return val; + } + } +} + + +/* + * Reset state to begin a fresh datastream. + */ + +METHODDEF(void) +reset_input_controller (j_decompress_ptr cinfo) +{ + my_inputctl_ptr inputctl = (my_inputctl_ptr) cinfo->inputctl; + + inputctl->pub.consume_input = consume_markers; + inputctl->pub.has_multiple_scans = FALSE; /* "unknown" would be better */ + inputctl->pub.eoi_reached = FALSE; + inputctl->inheaders = 1; + /* Reset other modules */ + (*cinfo->err->reset_error_mgr) ((j_common_ptr) cinfo); + (*cinfo->marker->reset_marker_reader) (cinfo); + /* Reset progression state -- would be cleaner if entropy decoder did this */ + cinfo->coef_bits = NULL; +} + + +/* + * Initialize the input controller module. + * This is called only once, when the decompression object is created. + */ + +GLOBAL(void) +jinit_input_controller (j_decompress_ptr cinfo) +{ + my_inputctl_ptr inputctl; + + /* Create subobject in permanent pool */ + inputctl = (my_inputctl_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_PERMANENT, SIZEOF(my_input_controller)); + cinfo->inputctl = &inputctl->pub; + /* Initialize method pointers */ + inputctl->pub.consume_input = consume_markers; + inputctl->pub.reset_input_controller = reset_input_controller; + inputctl->pub.start_input_pass = start_input_pass; + inputctl->pub.finish_input_pass = finish_input_pass; + /* Initialize state: can't use reset_input_controller since we don't + * want to try to reset other modules yet. + */ + inputctl->pub.has_multiple_scans = FALSE; /* "unknown" would be better */ + inputctl->pub.eoi_reached = FALSE; + inputctl->inheaders = 1; +} diff --git a/cmake/externals/libjpeg/jdmainct.c b/cmake/externals/libjpeg/jdmainct.c new file mode 100644 index 000000000..1cd66d853 --- /dev/null +++ b/cmake/externals/libjpeg/jdmainct.c @@ -0,0 +1,511 @@ +/* + * jdmainct.c + * + * Copyright (C) 1994-1996, Thomas G. Lane. + * Modified 2002-2020 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains the main buffer controller for decompression. + * The main buffer lies between the JPEG decompressor proper and the + * post-processor; it holds downsampled data in the JPEG colorspace. + * + * Note that this code is bypassed in raw-data mode, since the application + * supplies the equivalent of the main buffer in that case. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* + * In the current system design, the main buffer need never be a full-image + * buffer; any full-height buffers will be found inside the coefficient or + * postprocessing controllers. Nonetheless, the main controller is not + * trivial. Its responsibility is to provide context rows for upsampling/ + * rescaling, and doing this in an efficient fashion is a bit tricky. + * + * Postprocessor input data is counted in "row groups". A row group is + * defined to be (v_samp_factor * DCT_v_scaled_size / min_DCT_v_scaled_size) + * sample rows of each component. (We require DCT_scaled_size values to be + * chosen such that these numbers are integers. In practice DCT_scaled_size + * values will likely be powers of two, so we actually have the stronger + * condition that DCT_scaled_size / min_DCT_scaled_size is an integer.) + * Upsampling will typically produce max_v_samp_factor pixel rows from each + * row group (times any additional scale factor that the upsampler is + * applying). + * + * The coefficient controller will deliver data to us one iMCU row at a time; + * each iMCU row contains v_samp_factor * DCT_v_scaled_size sample rows, or + * exactly min_DCT_v_scaled_size row groups. (This amount of data corresponds + * to one row of MCUs when the image is fully interleaved.) Note that the + * number of sample rows varies across components, but the number of row + * groups does not. Some garbage sample rows may be included in the last iMCU + * row at the bottom of the image. + * + * Depending on the vertical scaling algorithm used, the upsampler may need + * access to the sample row(s) above and below its current input row group. + * The upsampler is required to set need_context_rows TRUE at global selection + * time if so. When need_context_rows is FALSE, this controller can simply + * obtain one iMCU row at a time from the coefficient controller and dole it + * out as row groups to the postprocessor. + * + * When need_context_rows is TRUE, this controller guarantees that the buffer + * passed to postprocessing contains at least one row group's worth of samples + * above and below the row group(s) being processed. Note that the context + * rows "above" the first passed row group appear at negative row offsets in + * the passed buffer. At the top and bottom of the image, the required + * context rows are manufactured by duplicating the first or last real sample + * row; this avoids having special cases in the upsampling inner loops. + * + * The amount of context is fixed at one row group just because that's a + * convenient number for this controller to work with. The existing + * upsamplers really only need one sample row of context. An upsampler + * supporting arbitrary output rescaling might wish for more than one row + * group of context when shrinking the image; tough, we don't handle that. + * (This is justified by the assumption that downsizing will be handled mostly + * by adjusting the DCT_scaled_size values, so that the actual scale factor at + * the upsample step needn't be much less than one.) + * + * To provide the desired context, we have to retain the last two row groups + * of one iMCU row while reading in the next iMCU row. (The last row group + * can't be processed until we have another row group for its below-context, + * and so we have to save the next-to-last group too for its above-context.) + * We could do this most simply by copying data around in our buffer, but + * that'd be very slow. We can avoid copying any data by creating a rather + * strange pointer structure. Here's how it works. We allocate a workspace + * consisting of M+2 row groups (where M = min_DCT_v_scaled_size is the number + * of row groups per iMCU row). We create two sets of redundant pointers to + * the workspace. Labeling the physical row groups 0 to M+1, the synthesized + * pointer lists look like this: + * M+1 M-1 + * master pointer --> 0 master pointer --> 0 + * 1 1 + * ... ... + * M-3 M-3 + * M-2 M + * M-1 M+1 + * M M-2 + * M+1 M-1 + * 0 0 + * We read alternate iMCU rows using each master pointer; thus the last two + * row groups of the previous iMCU row remain un-overwritten in the workspace. + * The pointer lists are set up so that the required context rows appear to + * be adjacent to the proper places when we pass the pointer lists to the + * upsampler. + * + * The above pictures describe the normal state of the pointer lists. + * At top and bottom of the image, we diddle the pointer lists to duplicate + * the first or last sample row as necessary (this is cheaper than copying + * sample rows around). + * + * This scheme breaks down if M < 2, ie, min_DCT_v_scaled_size is 1. In that + * situation each iMCU row provides only one row group so the buffering logic + * must be different (eg, we must read two iMCU rows before we can emit the + * first row group). For now, we simply do not support providing context + * rows when min_DCT_v_scaled_size is 1. That combination seems unlikely to + * be worth providing --- if someone wants a 1/8th-size preview, they probably + * want it quick and dirty, so a context-free upsampler is sufficient. + */ + + +/* Private buffer controller object */ + +typedef struct { + struct jpeg_d_main_controller pub; /* public fields */ + + /* Pointer to allocated workspace (M or M+2 row groups). */ + JSAMPARRAY buffer[MAX_COMPONENTS]; + + JDIMENSION rowgroup_ctr; /* counts row groups output to postprocessor */ + JDIMENSION rowgroups_avail; /* row groups available to postprocessor */ + + /* Remaining fields are only used in the context case. */ + + boolean buffer_full; /* Have we gotten an iMCU row from decoder? */ + + /* These are the master pointers to the funny-order pointer lists. */ + JSAMPIMAGE xbuffer[2]; /* pointers to weird pointer lists */ + + int whichptr; /* indicates which pointer set is now in use */ + int context_state; /* process_data state machine status */ + JDIMENSION iMCU_row_ctr; /* counts iMCU rows to detect image top/bot */ +} my_main_controller; + +typedef my_main_controller * my_main_ptr; + +/* context_state values: */ +#define CTX_PREPARE_FOR_IMCU 0 /* need to prepare for MCU row */ +#define CTX_PROCESS_IMCU 1 /* feeding iMCU to postprocessor */ +#define CTX_POSTPONED_ROW 2 /* feeding postponed row group */ + + +/* Forward declarations */ +METHODDEF(void) process_data_simple_main + JPP((j_decompress_ptr cinfo, JSAMPARRAY output_buf, + JDIMENSION *out_row_ctr, JDIMENSION out_rows_avail)); +METHODDEF(void) process_data_context_main + JPP((j_decompress_ptr cinfo, JSAMPARRAY output_buf, + JDIMENSION *out_row_ctr, JDIMENSION out_rows_avail)); +#ifdef QUANT_2PASS_SUPPORTED +METHODDEF(void) process_data_crank_post + JPP((j_decompress_ptr cinfo, JSAMPARRAY output_buf, + JDIMENSION *out_row_ctr, JDIMENSION out_rows_avail)); +#endif + + +LOCAL(void) +alloc_funny_pointers (j_decompress_ptr cinfo) +/* Allocate space for the funny pointer lists. + * This is done only once, not once per pass. + */ +{ + my_main_ptr mainp = (my_main_ptr) cinfo->main; + int ci, rgroup; + int M = cinfo->min_DCT_v_scaled_size; + jpeg_component_info *compptr; + JSAMPARRAY xbuf; + + /* Get top-level space for component array pointers. + * We alloc both arrays with one call to save a few cycles. + */ + mainp->xbuffer[0] = (JSAMPIMAGE) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, + cinfo->num_components * 2 * SIZEOF(JSAMPARRAY)); + mainp->xbuffer[1] = mainp->xbuffer[0] + cinfo->num_components; + + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + if (! compptr->component_needed) + continue; /* skip uninteresting component */ + rgroup = (compptr->v_samp_factor * compptr->DCT_v_scaled_size) / + cinfo->min_DCT_v_scaled_size; /* height of a row group of component */ + /* Get space for pointer lists --- M+4 row groups in each list. + * We alloc both pointer lists with one call to save a few cycles. + */ + xbuf = (JSAMPARRAY) (*cinfo->mem->alloc_small) ((j_common_ptr) cinfo, + JPOOL_IMAGE, 2 * (rgroup * (M + 4)) * SIZEOF(JSAMPROW)); + xbuf += rgroup; /* want one row group at negative offsets */ + mainp->xbuffer[0][ci] = xbuf; + xbuf += rgroup * (M + 4); + mainp->xbuffer[1][ci] = xbuf; + } +} + + +LOCAL(void) +make_funny_pointers (j_decompress_ptr cinfo) +/* Create the funny pointer lists discussed in the comments above. + * The actual workspace is already allocated (in mainp->buffer), + * and the space for the pointer lists is allocated too. + * This routine just fills in the curiously ordered lists. + * This will be repeated at the beginning of each pass. + */ +{ + my_main_ptr mainp = (my_main_ptr) cinfo->main; + int ci, i, rgroup; + int M = cinfo->min_DCT_v_scaled_size; + jpeg_component_info *compptr; + JSAMPARRAY buf, xbuf0, xbuf1; + + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + if (! compptr->component_needed) + continue; /* skip uninteresting component */ + rgroup = (compptr->v_samp_factor * compptr->DCT_v_scaled_size) / + cinfo->min_DCT_v_scaled_size; /* height of a row group of component */ + xbuf0 = mainp->xbuffer[0][ci]; + xbuf1 = mainp->xbuffer[1][ci]; + /* First copy the workspace pointers as-is */ + buf = mainp->buffer[ci]; + for (i = 0; i < rgroup * (M + 2); i++) { + xbuf0[i] = xbuf1[i] = buf[i]; + } + /* In the second list, put the last four row groups in swapped order */ + for (i = 0; i < rgroup * 2; i++) { + xbuf1[rgroup*(M-2) + i] = buf[rgroup*M + i]; + xbuf1[rgroup*M + i] = buf[rgroup*(M-2) + i]; + } + /* The wraparound pointers at top and bottom will be filled later + * (see set_wraparound_pointers, below). Initially we want the "above" + * pointers to duplicate the first actual data line. This only needs + * to happen in xbuffer[0]. + */ + for (i = 0; i < rgroup; i++) { + xbuf0[i - rgroup] = xbuf0[0]; + } + } +} + + +LOCAL(void) +set_wraparound_pointers (j_decompress_ptr cinfo) +/* Set up the "wraparound" pointers at top and bottom of the pointer lists. + * This changes the pointer list state from top-of-image to the normal state. + */ +{ + my_main_ptr mainp = (my_main_ptr) cinfo->main; + int ci, i, rgroup; + int M = cinfo->min_DCT_v_scaled_size; + jpeg_component_info *compptr; + JSAMPARRAY xbuf0, xbuf1; + + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + if (! compptr->component_needed) + continue; /* skip uninteresting component */ + rgroup = (compptr->v_samp_factor * compptr->DCT_v_scaled_size) / + cinfo->min_DCT_v_scaled_size; /* height of a row group of component */ + xbuf0 = mainp->xbuffer[0][ci]; + xbuf1 = mainp->xbuffer[1][ci]; + for (i = 0; i < rgroup; i++) { + xbuf0[i - rgroup] = xbuf0[rgroup*(M+1) + i]; + xbuf1[i - rgroup] = xbuf1[rgroup*(M+1) + i]; + xbuf0[rgroup*(M+2) + i] = xbuf0[i]; + xbuf1[rgroup*(M+2) + i] = xbuf1[i]; + } + } +} + + +LOCAL(void) +set_bottom_pointers (j_decompress_ptr cinfo) +/* Change the pointer lists to duplicate the last sample row at the bottom + * of the image. whichptr indicates which xbuffer holds the final iMCU row. + * Also sets rowgroups_avail to indicate number of nondummy row groups in row. + */ +{ + my_main_ptr mainp = (my_main_ptr) cinfo->main; + int ci, i, rgroup, iMCUheight, rows_left; + jpeg_component_info *compptr; + JSAMPARRAY xbuf; + + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + if (! compptr->component_needed) + continue; /* skip uninteresting component */ + /* Count sample rows in one iMCU row and in one row group */ + iMCUheight = compptr->v_samp_factor * compptr->DCT_v_scaled_size; + rgroup = iMCUheight / cinfo->min_DCT_v_scaled_size; + /* Count nondummy sample rows remaining for this component */ + rows_left = (int) (compptr->downsampled_height % (JDIMENSION) iMCUheight); + if (rows_left == 0) rows_left = iMCUheight; + /* Count nondummy row groups. Should get same answer for each component, + * so we need only do it once. + */ + if (ci == 0) { + mainp->rowgroups_avail = (JDIMENSION) ((rows_left-1) / rgroup + 1); + } + /* Duplicate the last real sample row rgroup*2 times; this pads out the + * last partial rowgroup and ensures at least one full rowgroup of context. + */ + xbuf = mainp->xbuffer[mainp->whichptr][ci]; + for (i = 0; i < rgroup * 2; i++) { + xbuf[rows_left + i] = xbuf[rows_left-1]; + } + } +} + + +/* + * Initialize for a processing pass. + */ + +METHODDEF(void) +start_pass_main (j_decompress_ptr cinfo, J_BUF_MODE pass_mode) +{ + my_main_ptr mainp = (my_main_ptr) cinfo->main; + + switch (pass_mode) { + case JBUF_PASS_THRU: + if (cinfo->upsample->need_context_rows) { + mainp->pub.process_data = process_data_context_main; + make_funny_pointers(cinfo); /* Create the xbuffer[] lists */ + mainp->whichptr = 0; /* Read first iMCU row into xbuffer[0] */ + mainp->context_state = CTX_PREPARE_FOR_IMCU; + mainp->iMCU_row_ctr = 0; + mainp->buffer_full = FALSE; /* Mark buffer empty */ + } else { + /* Simple case with no context needed */ + mainp->pub.process_data = process_data_simple_main; + mainp->rowgroup_ctr = mainp->rowgroups_avail; /* Mark buffer empty */ + } + break; +#ifdef QUANT_2PASS_SUPPORTED + case JBUF_CRANK_DEST: + /* For last pass of 2-pass quantization, just crank the postprocessor */ + mainp->pub.process_data = process_data_crank_post; + break; +#endif + default: + ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); + } +} + + +/* + * Process some data. + * This handles the simple case where no context is required. + */ + +METHODDEF(void) +process_data_simple_main (j_decompress_ptr cinfo, JSAMPARRAY output_buf, + JDIMENSION *out_row_ctr, JDIMENSION out_rows_avail) +{ + my_main_ptr mainp = (my_main_ptr) cinfo->main; + + /* Read input data if we haven't filled the main buffer yet */ + if (mainp->rowgroup_ctr >= mainp->rowgroups_avail) { + if (! (*cinfo->coef->decompress_data) (cinfo, mainp->buffer)) + return; /* suspension forced, can do nothing more */ + mainp->rowgroup_ctr = 0; /* OK, we have an iMCU row to work with */ + } + + /* Note: at the bottom of the image, we may pass extra garbage row groups + * to the postprocessor. The postprocessor has to check for bottom + * of image anyway (at row resolution), so no point in us doing it too. + */ + + /* Feed the postprocessor */ + (*cinfo->post->post_process_data) (cinfo, mainp->buffer, + &mainp->rowgroup_ctr, mainp->rowgroups_avail, + output_buf, out_row_ctr, out_rows_avail); +} + + +/* + * Process some data. + * This handles the case where context rows must be provided. + */ + +METHODDEF(void) +process_data_context_main (j_decompress_ptr cinfo, JSAMPARRAY output_buf, + JDIMENSION *out_row_ctr, JDIMENSION out_rows_avail) +{ + my_main_ptr mainp = (my_main_ptr) cinfo->main; + + /* Read input data if we haven't filled the main buffer yet */ + if (! mainp->buffer_full) { + if (! (*cinfo->coef->decompress_data) (cinfo, + mainp->xbuffer[mainp->whichptr])) + return; /* suspension forced, can do nothing more */ + mainp->buffer_full = TRUE; /* OK, we have an iMCU row to work with */ + mainp->iMCU_row_ctr++; /* count rows received */ + } + + /* Postprocessor typically will not swallow all the input data it is handed + * in one call (due to filling the output buffer first). Must be prepared + * to exit and restart. This switch lets us keep track of how far we got. + * Note that each case falls through to the next on successful completion. + */ + switch (mainp->context_state) { + case CTX_POSTPONED_ROW: + /* Call postprocessor using previously set pointers for postponed row */ + (*cinfo->post->post_process_data) (cinfo, mainp->xbuffer[mainp->whichptr], + &mainp->rowgroup_ctr, mainp->rowgroups_avail, + output_buf, out_row_ctr, out_rows_avail); + if (mainp->rowgroup_ctr < mainp->rowgroups_avail) + return; /* Need to suspend */ + mainp->context_state = CTX_PREPARE_FOR_IMCU; + if (*out_row_ctr >= out_rows_avail) + return; /* Postprocessor exactly filled output buf */ + /*FALLTHROUGH*/ + case CTX_PREPARE_FOR_IMCU: + /* Prepare to process first M-1 row groups of this iMCU row */ + mainp->rowgroup_ctr = 0; + mainp->rowgroups_avail = (JDIMENSION) (cinfo->min_DCT_v_scaled_size - 1); + /* Check for bottom of image: if so, tweak pointers to "duplicate" + * the last sample row, and adjust rowgroups_avail to ignore padding rows. + */ + if (mainp->iMCU_row_ctr == cinfo->total_iMCU_rows) + set_bottom_pointers(cinfo); + mainp->context_state = CTX_PROCESS_IMCU; + /*FALLTHROUGH*/ + case CTX_PROCESS_IMCU: + /* Call postprocessor using previously set pointers */ + (*cinfo->post->post_process_data) (cinfo, mainp->xbuffer[mainp->whichptr], + &mainp->rowgroup_ctr, mainp->rowgroups_avail, + output_buf, out_row_ctr, out_rows_avail); + if (mainp->rowgroup_ctr < mainp->rowgroups_avail) + return; /* Need to suspend */ + /* After the first iMCU, change wraparound pointers to normal state */ + if (mainp->iMCU_row_ctr == 1) + set_wraparound_pointers(cinfo); + /* Prepare to load new iMCU row using other xbuffer list */ + mainp->whichptr ^= 1; /* 0=>1 or 1=>0 */ + mainp->buffer_full = FALSE; + /* Still need to process last row group of this iMCU row, */ + /* which is saved at index M+1 of the other xbuffer */ + mainp->rowgroup_ctr = (JDIMENSION) (cinfo->min_DCT_v_scaled_size + 1); + mainp->rowgroups_avail = (JDIMENSION) (cinfo->min_DCT_v_scaled_size + 2); + mainp->context_state = CTX_POSTPONED_ROW; + } +} + + +/* + * Process some data. + * Final pass of two-pass quantization: just call the postprocessor. + * Source data will be the postprocessor controller's internal buffer. + */ + +#ifdef QUANT_2PASS_SUPPORTED + +METHODDEF(void) +process_data_crank_post (j_decompress_ptr cinfo, JSAMPARRAY output_buf, + JDIMENSION *out_row_ctr, JDIMENSION out_rows_avail) +{ + (*cinfo->post->post_process_data) (cinfo, (JSAMPIMAGE) NULL, + (JDIMENSION *) NULL, (JDIMENSION) 0, + output_buf, out_row_ctr, out_rows_avail); +} + +#endif /* QUANT_2PASS_SUPPORTED */ + + +/* + * Initialize main buffer controller. + */ + +GLOBAL(void) +jinit_d_main_controller (j_decompress_ptr cinfo, boolean need_full_buffer) +{ + my_main_ptr mainp; + int ci, rgroup, ngroups; + jpeg_component_info *compptr; + + mainp = (my_main_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_main_controller)); + cinfo->main = &mainp->pub; + mainp->pub.start_pass = start_pass_main; + + if (need_full_buffer) /* shouldn't happen */ + ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); + + /* Allocate the workspace. + * ngroups is the number of row groups we need. + */ + if (cinfo->upsample->need_context_rows) { + if (cinfo->min_DCT_v_scaled_size < 2) /* unsupported, see comments above */ + ERREXIT(cinfo, JERR_NOTIMPL); + alloc_funny_pointers(cinfo); /* Alloc space for xbuffer[] lists */ + ngroups = cinfo->min_DCT_v_scaled_size + 2; + } else { + /* There are always min_DCT_v_scaled_size row groups in an iMCU row. */ + ngroups = cinfo->min_DCT_v_scaled_size; + mainp->rowgroups_avail = (JDIMENSION) ngroups; + } + + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + if (! compptr->component_needed) + continue; /* skip uninteresting component */ + rgroup = (compptr->v_samp_factor * compptr->DCT_v_scaled_size) / + cinfo->min_DCT_v_scaled_size; /* height of a row group of component */ + mainp->buffer[ci] = (*cinfo->mem->alloc_sarray) + ((j_common_ptr) cinfo, JPOOL_IMAGE, + compptr->width_in_blocks * ((JDIMENSION) compptr->DCT_h_scaled_size), + (JDIMENSION) (rgroup * ngroups)); + } +} diff --git a/cmake/externals/libjpeg/jdmarker.c b/cmake/externals/libjpeg/jdmarker.c new file mode 100644 index 000000000..c10fde605 --- /dev/null +++ b/cmake/externals/libjpeg/jdmarker.c @@ -0,0 +1,1505 @@ +/* + * jdmarker.c + * + * Copyright (C) 1991-1998, Thomas G. Lane. + * Modified 2009-2019 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains routines to decode JPEG datastream markers. + * Most of the complexity arises from our desire to support input + * suspension: if not all of the data for a marker is available, + * we must exit back to the application. On resumption, we reprocess + * the marker. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +typedef enum { /* JPEG marker codes */ + M_SOF0 = 0xc0, + M_SOF1 = 0xc1, + M_SOF2 = 0xc2, + M_SOF3 = 0xc3, + + M_SOF5 = 0xc5, + M_SOF6 = 0xc6, + M_SOF7 = 0xc7, + + M_JPG = 0xc8, + M_SOF9 = 0xc9, + M_SOF10 = 0xca, + M_SOF11 = 0xcb, + + M_SOF13 = 0xcd, + M_SOF14 = 0xce, + M_SOF15 = 0xcf, + + M_DHT = 0xc4, + + M_DAC = 0xcc, + + M_RST0 = 0xd0, + M_RST1 = 0xd1, + M_RST2 = 0xd2, + M_RST3 = 0xd3, + M_RST4 = 0xd4, + M_RST5 = 0xd5, + M_RST6 = 0xd6, + M_RST7 = 0xd7, + + M_SOI = 0xd8, + M_EOI = 0xd9, + M_SOS = 0xda, + M_DQT = 0xdb, + M_DNL = 0xdc, + M_DRI = 0xdd, + M_DHP = 0xde, + M_EXP = 0xdf, + + M_APP0 = 0xe0, + M_APP1 = 0xe1, + M_APP2 = 0xe2, + M_APP3 = 0xe3, + M_APP4 = 0xe4, + M_APP5 = 0xe5, + M_APP6 = 0xe6, + M_APP7 = 0xe7, + M_APP8 = 0xe8, + M_APP9 = 0xe9, + M_APP10 = 0xea, + M_APP11 = 0xeb, + M_APP12 = 0xec, + M_APP13 = 0xed, + M_APP14 = 0xee, + M_APP15 = 0xef, + + M_JPG0 = 0xf0, + M_JPG8 = 0xf8, + M_JPG13 = 0xfd, + M_COM = 0xfe, + + M_TEM = 0x01, + + M_ERROR = 0x100 +} JPEG_MARKER; + + +/* Private state */ + +typedef struct { + struct jpeg_marker_reader pub; /* public fields */ + + /* Application-overridable marker processing methods */ + jpeg_marker_parser_method process_COM; + jpeg_marker_parser_method process_APPn[16]; + + /* Limit on marker data length to save for each marker type */ + unsigned int length_limit_COM; + unsigned int length_limit_APPn[16]; + + /* Status of COM/APPn marker saving */ + jpeg_saved_marker_ptr cur_marker; /* NULL if not processing a marker */ + unsigned int bytes_read; /* data bytes read so far in marker */ + /* Note: cur_marker is not linked into marker_list until it's all read. */ +} my_marker_reader; + +typedef my_marker_reader * my_marker_ptr; + + +/* + * Macros for fetching data from the data source module. + * + * At all times, cinfo->src->next_input_byte and ->bytes_in_buffer reflect + * the current restart point; we update them only when we have reached a + * suitable place to restart if a suspension occurs. + */ + +/* Declare and initialize local copies of input pointer/count */ +#define INPUT_VARS(cinfo) \ + struct jpeg_source_mgr * datasrc = (cinfo)->src; \ + const JOCTET * next_input_byte = datasrc->next_input_byte; \ + size_t bytes_in_buffer = datasrc->bytes_in_buffer + +/* Unload the local copies --- do this only at a restart boundary */ +#define INPUT_SYNC(cinfo) \ + ( datasrc->next_input_byte = next_input_byte, \ + datasrc->bytes_in_buffer = bytes_in_buffer ) + +/* Reload the local copies --- used only in MAKE_BYTE_AVAIL */ +#define INPUT_RELOAD(cinfo) \ + ( next_input_byte = datasrc->next_input_byte, \ + bytes_in_buffer = datasrc->bytes_in_buffer ) + +/* Internal macro for INPUT_BYTE and INPUT_2BYTES: make a byte available. + * Note we do *not* do INPUT_SYNC before calling fill_input_buffer, + * but we must reload the local copies after a successful fill. + */ +#define MAKE_BYTE_AVAIL(cinfo,action) \ + if (bytes_in_buffer == 0) { \ + if (! (*datasrc->fill_input_buffer) (cinfo)) \ + { action; } \ + INPUT_RELOAD(cinfo); \ + } + +/* Read a byte into variable V. + * If must suspend, take the specified action (typically "return FALSE"). + */ +#define INPUT_BYTE(cinfo,V,action) \ + MAKESTMT( MAKE_BYTE_AVAIL(cinfo,action); \ + bytes_in_buffer--; \ + V = GETJOCTET(*next_input_byte++); ) + +/* As above, but read two bytes interpreted as an unsigned 16-bit integer. + * V should be declared unsigned int or perhaps INT32. + */ +#define INPUT_2BYTES(cinfo,V,action) \ + MAKESTMT( MAKE_BYTE_AVAIL(cinfo,action); \ + bytes_in_buffer--; \ + V = ((unsigned int) GETJOCTET(*next_input_byte++)) << 8; \ + MAKE_BYTE_AVAIL(cinfo,action); \ + bytes_in_buffer--; \ + V += GETJOCTET(*next_input_byte++); ) + + +/* + * Routines to process JPEG markers. + * + * Entry condition: JPEG marker itself has been read and its code saved + * in cinfo->unread_marker; input restart point is just after the marker. + * + * Exit: if return TRUE, have read and processed any parameters, and have + * updated the restart point to point after the parameters. + * If return FALSE, was forced to suspend before reaching end of + * marker parameters; restart point has not been moved. Same routine + * will be called again after application supplies more input data. + * + * This approach to suspension assumes that all of a marker's parameters + * can fit into a single input bufferload. This should hold for "normal" + * markers. Some COM/APPn markers might have large parameter segments + * that might not fit. If we are simply dropping such a marker, we use + * skip_input_data to get past it, and thereby put the problem on the + * source manager's shoulders. If we are saving the marker's contents + * into memory, we use a slightly different convention: when forced to + * suspend, the marker processor updates the restart point to the end of + * what it's consumed (ie, the end of the buffer) before returning FALSE. + * On resumption, cinfo->unread_marker still contains the marker code, + * but the data source will point to the next chunk of marker data. + * The marker processor must retain internal state to deal with this. + * + * Note that we don't bother to avoid duplicate trace messages if a + * suspension occurs within marker parameters. Other side effects + * require more care. + */ + + +LOCAL(boolean) +get_soi (j_decompress_ptr cinfo) +/* Process an SOI marker */ +{ + int i; + + TRACEMS(cinfo, 1, JTRC_SOI); + + if (cinfo->marker->saw_SOI) + ERREXIT(cinfo, JERR_SOI_DUPLICATE); + + /* Reset all parameters that are defined to be reset by SOI */ + + for (i = 0; i < NUM_ARITH_TBLS; i++) { + cinfo->arith_dc_L[i] = 0; + cinfo->arith_dc_U[i] = 1; + cinfo->arith_ac_K[i] = 5; + } + cinfo->restart_interval = 0; + + /* Set initial assumptions for colorspace etc */ + + cinfo->jpeg_color_space = JCS_UNKNOWN; + cinfo->color_transform = JCT_NONE; + cinfo->CCIR601_sampling = FALSE; /* Assume non-CCIR sampling??? */ + + cinfo->saw_JFIF_marker = FALSE; + cinfo->JFIF_major_version = 1; /* set default JFIF APP0 values */ + cinfo->JFIF_minor_version = 1; + cinfo->density_unit = 0; + cinfo->X_density = 1; + cinfo->Y_density = 1; + cinfo->saw_Adobe_marker = FALSE; + cinfo->Adobe_transform = 0; + + cinfo->marker->saw_SOI = TRUE; + + return TRUE; +} + + +LOCAL(boolean) +get_sof (j_decompress_ptr cinfo, boolean is_baseline, boolean is_prog, + boolean is_arith) +/* Process a SOFn marker */ +{ + INT32 length; + int c, ci, i; + jpeg_component_info * compptr; + INPUT_VARS(cinfo); + + cinfo->is_baseline = is_baseline; + cinfo->progressive_mode = is_prog; + cinfo->arith_code = is_arith; + + INPUT_2BYTES(cinfo, length, return FALSE); + + INPUT_BYTE(cinfo, cinfo->data_precision, return FALSE); + INPUT_2BYTES(cinfo, cinfo->image_height, return FALSE); + INPUT_2BYTES(cinfo, cinfo->image_width, return FALSE); + INPUT_BYTE(cinfo, cinfo->num_components, return FALSE); + + length -= 8; + + TRACEMS4(cinfo, 1, JTRC_SOF, cinfo->unread_marker, + (int) cinfo->image_width, (int) cinfo->image_height, + cinfo->num_components); + + if (cinfo->marker->saw_SOF) + ERREXIT(cinfo, JERR_SOF_DUPLICATE); + + /* We don't support files in which the image height is initially specified */ + /* as 0 and is later redefined by DNL. As long as we have to check that, */ + /* might as well have a general sanity check. */ + if (cinfo->image_height <= 0 || cinfo->image_width <= 0 || + cinfo->num_components <= 0) + ERREXIT(cinfo, JERR_EMPTY_IMAGE); + + if (length != (cinfo->num_components * 3)) + ERREXIT(cinfo, JERR_BAD_LENGTH); + + if (cinfo->comp_info == NULL) /* do only once, even if suspend */ + cinfo->comp_info = (jpeg_component_info *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, + cinfo->num_components * SIZEOF(jpeg_component_info)); + + for (ci = 0; ci < cinfo->num_components; ci++) { + INPUT_BYTE(cinfo, c, return FALSE); + /* Check to see whether component id has already been seen */ + /* (in violation of the spec, but unfortunately seen in some */ + /* files). If so, create "fake" component id equal to the */ + /* max id seen so far + 1. */ + for (i = 0, compptr = cinfo->comp_info; i < ci; i++, compptr++) { + if (c == compptr->component_id) { + compptr = cinfo->comp_info; + c = compptr->component_id; + compptr++; + for (i = 1; i < ci; i++, compptr++) { + if (compptr->component_id > c) c = compptr->component_id; + } + c++; + break; + } + } + compptr->component_id = c; + compptr->component_index = ci; + INPUT_BYTE(cinfo, c, return FALSE); + compptr->h_samp_factor = (c >> 4) & 15; + compptr->v_samp_factor = (c ) & 15; + INPUT_BYTE(cinfo, compptr->quant_tbl_no, return FALSE); + + TRACEMS4(cinfo, 1, JTRC_SOF_COMPONENT, + compptr->component_id, compptr->h_samp_factor, + compptr->v_samp_factor, compptr->quant_tbl_no); + } + + cinfo->marker->saw_SOF = TRUE; + + INPUT_SYNC(cinfo); + return TRUE; +} + + +LOCAL(boolean) +get_sos (j_decompress_ptr cinfo) +/* Process a SOS marker */ +{ + INT32 length; + int c, ci, i, n; + jpeg_component_info * compptr; + INPUT_VARS(cinfo); + + if (! cinfo->marker->saw_SOF) + ERREXITS(cinfo, JERR_SOF_BEFORE, "SOS"); + + INPUT_2BYTES(cinfo, length, return FALSE); + + INPUT_BYTE(cinfo, n, return FALSE); /* Number of components */ + + TRACEMS1(cinfo, 1, JTRC_SOS, n); + + if (length != (n * 2 + 6) || n > MAX_COMPS_IN_SCAN || + (n == 0 && !cinfo->progressive_mode)) + /* pseudo SOS marker only allowed in progressive mode */ + ERREXIT(cinfo, JERR_BAD_LENGTH); + + cinfo->comps_in_scan = n; + + /* Collect the component-spec parameters */ + + for (i = 0; i < n; i++) { + INPUT_BYTE(cinfo, c, return FALSE); + + /* Detect the case where component id's are not unique, and, if so, */ + /* create a fake component id using the same logic as in get_sof. */ + /* Note: This also ensures that all of the SOF components are */ + /* referenced in the single scan case, which prevents access to */ + /* uninitialized memory in later decoding stages. */ + for (ci = 0; ci < i; ci++) { + if (c == cinfo->cur_comp_info[ci]->component_id) { + c = cinfo->cur_comp_info[0]->component_id; + for (ci = 1; ci < i; ci++) { + compptr = cinfo->cur_comp_info[ci]; + if (compptr->component_id > c) c = compptr->component_id; + } + c++; + break; + } + } + + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + if (c == compptr->component_id) + goto id_found; + } + + ERREXIT1(cinfo, JERR_BAD_COMPONENT_ID, c); + + id_found: + + cinfo->cur_comp_info[i] = compptr; + INPUT_BYTE(cinfo, c, return FALSE); + compptr->dc_tbl_no = (c >> 4) & 15; + compptr->ac_tbl_no = (c ) & 15; + + TRACEMS3(cinfo, 1, JTRC_SOS_COMPONENT, compptr->component_id, + compptr->dc_tbl_no, compptr->ac_tbl_no); + } + + /* Collect the additional scan parameters Ss, Se, Ah/Al. */ + INPUT_BYTE(cinfo, c, return FALSE); + cinfo->Ss = c; + INPUT_BYTE(cinfo, c, return FALSE); + cinfo->Se = c; + INPUT_BYTE(cinfo, c, return FALSE); + cinfo->Ah = (c >> 4) & 15; + cinfo->Al = (c ) & 15; + + TRACEMS4(cinfo, 1, JTRC_SOS_PARAMS, cinfo->Ss, cinfo->Se, + cinfo->Ah, cinfo->Al); + + /* Prepare to scan data & restart markers */ + cinfo->marker->next_restart_num = 0; + + /* Count another (non-pseudo) SOS marker */ + if (n) cinfo->input_scan_number++; + + INPUT_SYNC(cinfo); + return TRUE; +} + + +#ifdef D_ARITH_CODING_SUPPORTED + +LOCAL(boolean) +get_dac (j_decompress_ptr cinfo) +/* Process a DAC marker */ +{ + INT32 length; + int index, val; + INPUT_VARS(cinfo); + + INPUT_2BYTES(cinfo, length, return FALSE); + length -= 2; + + while (length > 0) { + INPUT_BYTE(cinfo, index, return FALSE); + INPUT_BYTE(cinfo, val, return FALSE); + + length -= 2; + + TRACEMS2(cinfo, 1, JTRC_DAC, index, val); + + if (index < 0 || index >= (2*NUM_ARITH_TBLS)) + ERREXIT1(cinfo, JERR_DAC_INDEX, index); + + if (index >= NUM_ARITH_TBLS) { /* define AC table */ + cinfo->arith_ac_K[index-NUM_ARITH_TBLS] = (UINT8) val; + } else { /* define DC table */ + cinfo->arith_dc_L[index] = (UINT8) (val & 0x0F); + cinfo->arith_dc_U[index] = (UINT8) (val >> 4); + if (cinfo->arith_dc_L[index] > cinfo->arith_dc_U[index]) + ERREXIT1(cinfo, JERR_DAC_VALUE, val); + } + } + + if (length != 0) + ERREXIT(cinfo, JERR_BAD_LENGTH); + + INPUT_SYNC(cinfo); + return TRUE; +} + +#else /* ! D_ARITH_CODING_SUPPORTED */ + +#define get_dac(cinfo) skip_variable(cinfo) + +#endif /* D_ARITH_CODING_SUPPORTED */ + + +LOCAL(boolean) +get_dht (j_decompress_ptr cinfo) +/* Process a DHT marker */ +{ + INT32 length; + UINT8 bits[17]; + UINT8 huffval[256]; + int i, index, count; + JHUFF_TBL **htblptr; + INPUT_VARS(cinfo); + + INPUT_2BYTES(cinfo, length, return FALSE); + length -= 2; + + while (length > 16) { + INPUT_BYTE(cinfo, index, return FALSE); + + TRACEMS1(cinfo, 1, JTRC_DHT, index); + + bits[0] = 0; + count = 0; + for (i = 1; i <= 16; i++) { + INPUT_BYTE(cinfo, bits[i], return FALSE); + count += bits[i]; + } + + length -= 1 + 16; + + TRACEMS8(cinfo, 2, JTRC_HUFFBITS, + bits[1], bits[2], bits[3], bits[4], + bits[5], bits[6], bits[7], bits[8]); + TRACEMS8(cinfo, 2, JTRC_HUFFBITS, + bits[9], bits[10], bits[11], bits[12], + bits[13], bits[14], bits[15], bits[16]); + + /* Here we just do minimal validation of the counts to avoid walking + * off the end of our table space. jdhuff.c will check more carefully. + */ + if (count > 256 || ((INT32) count) > length) + ERREXIT(cinfo, JERR_BAD_HUFF_TABLE); + + for (i = 0; i < count; i++) + INPUT_BYTE(cinfo, huffval[i], return FALSE); + + length -= count; + + if (index & 0x10) { /* AC table definition */ + index -= 0x10; + htblptr = &cinfo->ac_huff_tbl_ptrs[index]; + } else { /* DC table definition */ + htblptr = &cinfo->dc_huff_tbl_ptrs[index]; + } + + if (index < 0 || index >= NUM_HUFF_TBLS) + ERREXIT1(cinfo, JERR_DHT_INDEX, index); + + if (*htblptr == NULL) + *htblptr = jpeg_alloc_huff_table((j_common_ptr) cinfo); + + MEMCOPY((*htblptr)->bits, bits, SIZEOF((*htblptr)->bits)); + if (count > 0) + MEMCOPY((*htblptr)->huffval, huffval, count * SIZEOF(UINT8)); + } + + if (length != 0) + ERREXIT(cinfo, JERR_BAD_LENGTH); + + INPUT_SYNC(cinfo); + return TRUE; +} + + +LOCAL(boolean) +get_dqt (j_decompress_ptr cinfo) +/* Process a DQT marker */ +{ + INT32 length, count, i; + int n, prec; + unsigned int tmp; + JQUANT_TBL *quant_ptr; + const int *natural_order; + INPUT_VARS(cinfo); + + INPUT_2BYTES(cinfo, length, return FALSE); + length -= 2; + + while (length > 0) { + length--; + INPUT_BYTE(cinfo, n, return FALSE); + prec = n >> 4; + n &= 0x0F; + + TRACEMS2(cinfo, 1, JTRC_DQT, n, prec); + + if (n >= NUM_QUANT_TBLS) + ERREXIT1(cinfo, JERR_DQT_INDEX, n); + + if (cinfo->quant_tbl_ptrs[n] == NULL) + cinfo->quant_tbl_ptrs[n] = jpeg_alloc_quant_table((j_common_ptr) cinfo); + quant_ptr = cinfo->quant_tbl_ptrs[n]; + + if (prec) { + if (length < DCTSIZE2 * 2) { + /* Initialize full table for safety. */ + for (i = 0; i < DCTSIZE2; i++) { + quant_ptr->quantval[i] = 1; + } + count = length >> 1; + } else + count = DCTSIZE2; + } else { + if (length < DCTSIZE2) { + /* Initialize full table for safety. */ + for (i = 0; i < DCTSIZE2; i++) { + quant_ptr->quantval[i] = 1; + } + count = length; + } else + count = DCTSIZE2; + } + + switch ((int) count) { + case (2*2): natural_order = jpeg_natural_order2; break; + case (3*3): natural_order = jpeg_natural_order3; break; + case (4*4): natural_order = jpeg_natural_order4; break; + case (5*5): natural_order = jpeg_natural_order5; break; + case (6*6): natural_order = jpeg_natural_order6; break; + case (7*7): natural_order = jpeg_natural_order7; break; + default: natural_order = jpeg_natural_order; + } + + for (i = 0; i < count; i++) { + if (prec) + INPUT_2BYTES(cinfo, tmp, return FALSE); + else + INPUT_BYTE(cinfo, tmp, return FALSE); + /* We convert the zigzag-order table to natural array order. */ + quant_ptr->quantval[natural_order[i]] = (UINT16) tmp; + } + + if (cinfo->err->trace_level >= 2) { + for (i = 0; i < DCTSIZE2; i += 8) { + TRACEMS8(cinfo, 2, JTRC_QUANTVALS, + quant_ptr->quantval[i], quant_ptr->quantval[i+1], + quant_ptr->quantval[i+2], quant_ptr->quantval[i+3], + quant_ptr->quantval[i+4], quant_ptr->quantval[i+5], + quant_ptr->quantval[i+6], quant_ptr->quantval[i+7]); + } + } + + length -= count; + if (prec) length -= count; + } + + if (length != 0) + ERREXIT(cinfo, JERR_BAD_LENGTH); + + INPUT_SYNC(cinfo); + return TRUE; +} + + +LOCAL(boolean) +get_dri (j_decompress_ptr cinfo) +/* Process a DRI marker */ +{ + INT32 length; + unsigned int tmp; + INPUT_VARS(cinfo); + + INPUT_2BYTES(cinfo, length, return FALSE); + + if (length != 4) + ERREXIT(cinfo, JERR_BAD_LENGTH); + + INPUT_2BYTES(cinfo, tmp, return FALSE); + + TRACEMS1(cinfo, 1, JTRC_DRI, tmp); + + cinfo->restart_interval = tmp; + + INPUT_SYNC(cinfo); + return TRUE; +} + + +LOCAL(boolean) +get_lse (j_decompress_ptr cinfo) +/* Process an LSE marker */ +{ + INT32 length; + unsigned int tmp; + int cid; + INPUT_VARS(cinfo); + + if (! cinfo->marker->saw_SOF) + ERREXITS(cinfo, JERR_SOF_BEFORE, "LSE"); + + if (cinfo->num_components < 3) goto bad; + + INPUT_2BYTES(cinfo, length, return FALSE); + + if (length != 24) + ERREXIT(cinfo, JERR_BAD_LENGTH); + + INPUT_BYTE(cinfo, tmp, return FALSE); + if (tmp != 0x0D) /* ID inverse transform specification */ + ERREXIT1(cinfo, JERR_UNKNOWN_MARKER, cinfo->unread_marker); + INPUT_2BYTES(cinfo, tmp, return FALSE); + if (tmp != MAXJSAMPLE) goto bad; /* MAXTRANS */ + INPUT_BYTE(cinfo, tmp, return FALSE); + if (tmp != 3) goto bad; /* Nt=3 */ + INPUT_BYTE(cinfo, cid, return FALSE); + if (cid != cinfo->comp_info[1].component_id) goto bad; + INPUT_BYTE(cinfo, cid, return FALSE); + if (cid != cinfo->comp_info[0].component_id) goto bad; + INPUT_BYTE(cinfo, cid, return FALSE); + if (cid != cinfo->comp_info[2].component_id) goto bad; + INPUT_BYTE(cinfo, tmp, return FALSE); + if (tmp != 0x80) goto bad; /* F1: CENTER1=1, NORM1=0 */ + INPUT_2BYTES(cinfo, tmp, return FALSE); + if (tmp != 0) goto bad; /* A(1,1)=0 */ + INPUT_2BYTES(cinfo, tmp, return FALSE); + if (tmp != 0) goto bad; /* A(1,2)=0 */ + INPUT_BYTE(cinfo, tmp, return FALSE); + if (tmp != 0) goto bad; /* F2: CENTER2=0, NORM2=0 */ + INPUT_2BYTES(cinfo, tmp, return FALSE); + if (tmp != 1) goto bad; /* A(2,1)=1 */ + INPUT_2BYTES(cinfo, tmp, return FALSE); + if (tmp != 0) goto bad; /* A(2,2)=0 */ + INPUT_BYTE(cinfo, tmp, return FALSE); + if (tmp != 0) goto bad; /* F3: CENTER3=0, NORM3=0 */ + INPUT_2BYTES(cinfo, tmp, return FALSE); + if (tmp != 1) goto bad; /* A(3,1)=1 */ + INPUT_2BYTES(cinfo, tmp, return FALSE); + if (tmp != 0) { /* A(3,2)=0 */ + bad: + ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); + } + + /* OK, valid transform that we can handle. */ + cinfo->color_transform = JCT_SUBTRACT_GREEN; + + INPUT_SYNC(cinfo); + return TRUE; +} + + +/* + * Routines for processing APPn and COM markers. + * These are either saved in memory or discarded, per application request. + * APP0 and APP14 are specially checked to see if they are + * JFIF and Adobe markers, respectively. + */ + +#define APP0_DATA_LEN 14 /* Length of interesting data in APP0 */ +#define APP14_DATA_LEN 12 /* Length of interesting data in APP14 */ +#define APPN_DATA_LEN 14 /* Must be the largest of the above!! */ + + +LOCAL(void) +examine_app0 (j_decompress_ptr cinfo, JOCTET FAR * data, + unsigned int datalen, INT32 remaining) +/* Examine first few bytes from an APP0. + * Take appropriate action if it is a JFIF marker. + * datalen is # of bytes at data[], remaining is length of rest of marker data. + */ +{ + INT32 totallen = (INT32) datalen + remaining; + + if (datalen >= APP0_DATA_LEN && + GETJOCTET(data[0]) == 0x4A && + GETJOCTET(data[1]) == 0x46 && + GETJOCTET(data[2]) == 0x49 && + GETJOCTET(data[3]) == 0x46 && + GETJOCTET(data[4]) == 0) { + /* Found JFIF APP0 marker: save info */ + cinfo->saw_JFIF_marker = TRUE; + cinfo->JFIF_major_version = GETJOCTET(data[5]); + cinfo->JFIF_minor_version = GETJOCTET(data[6]); + cinfo->density_unit = GETJOCTET(data[7]); + cinfo->X_density = (GETJOCTET(data[8]) << 8) + GETJOCTET(data[9]); + cinfo->Y_density = (GETJOCTET(data[10]) << 8) + GETJOCTET(data[11]); + /* Check version. + * Major version must be 1 or 2, anything else signals an incompatible + * change. + * (We used to treat this as an error, but now it's a nonfatal warning, + * because some bozo at Hijaak couldn't read the spec.) + * Minor version should be 0..2, but process anyway if newer. + */ + if (cinfo->JFIF_major_version != 1 && cinfo->JFIF_major_version != 2) + WARNMS2(cinfo, JWRN_JFIF_MAJOR, + cinfo->JFIF_major_version, cinfo->JFIF_minor_version); + /* Generate trace messages */ + TRACEMS5(cinfo, 1, JTRC_JFIF, + cinfo->JFIF_major_version, cinfo->JFIF_minor_version, + cinfo->X_density, cinfo->Y_density, cinfo->density_unit); + /* Validate thumbnail dimensions and issue appropriate messages */ + if (GETJOCTET(data[12]) | GETJOCTET(data[13])) + TRACEMS2(cinfo, 1, JTRC_JFIF_THUMBNAIL, + GETJOCTET(data[12]), GETJOCTET(data[13])); + totallen -= APP0_DATA_LEN; + if (totallen != + ((INT32)GETJOCTET(data[12]) * (INT32)GETJOCTET(data[13]) * (INT32) 3)) + TRACEMS1(cinfo, 1, JTRC_JFIF_BADTHUMBNAILSIZE, (int) totallen); + } else if (datalen >= 6 && + GETJOCTET(data[0]) == 0x4A && + GETJOCTET(data[1]) == 0x46 && + GETJOCTET(data[2]) == 0x58 && + GETJOCTET(data[3]) == 0x58 && + GETJOCTET(data[4]) == 0) { + /* Found JFIF "JFXX" extension APP0 marker */ + /* The library doesn't actually do anything with these, + * but we try to produce a helpful trace message. + */ + switch (GETJOCTET(data[5])) { + case 0x10: + TRACEMS1(cinfo, 1, JTRC_THUMB_JPEG, (int) totallen); + break; + case 0x11: + TRACEMS1(cinfo, 1, JTRC_THUMB_PALETTE, (int) totallen); + break; + case 0x13: + TRACEMS1(cinfo, 1, JTRC_THUMB_RGB, (int) totallen); + break; + default: + TRACEMS2(cinfo, 1, JTRC_JFIF_EXTENSION, + GETJOCTET(data[5]), (int) totallen); + } + } else { + /* Start of APP0 does not match "JFIF" or "JFXX", or too short */ + TRACEMS1(cinfo, 1, JTRC_APP0, (int) totallen); + } +} + + +LOCAL(void) +examine_app14 (j_decompress_ptr cinfo, JOCTET FAR * data, + unsigned int datalen, INT32 remaining) +/* Examine first few bytes from an APP14. + * Take appropriate action if it is an Adobe marker. + * datalen is # of bytes at data[], remaining is length of rest of marker data. + */ +{ + unsigned int version, flags0, flags1, transform; + + if (datalen >= APP14_DATA_LEN && + GETJOCTET(data[0]) == 0x41 && + GETJOCTET(data[1]) == 0x64 && + GETJOCTET(data[2]) == 0x6F && + GETJOCTET(data[3]) == 0x62 && + GETJOCTET(data[4]) == 0x65) { + /* Found Adobe APP14 marker */ + version = (GETJOCTET(data[5]) << 8) + GETJOCTET(data[6]); + flags0 = (GETJOCTET(data[7]) << 8) + GETJOCTET(data[8]); + flags1 = (GETJOCTET(data[9]) << 8) + GETJOCTET(data[10]); + transform = GETJOCTET(data[11]); + TRACEMS4(cinfo, 1, JTRC_ADOBE, version, flags0, flags1, transform); + cinfo->saw_Adobe_marker = TRUE; + cinfo->Adobe_transform = (UINT8) transform; + } else { + /* Start of APP14 does not match "Adobe", or too short */ + TRACEMS1(cinfo, 1, JTRC_APP14, (int) (datalen + remaining)); + } +} + + +METHODDEF(boolean) +get_interesting_appn (j_decompress_ptr cinfo) +/* Process an APP0 or APP14 marker without saving it */ +{ + INT32 length; + JOCTET b[APPN_DATA_LEN]; + unsigned int i, numtoread; + INPUT_VARS(cinfo); + + INPUT_2BYTES(cinfo, length, return FALSE); + length -= 2; + + /* get the interesting part of the marker data */ + if (length >= APPN_DATA_LEN) + numtoread = APPN_DATA_LEN; + else if (length > 0) + numtoread = (unsigned int) length; + else + numtoread = 0; + for (i = 0; i < numtoread; i++) + INPUT_BYTE(cinfo, b[i], return FALSE); + length -= numtoread; + + /* process it */ + switch (cinfo->unread_marker) { + case M_APP0: + examine_app0(cinfo, (JOCTET FAR *) b, numtoread, length); + break; + case M_APP14: + examine_app14(cinfo, (JOCTET FAR *) b, numtoread, length); + break; + default: + /* can't get here unless jpeg_save_markers chooses wrong processor */ + ERREXIT1(cinfo, JERR_UNKNOWN_MARKER, cinfo->unread_marker); + } + + /* skip any remaining data -- could be lots */ + INPUT_SYNC(cinfo); + if (length > 0) + (*cinfo->src->skip_input_data) (cinfo, (long) length); + + return TRUE; +} + + +#ifdef SAVE_MARKERS_SUPPORTED + +METHODDEF(boolean) +save_marker (j_decompress_ptr cinfo) +/* Save an APPn or COM marker into the marker list */ +{ + my_marker_ptr marker = (my_marker_ptr) cinfo->marker; + jpeg_saved_marker_ptr cur_marker = marker->cur_marker; + unsigned int bytes_read, data_length; + JOCTET FAR * data; + INT32 length = 0; + INPUT_VARS(cinfo); + + if (cur_marker == NULL) { + /* begin reading a marker */ + INPUT_2BYTES(cinfo, length, return FALSE); + length -= 2; + if (length >= 0) { /* watch out for bogus length word */ + /* figure out how much we want to save */ + unsigned int limit; + if (cinfo->unread_marker == (int) M_COM) + limit = marker->length_limit_COM; + else + limit = marker->length_limit_APPn[cinfo->unread_marker - (int) M_APP0]; + if ((unsigned int) length < limit) + limit = (unsigned int) length; + /* allocate and initialize the marker item */ + cur_marker = (jpeg_saved_marker_ptr) + (*cinfo->mem->alloc_large) ((j_common_ptr) cinfo, JPOOL_IMAGE, + SIZEOF(struct jpeg_marker_struct) + limit); + cur_marker->next = NULL; + cur_marker->marker = (UINT8) cinfo->unread_marker; + cur_marker->original_length = (unsigned int) length; + cur_marker->data_length = limit; + /* data area is just beyond the jpeg_marker_struct */ + data = cur_marker->data = (JOCTET FAR *) (cur_marker + 1); + marker->cur_marker = cur_marker; + marker->bytes_read = 0; + bytes_read = 0; + data_length = limit; + } else { + /* deal with bogus length word */ + bytes_read = data_length = 0; + data = NULL; + } + } else { + /* resume reading a marker */ + bytes_read = marker->bytes_read; + data_length = cur_marker->data_length; + data = cur_marker->data + bytes_read; + } + + while (bytes_read < data_length) { + INPUT_SYNC(cinfo); /* move the restart point to here */ + marker->bytes_read = bytes_read; + /* If there's not at least one byte in buffer, suspend */ + MAKE_BYTE_AVAIL(cinfo, return FALSE); + /* Copy bytes with reasonable rapidity */ + while (bytes_read < data_length && bytes_in_buffer > 0) { + *data++ = *next_input_byte++; + bytes_in_buffer--; + bytes_read++; + } + } + + /* Done reading what we want to read */ + if (cur_marker != NULL) { /* will be NULL if bogus length word */ + /* Add new marker to end of list */ + if (cinfo->marker_list == NULL) { + cinfo->marker_list = cur_marker; + } else { + jpeg_saved_marker_ptr prev = cinfo->marker_list; + while (prev->next != NULL) + prev = prev->next; + prev->next = cur_marker; + } + /* Reset pointer & calc remaining data length */ + data = cur_marker->data; + length = cur_marker->original_length - data_length; + } + /* Reset to initial state for next marker */ + marker->cur_marker = NULL; + + /* Process the marker if interesting; else just make a generic trace msg */ + switch (cinfo->unread_marker) { + case M_APP0: + examine_app0(cinfo, data, data_length, length); + break; + case M_APP14: + examine_app14(cinfo, data, data_length, length); + break; + default: + TRACEMS2(cinfo, 1, JTRC_MISC_MARKER, cinfo->unread_marker, + (int) (data_length + length)); + } + + /* skip any remaining data -- could be lots */ + INPUT_SYNC(cinfo); /* do before skip_input_data */ + if (length > 0) + (*cinfo->src->skip_input_data) (cinfo, (long) length); + + return TRUE; +} + +#endif /* SAVE_MARKERS_SUPPORTED */ + + +METHODDEF(boolean) +skip_variable (j_decompress_ptr cinfo) +/* Skip over an unknown or uninteresting variable-length marker */ +{ + INT32 length; + INPUT_VARS(cinfo); + + INPUT_2BYTES(cinfo, length, return FALSE); + length -= 2; + + TRACEMS2(cinfo, 1, JTRC_MISC_MARKER, cinfo->unread_marker, (int) length); + + INPUT_SYNC(cinfo); /* do before skip_input_data */ + if (length > 0) + (*cinfo->src->skip_input_data) (cinfo, (long) length); + + return TRUE; +} + + +/* + * Find the next JPEG marker, save it in cinfo->unread_marker. + * Returns FALSE if had to suspend before reaching a marker; + * in that case cinfo->unread_marker is unchanged. + * + * Note that the result might not be a valid marker code, + * but it will never be 0 or FF. + */ + +LOCAL(boolean) +next_marker (j_decompress_ptr cinfo) +{ + int c; + INPUT_VARS(cinfo); + + for (;;) { + INPUT_BYTE(cinfo, c, return FALSE); + /* Skip any non-FF bytes. + * This may look a bit inefficient, but it will not occur in a valid file. + * We sync after each discarded byte so that a suspending data source + * can discard the byte from its buffer. + */ + while (c != 0xFF) { + cinfo->marker->discarded_bytes++; + INPUT_SYNC(cinfo); + INPUT_BYTE(cinfo, c, return FALSE); + } + /* This loop swallows any duplicate FF bytes. Extra FFs are legal as + * pad bytes, so don't count them in discarded_bytes. We assume there + * will not be so many consecutive FF bytes as to overflow a suspending + * data source's input buffer. + */ + do { + INPUT_BYTE(cinfo, c, return FALSE); + } while (c == 0xFF); + if (c != 0) + break; /* found a valid marker, exit loop */ + /* Reach here if we found a stuffed-zero data sequence (FF/00). + * Discard it and loop back to try again. + */ + cinfo->marker->discarded_bytes += 2; + INPUT_SYNC(cinfo); + } + + if (cinfo->marker->discarded_bytes != 0) { + WARNMS2(cinfo, JWRN_EXTRANEOUS_DATA, cinfo->marker->discarded_bytes, c); + cinfo->marker->discarded_bytes = 0; + } + + cinfo->unread_marker = c; + + INPUT_SYNC(cinfo); + return TRUE; +} + + +LOCAL(boolean) +first_marker (j_decompress_ptr cinfo) +/* Like next_marker, but used to obtain the initial SOI marker. */ +/* For this marker, we do not allow preceding garbage or fill; otherwise, + * we might well scan an entire input file before realizing it ain't JPEG. + * If an application wants to process non-JFIF files, it must seek to the + * SOI before calling the JPEG library. + */ +{ + int c, c2; + INPUT_VARS(cinfo); + + INPUT_BYTE(cinfo, c, return FALSE); + INPUT_BYTE(cinfo, c2, return FALSE); + if (c != 0xFF || c2 != (int) M_SOI) + ERREXIT2(cinfo, JERR_NO_SOI, c, c2); + + cinfo->unread_marker = c2; + + INPUT_SYNC(cinfo); + return TRUE; +} + + +/* + * Read markers until SOS or EOI. + * + * Returns same codes as are defined for jpeg_consume_input: + * JPEG_SUSPENDED, JPEG_REACHED_SOS, or JPEG_REACHED_EOI. + * + * Note: This function may return a pseudo SOS marker (with zero + * component number) for treat by input controller's consume_input. + * consume_input itself should filter out (skip) the pseudo marker + * after processing for the caller. + */ + +METHODDEF(int) +read_markers (j_decompress_ptr cinfo) +{ + /* Outer loop repeats once for each marker. */ + for (;;) { + /* Collect the marker proper, unless we already did. */ + /* NB: first_marker() enforces the requirement that SOI appear first. */ + if (cinfo->unread_marker == 0) { + if (! cinfo->marker->saw_SOI) { + if (! first_marker(cinfo)) + return JPEG_SUSPENDED; + } else { + if (! next_marker(cinfo)) + return JPEG_SUSPENDED; + } + } + /* At this point cinfo->unread_marker contains the marker code and the + * input point is just past the marker proper, but before any parameters. + * A suspension will cause us to return with this state still true. + */ + switch (cinfo->unread_marker) { + case M_SOI: + if (! get_soi(cinfo)) + return JPEG_SUSPENDED; + break; + + case M_SOF0: /* Baseline */ + if (! get_sof(cinfo, TRUE, FALSE, FALSE)) + return JPEG_SUSPENDED; + break; + + case M_SOF1: /* Extended sequential, Huffman */ + if (! get_sof(cinfo, FALSE, FALSE, FALSE)) + return JPEG_SUSPENDED; + break; + + case M_SOF2: /* Progressive, Huffman */ + if (! get_sof(cinfo, FALSE, TRUE, FALSE)) + return JPEG_SUSPENDED; + break; + + case M_SOF9: /* Extended sequential, arithmetic */ + if (! get_sof(cinfo, FALSE, FALSE, TRUE)) + return JPEG_SUSPENDED; + break; + + case M_SOF10: /* Progressive, arithmetic */ + if (! get_sof(cinfo, FALSE, TRUE, TRUE)) + return JPEG_SUSPENDED; + break; + + /* Currently unsupported SOFn types */ + case M_SOF3: /* Lossless, Huffman */ + case M_SOF5: /* Differential sequential, Huffman */ + case M_SOF6: /* Differential progressive, Huffman */ + case M_SOF7: /* Differential lossless, Huffman */ + case M_JPG: /* Reserved for JPEG extensions */ + case M_SOF11: /* Lossless, arithmetic */ + case M_SOF13: /* Differential sequential, arithmetic */ + case M_SOF14: /* Differential progressive, arithmetic */ + case M_SOF15: /* Differential lossless, arithmetic */ + ERREXIT1(cinfo, JERR_SOF_UNSUPPORTED, cinfo->unread_marker); + break; + + case M_SOS: + if (! get_sos(cinfo)) + return JPEG_SUSPENDED; + cinfo->unread_marker = 0; /* processed the marker */ + return JPEG_REACHED_SOS; + + case M_EOI: + TRACEMS(cinfo, 1, JTRC_EOI); + cinfo->unread_marker = 0; /* processed the marker */ + return JPEG_REACHED_EOI; + + case M_DAC: + if (! get_dac(cinfo)) + return JPEG_SUSPENDED; + break; + + case M_DHT: + if (! get_dht(cinfo)) + return JPEG_SUSPENDED; + break; + + case M_DQT: + if (! get_dqt(cinfo)) + return JPEG_SUSPENDED; + break; + + case M_DRI: + if (! get_dri(cinfo)) + return JPEG_SUSPENDED; + break; + + case M_JPG8: + if (! get_lse(cinfo)) + return JPEG_SUSPENDED; + break; + + case M_APP0: + case M_APP1: + case M_APP2: + case M_APP3: + case M_APP4: + case M_APP5: + case M_APP6: + case M_APP7: + case M_APP8: + case M_APP9: + case M_APP10: + case M_APP11: + case M_APP12: + case M_APP13: + case M_APP14: + case M_APP15: + if (! (*((my_marker_ptr) cinfo->marker)->process_APPn[ + cinfo->unread_marker - (int) M_APP0]) (cinfo)) + return JPEG_SUSPENDED; + break; + + case M_COM: + if (! (*((my_marker_ptr) cinfo->marker)->process_COM) (cinfo)) + return JPEG_SUSPENDED; + break; + + case M_RST0: /* these are all parameterless */ + case M_RST1: + case M_RST2: + case M_RST3: + case M_RST4: + case M_RST5: + case M_RST6: + case M_RST7: + case M_TEM: + TRACEMS1(cinfo, 1, JTRC_PARMLESS_MARKER, cinfo->unread_marker); + break; + + case M_DNL: /* Ignore DNL ... perhaps the wrong thing */ + if (! skip_variable(cinfo)) + return JPEG_SUSPENDED; + break; + + default: /* must be DHP, EXP, JPGn, or RESn */ + /* For now, we treat the reserved markers as fatal errors since they are + * likely to be used to signal incompatible JPEG Part 3 extensions. + * Once the JPEG 3 version-number marker is well defined, this code + * ought to change! + */ + ERREXIT1(cinfo, JERR_UNKNOWN_MARKER, cinfo->unread_marker); + } + /* Successfully processed marker, so reset state variable */ + cinfo->unread_marker = 0; + } /* end loop */ +} + + +/* + * Read a restart marker, which is expected to appear next in the datastream; + * if the marker is not there, take appropriate recovery action. + * Returns FALSE if suspension is required. + * + * This is called by the entropy decoder after it has read an appropriate + * number of MCUs. cinfo->unread_marker may be nonzero if the entropy decoder + * has already read a marker from the data source. Under normal conditions + * cinfo->unread_marker will be reset to 0 before returning; if not reset, + * it holds a marker which the decoder will be unable to read past. + */ + +METHODDEF(boolean) +read_restart_marker (j_decompress_ptr cinfo) +{ + /* Obtain a marker unless we already did. */ + /* Note that next_marker will complain if it skips any data. */ + if (cinfo->unread_marker == 0) { + if (! next_marker(cinfo)) + return FALSE; + } + + if (cinfo->unread_marker == + ((int) M_RST0 + cinfo->marker->next_restart_num)) { + /* Normal case --- swallow the marker and let entropy decoder continue */ + TRACEMS1(cinfo, 3, JTRC_RST, cinfo->marker->next_restart_num); + cinfo->unread_marker = 0; + } else { + /* Uh-oh, the restart markers have been messed up. */ + /* Let the data source manager determine how to resync. */ + if (! (*cinfo->src->resync_to_restart) (cinfo, + cinfo->marker->next_restart_num)) + return FALSE; + } + + /* Update next-restart state */ + cinfo->marker->next_restart_num = (cinfo->marker->next_restart_num + 1) & 7; + + return TRUE; +} + + +/* + * This is the default resync_to_restart method for data source managers + * to use if they don't have any better approach. Some data source managers + * may be able to back up, or may have additional knowledge about the data + * which permits a more intelligent recovery strategy; such managers would + * presumably supply their own resync method. + * + * read_restart_marker calls resync_to_restart if it finds a marker other than + * the restart marker it was expecting. (This code is *not* used unless + * a nonzero restart interval has been declared.) cinfo->unread_marker is + * the marker code actually found (might be anything, except 0 or FF). + * The desired restart marker number (0..7) is passed as a parameter. + * This routine is supposed to apply whatever error recovery strategy seems + * appropriate in order to position the input stream to the next data segment. + * Note that cinfo->unread_marker is treated as a marker appearing before + * the current data-source input point; usually it should be reset to zero + * before returning. + * Returns FALSE if suspension is required. + * + * This implementation is substantially constrained by wanting to treat the + * input as a data stream; this means we can't back up. Therefore, we have + * only the following actions to work with: + * 1. Simply discard the marker and let the entropy decoder resume at next + * byte of file. + * 2. Read forward until we find another marker, discarding intervening + * data. (In theory we could look ahead within the current bufferload, + * without having to discard data if we don't find the desired marker. + * This idea is not implemented here, in part because it makes behavior + * dependent on buffer size and chance buffer-boundary positions.) + * 3. Leave the marker unread (by failing to zero cinfo->unread_marker). + * This will cause the entropy decoder to process an empty data segment, + * inserting dummy zeroes, and then we will reprocess the marker. + * + * #2 is appropriate if we think the desired marker lies ahead, while #3 is + * appropriate if the found marker is a future restart marker (indicating + * that we have missed the desired restart marker, probably because it got + * corrupted). + * We apply #2 or #3 if the found marker is a restart marker no more than + * two counts behind or ahead of the expected one. We also apply #2 if the + * found marker is not a legal JPEG marker code (it's certainly bogus data). + * If the found marker is a restart marker more than 2 counts away, we do #1 + * (too much risk that the marker is erroneous; with luck we will be able to + * resync at some future point). + * For any valid non-restart JPEG marker, we apply #3. This keeps us from + * overrunning the end of a scan. An implementation limited to single-scan + * files might find it better to apply #2 for markers other than EOI, since + * any other marker would have to be bogus data in that case. + */ + +GLOBAL(boolean) +jpeg_resync_to_restart (j_decompress_ptr cinfo, int desired) +{ + int marker = cinfo->unread_marker; + int action = 1; + + /* Always put up a warning. */ + WARNMS2(cinfo, JWRN_MUST_RESYNC, marker, desired); + + /* Outer loop handles repeated decision after scanning forward. */ + for (;;) { + if (marker < (int) M_SOF0) + action = 2; /* invalid marker */ + else if (marker < (int) M_RST0 || marker > (int) M_RST7) + action = 3; /* valid non-restart marker */ + else { + if (marker == ((int) M_RST0 + ((desired+1) & 7)) || + marker == ((int) M_RST0 + ((desired+2) & 7))) + action = 3; /* one of the next two expected restarts */ + else if (marker == ((int) M_RST0 + ((desired-1) & 7)) || + marker == ((int) M_RST0 + ((desired-2) & 7))) + action = 2; /* a prior restart, so advance */ + else + action = 1; /* desired restart or too far away */ + } + TRACEMS2(cinfo, 4, JTRC_RECOVERY_ACTION, marker, action); + switch (action) { + case 1: + /* Discard marker and let entropy decoder resume processing. */ + cinfo->unread_marker = 0; + return TRUE; + case 2: + /* Scan to the next marker, and repeat the decision loop. */ + if (! next_marker(cinfo)) + return FALSE; + marker = cinfo->unread_marker; + break; + case 3: + /* Return without advancing past this marker. */ + /* Entropy decoder will be forced to process an empty segment. */ + return TRUE; + } + } /* end loop */ +} + + +/* + * Reset marker processing state to begin a fresh datastream. + */ + +METHODDEF(void) +reset_marker_reader (j_decompress_ptr cinfo) +{ + my_marker_ptr marker = (my_marker_ptr) cinfo->marker; + + cinfo->comp_info = NULL; /* until allocated by get_sof */ + cinfo->input_scan_number = 0; /* no SOS seen yet */ + cinfo->unread_marker = 0; /* no pending marker */ + marker->pub.saw_SOI = FALSE; /* set internal state too */ + marker->pub.saw_SOF = FALSE; + marker->pub.discarded_bytes = 0; + marker->cur_marker = NULL; +} + + +/* + * Initialize the marker reader module. + * This is called only once, when the decompression object is created. + */ + +GLOBAL(void) +jinit_marker_reader (j_decompress_ptr cinfo) +{ + my_marker_ptr marker; + int i; + + /* Create subobject in permanent pool */ + marker = (my_marker_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_PERMANENT, SIZEOF(my_marker_reader)); + cinfo->marker = &marker->pub; + /* Initialize public method pointers */ + marker->pub.reset_marker_reader = reset_marker_reader; + marker->pub.read_markers = read_markers; + marker->pub.read_restart_marker = read_restart_marker; + /* Initialize COM/APPn processing. + * By default, we examine and then discard APP0 and APP14, + * but simply discard COM and all other APPn. + */ + marker->process_COM = skip_variable; + marker->length_limit_COM = 0; + for (i = 0; i < 16; i++) { + marker->process_APPn[i] = skip_variable; + marker->length_limit_APPn[i] = 0; + } + marker->process_APPn[0] = get_interesting_appn; + marker->process_APPn[14] = get_interesting_appn; + /* Reset marker processing state */ + reset_marker_reader(cinfo); +} + + +/* + * Control saving of COM and APPn markers into marker_list. + */ + +#ifdef SAVE_MARKERS_SUPPORTED + +GLOBAL(void) +jpeg_save_markers (j_decompress_ptr cinfo, int marker_code, + unsigned int length_limit) +{ + my_marker_ptr marker = (my_marker_ptr) cinfo->marker; + long maxlength; + jpeg_marker_parser_method processor; + + /* Length limit mustn't be larger than what we can allocate + * (should only be a concern in a 16-bit environment). + */ + maxlength = cinfo->mem->max_alloc_chunk - SIZEOF(struct jpeg_marker_struct); + if (((long) length_limit) > maxlength) + length_limit = (unsigned int) maxlength; + + /* Choose processor routine to use. + * APP0/APP14 have special requirements. + */ + if (length_limit) { + processor = save_marker; + /* If saving APP0/APP14, save at least enough for our internal use. */ + if (marker_code == (int) M_APP0 && length_limit < APP0_DATA_LEN) + length_limit = APP0_DATA_LEN; + else if (marker_code == (int) M_APP14 && length_limit < APP14_DATA_LEN) + length_limit = APP14_DATA_LEN; + } else { + processor = skip_variable; + /* If discarding APP0/APP14, use our regular on-the-fly processor. */ + if (marker_code == (int) M_APP0 || marker_code == (int) M_APP14) + processor = get_interesting_appn; + } + + if (marker_code == (int) M_COM) { + marker->process_COM = processor; + marker->length_limit_COM = length_limit; + } else if (marker_code >= (int) M_APP0 && marker_code <= (int) M_APP15) { + marker->process_APPn[marker_code - (int) M_APP0] = processor; + marker->length_limit_APPn[marker_code - (int) M_APP0] = length_limit; + } else + ERREXIT1(cinfo, JERR_UNKNOWN_MARKER, marker_code); +} + +#endif /* SAVE_MARKERS_SUPPORTED */ + + +/* + * Install a special processing method for COM or APPn markers. + */ + +GLOBAL(void) +jpeg_set_marker_processor (j_decompress_ptr cinfo, int marker_code, + jpeg_marker_parser_method routine) +{ + my_marker_ptr marker = (my_marker_ptr) cinfo->marker; + + if (marker_code == (int) M_COM) + marker->process_COM = routine; + else if (marker_code >= (int) M_APP0 && marker_code <= (int) M_APP15) + marker->process_APPn[marker_code - (int) M_APP0] = routine; + else + ERREXIT1(cinfo, JERR_UNKNOWN_MARKER, marker_code); +} diff --git a/cmake/externals/libjpeg/jdmaster.c b/cmake/externals/libjpeg/jdmaster.c new file mode 100644 index 000000000..3070b7bb4 --- /dev/null +++ b/cmake/externals/libjpeg/jdmaster.c @@ -0,0 +1,532 @@ +/* + * jdmaster.c + * + * Copyright (C) 1991-1997, Thomas G. Lane. + * Modified 2002-2020 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains master control logic for the JPEG decompressor. + * These routines are concerned with selecting the modules to be executed + * and with determining the number of passes and the work to be done in each + * pass. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* Private state */ + +typedef struct { + struct jpeg_decomp_master pub; /* public fields */ + + int pass_number; /* # of passes completed */ + + boolean using_merged_upsample; /* TRUE if using merged upsample/cconvert */ + + /* Saved references to initialized quantizer modules, + * in case we need to switch modes. + */ + struct jpeg_color_quantizer * quantizer_1pass; + struct jpeg_color_quantizer * quantizer_2pass; +} my_decomp_master; + +typedef my_decomp_master * my_master_ptr; + + +/* + * Determine whether merged upsample/color conversion should be used. + * CRUCIAL: this must match the actual capabilities of jdmerge.c! + */ + +LOCAL(boolean) +use_merged_upsample (j_decompress_ptr cinfo) +{ +#ifdef UPSAMPLE_MERGING_SUPPORTED + /* Merging is the equivalent of plain box-filter upsampling. */ + /* The following condition is only needed if fancy shall select + * a different upsampling method. In our current implementation + * fancy only affects the DCT scaling, thus we can use fancy + * upsampling and merged upsample simultaneously, in particular + * with scaled DCT sizes larger than the default DCTSIZE. + */ +#if 0 + if (cinfo->do_fancy_upsampling) + return FALSE; +#endif + if (cinfo->CCIR601_sampling) + return FALSE; + /* jdmerge.c only supports YCC=>RGB color conversion */ + if ((cinfo->jpeg_color_space != JCS_YCbCr && + cinfo->jpeg_color_space != JCS_BG_YCC) || + cinfo->num_components != 3 || + cinfo->out_color_space != JCS_RGB || + cinfo->out_color_components != RGB_PIXELSIZE || + cinfo->color_transform) + return FALSE; + /* and it only handles 2h1v or 2h2v sampling ratios */ + if (cinfo->comp_info[0].h_samp_factor != 2 || + cinfo->comp_info[1].h_samp_factor != 1 || + cinfo->comp_info[2].h_samp_factor != 1 || + cinfo->comp_info[0].v_samp_factor > 2 || + cinfo->comp_info[1].v_samp_factor != 1 || + cinfo->comp_info[2].v_samp_factor != 1) + return FALSE; + /* furthermore, it doesn't work if we've scaled the IDCTs differently */ + if (cinfo->comp_info[0].DCT_h_scaled_size != cinfo->min_DCT_h_scaled_size || + cinfo->comp_info[1].DCT_h_scaled_size != cinfo->min_DCT_h_scaled_size || + cinfo->comp_info[2].DCT_h_scaled_size != cinfo->min_DCT_h_scaled_size || + cinfo->comp_info[0].DCT_v_scaled_size != cinfo->min_DCT_v_scaled_size || + cinfo->comp_info[1].DCT_v_scaled_size != cinfo->min_DCT_v_scaled_size || + cinfo->comp_info[2].DCT_v_scaled_size != cinfo->min_DCT_v_scaled_size) + return FALSE; + /* ??? also need to test for upsample-time rescaling, when & if supported */ + return TRUE; /* by golly, it'll work... */ +#else + return FALSE; +#endif +} + + +/* + * Compute output image dimensions and related values. + * NOTE: this is exported for possible use by application. + * Hence it mustn't do anything that can't be done twice. + * Also note that it may be called before the master module is initialized! + */ + +GLOBAL(void) +jpeg_calc_output_dimensions (j_decompress_ptr cinfo) +/* Do computations that are needed before master selection phase. + * This function is used for full decompression. + */ +{ + int ci, i; + jpeg_component_info *compptr; + + /* Prevent application from calling me at wrong times */ + if (cinfo->global_state != DSTATE_READY) + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + + /* Compute core output image dimensions and DCT scaling choices. */ + jpeg_core_output_dimensions(cinfo); + +#ifdef IDCT_SCALING_SUPPORTED + + /* In selecting the actual DCT scaling for each component, we try to + * scale up the chroma components via IDCT scaling rather than upsampling. + * This saves time if the upsampler gets to use 1:1 scaling. + * Note this code adapts subsampling ratios which are powers of 2. + */ + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + int ssize = 1; + if (! cinfo->raw_data_out) + while (cinfo->min_DCT_h_scaled_size * ssize <= + (cinfo->do_fancy_upsampling ? DCTSIZE : DCTSIZE / 2) && + (cinfo->max_h_samp_factor % (compptr->h_samp_factor * ssize * 2)) == + 0) { + ssize = ssize * 2; + } + compptr->DCT_h_scaled_size = cinfo->min_DCT_h_scaled_size * ssize; + ssize = 1; + if (! cinfo->raw_data_out) + while (cinfo->min_DCT_v_scaled_size * ssize <= + (cinfo->do_fancy_upsampling ? DCTSIZE : DCTSIZE / 2) && + (cinfo->max_v_samp_factor % (compptr->v_samp_factor * ssize * 2)) == + 0) { + ssize = ssize * 2; + } + compptr->DCT_v_scaled_size = cinfo->min_DCT_v_scaled_size * ssize; + + /* We don't support IDCT ratios larger than 2. */ + if (compptr->DCT_h_scaled_size > compptr->DCT_v_scaled_size * 2) + compptr->DCT_h_scaled_size = compptr->DCT_v_scaled_size * 2; + else if (compptr->DCT_v_scaled_size > compptr->DCT_h_scaled_size * 2) + compptr->DCT_v_scaled_size = compptr->DCT_h_scaled_size * 2; + + /* Recompute downsampled dimensions of components; + * application needs to know these if using raw downsampled data. + */ + /* Size in samples, after IDCT scaling */ + compptr->downsampled_width = (JDIMENSION) + jdiv_round_up((long) cinfo->image_width * + (long) (compptr->h_samp_factor * compptr->DCT_h_scaled_size), + (long) (cinfo->max_h_samp_factor * cinfo->block_size)); + compptr->downsampled_height = (JDIMENSION) + jdiv_round_up((long) cinfo->image_height * + (long) (compptr->v_samp_factor * compptr->DCT_v_scaled_size), + (long) (cinfo->max_v_samp_factor * cinfo->block_size)); + } + +#endif /* IDCT_SCALING_SUPPORTED */ + + /* Report number of components in selected colorspace. */ + /* This should correspond to the actual code in the color conversion module. */ + switch (cinfo->out_color_space) { + case JCS_GRAYSCALE: + cinfo->out_color_components = 1; + break; + case JCS_RGB: + case JCS_BG_RGB: + cinfo->out_color_components = RGB_PIXELSIZE; + break; + default: /* YCCK <=> CMYK conversion or same colorspace as in file */ + i = 0; + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) + if (compptr->component_needed) + i++; /* count output color components */ + cinfo->out_color_components = i; + } + cinfo->output_components = (cinfo->quantize_colors ? 1 : + cinfo->out_color_components); + + /* See if upsampler will want to emit more than one row at a time */ + if (use_merged_upsample(cinfo)) + cinfo->rec_outbuf_height = cinfo->max_v_samp_factor; + else + cinfo->rec_outbuf_height = 1; +} + + +/* + * Several decompression processes need to range-limit values to the range + * 0..MAXJSAMPLE; the input value may fall somewhat outside this range + * due to noise introduced by quantization, roundoff error, etc. These + * processes are inner loops and need to be as fast as possible. On most + * machines, particularly CPUs with pipelines or instruction prefetch, + * a (subscript-check-less) C table lookup + * x = sample_range_limit[x]; + * is faster than explicit tests + * if (x < 0) x = 0; + * else if (x > MAXJSAMPLE) x = MAXJSAMPLE; + * These processes all use a common table prepared by the routine below. + * + * For most steps we can mathematically guarantee that the initial value + * of x is within 2*(MAXJSAMPLE+1) of the legal range, so a table running + * from -2*(MAXJSAMPLE+1) to 3*MAXJSAMPLE+2 is sufficient. But for the + * initial limiting step (just after the IDCT), a wildly out-of-range value + * is possible if the input data is corrupt. To avoid any chance of indexing + * off the end of memory and getting a bad-pointer trap, we perform the + * post-IDCT limiting thus: + * x = (sample_range_limit - SUBSET)[(x + CENTER) & MASK]; + * where MASK is 2 bits wider than legal sample data, ie 10 bits for 8-bit + * samples. Under normal circumstances this is more than enough range and + * a correct output will be generated; with bogus input data the mask will + * cause wraparound, and we will safely generate a bogus-but-in-range output. + * For the post-IDCT step, we want to convert the data from signed to unsigned + * representation by adding CENTERJSAMPLE at the same time that we limit it. + * This is accomplished with SUBSET = CENTER - CENTERJSAMPLE. + * + * Note that the table is allocated in near data space on PCs; it's small + * enough and used often enough to justify this. + */ + +LOCAL(void) +prepare_range_limit_table (j_decompress_ptr cinfo) +/* Allocate and fill in the sample_range_limit table */ +{ + JSAMPLE * table; + int i; + + table = (JSAMPLE *) (*cinfo->mem->alloc_small) ((j_common_ptr) cinfo, + JPOOL_IMAGE, (RANGE_CENTER * 2 + MAXJSAMPLE + 1) * SIZEOF(JSAMPLE)); + /* First segment of range limit table: limit[x] = 0 for x < 0 */ + MEMZERO(table, RANGE_CENTER * SIZEOF(JSAMPLE)); + table += RANGE_CENTER; /* allow negative subscripts of table */ + cinfo->sample_range_limit = table; + /* Main part of range limit table: limit[x] = x */ + for (i = 0; i <= MAXJSAMPLE; i++) + table[i] = (JSAMPLE) i; + /* End of range limit table: limit[x] = MAXJSAMPLE for x > MAXJSAMPLE */ + for (; i <= MAXJSAMPLE + RANGE_CENTER; i++) + table[i] = MAXJSAMPLE; +} + + +/* + * Master selection of decompression modules. + * This is done once at jpeg_start_decompress time. We determine + * which modules will be used and give them appropriate initialization calls. + * We also initialize the decompressor input side to begin consuming data. + * + * Since jpeg_read_header has finished, we know what is in the SOF + * and (first) SOS markers. We also have all the application parameter + * settings. + */ + +LOCAL(void) +master_selection (j_decompress_ptr cinfo) +{ + my_master_ptr master = (my_master_ptr) cinfo->master; + boolean use_c_buffer; + long samplesperrow; + JDIMENSION jd_samplesperrow; + + /* For now, precision must match compiled-in value... */ + if (cinfo->data_precision != BITS_IN_JSAMPLE) + ERREXIT1(cinfo, JERR_BAD_PRECISION, cinfo->data_precision); + + /* Initialize dimensions and other stuff */ + jpeg_calc_output_dimensions(cinfo); + prepare_range_limit_table(cinfo); + + /* Sanity check on image dimensions */ + if (cinfo->output_height <= 0 || cinfo->output_width <= 0 || + cinfo->out_color_components <= 0) + ERREXIT(cinfo, JERR_EMPTY_IMAGE); + + /* Width of an output scanline must be representable as JDIMENSION. */ + samplesperrow = (long) cinfo->output_width * (long) cinfo->out_color_components; + jd_samplesperrow = (JDIMENSION) samplesperrow; + if ((long) jd_samplesperrow != samplesperrow) + ERREXIT(cinfo, JERR_WIDTH_OVERFLOW); + + /* Initialize my private state */ + master->pass_number = 0; + master->using_merged_upsample = use_merged_upsample(cinfo); + + /* Color quantizer selection */ + master->quantizer_1pass = NULL; + master->quantizer_2pass = NULL; + /* No mode changes if not using buffered-image mode. */ + if (! cinfo->quantize_colors || ! cinfo->buffered_image) { + cinfo->enable_1pass_quant = FALSE; + cinfo->enable_external_quant = FALSE; + cinfo->enable_2pass_quant = FALSE; + } + if (cinfo->quantize_colors) { + if (cinfo->raw_data_out) + ERREXIT(cinfo, JERR_NOTIMPL); + /* 2-pass quantizer only works in 3-component color space. */ + if (cinfo->out_color_components != 3) { + cinfo->enable_1pass_quant = TRUE; + cinfo->enable_external_quant = FALSE; + cinfo->enable_2pass_quant = FALSE; + cinfo->colormap = NULL; + } else if (cinfo->colormap != NULL) { + cinfo->enable_external_quant = TRUE; + } else if (cinfo->two_pass_quantize) { + cinfo->enable_2pass_quant = TRUE; + } else { + cinfo->enable_1pass_quant = TRUE; + } + + if (cinfo->enable_1pass_quant) { +#ifdef QUANT_1PASS_SUPPORTED + jinit_1pass_quantizer(cinfo); + master->quantizer_1pass = cinfo->cquantize; +#else + ERREXIT(cinfo, JERR_NOT_COMPILED); +#endif + } + + /* We use the 2-pass code to map to external colormaps. */ + if (cinfo->enable_2pass_quant || cinfo->enable_external_quant) { +#ifdef QUANT_2PASS_SUPPORTED + jinit_2pass_quantizer(cinfo); + master->quantizer_2pass = cinfo->cquantize; +#else + ERREXIT(cinfo, JERR_NOT_COMPILED); +#endif + } + /* If both quantizers are initialized, the 2-pass one is left active; + * this is necessary for starting with quantization to an external map. + */ + } + + /* Post-processing: in particular, color conversion first */ + if (! cinfo->raw_data_out) { + if (master->using_merged_upsample) { +#ifdef UPSAMPLE_MERGING_SUPPORTED + jinit_merged_upsampler(cinfo); /* does color conversion too */ +#else + ERREXIT(cinfo, JERR_NOT_COMPILED); +#endif + } else { + jinit_color_deconverter(cinfo); + jinit_upsampler(cinfo); + } + jinit_d_post_controller(cinfo, cinfo->enable_2pass_quant); + } + /* Inverse DCT */ + jinit_inverse_dct(cinfo); + /* Entropy decoding: either Huffman or arithmetic coding. */ + if (cinfo->arith_code) + jinit_arith_decoder(cinfo); + else { + jinit_huff_decoder(cinfo); + } + + /* Initialize principal buffer controllers. */ + use_c_buffer = cinfo->inputctl->has_multiple_scans || cinfo->buffered_image; + jinit_d_coef_controller(cinfo, use_c_buffer); + + if (! cinfo->raw_data_out) + jinit_d_main_controller(cinfo, FALSE /* never need full buffer here */); + + /* We can now tell the memory manager to allocate virtual arrays. */ + (*cinfo->mem->realize_virt_arrays) ((j_common_ptr) cinfo); + + /* Initialize input side of decompressor to consume first scan. */ + (*cinfo->inputctl->start_input_pass) (cinfo); + +#ifdef D_MULTISCAN_FILES_SUPPORTED + /* If jpeg_start_decompress will read the whole file, initialize + * progress monitoring appropriately. The input step is counted + * as one pass. + */ + if (cinfo->progress != NULL && ! cinfo->buffered_image && + cinfo->inputctl->has_multiple_scans) { + int nscans; + /* Estimate number of scans to set pass_limit. */ + if (cinfo->progressive_mode) { + /* Arbitrarily estimate 2 interleaved DC scans + 3 AC scans/component. */ + nscans = 2 + 3 * cinfo->num_components; + } else { + /* For a nonprogressive multiscan file, estimate 1 scan per component. */ + nscans = cinfo->num_components; + } + cinfo->progress->pass_counter = 0L; + cinfo->progress->pass_limit = (long) cinfo->total_iMCU_rows * nscans; + cinfo->progress->completed_passes = 0; + cinfo->progress->total_passes = (cinfo->enable_2pass_quant ? 3 : 2); + /* Count the input pass as done */ + master->pass_number++; + } +#endif /* D_MULTISCAN_FILES_SUPPORTED */ +} + + +/* + * Per-pass setup. + * This is called at the beginning of each output pass. We determine which + * modules will be active during this pass and give them appropriate + * start_pass calls. We also set is_dummy_pass to indicate whether this + * is a "real" output pass or a dummy pass for color quantization. + * (In the latter case, jdapistd.c will crank the pass to completion.) + */ + +METHODDEF(void) +prepare_for_output_pass (j_decompress_ptr cinfo) +{ + my_master_ptr master = (my_master_ptr) cinfo->master; + + if (master->pub.is_dummy_pass) { +#ifdef QUANT_2PASS_SUPPORTED + /* Final pass of 2-pass quantization */ + master->pub.is_dummy_pass = FALSE; + (*cinfo->cquantize->start_pass) (cinfo, FALSE); + (*cinfo->post->start_pass) (cinfo, JBUF_CRANK_DEST); + (*cinfo->main->start_pass) (cinfo, JBUF_CRANK_DEST); +#else + ERREXIT(cinfo, JERR_NOT_COMPILED); +#endif /* QUANT_2PASS_SUPPORTED */ + } else { + if (cinfo->quantize_colors && cinfo->colormap == NULL) { + /* Select new quantization method */ + if (cinfo->two_pass_quantize && cinfo->enable_2pass_quant) { + cinfo->cquantize = master->quantizer_2pass; + master->pub.is_dummy_pass = TRUE; + } else if (cinfo->enable_1pass_quant) { + cinfo->cquantize = master->quantizer_1pass; + } else { + ERREXIT(cinfo, JERR_MODE_CHANGE); + } + } + (*cinfo->idct->start_pass) (cinfo); + (*cinfo->coef->start_output_pass) (cinfo); + if (! cinfo->raw_data_out) { + if (! master->using_merged_upsample) + (*cinfo->cconvert->start_pass) (cinfo); + (*cinfo->upsample->start_pass) (cinfo); + if (cinfo->quantize_colors) + (*cinfo->cquantize->start_pass) (cinfo, master->pub.is_dummy_pass); + (*cinfo->post->start_pass) (cinfo, + (master->pub.is_dummy_pass ? JBUF_SAVE_AND_PASS : JBUF_PASS_THRU)); + (*cinfo->main->start_pass) (cinfo, JBUF_PASS_THRU); + } + } + + /* Set up progress monitor's pass info if present */ + if (cinfo->progress != NULL) { + cinfo->progress->completed_passes = master->pass_number; + cinfo->progress->total_passes = master->pass_number + + (master->pub.is_dummy_pass ? 2 : 1); + /* In buffered-image mode, we assume one more output pass if EOI not + * yet reached, but no more passes if EOI has been reached. + */ + if (cinfo->buffered_image && ! cinfo->inputctl->eoi_reached) { + cinfo->progress->total_passes += (cinfo->enable_2pass_quant ? 2 : 1); + } + } +} + + +/* + * Finish up at end of an output pass. + */ + +METHODDEF(void) +finish_output_pass (j_decompress_ptr cinfo) +{ + my_master_ptr master = (my_master_ptr) cinfo->master; + + if (cinfo->quantize_colors) + (*cinfo->cquantize->finish_pass) (cinfo); + master->pass_number++; +} + + +#ifdef D_MULTISCAN_FILES_SUPPORTED + +/* + * Switch to a new external colormap between output passes. + */ + +GLOBAL(void) +jpeg_new_colormap (j_decompress_ptr cinfo) +{ + my_master_ptr master = (my_master_ptr) cinfo->master; + + /* Prevent application from calling me at wrong times */ + if (cinfo->global_state != DSTATE_BUFIMAGE) + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + + if (cinfo->quantize_colors && cinfo->enable_external_quant && + cinfo->colormap != NULL) { + /* Select 2-pass quantizer for external colormap use */ + cinfo->cquantize = master->quantizer_2pass; + /* Notify quantizer of colormap change */ + (*cinfo->cquantize->new_color_map) (cinfo); + master->pub.is_dummy_pass = FALSE; /* just in case */ + } else + ERREXIT(cinfo, JERR_MODE_CHANGE); +} + +#endif /* D_MULTISCAN_FILES_SUPPORTED */ + + +/* + * Initialize master decompression control and select active modules. + * This is performed at the start of jpeg_start_decompress. + */ + +GLOBAL(void) +jinit_master_decompress (j_decompress_ptr cinfo) +{ + my_master_ptr master; + + master = (my_master_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_decomp_master)); + cinfo->master = &master->pub; + master->pub.prepare_for_output_pass = prepare_for_output_pass; + master->pub.finish_output_pass = finish_output_pass; + + master->pub.is_dummy_pass = FALSE; + + master_selection(cinfo); +} diff --git a/cmake/externals/libjpeg/jdmerge.c b/cmake/externals/libjpeg/jdmerge.c new file mode 100644 index 000000000..0d16821be --- /dev/null +++ b/cmake/externals/libjpeg/jdmerge.c @@ -0,0 +1,437 @@ +/* + * jdmerge.c + * + * Copyright (C) 1994-1996, Thomas G. Lane. + * Modified 2013-2022 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains code for merged upsampling/color conversion. + * + * This file combines functions from jdsample.c and jdcolor.c; + * read those files first to understand what's going on. + * + * When the chroma components are to be upsampled by simple replication + * (ie, box filtering), we can save some work in color conversion by + * calculating all the output pixels corresponding to a pair of chroma + * samples at one time. In the conversion equations + * R = Y + K1 * Cr + * G = Y + K2 * Cb + K3 * Cr + * B = Y + K4 * Cb + * only the Y term varies among the group of pixels corresponding to a pair + * of chroma samples, so the rest of the terms can be calculated just once. + * At typical sampling ratios, this eliminates half or three-quarters + * of the multiplications needed for color conversion. + * + * This file currently provides implementations for the following cases: + * YCC => RGB color conversion only (YCbCr or BG_YCC). + * Sampling ratios of 2h1v or 2h2v. + * No scaling needed at upsample time. + * Corner-aligned (non-CCIR601) sampling alignment. + * Other special cases could be added, but in most applications these + * are the only common cases. (For uncommon cases we fall back on + * the more general code in jdsample.c and jdcolor.c.) + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + +#ifdef UPSAMPLE_MERGING_SUPPORTED + + +#if RANGE_BITS < 2 + /* Deliberate syntax err */ + Sorry, this code requires 2 or more range extension bits. +#endif + + +/* Private subobject */ + +typedef struct { + struct jpeg_upsampler pub; /* public fields */ + + /* Pointer to routine to do actual upsampling/conversion of one row group */ + JMETHOD(void, upmethod, (j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION in_row_group_ctr, + JSAMPARRAY output_buf)); + + /* Private state for YCC->RGB conversion */ + int * Cr_r_tab; /* => table for Cr to R conversion */ + int * Cb_b_tab; /* => table for Cb to B conversion */ + INT32 * Cr_g_tab; /* => table for Cr to G conversion */ + INT32 * Cb_g_tab; /* => table for Cb to G conversion */ + + /* For 2:1 vertical sampling, we produce two output rows at a time. + * We need a "spare" row buffer to hold the second output row if the + * application provides just a one-row buffer; we also use the spare + * to discard the dummy last row if the image height is odd. + */ + JSAMPROW spare_row; + boolean spare_full; /* T if spare buffer is occupied */ + + JDIMENSION out_row_width; /* samples per output row */ + JDIMENSION rows_to_go; /* counts rows remaining in image */ +} my_upsampler; + +typedef my_upsampler * my_upsample_ptr; + +#define SCALEBITS 16 /* speediest right-shift on some machines */ +#define ONE_HALF ((INT32) 1 << (SCALEBITS-1)) +#define FIX(x) ((INT32) ((x) * (1L<RGB and BG_YCC->RGB colorspace conversion. + * This is taken directly from jdcolor.c; see that file for more info. + */ + +LOCAL(void) +build_ycc_rgb_table (j_decompress_ptr cinfo) +/* Normal case, sYCC */ +{ + my_upsample_ptr upsample = (my_upsample_ptr) cinfo->upsample; + int i; + INT32 x; + SHIFT_TEMPS + + upsample->Cr_r_tab = (int *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(int)); + upsample->Cb_b_tab = (int *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(int)); + upsample->Cr_g_tab = (INT32 *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(INT32)); + upsample->Cb_g_tab = (INT32 *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(INT32)); + + for (i = 0, x = -CENTERJSAMPLE; i <= MAXJSAMPLE; i++, x++) { + /* i is the actual input pixel value, in the range 0..MAXJSAMPLE */ + /* The Cb or Cr value we are thinking of is x = i - CENTERJSAMPLE */ + /* Cr=>R value is nearest int to 1.402 * x */ + upsample->Cr_r_tab[i] = (int) DESCALE(FIX(1.402) * x, SCALEBITS); + /* Cb=>B value is nearest int to 1.772 * x */ + upsample->Cb_b_tab[i] = (int) DESCALE(FIX(1.772) * x, SCALEBITS); + /* Cr=>G value is scaled-up -0.714136286 * x */ + upsample->Cr_g_tab[i] = (- FIX(0.714136286)) * x; + /* Cb=>G value is scaled-up -0.344136286 * x */ + /* We also add in ONE_HALF so that need not do it in inner loop */ + upsample->Cb_g_tab[i] = (- FIX(0.344136286)) * x + ONE_HALF; + } +} + + +LOCAL(void) +build_bg_ycc_rgb_table (j_decompress_ptr cinfo) +/* Wide gamut case, bg-sYCC */ +{ + my_upsample_ptr upsample = (my_upsample_ptr) cinfo->upsample; + int i; + INT32 x; + SHIFT_TEMPS + + upsample->Cr_r_tab = (int *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(int)); + upsample->Cb_b_tab = (int *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(int)); + upsample->Cr_g_tab = (INT32 *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(INT32)); + upsample->Cb_g_tab = (INT32 *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(INT32)); + + for (i = 0, x = -CENTERJSAMPLE; i <= MAXJSAMPLE; i++, x++) { + /* i is the actual input pixel value, in the range 0..MAXJSAMPLE */ + /* The Cb or Cr value we are thinking of is x = i - CENTERJSAMPLE */ + /* Cr=>R value is nearest int to 2.804 * x */ + upsample->Cr_r_tab[i] = (int) DESCALE(FIX(2.804) * x, SCALEBITS); + /* Cb=>B value is nearest int to 3.544 * x */ + upsample->Cb_b_tab[i] = (int) DESCALE(FIX(3.544) * x, SCALEBITS); + /* Cr=>G value is scaled-up -1.428272572 * x */ + upsample->Cr_g_tab[i] = (- FIX(1.428272572)) * x; + /* Cb=>G value is scaled-up -0.688272572 * x */ + /* We also add in ONE_HALF so that need not do it in inner loop */ + upsample->Cb_g_tab[i] = (- FIX(0.688272572)) * x + ONE_HALF; + } +} + + +/* + * Initialize for an upsampling pass. + */ + +METHODDEF(void) +start_pass_merged_upsample (j_decompress_ptr cinfo) +{ + my_upsample_ptr upsample = (my_upsample_ptr) cinfo->upsample; + + /* Mark the spare buffer empty */ + upsample->spare_full = FALSE; + /* Initialize total-height counter for detecting bottom of image */ + upsample->rows_to_go = cinfo->output_height; +} + + +/* + * Control routine to do upsampling (and color conversion). + * + * The control routine just handles the row buffering considerations. + */ + +METHODDEF(void) +merged_2v_upsample (j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION *in_row_group_ctr, + JDIMENSION in_row_groups_avail, + JSAMPARRAY output_buf, JDIMENSION *out_row_ctr, + JDIMENSION out_rows_avail) +/* 2:1 vertical sampling case: may need a spare row. */ +{ + my_upsample_ptr upsample = (my_upsample_ptr) cinfo->upsample; + JSAMPROW work_ptrs[2]; + JDIMENSION num_rows; /* number of rows returned to caller */ + + if (upsample->spare_full) { + /* If we have a spare row saved from a previous cycle, just return it. */ + jcopy_sample_rows(& upsample->spare_row, output_buf + *out_row_ctr, + 1, upsample->out_row_width); + num_rows = 1; + upsample->spare_full = FALSE; + } else { + /* Figure number of rows to return to caller. */ + num_rows = 2; + /* Not more than the distance to the end of the image. */ + if (num_rows > upsample->rows_to_go) + num_rows = upsample->rows_to_go; + /* And not more than what the client can accept: */ + out_rows_avail -= *out_row_ctr; + if (num_rows > out_rows_avail) + num_rows = out_rows_avail; + /* Create output pointer array for upsampler. */ + work_ptrs[0] = output_buf[*out_row_ctr]; + if (num_rows > 1) { + work_ptrs[1] = output_buf[*out_row_ctr + 1]; + } else { + work_ptrs[1] = upsample->spare_row; + upsample->spare_full = TRUE; + } + /* Now do the upsampling. */ + (*upsample->upmethod) (cinfo, input_buf, *in_row_group_ctr, work_ptrs); + } + + /* Adjust counts */ + *out_row_ctr += num_rows; + upsample->rows_to_go -= num_rows; + /* When the buffer is emptied, declare this input row group consumed */ + if (! upsample->spare_full) + (*in_row_group_ctr)++; +} + + +METHODDEF(void) +merged_1v_upsample (j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION *in_row_group_ctr, + JDIMENSION in_row_groups_avail, + JSAMPARRAY output_buf, JDIMENSION *out_row_ctr, + JDIMENSION out_rows_avail) +/* 1:1 vertical sampling case: much easier, never need a spare row. */ +{ + my_upsample_ptr upsample = (my_upsample_ptr) cinfo->upsample; + + /* Just do the upsampling. */ + (*upsample->upmethod) (cinfo, input_buf, *in_row_group_ctr, + output_buf + *out_row_ctr); + /* Adjust counts */ + (*out_row_ctr)++; + (*in_row_group_ctr)++; +} + + +/* + * These are the routines invoked by the control routines to do + * the actual upsampling/conversion. One row group is processed per call. + * + * Note: since we may be writing directly into application-supplied buffers, + * we have to be honest about the output width; we can't assume the buffer + * has been rounded up to an even width. + */ + + +/* + * Upsample and color convert for the case of 2:1 horizontal and 1:1 vertical. + */ + +METHODDEF(void) +h2v1_merged_upsample (j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION in_row_group_ctr, + JSAMPARRAY output_buf) +{ + my_upsample_ptr upsample = (my_upsample_ptr) cinfo->upsample; + register int y, cred, cgreen, cblue; + int cb, cr; + register JSAMPROW outptr; + JSAMPROW inptr0, inptr1, inptr2; + JDIMENSION col; + /* copy these pointers into registers if possible */ + register JSAMPLE * range_limit = cinfo->sample_range_limit; + int * Crrtab = upsample->Cr_r_tab; + int * Cbbtab = upsample->Cb_b_tab; + INT32 * Crgtab = upsample->Cr_g_tab; + INT32 * Cbgtab = upsample->Cb_g_tab; + SHIFT_TEMPS + + inptr0 = input_buf[0][in_row_group_ctr]; + inptr1 = input_buf[1][in_row_group_ctr]; + inptr2 = input_buf[2][in_row_group_ctr]; + outptr = output_buf[0]; + /* Loop for each pair of output pixels */ + for (col = cinfo->output_width >> 1; col > 0; col--) { + /* Do the chroma part of the calculation */ + cb = GETJSAMPLE(*inptr1++); + cr = GETJSAMPLE(*inptr2++); + cgreen = (int) RIGHT_SHIFT(Cbgtab[cb] + Crgtab[cr], SCALEBITS); + cblue = Cbbtab[cb]; + cred = Crrtab[cr]; + /* Fetch 2 Y values and emit 2 pixels */ + y = GETJSAMPLE(*inptr0++); + outptr[RGB_RED] = range_limit[y + cred]; + outptr[RGB_GREEN] = range_limit[y + cgreen]; + outptr[RGB_BLUE] = range_limit[y + cblue]; + outptr += RGB_PIXELSIZE; + y = GETJSAMPLE(*inptr0++); + outptr[RGB_RED] = range_limit[y + cred]; + outptr[RGB_GREEN] = range_limit[y + cgreen]; + outptr[RGB_BLUE] = range_limit[y + cblue]; + outptr += RGB_PIXELSIZE; + } + /* If image width is odd, do the last output column separately */ + if (cinfo->output_width & 1) { + y = GETJSAMPLE(*inptr0); + cb = GETJSAMPLE(*inptr1); + cr = GETJSAMPLE(*inptr2); + outptr[RGB_RED] = range_limit[y + Crrtab[cr]]; + outptr[RGB_GREEN] = range_limit[y + + ((int) RIGHT_SHIFT(Cbgtab[cb] + Crgtab[cr], + SCALEBITS))]; + outptr[RGB_BLUE] = range_limit[y + Cbbtab[cb]]; + } +} + + +/* + * Upsample and color convert for the case of 2:1 horizontal and 2:1 vertical. + */ + +METHODDEF(void) +h2v2_merged_upsample (j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION in_row_group_ctr, + JSAMPARRAY output_buf) +{ + my_upsample_ptr upsample = (my_upsample_ptr) cinfo->upsample; + register int y, cred, cgreen, cblue; + int cb, cr; + register JSAMPROW outptr0, outptr1; + JSAMPROW inptr00, inptr01, inptr1, inptr2; + JDIMENSION col; + /* copy these pointers into registers if possible */ + register JSAMPLE * range_limit = cinfo->sample_range_limit; + int * Crrtab = upsample->Cr_r_tab; + int * Cbbtab = upsample->Cb_b_tab; + INT32 * Crgtab = upsample->Cr_g_tab; + INT32 * Cbgtab = upsample->Cb_g_tab; + SHIFT_TEMPS + + inptr00 = input_buf[0][in_row_group_ctr*2]; + inptr01 = input_buf[0][in_row_group_ctr*2 + 1]; + inptr1 = input_buf[1][in_row_group_ctr]; + inptr2 = input_buf[2][in_row_group_ctr]; + outptr0 = output_buf[0]; + outptr1 = output_buf[1]; + /* Loop for each group of output pixels */ + for (col = cinfo->output_width >> 1; col > 0; col--) { + /* Do the chroma part of the calculation */ + cb = GETJSAMPLE(*inptr1++); + cr = GETJSAMPLE(*inptr2++); + cgreen = (int) RIGHT_SHIFT(Cbgtab[cb] + Crgtab[cr], SCALEBITS); + cblue = Cbbtab[cb]; + cred = Crrtab[cr]; + /* Fetch 4 Y values and emit 4 pixels */ + y = GETJSAMPLE(*inptr00++); + outptr0[RGB_RED] = range_limit[y + cred]; + outptr0[RGB_GREEN] = range_limit[y + cgreen]; + outptr0[RGB_BLUE] = range_limit[y + cblue]; + outptr0 += RGB_PIXELSIZE; + y = GETJSAMPLE(*inptr00++); + outptr0[RGB_RED] = range_limit[y + cred]; + outptr0[RGB_GREEN] = range_limit[y + cgreen]; + outptr0[RGB_BLUE] = range_limit[y + cblue]; + outptr0 += RGB_PIXELSIZE; + y = GETJSAMPLE(*inptr01++); + outptr1[RGB_RED] = range_limit[y + cred]; + outptr1[RGB_GREEN] = range_limit[y + cgreen]; + outptr1[RGB_BLUE] = range_limit[y + cblue]; + outptr1 += RGB_PIXELSIZE; + y = GETJSAMPLE(*inptr01++); + outptr1[RGB_RED] = range_limit[y + cred]; + outptr1[RGB_GREEN] = range_limit[y + cgreen]; + outptr1[RGB_BLUE] = range_limit[y + cblue]; + outptr1 += RGB_PIXELSIZE; + } + /* If image width is odd, do the last output column separately */ + if (cinfo->output_width & 1) { + cb = GETJSAMPLE(*inptr1); + cr = GETJSAMPLE(*inptr2); + cgreen = (int) RIGHT_SHIFT(Cbgtab[cb] + Crgtab[cr], SCALEBITS); + cblue = Cbbtab[cb]; + cred = Crrtab[cr]; + y = GETJSAMPLE(*inptr00); + outptr0[RGB_RED] = range_limit[y + cred]; + outptr0[RGB_GREEN] = range_limit[y + cgreen]; + outptr0[RGB_BLUE] = range_limit[y + cblue]; + y = GETJSAMPLE(*inptr01); + outptr1[RGB_RED] = range_limit[y + cred]; + outptr1[RGB_GREEN] = range_limit[y + cgreen]; + outptr1[RGB_BLUE] = range_limit[y + cblue]; + } +} + + +/* + * Module initialization routine for merged upsampling/color conversion. + * + * NB: this is called under the conditions determined by use_merged_upsample() + * in jdmaster.c. That routine MUST correspond to the actual capabilities + * of this module; no safety checks are made here. + */ + +GLOBAL(void) +jinit_merged_upsampler (j_decompress_ptr cinfo) +{ + my_upsample_ptr upsample; + + upsample = (my_upsample_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_upsampler)); + cinfo->upsample = &upsample->pub; + upsample->pub.start_pass = start_pass_merged_upsample; + upsample->pub.need_context_rows = FALSE; + + upsample->out_row_width = cinfo->output_width * cinfo->out_color_components; + + if (cinfo->max_v_samp_factor == 2) { + upsample->pub.upsample = merged_2v_upsample; + upsample->upmethod = h2v2_merged_upsample; + /* Allocate a spare row buffer */ + upsample->spare_row = (JSAMPROW) (*cinfo->mem->alloc_large) + ((j_common_ptr) cinfo, JPOOL_IMAGE, + (size_t) upsample->out_row_width * SIZEOF(JSAMPLE)); + } else { + upsample->pub.upsample = merged_1v_upsample; + upsample->upmethod = h2v1_merged_upsample; + /* No spare row needed */ + upsample->spare_row = NULL; + } + + if (cinfo->jpeg_color_space == JCS_BG_YCC) + build_bg_ycc_rgb_table(cinfo); + else + build_ycc_rgb_table(cinfo); +} + +#endif /* UPSAMPLE_MERGING_SUPPORTED */ diff --git a/cmake/externals/libjpeg/jdpostct.c b/cmake/externals/libjpeg/jdpostct.c new file mode 100644 index 000000000..571563d72 --- /dev/null +++ b/cmake/externals/libjpeg/jdpostct.c @@ -0,0 +1,290 @@ +/* + * jdpostct.c + * + * Copyright (C) 1994-1996, Thomas G. Lane. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains the decompression postprocessing controller. + * This controller manages the upsampling, color conversion, and color + * quantization/reduction steps; specifically, it controls the buffering + * between upsample/color conversion and color quantization/reduction. + * + * If no color quantization/reduction is required, then this module has no + * work to do, and it just hands off to the upsample/color conversion code. + * An integrated upsample/convert/quantize process would replace this module + * entirely. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* Private buffer controller object */ + +typedef struct { + struct jpeg_d_post_controller pub; /* public fields */ + + /* Color quantization source buffer: this holds output data from + * the upsample/color conversion step to be passed to the quantizer. + * For two-pass color quantization, we need a full-image buffer; + * for one-pass operation, a strip buffer is sufficient. + */ + jvirt_sarray_ptr whole_image; /* virtual array, or NULL if one-pass */ + JSAMPARRAY buffer; /* strip buffer, or current strip of virtual */ + JDIMENSION strip_height; /* buffer size in rows */ + /* for two-pass mode only: */ + JDIMENSION starting_row; /* row # of first row in current strip */ + JDIMENSION next_row; /* index of next row to fill/empty in strip */ +} my_post_controller; + +typedef my_post_controller * my_post_ptr; + + +/* Forward declarations */ +METHODDEF(void) post_process_1pass + JPP((j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION *in_row_group_ctr, + JDIMENSION in_row_groups_avail, + JSAMPARRAY output_buf, JDIMENSION *out_row_ctr, + JDIMENSION out_rows_avail)); +#ifdef QUANT_2PASS_SUPPORTED +METHODDEF(void) post_process_prepass + JPP((j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION *in_row_group_ctr, + JDIMENSION in_row_groups_avail, + JSAMPARRAY output_buf, JDIMENSION *out_row_ctr, + JDIMENSION out_rows_avail)); +METHODDEF(void) post_process_2pass + JPP((j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION *in_row_group_ctr, + JDIMENSION in_row_groups_avail, + JSAMPARRAY output_buf, JDIMENSION *out_row_ctr, + JDIMENSION out_rows_avail)); +#endif + + +/* + * Initialize for a processing pass. + */ + +METHODDEF(void) +start_pass_dpost (j_decompress_ptr cinfo, J_BUF_MODE pass_mode) +{ + my_post_ptr post = (my_post_ptr) cinfo->post; + + switch (pass_mode) { + case JBUF_PASS_THRU: + if (cinfo->quantize_colors) { + /* Single-pass processing with color quantization. */ + post->pub.post_process_data = post_process_1pass; + /* We could be doing buffered-image output before starting a 2-pass + * color quantization; in that case, jinit_d_post_controller did not + * allocate a strip buffer. Use the virtual-array buffer as workspace. + */ + if (post->buffer == NULL) { + post->buffer = (*cinfo->mem->access_virt_sarray) + ((j_common_ptr) cinfo, post->whole_image, + (JDIMENSION) 0, post->strip_height, TRUE); + } + } else { + /* For single-pass processing without color quantization, + * I have no work to do; just call the upsampler directly. + */ + post->pub.post_process_data = cinfo->upsample->upsample; + } + break; +#ifdef QUANT_2PASS_SUPPORTED + case JBUF_SAVE_AND_PASS: + /* First pass of 2-pass quantization */ + if (post->whole_image == NULL) + ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); + post->pub.post_process_data = post_process_prepass; + break; + case JBUF_CRANK_DEST: + /* Second pass of 2-pass quantization */ + if (post->whole_image == NULL) + ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); + post->pub.post_process_data = post_process_2pass; + break; +#endif /* QUANT_2PASS_SUPPORTED */ + default: + ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); + break; + } + post->starting_row = post->next_row = 0; +} + + +/* + * Process some data in the one-pass (strip buffer) case. + * This is used for color precision reduction as well as one-pass quantization. + */ + +METHODDEF(void) +post_process_1pass (j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION *in_row_group_ctr, + JDIMENSION in_row_groups_avail, + JSAMPARRAY output_buf, JDIMENSION *out_row_ctr, + JDIMENSION out_rows_avail) +{ + my_post_ptr post = (my_post_ptr) cinfo->post; + JDIMENSION num_rows, max_rows; + + /* Fill the buffer, but not more than what we can dump out in one go. */ + /* Note we rely on the upsampler to detect bottom of image. */ + max_rows = out_rows_avail - *out_row_ctr; + if (max_rows > post->strip_height) + max_rows = post->strip_height; + num_rows = 0; + (*cinfo->upsample->upsample) (cinfo, + input_buf, in_row_group_ctr, in_row_groups_avail, + post->buffer, &num_rows, max_rows); + /* Quantize and emit data. */ + (*cinfo->cquantize->color_quantize) (cinfo, + post->buffer, output_buf + *out_row_ctr, (int) num_rows); + *out_row_ctr += num_rows; +} + + +#ifdef QUANT_2PASS_SUPPORTED + +/* + * Process some data in the first pass of 2-pass quantization. + */ + +METHODDEF(void) +post_process_prepass (j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION *in_row_group_ctr, + JDIMENSION in_row_groups_avail, + JSAMPARRAY output_buf, JDIMENSION *out_row_ctr, + JDIMENSION out_rows_avail) +{ + my_post_ptr post = (my_post_ptr) cinfo->post; + JDIMENSION old_next_row, num_rows; + + /* Reposition virtual buffer if at start of strip. */ + if (post->next_row == 0) { + post->buffer = (*cinfo->mem->access_virt_sarray) + ((j_common_ptr) cinfo, post->whole_image, + post->starting_row, post->strip_height, TRUE); + } + + /* Upsample some data (up to a strip height's worth). */ + old_next_row = post->next_row; + (*cinfo->upsample->upsample) (cinfo, + input_buf, in_row_group_ctr, in_row_groups_avail, + post->buffer, &post->next_row, post->strip_height); + + /* Allow quantizer to scan new data. No data is emitted, */ + /* but we advance out_row_ctr so outer loop can tell when we're done. */ + if (post->next_row > old_next_row) { + num_rows = post->next_row - old_next_row; + (*cinfo->cquantize->color_quantize) (cinfo, post->buffer + old_next_row, + (JSAMPARRAY) NULL, (int) num_rows); + *out_row_ctr += num_rows; + } + + /* Advance if we filled the strip. */ + if (post->next_row >= post->strip_height) { + post->starting_row += post->strip_height; + post->next_row = 0; + } +} + + +/* + * Process some data in the second pass of 2-pass quantization. + */ + +METHODDEF(void) +post_process_2pass (j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION *in_row_group_ctr, + JDIMENSION in_row_groups_avail, + JSAMPARRAY output_buf, JDIMENSION *out_row_ctr, + JDIMENSION out_rows_avail) +{ + my_post_ptr post = (my_post_ptr) cinfo->post; + JDIMENSION num_rows, max_rows; + + /* Reposition virtual buffer if at start of strip. */ + if (post->next_row == 0) { + post->buffer = (*cinfo->mem->access_virt_sarray) + ((j_common_ptr) cinfo, post->whole_image, + post->starting_row, post->strip_height, FALSE); + } + + /* Determine number of rows to emit. */ + num_rows = post->strip_height - post->next_row; /* available in strip */ + max_rows = out_rows_avail - *out_row_ctr; /* available in output area */ + if (num_rows > max_rows) + num_rows = max_rows; + /* We have to check bottom of image here, can't depend on upsampler. */ + max_rows = cinfo->output_height - post->starting_row; + if (num_rows > max_rows) + num_rows = max_rows; + + /* Quantize and emit data. */ + (*cinfo->cquantize->color_quantize) (cinfo, + post->buffer + post->next_row, output_buf + *out_row_ctr, + (int) num_rows); + *out_row_ctr += num_rows; + + /* Advance if we filled the strip. */ + post->next_row += num_rows; + if (post->next_row >= post->strip_height) { + post->starting_row += post->strip_height; + post->next_row = 0; + } +} + +#endif /* QUANT_2PASS_SUPPORTED */ + + +/* + * Initialize postprocessing controller. + */ + +GLOBAL(void) +jinit_d_post_controller (j_decompress_ptr cinfo, boolean need_full_buffer) +{ + my_post_ptr post; + + post = (my_post_ptr) + (*cinfo->mem->alloc_small) ((j_common_ptr) cinfo, JPOOL_IMAGE, + SIZEOF(my_post_controller)); + cinfo->post = (struct jpeg_d_post_controller *) post; + post->pub.start_pass = start_pass_dpost; + post->whole_image = NULL; /* flag for no virtual arrays */ + post->buffer = NULL; /* flag for no strip buffer */ + + /* Create the quantization buffer, if needed */ + if (cinfo->quantize_colors) { + /* The buffer strip height is max_v_samp_factor, which is typically + * an efficient number of rows for upsampling to return. + * (In the presence of output rescaling, we might want to be smarter?) + */ + post->strip_height = (JDIMENSION) cinfo->max_v_samp_factor; + if (need_full_buffer) { + /* Two-pass color quantization: need full-image storage. */ + /* We round up the number of rows to a multiple of the strip height. */ +#ifdef QUANT_2PASS_SUPPORTED + post->whole_image = (*cinfo->mem->request_virt_sarray) + ((j_common_ptr) cinfo, JPOOL_IMAGE, FALSE, + cinfo->output_width * cinfo->out_color_components, + (JDIMENSION) jround_up((long) cinfo->output_height, + (long) post->strip_height), + post->strip_height); +#else + ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); +#endif /* QUANT_2PASS_SUPPORTED */ + } else { + /* One-pass color quantization: just make a strip buffer. */ + post->buffer = (*cinfo->mem->alloc_sarray) + ((j_common_ptr) cinfo, JPOOL_IMAGE, + cinfo->output_width * cinfo->out_color_components, + post->strip_height); + } + } +} diff --git a/cmake/externals/libjpeg/jdsample.c b/cmake/externals/libjpeg/jdsample.c new file mode 100644 index 000000000..15afeafe3 --- /dev/null +++ b/cmake/externals/libjpeg/jdsample.c @@ -0,0 +1,341 @@ +/* + * jdsample.c + * + * Copyright (C) 1991-1996, Thomas G. Lane. + * Modified 2002-2020 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains upsampling routines. + * + * Upsampling input data is counted in "row groups". A row group + * is defined to be (v_samp_factor * DCT_v_scaled_size / min_DCT_v_scaled_size) + * sample rows of each component. Upsampling will normally produce + * max_v_samp_factor pixel rows from each row group (but this could vary + * if the upsampler is applying a scale factor of its own). + * + * An excellent reference for image resampling is + * Digital Image Warping, George Wolberg, 1990. + * Pub. by IEEE Computer Society Press, Los Alamitos, CA. ISBN 0-8186-8944-7. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* Pointer to routine to upsample a single component */ +typedef JMETHOD(void, upsample1_ptr, + (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JSAMPARRAY input_data, JSAMPIMAGE output_data_ptr)); + +/* Private subobject */ + +typedef struct { + struct jpeg_upsampler pub; /* public fields */ + + /* Color conversion buffer. When using separate upsampling and color + * conversion steps, this buffer holds one upsampled row group until it + * has been color converted and output. + * Note: we do not allocate any storage for component(s) which are full-size, + * ie do not need rescaling. The corresponding entry of color_buf[] is + * simply set to point to the input data array, thereby avoiding copying. + */ + JSAMPARRAY color_buf[MAX_COMPONENTS]; + + /* Per-component upsampling method pointers */ + upsample1_ptr methods[MAX_COMPONENTS]; + + int next_row_out; /* counts rows emitted from color_buf */ + JDIMENSION rows_to_go; /* counts rows remaining in image */ + + /* Height of an input row group for each component. */ + int rowgroup_height[MAX_COMPONENTS]; + + /* These arrays save pixel expansion factors so that int_expand need not + * recompute them each time. They are unused for other upsampling methods. + */ + UINT8 h_expand[MAX_COMPONENTS]; + UINT8 v_expand[MAX_COMPONENTS]; +} my_upsampler; + +typedef my_upsampler * my_upsample_ptr; + + +/* + * Initialize for an upsampling pass. + */ + +METHODDEF(void) +start_pass_upsample (j_decompress_ptr cinfo) +{ + my_upsample_ptr upsample = (my_upsample_ptr) cinfo->upsample; + + /* Mark the conversion buffer empty */ + upsample->next_row_out = cinfo->max_v_samp_factor; + /* Initialize total-height counter for detecting bottom of image */ + upsample->rows_to_go = cinfo->output_height; +} + + +/* + * Control routine to do upsampling (and color conversion). + * + * In this version we upsample each component independently. + * We upsample one row group into the conversion buffer, then apply + * color conversion a row at a time. + */ + +METHODDEF(void) +sep_upsample (j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION *in_row_group_ctr, + JDIMENSION in_row_groups_avail, + JSAMPARRAY output_buf, JDIMENSION *out_row_ctr, + JDIMENSION out_rows_avail) +{ + my_upsample_ptr upsample = (my_upsample_ptr) cinfo->upsample; + int ci; + jpeg_component_info * compptr; + JDIMENSION num_rows; + + /* Fill the conversion buffer, if it's empty */ + if (upsample->next_row_out >= cinfo->max_v_samp_factor) { + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + /* Don't bother to upsample an uninteresting component. */ + if (! compptr->component_needed) + continue; + /* Invoke per-component upsample method. Notice we pass a POINTER + * to color_buf[ci], so that fullsize_upsample can change it. + */ + (*upsample->methods[ci]) (cinfo, compptr, + input_buf[ci] + (*in_row_group_ctr * upsample->rowgroup_height[ci]), + upsample->color_buf + ci); + } + upsample->next_row_out = 0; + } + + /* Color-convert and emit rows */ + + /* How many we have in the buffer: */ + num_rows = (JDIMENSION) (cinfo->max_v_samp_factor - upsample->next_row_out); + /* Not more than the distance to the end of the image. Need this test + * in case the image height is not a multiple of max_v_samp_factor: + */ + if (num_rows > upsample->rows_to_go) + num_rows = upsample->rows_to_go; + /* And not more than what the client can accept: */ + out_rows_avail -= *out_row_ctr; + if (num_rows > out_rows_avail) + num_rows = out_rows_avail; + + (*cinfo->cconvert->color_convert) (cinfo, upsample->color_buf, + (JDIMENSION) upsample->next_row_out, + output_buf + *out_row_ctr, + (int) num_rows); + + /* Adjust counts */ + *out_row_ctr += num_rows; + upsample->rows_to_go -= num_rows; + upsample->next_row_out += num_rows; + /* When the buffer is emptied, declare this input row group consumed */ + if (upsample->next_row_out >= cinfo->max_v_samp_factor) + (*in_row_group_ctr)++; +} + + +/* + * These are the routines invoked by sep_upsample to upsample pixel values + * of a single component. One row group is processed per call. + */ + + +/* + * For full-size components, we just make color_buf[ci] point at the + * input buffer, and thus avoid copying any data. Note that this is + * safe only because sep_upsample doesn't declare the input row group + * "consumed" until we are done color converting and emitting it. + */ + +METHODDEF(void) +fullsize_upsample (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JSAMPARRAY input_data, JSAMPIMAGE output_data_ptr) +{ + *output_data_ptr = input_data; +} + + +/* + * This version handles any integral sampling ratios. + * This is not used for typical JPEG files, so it need not be fast. + * Nor, for that matter, is it particularly accurate: the algorithm is + * simple replication of the input pixel onto the corresponding output + * pixels. The hi-falutin sampling literature refers to this as a + * "box filter". A box filter tends to introduce visible artifacts, + * so if you are actually going to use 3:1 or 4:1 sampling ratios + * you would be well advised to improve this code. + */ + +METHODDEF(void) +int_upsample (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JSAMPARRAY input_data, JSAMPIMAGE output_data_ptr) +{ + my_upsample_ptr upsample = (my_upsample_ptr) cinfo->upsample; + JSAMPARRAY output_data, output_end; + register JSAMPROW inptr, outptr; + register JSAMPLE invalue; + register int h; + JSAMPROW outend; + int h_expand, v_expand; + + h_expand = upsample->h_expand[compptr->component_index]; + v_expand = upsample->v_expand[compptr->component_index]; + + output_data = *output_data_ptr; + output_end = output_data + cinfo->max_v_samp_factor; + for (; output_data < output_end; output_data += v_expand) { + /* Generate one output row with proper horizontal expansion */ + inptr = *input_data++; + outptr = *output_data; + outend = outptr + cinfo->output_width; + while (outptr < outend) { + invalue = *inptr++; /* don't need GETJSAMPLE() here */ + for (h = h_expand; h > 0; h--) { + *outptr++ = invalue; + } + } + /* Generate any additional output rows by duplicating the first one */ + if (v_expand > 1) { + jcopy_sample_rows(output_data, output_data + 1, + v_expand - 1, cinfo->output_width); + } + } +} + + +/* + * Fast processing for the common case of 2:1 horizontal and 1:1 vertical. + * It's still a box filter. + */ + +METHODDEF(void) +h2v1_upsample (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JSAMPARRAY input_data, JSAMPIMAGE output_data_ptr) +{ + JSAMPARRAY output_data = *output_data_ptr; + register JSAMPROW inptr, outptr; + register JSAMPLE invalue; + JSAMPROW outend; + int outrow; + + for (outrow = 0; outrow < cinfo->max_v_samp_factor; outrow++) { + inptr = input_data[outrow]; + outptr = output_data[outrow]; + outend = outptr + cinfo->output_width; + while (outptr < outend) { + invalue = *inptr++; /* don't need GETJSAMPLE() here */ + *outptr++ = invalue; + *outptr++ = invalue; + } + } +} + + +/* + * Fast processing for the common case of 2:1 horizontal and 2:1 vertical. + * It's still a box filter. + */ + +METHODDEF(void) +h2v2_upsample (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JSAMPARRAY input_data, JSAMPIMAGE output_data_ptr) +{ + JSAMPARRAY output_data, output_end; + register JSAMPROW inptr, outptr; + register JSAMPLE invalue; + JSAMPROW outend; + + output_data = *output_data_ptr; + output_end = output_data + cinfo->max_v_samp_factor; + for (; output_data < output_end; output_data += 2) { + inptr = *input_data++; + outptr = *output_data; + outend = outptr + cinfo->output_width; + while (outptr < outend) { + invalue = *inptr++; /* don't need GETJSAMPLE() here */ + *outptr++ = invalue; + *outptr++ = invalue; + } + jcopy_sample_rows(output_data, output_data + 1, + 1, cinfo->output_width); + } +} + + +/* + * Module initialization routine for upsampling. + */ + +GLOBAL(void) +jinit_upsampler (j_decompress_ptr cinfo) +{ + my_upsample_ptr upsample; + int ci; + jpeg_component_info * compptr; + int h_in_group, v_in_group, h_out_group, v_out_group; + + upsample = (my_upsample_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_upsampler)); + cinfo->upsample = &upsample->pub; + upsample->pub.start_pass = start_pass_upsample; + upsample->pub.upsample = sep_upsample; + upsample->pub.need_context_rows = FALSE; /* until we find out differently */ + + if (cinfo->CCIR601_sampling) /* this isn't supported */ + ERREXIT(cinfo, JERR_CCIR601_NOTIMPL); + + /* Verify we can handle the sampling factors, select per-component methods, + * and create storage as needed. + */ + for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; + ci++, compptr++) { + /* Don't bother to upsample an uninteresting component. */ + if (! compptr->component_needed) + continue; + /* Compute size of an "input group" after IDCT scaling. This many samples + * are to be converted to max_h_samp_factor * max_v_samp_factor pixels. + */ + h_in_group = (compptr->h_samp_factor * compptr->DCT_h_scaled_size) / + cinfo->min_DCT_h_scaled_size; + v_in_group = (compptr->v_samp_factor * compptr->DCT_v_scaled_size) / + cinfo->min_DCT_v_scaled_size; + h_out_group = cinfo->max_h_samp_factor; + v_out_group = cinfo->max_v_samp_factor; + upsample->rowgroup_height[ci] = v_in_group; /* save for use later */ + if (h_in_group == h_out_group && v_in_group == v_out_group) { + /* Fullsize components can be processed without any work. */ + upsample->methods[ci] = fullsize_upsample; + continue; /* don't need to allocate buffer */ + } + if (h_in_group * 2 == h_out_group && v_in_group == v_out_group) { + /* Special case for 2h1v upsampling */ + upsample->methods[ci] = h2v1_upsample; + } else if (h_in_group * 2 == h_out_group && + v_in_group * 2 == v_out_group) { + /* Special case for 2h2v upsampling */ + upsample->methods[ci] = h2v2_upsample; + } else if ((h_out_group % h_in_group) == 0 && + (v_out_group % v_in_group) == 0) { + /* Generic integral-factors upsampling method */ + upsample->methods[ci] = int_upsample; + upsample->h_expand[ci] = (UINT8) (h_out_group / h_in_group); + upsample->v_expand[ci] = (UINT8) (v_out_group / v_in_group); + } else + ERREXIT(cinfo, JERR_FRACT_SAMPLE_NOTIMPL); + upsample->color_buf[ci] = (*cinfo->mem->alloc_sarray) + ((j_common_ptr) cinfo, JPOOL_IMAGE, + (JDIMENSION) jround_up((long) cinfo->output_width, + (long) cinfo->max_h_samp_factor), + (JDIMENSION) cinfo->max_v_samp_factor); + } +} diff --git a/cmake/externals/libjpeg/jdtrans.c b/cmake/externals/libjpeg/jdtrans.c new file mode 100644 index 000000000..22dd47fb5 --- /dev/null +++ b/cmake/externals/libjpeg/jdtrans.c @@ -0,0 +1,140 @@ +/* + * jdtrans.c + * + * Copyright (C) 1995-1997, Thomas G. Lane. + * Modified 2000-2009 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains library routines for transcoding decompression, + * that is, reading raw DCT coefficient arrays from an input JPEG file. + * The routines in jdapimin.c will also be needed by a transcoder. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* Forward declarations */ +LOCAL(void) transdecode_master_selection JPP((j_decompress_ptr cinfo)); + + +/* + * Read the coefficient arrays from a JPEG file. + * jpeg_read_header must be completed before calling this. + * + * The entire image is read into a set of virtual coefficient-block arrays, + * one per component. The return value is a pointer to the array of + * virtual-array descriptors. These can be manipulated directly via the + * JPEG memory manager, or handed off to jpeg_write_coefficients(). + * To release the memory occupied by the virtual arrays, call + * jpeg_finish_decompress() when done with the data. + * + * An alternative usage is to simply obtain access to the coefficient arrays + * during a buffered-image-mode decompression operation. This is allowed + * after any jpeg_finish_output() call. The arrays can be accessed until + * jpeg_finish_decompress() is called. (Note that any call to the library + * may reposition the arrays, so don't rely on access_virt_barray() results + * to stay valid across library calls.) + * + * Returns NULL if suspended. This case need be checked only if + * a suspending data source is used. + */ + +GLOBAL(jvirt_barray_ptr *) +jpeg_read_coefficients (j_decompress_ptr cinfo) +{ + if (cinfo->global_state == DSTATE_READY) { + /* First call: initialize active modules */ + transdecode_master_selection(cinfo); + cinfo->global_state = DSTATE_RDCOEFS; + } + if (cinfo->global_state == DSTATE_RDCOEFS) { + /* Absorb whole file into the coef buffer */ + for (;;) { + int retcode; + /* Call progress monitor hook if present */ + if (cinfo->progress != NULL) + (*cinfo->progress->progress_monitor) ((j_common_ptr) cinfo); + /* Absorb some more input */ + retcode = (*cinfo->inputctl->consume_input) (cinfo); + if (retcode == JPEG_SUSPENDED) + return NULL; + if (retcode == JPEG_REACHED_EOI) + break; + /* Advance progress counter if appropriate */ + if (cinfo->progress != NULL && + (retcode == JPEG_ROW_COMPLETED || retcode == JPEG_REACHED_SOS)) { + if (++cinfo->progress->pass_counter >= cinfo->progress->pass_limit) { + /* startup underestimated number of scans; ratchet up one scan */ + cinfo->progress->pass_limit += (long) cinfo->total_iMCU_rows; + } + } + } + /* Set state so that jpeg_finish_decompress does the right thing */ + cinfo->global_state = DSTATE_STOPPING; + } + /* At this point we should be in state DSTATE_STOPPING if being used + * standalone, or in state DSTATE_BUFIMAGE if being invoked to get access + * to the coefficients during a full buffered-image-mode decompression. + */ + if ((cinfo->global_state == DSTATE_STOPPING || + cinfo->global_state == DSTATE_BUFIMAGE) && cinfo->buffered_image) { + return cinfo->coef->coef_arrays; + } + /* Oops, improper usage */ + ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); + return NULL; /* keep compiler happy */ +} + + +/* + * Master selection of decompression modules for transcoding. + * This substitutes for jdmaster.c's initialization of the full decompressor. + */ + +LOCAL(void) +transdecode_master_selection (j_decompress_ptr cinfo) +{ + /* This is effectively a buffered-image operation. */ + cinfo->buffered_image = TRUE; + + /* Compute output image dimensions and related values. */ + jpeg_core_output_dimensions(cinfo); + + /* Entropy decoding: either Huffman or arithmetic coding. */ + if (cinfo->arith_code) + jinit_arith_decoder(cinfo); + else { + jinit_huff_decoder(cinfo); + } + + /* Always get a full-image coefficient buffer. */ + jinit_d_coef_controller(cinfo, TRUE); + + /* We can now tell the memory manager to allocate virtual arrays. */ + (*cinfo->mem->realize_virt_arrays) ((j_common_ptr) cinfo); + + /* Initialize input side of decompressor to consume first scan. */ + (*cinfo->inputctl->start_input_pass) (cinfo); + + /* Initialize progress monitoring. */ + if (cinfo->progress != NULL) { + int nscans; + /* Estimate number of scans to set pass_limit. */ + if (cinfo->progressive_mode) { + /* Arbitrarily estimate 2 interleaved DC scans + 3 AC scans/component. */ + nscans = 2 + 3 * cinfo->num_components; + } else if (cinfo->inputctl->has_multiple_scans) { + /* For a nonprogressive multiscan file, estimate 1 scan per component. */ + nscans = cinfo->num_components; + } else { + nscans = 1; + } + cinfo->progress->pass_counter = 0L; + cinfo->progress->pass_limit = (long) cinfo->total_iMCU_rows * nscans; + cinfo->progress->completed_passes = 0; + cinfo->progress->total_passes = 1; + } +} diff --git a/cmake/externals/libjpeg/jerror.c b/cmake/externals/libjpeg/jerror.c new file mode 100644 index 000000000..7163af699 --- /dev/null +++ b/cmake/externals/libjpeg/jerror.c @@ -0,0 +1,253 @@ +/* + * jerror.c + * + * Copyright (C) 1991-1998, Thomas G. Lane. + * Modified 2012-2015 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains simple error-reporting and trace-message routines. + * These are suitable for Unix-like systems and others where writing to + * stderr is the right thing to do. Many applications will want to replace + * some or all of these routines. + * + * If you define USE_WINDOWS_MESSAGEBOX in jconfig.h or in the makefile, + * you get a Windows-specific hack to display error messages in a dialog box. + * It ain't much, but it beats dropping error messages into the bit bucket, + * which is what happens to output to stderr under most Windows C compilers. + * + * These routines are used by both the compression and decompression code. + */ + +#ifdef USE_WINDOWS_MESSAGEBOX +#include +#endif + +/* this is not a core library module, so it doesn't define JPEG_INTERNALS */ +#include "jinclude.h" +#include "jpeglib.h" +#include "jversion.h" +#include "jerror.h" + +#ifndef EXIT_FAILURE /* define exit() codes if not provided */ +#define EXIT_FAILURE 1 +#endif + + +/* + * Create the message string table. + * We do this from the master message list in jerror.h by re-reading + * jerror.h with a suitable definition for macro JMESSAGE. + * The message table is made an external symbol just in case any applications + * want to refer to it directly. + */ + +#ifdef NEED_SHORT_EXTERNAL_NAMES +#define jpeg_std_message_table jMsgTable +#endif + +#define JMESSAGE(code,string) string , + +const char * const jpeg_std_message_table[] = { +#include "jerror.h" + NULL +}; + + +/* + * Error exit handler: must not return to caller. + * + * Applications may override this if they want to get control back after + * an error. Typically one would longjmp somewhere instead of exiting. + * The setjmp buffer can be made a private field within an expanded error + * handler object. Note that the info needed to generate an error message + * is stored in the error object, so you can generate the message now or + * later, at your convenience. + * You should make sure that the JPEG object is cleaned up (with jpeg_abort + * or jpeg_destroy) at some point. + */ + +METHODDEF(noreturn_t) +error_exit (j_common_ptr cinfo) +{ + /* Always display the message */ + (*cinfo->err->output_message) (cinfo); + + /* Let the memory manager delete any temp files before we die */ + jpeg_destroy(cinfo); + + exit(EXIT_FAILURE); +} + + +/* + * Actual output of an error or trace message. + * Applications may override this method to send JPEG messages somewhere + * other than stderr. + * + * On Windows, printing to stderr is generally completely useless, + * so we provide optional code to produce an error-dialog popup. + * Most Windows applications will still prefer to override this routine, + * but if they don't, it'll do something at least marginally useful. + * + * NOTE: to use the library in an environment that doesn't support the + * C stdio library, you may have to delete the call to fprintf() entirely, + * not just not use this routine. + */ + +METHODDEF(void) +output_message (j_common_ptr cinfo) +{ + char buffer[JMSG_LENGTH_MAX]; + + /* Create the message */ + (*cinfo->err->format_message) (cinfo, buffer); + +#ifdef USE_WINDOWS_MESSAGEBOX + /* Display it in a message dialog box */ + MessageBox(GetActiveWindow(), buffer, "JPEG Library Error", + MB_OK | MB_ICONERROR); +#else + /* Send it to stderr, adding a newline */ + fprintf(stderr, "%s\n", buffer); +#endif +} + + +/* + * Decide whether to emit a trace or warning message. + * msg_level is one of: + * -1: recoverable corrupt-data warning, may want to abort. + * 0: important advisory messages (always display to user). + * 1: first level of tracing detail. + * 2,3,...: successively more detailed tracing messages. + * An application might override this method if it wanted to abort on warnings + * or change the policy about which messages to display. + */ + +METHODDEF(void) +emit_message (j_common_ptr cinfo, int msg_level) +{ + struct jpeg_error_mgr * err = cinfo->err; + + if (msg_level < 0) { + /* It's a warning message. Since corrupt files may generate many warnings, + * the policy implemented here is to show only the first warning, + * unless trace_level >= 3. + */ + if (err->num_warnings == 0 || err->trace_level >= 3) + (*err->output_message) (cinfo); + /* Always count warnings in num_warnings. */ + err->num_warnings++; + } else { + /* It's a trace message. Show it if trace_level >= msg_level. */ + if (err->trace_level >= msg_level) + (*err->output_message) (cinfo); + } +} + + +/* + * Format a message string for the most recent JPEG error or message. + * The message is stored into buffer, which should be at least JMSG_LENGTH_MAX + * characters. Note that no '\n' character is added to the string. + * Few applications should need to override this method. + */ + +METHODDEF(void) +format_message (j_common_ptr cinfo, char * buffer) +{ + struct jpeg_error_mgr * err = cinfo->err; + int msg_code = err->msg_code; + const char * msgtext = NULL; + const char * msgptr; + char ch; + boolean isstring; + + /* Look up message string in proper table */ + if (msg_code > 0 && msg_code <= err->last_jpeg_message) { + msgtext = err->jpeg_message_table[msg_code]; + } else if (err->addon_message_table != NULL && + msg_code >= err->first_addon_message && + msg_code <= err->last_addon_message) { + msgtext = err->addon_message_table[msg_code - err->first_addon_message]; + } + + /* Defend against bogus message number */ + if (msgtext == NULL) { + err->msg_parm.i[0] = msg_code; + msgtext = err->jpeg_message_table[0]; + } + + /* Check for string parameter, as indicated by %s in the message text */ + isstring = FALSE; + msgptr = msgtext; + while ((ch = *msgptr++) != '\0') { + if (ch == '%') { + if (*msgptr == 's') isstring = TRUE; + break; + } + } + + /* Format the message into the passed buffer */ + if (isstring) + sprintf(buffer, msgtext, err->msg_parm.s); + else + sprintf(buffer, msgtext, + err->msg_parm.i[0], err->msg_parm.i[1], + err->msg_parm.i[2], err->msg_parm.i[3], + err->msg_parm.i[4], err->msg_parm.i[5], + err->msg_parm.i[6], err->msg_parm.i[7]); +} + + +/* + * Reset error state variables at start of a new image. + * This is called during compression startup to reset trace/error + * processing to default state, without losing any application-specific + * method pointers. An application might possibly want to override + * this method if it has additional error processing state. + */ + +METHODDEF(void) +reset_error_mgr (j_common_ptr cinfo) +{ + cinfo->err->num_warnings = 0; + /* trace_level is not reset since it is an application-supplied parameter */ + cinfo->err->msg_code = 0; /* may be useful as a flag for "no error" */ +} + + +/* + * Fill in the standard error-handling methods in a jpeg_error_mgr object. + * Typical call is: + * struct jpeg_compress_struct cinfo; + * struct jpeg_error_mgr err; + * + * cinfo.err = jpeg_std_error(&err); + * after which the application may override some of the methods. + */ + +GLOBAL(struct jpeg_error_mgr *) +jpeg_std_error (struct jpeg_error_mgr * err) +{ + err->error_exit = error_exit; + err->emit_message = emit_message; + err->output_message = output_message; + err->format_message = format_message; + err->reset_error_mgr = reset_error_mgr; + + err->trace_level = 0; /* default = no tracing */ + err->num_warnings = 0; /* no warnings emitted yet */ + err->msg_code = 0; /* may be useful as a flag for "no error" */ + + /* Initialize message table pointers */ + err->jpeg_message_table = jpeg_std_message_table; + err->last_jpeg_message = (int) JMSG_LASTMSGCODE - 1; + + err->addon_message_table = NULL; + err->first_addon_message = 0; /* for safety */ + err->last_addon_message = 0; + + return err; +} diff --git a/cmake/externals/libjpeg/jerror.h b/cmake/externals/libjpeg/jerror.h new file mode 100644 index 000000000..db608b9c0 --- /dev/null +++ b/cmake/externals/libjpeg/jerror.h @@ -0,0 +1,304 @@ +/* + * jerror.h + * + * Copyright (C) 1994-1997, Thomas G. Lane. + * Modified 1997-2018 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file defines the error and message codes for the JPEG library. + * Edit this file to add new codes, or to translate the message strings to + * some other language. + * A set of error-reporting macros are defined too. Some applications using + * the JPEG library may wish to include this file to get the error codes + * and/or the macros. + */ + +/* + * To define the enum list of message codes, include this file without + * defining macro JMESSAGE. To create a message string table, include it + * again with a suitable JMESSAGE definition (see jerror.c for an example). + */ +#ifndef JMESSAGE +#ifndef JERROR_H +/* First time through, define the enum list */ +#define JMAKE_ENUM_LIST +#else +/* Repeated inclusions of this file are no-ops unless JMESSAGE is defined */ +#define JMESSAGE(code,string) +#endif /* JERROR_H */ +#endif /* JMESSAGE */ + +#ifdef JMAKE_ENUM_LIST + +typedef enum { + +#define JMESSAGE(code,string) code , + +#endif /* JMAKE_ENUM_LIST */ + +JMESSAGE(JMSG_NOMESSAGE, "Bogus message code %d") /* Must be first entry! */ + +/* For maintenance convenience, list is alphabetical by message code name */ +JMESSAGE(JERR_BAD_ALIGN_TYPE, "ALIGN_TYPE is wrong, please fix") +JMESSAGE(JERR_BAD_ALLOC_CHUNK, "MAX_ALLOC_CHUNK is wrong, please fix") +JMESSAGE(JERR_BAD_BUFFER_MODE, "Bogus buffer control mode") +JMESSAGE(JERR_BAD_COMPONENT_ID, "Invalid component ID %d in SOS") +JMESSAGE(JERR_BAD_CROP_SPEC, "Invalid crop request") +JMESSAGE(JERR_BAD_DCT_COEF, "DCT coefficient out of range") +JMESSAGE(JERR_BAD_DCTSIZE, "DCT scaled block size %dx%d not supported") +JMESSAGE(JERR_BAD_DROP_SAMPLING, + "Component index %d: mismatching sampling ratio %d:%d, %d:%d, %c") +JMESSAGE(JERR_BAD_HUFF_TABLE, "Bogus Huffman table definition") +JMESSAGE(JERR_BAD_IN_COLORSPACE, "Bogus input colorspace") +JMESSAGE(JERR_BAD_J_COLORSPACE, "Bogus JPEG colorspace") +JMESSAGE(JERR_BAD_LENGTH, "Bogus marker length") +JMESSAGE(JERR_BAD_LIB_VERSION, + "Wrong JPEG library version: library is %d, caller expects %d") +JMESSAGE(JERR_BAD_MCU_SIZE, "Sampling factors too large for interleaved scan") +JMESSAGE(JERR_BAD_POOL_ID, "Invalid memory pool code %d") +JMESSAGE(JERR_BAD_PRECISION, "Unsupported JPEG data precision %d") +JMESSAGE(JERR_BAD_PROGRESSION, + "Invalid progressive parameters Ss=%d Se=%d Ah=%d Al=%d") +JMESSAGE(JERR_BAD_PROG_SCRIPT, + "Invalid progressive parameters at scan script entry %d") +JMESSAGE(JERR_BAD_SAMPLING, "Bogus sampling factors") +JMESSAGE(JERR_BAD_SCAN_SCRIPT, "Invalid scan script at entry %d") +JMESSAGE(JERR_BAD_STATE, "Improper call to JPEG library in state %d") +JMESSAGE(JERR_BAD_STRUCT_SIZE, + "JPEG parameter struct mismatch: library thinks size is %u, caller expects %u") +JMESSAGE(JERR_BAD_VIRTUAL_ACCESS, "Bogus virtual array access") +JMESSAGE(JERR_BUFFER_SIZE, "Buffer passed to JPEG library is too small") +JMESSAGE(JERR_CANT_SUSPEND, "Suspension not allowed here") +JMESSAGE(JERR_CCIR601_NOTIMPL, "CCIR601 sampling not implemented yet") +JMESSAGE(JERR_COMPONENT_COUNT, "Too many color components: %d, max %d") +JMESSAGE(JERR_CONVERSION_NOTIMPL, "Unsupported color conversion request") +JMESSAGE(JERR_DAC_INDEX, "Bogus DAC index %d") +JMESSAGE(JERR_DAC_VALUE, "Bogus DAC value 0x%x") +JMESSAGE(JERR_DHT_INDEX, "Bogus DHT index %d") +JMESSAGE(JERR_DQT_INDEX, "Bogus DQT index %d") +JMESSAGE(JERR_EMPTY_IMAGE, "Empty JPEG image (DNL not supported)") +JMESSAGE(JERR_EMS_READ, "Read from EMS failed") +JMESSAGE(JERR_EMS_WRITE, "Write to EMS failed") +JMESSAGE(JERR_EOI_EXPECTED, "Didn't expect more than one scan") +JMESSAGE(JERR_FILE_READ, "Input file read error") +JMESSAGE(JERR_FILE_WRITE, "Output file write error --- out of disk space?") +JMESSAGE(JERR_FRACT_SAMPLE_NOTIMPL, "Fractional sampling not implemented yet") +JMESSAGE(JERR_HUFF_CLEN_OUTOFBOUNDS, "Huffman code size table out of bounds") +JMESSAGE(JERR_HUFF_MISSING_CODE, "Missing Huffman code table entry") +JMESSAGE(JERR_IMAGE_TOO_BIG, "Maximum supported image dimension is %u pixels") +JMESSAGE(JERR_INPUT_EMPTY, "Empty input file") +JMESSAGE(JERR_INPUT_EOF, "Premature end of input file") +JMESSAGE(JERR_MISMATCHED_QUANT_TABLE, + "Cannot transcode due to multiple use of quantization table %d") +JMESSAGE(JERR_MISSING_DATA, "Scan script does not transmit all data") +JMESSAGE(JERR_MODE_CHANGE, "Invalid color quantization mode change") +JMESSAGE(JERR_NOTIMPL, "Not implemented yet") +JMESSAGE(JERR_NOT_COMPILED, "Requested feature was omitted at compile time") +JMESSAGE(JERR_NO_ARITH_TABLE, "Arithmetic table 0x%02x was not defined") +JMESSAGE(JERR_NO_BACKING_STORE, "Backing store not supported") +JMESSAGE(JERR_NO_HUFF_TABLE, "Huffman table 0x%02x was not defined") +JMESSAGE(JERR_NO_IMAGE, "JPEG datastream contains no image") +JMESSAGE(JERR_NO_QUANT_TABLE, "Quantization table 0x%02x was not defined") +JMESSAGE(JERR_NO_SOI, "Not a JPEG file: starts with 0x%02x 0x%02x") +JMESSAGE(JERR_OUT_OF_MEMORY, "Insufficient memory (case %d)") +JMESSAGE(JERR_QUANT_COMPONENTS, + "Cannot quantize more than %d color components") +JMESSAGE(JERR_QUANT_FEW_COLORS, "Cannot quantize to fewer than %d colors") +JMESSAGE(JERR_QUANT_MANY_COLORS, "Cannot quantize to more than %d colors") +JMESSAGE(JERR_SOF_BEFORE, "Invalid JPEG file structure: %s before SOF") +JMESSAGE(JERR_SOF_DUPLICATE, "Invalid JPEG file structure: two SOF markers") +JMESSAGE(JERR_SOF_NO_SOS, "Invalid JPEG file structure: missing SOS marker") +JMESSAGE(JERR_SOF_UNSUPPORTED, "Unsupported JPEG process: SOF type 0x%02x") +JMESSAGE(JERR_SOI_DUPLICATE, "Invalid JPEG file structure: two SOI markers") +JMESSAGE(JERR_TFILE_CREATE, "Failed to create temporary file %s") +JMESSAGE(JERR_TFILE_READ, "Read failed on temporary file") +JMESSAGE(JERR_TFILE_SEEK, "Seek failed on temporary file") +JMESSAGE(JERR_TFILE_WRITE, + "Write failed on temporary file --- out of disk space?") +JMESSAGE(JERR_TOO_LITTLE_DATA, "Application transferred too few scanlines") +JMESSAGE(JERR_UNKNOWN_MARKER, "Unsupported marker type 0x%02x") +JMESSAGE(JERR_VIRTUAL_BUG, "Virtual array controller messed up") +JMESSAGE(JERR_WIDTH_OVERFLOW, "Image too wide for this implementation") +JMESSAGE(JERR_XMS_READ, "Read from XMS failed") +JMESSAGE(JERR_XMS_WRITE, "Write to XMS failed") +JMESSAGE(JMSG_COPYRIGHT, JCOPYRIGHT) +JMESSAGE(JMSG_VERSION, JVERSION) +JMESSAGE(JTRC_16BIT_TABLES, + "Caution: quantization tables are too coarse for baseline JPEG") +JMESSAGE(JTRC_ADOBE, + "Adobe APP14 marker: version %d, flags 0x%04x 0x%04x, transform %d") +JMESSAGE(JTRC_APP0, "Unknown APP0 marker (not JFIF), length %u") +JMESSAGE(JTRC_APP14, "Unknown APP14 marker (not Adobe), length %u") +JMESSAGE(JTRC_DAC, "Define Arithmetic Table 0x%02x: 0x%02x") +JMESSAGE(JTRC_DHT, "Define Huffman Table 0x%02x") +JMESSAGE(JTRC_DQT, "Define Quantization Table %d precision %d") +JMESSAGE(JTRC_DRI, "Define Restart Interval %u") +JMESSAGE(JTRC_EMS_CLOSE, "Freed EMS handle %u") +JMESSAGE(JTRC_EMS_OPEN, "Obtained EMS handle %u") +JMESSAGE(JTRC_EOI, "End Of Image") +JMESSAGE(JTRC_HUFFBITS, " %3d %3d %3d %3d %3d %3d %3d %3d") +JMESSAGE(JTRC_JFIF, "JFIF APP0 marker: version %d.%02d, density %dx%d %d") +JMESSAGE(JTRC_JFIF_BADTHUMBNAILSIZE, + "Warning: thumbnail image size does not match data length %u") +JMESSAGE(JTRC_JFIF_EXTENSION, + "JFIF extension marker: type 0x%02x, length %u") +JMESSAGE(JTRC_JFIF_THUMBNAIL, " with %d x %d thumbnail image") +JMESSAGE(JTRC_MISC_MARKER, "Miscellaneous marker 0x%02x, length %u") +JMESSAGE(JTRC_PARMLESS_MARKER, "Unexpected marker 0x%02x") +JMESSAGE(JTRC_QUANTVALS, " %4u %4u %4u %4u %4u %4u %4u %4u") +JMESSAGE(JTRC_QUANT_3_NCOLORS, "Quantizing to %d = %d*%d*%d colors") +JMESSAGE(JTRC_QUANT_NCOLORS, "Quantizing to %d colors") +JMESSAGE(JTRC_QUANT_SELECTED, "Selected %d colors for quantization") +JMESSAGE(JTRC_RECOVERY_ACTION, "At marker 0x%02x, recovery action %d") +JMESSAGE(JTRC_RST, "RST%d") +JMESSAGE(JTRC_SMOOTH_NOTIMPL, + "Smoothing not supported with nonstandard sampling ratios") +JMESSAGE(JTRC_SOF, "Start Of Frame 0x%02x: width=%u, height=%u, components=%d") +JMESSAGE(JTRC_SOF_COMPONENT, " Component %d: %dhx%dv q=%d") +JMESSAGE(JTRC_SOI, "Start of Image") +JMESSAGE(JTRC_SOS, "Start Of Scan: %d components") +JMESSAGE(JTRC_SOS_COMPONENT, " Component %d: dc=%d ac=%d") +JMESSAGE(JTRC_SOS_PARAMS, " Ss=%d, Se=%d, Ah=%d, Al=%d") +JMESSAGE(JTRC_TFILE_CLOSE, "Closed temporary file %s") +JMESSAGE(JTRC_TFILE_OPEN, "Opened temporary file %s") +JMESSAGE(JTRC_THUMB_JPEG, + "JFIF extension marker: JPEG-compressed thumbnail image, length %u") +JMESSAGE(JTRC_THUMB_PALETTE, + "JFIF extension marker: palette thumbnail image, length %u") +JMESSAGE(JTRC_THUMB_RGB, + "JFIF extension marker: RGB thumbnail image, length %u") +JMESSAGE(JTRC_UNKNOWN_IDS, + "Unrecognized component IDs %d %d %d, assuming YCbCr") +JMESSAGE(JTRC_XMS_CLOSE, "Freed XMS handle %u") +JMESSAGE(JTRC_XMS_OPEN, "Obtained XMS handle %u") +JMESSAGE(JWRN_ADOBE_XFORM, "Unknown Adobe color transform code %d") +JMESSAGE(JWRN_ARITH_BAD_CODE, "Corrupt JPEG data: bad arithmetic code") +JMESSAGE(JWRN_BOGUS_PROGRESSION, + "Inconsistent progression sequence for component %d coefficient %d") +JMESSAGE(JWRN_EXTRANEOUS_DATA, + "Corrupt JPEG data: %u extraneous bytes before marker 0x%02x") +JMESSAGE(JWRN_HIT_MARKER, "Corrupt JPEG data: premature end of data segment") +JMESSAGE(JWRN_HUFF_BAD_CODE, "Corrupt JPEG data: bad Huffman code") +JMESSAGE(JWRN_JFIF_MAJOR, "Warning: unknown JFIF revision number %d.%02d") +JMESSAGE(JWRN_JPEG_EOF, "Premature end of JPEG file") +JMESSAGE(JWRN_MUST_RESYNC, + "Corrupt JPEG data: found marker 0x%02x instead of RST%d") +JMESSAGE(JWRN_NOT_SEQUENTIAL, "Invalid SOS parameters for sequential JPEG") +JMESSAGE(JWRN_TOO_MUCH_DATA, "Application transferred too many scanlines") + +#ifdef JMAKE_ENUM_LIST + + JMSG_LASTMSGCODE +} J_MESSAGE_CODE; + +#undef JMAKE_ENUM_LIST +#endif /* JMAKE_ENUM_LIST */ + +/* Zap JMESSAGE macro so that future re-inclusions do nothing by default */ +#undef JMESSAGE + + +#ifndef JERROR_H +#define JERROR_H + +/* Macros to simplify using the error and trace message stuff */ +/* The first parameter is either type of cinfo pointer */ + +/* Fatal errors (print message and exit) */ +#define ERREXIT(cinfo,code) \ + ((cinfo)->err->msg_code = (code), \ + (*(cinfo)->err->error_exit) ((j_common_ptr) (cinfo))) +#define ERREXIT1(cinfo,code,p1) \ + ((cinfo)->err->msg_code = (code), \ + (cinfo)->err->msg_parm.i[0] = (p1), \ + (*(cinfo)->err->error_exit) ((j_common_ptr) (cinfo))) +#define ERREXIT2(cinfo,code,p1,p2) \ + ((cinfo)->err->msg_code = (code), \ + (cinfo)->err->msg_parm.i[0] = (p1), \ + (cinfo)->err->msg_parm.i[1] = (p2), \ + (*(cinfo)->err->error_exit) ((j_common_ptr) (cinfo))) +#define ERREXIT3(cinfo,code,p1,p2,p3) \ + ((cinfo)->err->msg_code = (code), \ + (cinfo)->err->msg_parm.i[0] = (p1), \ + (cinfo)->err->msg_parm.i[1] = (p2), \ + (cinfo)->err->msg_parm.i[2] = (p3), \ + (*(cinfo)->err->error_exit) ((j_common_ptr) (cinfo))) +#define ERREXIT4(cinfo,code,p1,p2,p3,p4) \ + ((cinfo)->err->msg_code = (code), \ + (cinfo)->err->msg_parm.i[0] = (p1), \ + (cinfo)->err->msg_parm.i[1] = (p2), \ + (cinfo)->err->msg_parm.i[2] = (p3), \ + (cinfo)->err->msg_parm.i[3] = (p4), \ + (*(cinfo)->err->error_exit) ((j_common_ptr) (cinfo))) +#define ERREXIT6(cinfo,code,p1,p2,p3,p4,p5,p6) \ + ((cinfo)->err->msg_code = (code), \ + (cinfo)->err->msg_parm.i[0] = (p1), \ + (cinfo)->err->msg_parm.i[1] = (p2), \ + (cinfo)->err->msg_parm.i[2] = (p3), \ + (cinfo)->err->msg_parm.i[3] = (p4), \ + (cinfo)->err->msg_parm.i[4] = (p5), \ + (cinfo)->err->msg_parm.i[5] = (p6), \ + (*(cinfo)->err->error_exit) ((j_common_ptr) (cinfo))) +#define ERREXITS(cinfo,code,str) \ + ((cinfo)->err->msg_code = (code), \ + strncpy((cinfo)->err->msg_parm.s, (str), JMSG_STR_PARM_MAX), \ + (*(cinfo)->err->error_exit) ((j_common_ptr) (cinfo))) + +#define MAKESTMT(stuff) do { stuff } while (0) + +/* Nonfatal errors (we can keep going, but the data is probably corrupt) */ +#define WARNMS(cinfo,code) \ + ((cinfo)->err->msg_code = (code), \ + (*(cinfo)->err->emit_message) ((j_common_ptr) (cinfo), -1)) +#define WARNMS1(cinfo,code,p1) \ + ((cinfo)->err->msg_code = (code), \ + (cinfo)->err->msg_parm.i[0] = (p1), \ + (*(cinfo)->err->emit_message) ((j_common_ptr) (cinfo), -1)) +#define WARNMS2(cinfo,code,p1,p2) \ + ((cinfo)->err->msg_code = (code), \ + (cinfo)->err->msg_parm.i[0] = (p1), \ + (cinfo)->err->msg_parm.i[1] = (p2), \ + (*(cinfo)->err->emit_message) ((j_common_ptr) (cinfo), -1)) + +/* Informational/debugging messages */ +#define TRACEMS(cinfo,lvl,code) \ + ((cinfo)->err->msg_code = (code), \ + (*(cinfo)->err->emit_message) ((j_common_ptr) (cinfo), (lvl))) +#define TRACEMS1(cinfo,lvl,code,p1) \ + ((cinfo)->err->msg_code = (code), \ + (cinfo)->err->msg_parm.i[0] = (p1), \ + (*(cinfo)->err->emit_message) ((j_common_ptr) (cinfo), (lvl))) +#define TRACEMS2(cinfo,lvl,code,p1,p2) \ + ((cinfo)->err->msg_code = (code), \ + (cinfo)->err->msg_parm.i[0] = (p1), \ + (cinfo)->err->msg_parm.i[1] = (p2), \ + (*(cinfo)->err->emit_message) ((j_common_ptr) (cinfo), (lvl))) +#define TRACEMS3(cinfo,lvl,code,p1,p2,p3) \ + MAKESTMT(int * _mp = (cinfo)->err->msg_parm.i; \ + _mp[0] = (p1); _mp[1] = (p2); _mp[2] = (p3); \ + (cinfo)->err->msg_code = (code); \ + (*(cinfo)->err->emit_message) ((j_common_ptr) (cinfo), (lvl)); ) +#define TRACEMS4(cinfo,lvl,code,p1,p2,p3,p4) \ + MAKESTMT(int * _mp = (cinfo)->err->msg_parm.i; \ + _mp[0] = (p1); _mp[1] = (p2); _mp[2] = (p3); _mp[3] = (p4); \ + (cinfo)->err->msg_code = (code); \ + (*(cinfo)->err->emit_message) ((j_common_ptr) (cinfo), (lvl)); ) +#define TRACEMS5(cinfo,lvl,code,p1,p2,p3,p4,p5) \ + MAKESTMT(int * _mp = (cinfo)->err->msg_parm.i; \ + _mp[0] = (p1); _mp[1] = (p2); _mp[2] = (p3); _mp[3] = (p4); \ + _mp[4] = (p5); \ + (cinfo)->err->msg_code = (code); \ + (*(cinfo)->err->emit_message) ((j_common_ptr) (cinfo), (lvl)); ) +#define TRACEMS8(cinfo,lvl,code,p1,p2,p3,p4,p5,p6,p7,p8) \ + MAKESTMT(int * _mp = (cinfo)->err->msg_parm.i; \ + _mp[0] = (p1); _mp[1] = (p2); _mp[2] = (p3); _mp[3] = (p4); \ + _mp[4] = (p5); _mp[5] = (p6); _mp[6] = (p7); _mp[7] = (p8); \ + (cinfo)->err->msg_code = (code); \ + (*(cinfo)->err->emit_message) ((j_common_ptr) (cinfo), (lvl)); ) +#define TRACEMSS(cinfo,lvl,code,str) \ + ((cinfo)->err->msg_code = (code), \ + strncpy((cinfo)->err->msg_parm.s, (str), JMSG_STR_PARM_MAX), \ + (*(cinfo)->err->emit_message) ((j_common_ptr) (cinfo), (lvl))) + +#endif /* JERROR_H */ diff --git a/cmake/externals/libjpeg/jfdctflt.c b/cmake/externals/libjpeg/jfdctflt.c new file mode 100644 index 000000000..013f29e0c --- /dev/null +++ b/cmake/externals/libjpeg/jfdctflt.c @@ -0,0 +1,176 @@ +/* + * jfdctflt.c + * + * Copyright (C) 1994-1996, Thomas G. Lane. + * Modified 2003-2017 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains a floating-point implementation of the + * forward DCT (Discrete Cosine Transform). + * + * This implementation should be more accurate than either of the integer + * DCT implementations. However, it may not give the same results on all + * machines because of differences in roundoff behavior. Speed will depend + * on the hardware's floating point capacity. + * + * A 2-D DCT can be done by 1-D DCT on each row followed by 1-D DCT + * on each column. Direct algorithms are also available, but they are + * much more complex and seem not to be any faster when reduced to code. + * + * This implementation is based on Arai, Agui, and Nakajima's algorithm for + * scaled DCT. Their original paper (Trans. IEICE E-71(11):1095) is in + * Japanese, but the algorithm is described in the Pennebaker & Mitchell + * JPEG textbook (see REFERENCES section in file README). The following code + * is based directly on figure 4-8 in P&M. + * While an 8-point DCT cannot be done in less than 11 multiplies, it is + * possible to arrange the computation so that many of the multiplies are + * simple scalings of the final outputs. These multiplies can then be + * folded into the multiplications or divisions by the JPEG quantization + * table entries. The AA&N method leaves only 5 multiplies and 29 adds + * to be done in the DCT itself. + * The primary disadvantage of this method is that with a fixed-point + * implementation, accuracy is lost due to imprecise representation of the + * scaled quantization values. However, that problem does not arise if + * we use floating point arithmetic. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" +#include "jdct.h" /* Private declarations for DCT subsystem */ + +#ifdef DCT_FLOAT_SUPPORTED + + +/* + * This module is specialized to the case DCTSIZE = 8. + */ + +#if DCTSIZE != 8 + Sorry, this code only copes with 8x8 DCT blocks. /* deliberate syntax err */ +#endif + + +/* + * Perform the forward DCT on one block of samples. + * + * cK represents cos(K*pi/16). + */ + +GLOBAL(void) +jpeg_fdct_float (FAST_FLOAT * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + FAST_FLOAT tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7; + FAST_FLOAT tmp10, tmp11, tmp12, tmp13; + FAST_FLOAT z1, z2, z3, z4, z5, z11, z13; + FAST_FLOAT *dataptr; + JSAMPROW elemptr; + int ctr; + + /* Pass 1: process rows. */ + + dataptr = data; + for (ctr = 0; ctr < DCTSIZE; ctr++) { + elemptr = sample_data[ctr] + start_col; + + /* Load data into workspace */ + tmp0 = (FAST_FLOAT) (GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[7])); + tmp7 = (FAST_FLOAT) (GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[7])); + tmp1 = (FAST_FLOAT) (GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[6])); + tmp6 = (FAST_FLOAT) (GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[6])); + tmp2 = (FAST_FLOAT) (GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[5])); + tmp5 = (FAST_FLOAT) (GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[5])); + tmp3 = (FAST_FLOAT) (GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[4])); + tmp4 = (FAST_FLOAT) (GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[4])); + + /* Even part */ + + tmp10 = tmp0 + tmp3; /* phase 2 */ + tmp13 = tmp0 - tmp3; + tmp11 = tmp1 + tmp2; + tmp12 = tmp1 - tmp2; + + /* Apply unsigned->signed conversion. */ + dataptr[0] = tmp10 + tmp11 - 8 * CENTERJSAMPLE; /* phase 3 */ + dataptr[4] = tmp10 - tmp11; + + z1 = (tmp12 + tmp13) * ((FAST_FLOAT) 0.707106781); /* c4 */ + dataptr[2] = tmp13 + z1; /* phase 5 */ + dataptr[6] = tmp13 - z1; + + /* Odd part */ + + tmp10 = tmp4 + tmp5; /* phase 2 */ + tmp11 = tmp5 + tmp6; + tmp12 = tmp6 + tmp7; + + /* The rotator is modified from fig 4-8 to avoid extra negations. */ + z5 = (tmp10 - tmp12) * ((FAST_FLOAT) 0.382683433); /* c6 */ + z2 = ((FAST_FLOAT) 0.541196100) * tmp10 + z5; /* c2-c6 */ + z4 = ((FAST_FLOAT) 1.306562965) * tmp12 + z5; /* c2+c6 */ + z3 = tmp11 * ((FAST_FLOAT) 0.707106781); /* c4 */ + + z11 = tmp7 + z3; /* phase 5 */ + z13 = tmp7 - z3; + + dataptr[5] = z13 + z2; /* phase 6 */ + dataptr[3] = z13 - z2; + dataptr[1] = z11 + z4; + dataptr[7] = z11 - z4; + + dataptr += DCTSIZE; /* advance pointer to next row */ + } + + /* Pass 2: process columns. */ + + dataptr = data; + for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { + tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*7]; + tmp7 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*7]; + tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*6]; + tmp6 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*6]; + tmp2 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*5]; + tmp5 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*5]; + tmp3 = dataptr[DCTSIZE*3] + dataptr[DCTSIZE*4]; + tmp4 = dataptr[DCTSIZE*3] - dataptr[DCTSIZE*4]; + + /* Even part */ + + tmp10 = tmp0 + tmp3; /* phase 2 */ + tmp13 = tmp0 - tmp3; + tmp11 = tmp1 + tmp2; + tmp12 = tmp1 - tmp2; + + dataptr[DCTSIZE*0] = tmp10 + tmp11; /* phase 3 */ + dataptr[DCTSIZE*4] = tmp10 - tmp11; + + z1 = (tmp12 + tmp13) * ((FAST_FLOAT) 0.707106781); /* c4 */ + dataptr[DCTSIZE*2] = tmp13 + z1; /* phase 5 */ + dataptr[DCTSIZE*6] = tmp13 - z1; + + /* Odd part */ + + tmp10 = tmp4 + tmp5; /* phase 2 */ + tmp11 = tmp5 + tmp6; + tmp12 = tmp6 + tmp7; + + /* The rotator is modified from fig 4-8 to avoid extra negations. */ + z5 = (tmp10 - tmp12) * ((FAST_FLOAT) 0.382683433); /* c6 */ + z2 = ((FAST_FLOAT) 0.541196100) * tmp10 + z5; /* c2-c6 */ + z4 = ((FAST_FLOAT) 1.306562965) * tmp12 + z5; /* c2+c6 */ + z3 = tmp11 * ((FAST_FLOAT) 0.707106781); /* c4 */ + + z11 = tmp7 + z3; /* phase 5 */ + z13 = tmp7 - z3; + + dataptr[DCTSIZE*5] = z13 + z2; /* phase 6 */ + dataptr[DCTSIZE*3] = z13 - z2; + dataptr[DCTSIZE*1] = z11 + z4; + dataptr[DCTSIZE*7] = z11 - z4; + + dataptr++; /* advance pointer to next column */ + } +} + +#endif /* DCT_FLOAT_SUPPORTED */ diff --git a/cmake/externals/libjpeg/jfdctfst.c b/cmake/externals/libjpeg/jfdctfst.c new file mode 100644 index 000000000..5e4e01722 --- /dev/null +++ b/cmake/externals/libjpeg/jfdctfst.c @@ -0,0 +1,232 @@ +/* + * jfdctfst.c + * + * Copyright (C) 1994-1996, Thomas G. Lane. + * Modified 2003-2017 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains a fast, not so accurate integer implementation of the + * forward DCT (Discrete Cosine Transform). + * + * A 2-D DCT can be done by 1-D DCT on each row followed by 1-D DCT + * on each column. Direct algorithms are also available, but they are + * much more complex and seem not to be any faster when reduced to code. + * + * This implementation is based on Arai, Agui, and Nakajima's algorithm for + * scaled DCT. Their original paper (Trans. IEICE E-71(11):1095) is in + * Japanese, but the algorithm is described in the Pennebaker & Mitchell + * JPEG textbook (see REFERENCES section in file README). The following code + * is based directly on figure 4-8 in P&M. + * While an 8-point DCT cannot be done in less than 11 multiplies, it is + * possible to arrange the computation so that many of the multiplies are + * simple scalings of the final outputs. These multiplies can then be + * folded into the multiplications or divisions by the JPEG quantization + * table entries. The AA&N method leaves only 5 multiplies and 29 adds + * to be done in the DCT itself. + * The primary disadvantage of this method is that with fixed-point math, + * accuracy is lost due to imprecise representation of the scaled + * quantization values. The smaller the quantization table entry, the less + * precise the scaled value, so this implementation does worse with high- + * quality-setting files than with low-quality ones. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" +#include "jdct.h" /* Private declarations for DCT subsystem */ + +#ifdef DCT_IFAST_SUPPORTED + + +/* + * This module is specialized to the case DCTSIZE = 8. + */ + +#if DCTSIZE != 8 + Sorry, this code only copes with 8x8 DCT blocks. /* deliberate syntax err */ +#endif + + +/* Scaling decisions are generally the same as in the LL&M algorithm; + * see jfdctint.c for more details. However, we choose to descale + * (right shift) multiplication products as soon as they are formed, + * rather than carrying additional fractional bits into subsequent additions. + * This compromises accuracy slightly, but it lets us save a few shifts. + * More importantly, 16-bit arithmetic is then adequate (for 8-bit samples) + * everywhere except in the multiplications proper; this saves a good deal + * of work on 16-bit-int machines. + * + * Again to save a few shifts, the intermediate results between pass 1 and + * pass 2 are not upscaled, but are represented only to integral precision. + * + * A final compromise is to represent the multiplicative constants to only + * 8 fractional bits, rather than 13. This saves some shifting work on some + * machines, and may also reduce the cost of multiplication (since there + * are fewer one-bits in the constants). + */ + +#define CONST_BITS 8 + + +/* Some C compilers fail to reduce "FIX(constant)" at compile time, thus + * causing a lot of useless floating-point operations at run time. + * To get around this we use the following pre-calculated constants. + * If you change CONST_BITS you may want to add appropriate values. + * (With a reasonable C compiler, you can just rely on the FIX() macro...) + */ + +#if CONST_BITS == 8 +#define FIX_0_382683433 ((INT32) 98) /* FIX(0.382683433) */ +#define FIX_0_541196100 ((INT32) 139) /* FIX(0.541196100) */ +#define FIX_0_707106781 ((INT32) 181) /* FIX(0.707106781) */ +#define FIX_1_306562965 ((INT32) 334) /* FIX(1.306562965) */ +#else +#define FIX_0_382683433 FIX(0.382683433) +#define FIX_0_541196100 FIX(0.541196100) +#define FIX_0_707106781 FIX(0.707106781) +#define FIX_1_306562965 FIX(1.306562965) +#endif + + +/* We can gain a little more speed, with a further compromise in accuracy, + * by omitting the addition in a descaling shift. This yields an incorrectly + * rounded result half the time... + */ + +#ifndef USE_ACCURATE_ROUNDING +#undef DESCALE +#define DESCALE(x,n) RIGHT_SHIFT(x, n) +#endif + + +/* Multiply a DCTELEM variable by an INT32 constant, and immediately + * descale to yield a DCTELEM result. + */ + +#define MULTIPLY(var,const) ((DCTELEM) DESCALE((var) * (const), CONST_BITS)) + + +/* + * Perform the forward DCT on one block of samples. + * + * cK represents cos(K*pi/16). + */ + +GLOBAL(void) +jpeg_fdct_ifast (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + DCTELEM tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7; + DCTELEM tmp10, tmp11, tmp12, tmp13; + DCTELEM z1, z2, z3, z4, z5, z11, z13; + DCTELEM *dataptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pass 1: process rows. */ + + dataptr = data; + for (ctr = 0; ctr < DCTSIZE; ctr++) { + elemptr = sample_data[ctr] + start_col; + + /* Load data into workspace */ + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[7]); + tmp7 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[7]); + tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[6]); + tmp6 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[6]); + tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[5]); + tmp5 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[5]); + tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[4]); + tmp4 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[4]); + + /* Even part */ + + tmp10 = tmp0 + tmp3; /* phase 2 */ + tmp13 = tmp0 - tmp3; + tmp11 = tmp1 + tmp2; + tmp12 = tmp1 - tmp2; + + /* Apply unsigned->signed conversion. */ + dataptr[0] = tmp10 + tmp11 - 8 * CENTERJSAMPLE; /* phase 3 */ + dataptr[4] = tmp10 - tmp11; + + z1 = MULTIPLY(tmp12 + tmp13, FIX_0_707106781); /* c4 */ + dataptr[2] = tmp13 + z1; /* phase 5 */ + dataptr[6] = tmp13 - z1; + + /* Odd part */ + + tmp10 = tmp4 + tmp5; /* phase 2 */ + tmp11 = tmp5 + tmp6; + tmp12 = tmp6 + tmp7; + + /* The rotator is modified from fig 4-8 to avoid extra negations. */ + z5 = MULTIPLY(tmp10 - tmp12, FIX_0_382683433); /* c6 */ + z2 = MULTIPLY(tmp10, FIX_0_541196100) + z5; /* c2-c6 */ + z4 = MULTIPLY(tmp12, FIX_1_306562965) + z5; /* c2+c6 */ + z3 = MULTIPLY(tmp11, FIX_0_707106781); /* c4 */ + + z11 = tmp7 + z3; /* phase 5 */ + z13 = tmp7 - z3; + + dataptr[5] = z13 + z2; /* phase 6 */ + dataptr[3] = z13 - z2; + dataptr[1] = z11 + z4; + dataptr[7] = z11 - z4; + + dataptr += DCTSIZE; /* advance pointer to next row */ + } + + /* Pass 2: process columns. */ + + dataptr = data; + for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { + tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*7]; + tmp7 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*7]; + tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*6]; + tmp6 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*6]; + tmp2 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*5]; + tmp5 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*5]; + tmp3 = dataptr[DCTSIZE*3] + dataptr[DCTSIZE*4]; + tmp4 = dataptr[DCTSIZE*3] - dataptr[DCTSIZE*4]; + + /* Even part */ + + tmp10 = tmp0 + tmp3; /* phase 2 */ + tmp13 = tmp0 - tmp3; + tmp11 = tmp1 + tmp2; + tmp12 = tmp1 - tmp2; + + dataptr[DCTSIZE*0] = tmp10 + tmp11; /* phase 3 */ + dataptr[DCTSIZE*4] = tmp10 - tmp11; + + z1 = MULTIPLY(tmp12 + tmp13, FIX_0_707106781); /* c4 */ + dataptr[DCTSIZE*2] = tmp13 + z1; /* phase 5 */ + dataptr[DCTSIZE*6] = tmp13 - z1; + + /* Odd part */ + + tmp10 = tmp4 + tmp5; /* phase 2 */ + tmp11 = tmp5 + tmp6; + tmp12 = tmp6 + tmp7; + + /* The rotator is modified from fig 4-8 to avoid extra negations. */ + z5 = MULTIPLY(tmp10 - tmp12, FIX_0_382683433); /* c6 */ + z2 = MULTIPLY(tmp10, FIX_0_541196100) + z5; /* c2-c6 */ + z4 = MULTIPLY(tmp12, FIX_1_306562965) + z5; /* c2+c6 */ + z3 = MULTIPLY(tmp11, FIX_0_707106781); /* c4 */ + + z11 = tmp7 + z3; /* phase 5 */ + z13 = tmp7 - z3; + + dataptr[DCTSIZE*5] = z13 + z2; /* phase 6 */ + dataptr[DCTSIZE*3] = z13 - z2; + dataptr[DCTSIZE*1] = z11 + z4; + dataptr[DCTSIZE*7] = z11 - z4; + + dataptr++; /* advance pointer to next column */ + } +} + +#endif /* DCT_IFAST_SUPPORTED */ diff --git a/cmake/externals/libjpeg/jfdctint.c b/cmake/externals/libjpeg/jfdctint.c new file mode 100644 index 000000000..05df4750b --- /dev/null +++ b/cmake/externals/libjpeg/jfdctint.c @@ -0,0 +1,4415 @@ +/* + * jfdctint.c + * + * Copyright (C) 1991-1996, Thomas G. Lane. + * Modification developed 2003-2018 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains a slow-but-accurate integer implementation of the + * forward DCT (Discrete Cosine Transform). + * + * A 2-D DCT can be done by 1-D DCT on each row followed by 1-D DCT + * on each column. Direct algorithms are also available, but they are + * much more complex and seem not to be any faster when reduced to code. + * + * This implementation is based on an algorithm described in + * C. Loeffler, A. Ligtenberg and G. Moschytz, "Practical Fast 1-D DCT + * Algorithms with 11 Multiplications", Proc. Int'l. Conf. on Acoustics, + * Speech, and Signal Processing 1989 (ICASSP '89), pp. 988-991. + * The primary algorithm described there uses 11 multiplies and 29 adds. + * We use their alternate method with 12 multiplies and 32 adds. + * The advantage of this method is that no data path contains more than one + * multiplication; this allows a very simple and accurate implementation in + * scaled fixed-point arithmetic, with a minimal number of shifts. + * + * We also provide FDCT routines with various input sample block sizes for + * direct resolution reduction or enlargement and for direct resolving the + * common 2x1 and 1x2 subsampling cases without additional resampling: NxN + * (N=1...16), 2NxN, and Nx2N (N=1...8) pixels for one 8x8 output DCT block. + * + * For N<8 we fill the remaining block coefficients with zero. + * For N>8 we apply a partial N-point FDCT on the input samples, computing + * just the lower 8 frequency coefficients and discarding the rest. + * + * We must scale the output coefficients of the N-point FDCT appropriately + * to the standard 8-point FDCT level by 8/N per 1-D pass. This scaling + * is folded into the constant multipliers (pass 2) and/or final/initial + * shifting. + * + * CAUTION: We rely on the FIX() macro except for the N=1,2,4,8 cases + * since there would be too many additional constants to pre-calculate. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" +#include "jdct.h" /* Private declarations for DCT subsystem */ + +#ifdef DCT_ISLOW_SUPPORTED + + +/* + * This module is specialized to the case DCTSIZE = 8. + */ + +#if DCTSIZE != 8 + Sorry, this code only copes with 8x8 DCT blocks. /* deliberate syntax err */ +#endif + + +/* + * The poop on this scaling stuff is as follows: + * + * Each 1-D DCT step produces outputs which are a factor of sqrt(N) + * larger than the true DCT outputs. The final outputs are therefore + * a factor of N larger than desired; since N=8 this can be cured by + * a simple right shift at the end of the algorithm. The advantage of + * this arrangement is that we save two multiplications per 1-D DCT, + * because the y0 and y4 outputs need not be divided by sqrt(N). + * In the IJG code, this factor of 8 is removed by the quantization step + * (in jcdctmgr.c), NOT in this module. + * + * We have to do addition and subtraction of the integer inputs, which + * is no problem, and multiplication by fractional constants, which is + * a problem to do in integer arithmetic. We multiply all the constants + * by CONST_SCALE and convert them to integer constants (thus retaining + * CONST_BITS bits of precision in the constants). After doing a + * multiplication we have to divide the product by CONST_SCALE, with proper + * rounding, to produce the correct output. This division can be done + * cheaply as a right shift of CONST_BITS bits. We postpone shifting + * as long as possible so that partial sums can be added together with + * full fractional precision. + * + * The outputs of the first pass are scaled up by PASS1_BITS bits so that + * they are represented to better-than-integral precision. These outputs + * require BITS_IN_JSAMPLE + PASS1_BITS + 3 bits; this fits in a 16-bit word + * with the recommended scaling. (For 12-bit sample data, the intermediate + * array is INT32 anyway.) + * + * To avoid overflow of the 32-bit intermediate results in pass 2, we must + * have BITS_IN_JSAMPLE + CONST_BITS + PASS1_BITS <= 26. Error analysis + * shows that the values given below are the most effective. + */ + +#if BITS_IN_JSAMPLE == 8 +#define CONST_BITS 13 +#define PASS1_BITS 2 +#else +#define CONST_BITS 13 +#define PASS1_BITS 1 /* lose a little precision to avoid overflow */ +#endif + +/* Some C compilers fail to reduce "FIX(constant)" at compile time, thus + * causing a lot of useless floating-point operations at run time. + * To get around this we use the following pre-calculated constants. + * If you change CONST_BITS you may want to add appropriate values. + * (With a reasonable C compiler, you can just rely on the FIX() macro...) + */ + +#if CONST_BITS == 13 +#define FIX_0_298631336 ((INT32) 2446) /* FIX(0.298631336) */ +#define FIX_0_390180644 ((INT32) 3196) /* FIX(0.390180644) */ +#define FIX_0_541196100 ((INT32) 4433) /* FIX(0.541196100) */ +#define FIX_0_765366865 ((INT32) 6270) /* FIX(0.765366865) */ +#define FIX_0_899976223 ((INT32) 7373) /* FIX(0.899976223) */ +#define FIX_1_175875602 ((INT32) 9633) /* FIX(1.175875602) */ +#define FIX_1_501321110 ((INT32) 12299) /* FIX(1.501321110) */ +#define FIX_1_847759065 ((INT32) 15137) /* FIX(1.847759065) */ +#define FIX_1_961570560 ((INT32) 16069) /* FIX(1.961570560) */ +#define FIX_2_053119869 ((INT32) 16819) /* FIX(2.053119869) */ +#define FIX_2_562915447 ((INT32) 20995) /* FIX(2.562915447) */ +#define FIX_3_072711026 ((INT32) 25172) /* FIX(3.072711026) */ +#else +#define FIX_0_298631336 FIX(0.298631336) +#define FIX_0_390180644 FIX(0.390180644) +#define FIX_0_541196100 FIX(0.541196100) +#define FIX_0_765366865 FIX(0.765366865) +#define FIX_0_899976223 FIX(0.899976223) +#define FIX_1_175875602 FIX(1.175875602) +#define FIX_1_501321110 FIX(1.501321110) +#define FIX_1_847759065 FIX(1.847759065) +#define FIX_1_961570560 FIX(1.961570560) +#define FIX_2_053119869 FIX(2.053119869) +#define FIX_2_562915447 FIX(2.562915447) +#define FIX_3_072711026 FIX(3.072711026) +#endif + + +/* Multiply an INT32 variable by an INT32 constant to yield an INT32 result. + * For 8-bit samples with the recommended scaling, all the variable + * and constant values involved are no more than 16 bits wide, so a + * 16x16->32 bit multiply can be used instead of a full 32x32 multiply. + * For 12-bit samples, a full 32-bit multiplication will be needed. + */ + +#if BITS_IN_JSAMPLE == 8 +#define MULTIPLY(var,const) MULTIPLY16C16(var,const) +#else +#define MULTIPLY(var,const) ((var) * (const)) +#endif + + +/* + * Perform the forward DCT on one block of samples. + */ + +GLOBAL(void) +jpeg_fdct_islow (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3; + INT32 tmp10, tmp11, tmp12, tmp13; + INT32 z1; + DCTELEM *dataptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT; + * furthermore, we scale the results by 2**PASS1_BITS. + * cK represents sqrt(2) * cos(K*pi/16). + */ + + dataptr = data; + for (ctr = 0; ctr < DCTSIZE; ctr++) { + elemptr = sample_data[ctr] + start_col; + + /* Even part per LL&M figure 1 --- note that published figure is faulty; + * rotator "c1" should be "c6". + */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[7]); + tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[6]); + tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[5]); + tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[4]); + + tmp10 = tmp0 + tmp3; + tmp12 = tmp0 - tmp3; + tmp11 = tmp1 + tmp2; + tmp13 = tmp1 - tmp2; + + tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[7]); + tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[6]); + tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[5]); + tmp3 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[4]); + + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) ((tmp10 + tmp11 - 8 * CENTERJSAMPLE) << PASS1_BITS); + dataptr[4] = (DCTELEM) ((tmp10 - tmp11) << PASS1_BITS); + + z1 = MULTIPLY(tmp12 + tmp13, FIX_0_541196100); /* c6 */ + /* Add fudge factor here for final descale. */ + z1 += ONE << (CONST_BITS-PASS1_BITS-1); + + dataptr[2] = (DCTELEM) + RIGHT_SHIFT(z1 + MULTIPLY(tmp12, FIX_0_765366865), /* c2-c6 */ + CONST_BITS-PASS1_BITS); + dataptr[6] = (DCTELEM) + RIGHT_SHIFT(z1 - MULTIPLY(tmp13, FIX_1_847759065), /* c2+c6 */ + CONST_BITS-PASS1_BITS); + + /* Odd part per figure 8 --- note paper omits factor of sqrt(2). + * i0..i3 in the paper are tmp0..tmp3 here. + */ + + tmp12 = tmp0 + tmp2; + tmp13 = tmp1 + tmp3; + + z1 = MULTIPLY(tmp12 + tmp13, FIX_1_175875602); /* c3 */ + /* Add fudge factor here for final descale. */ + z1 += ONE << (CONST_BITS-PASS1_BITS-1); + + tmp12 = MULTIPLY(tmp12, - FIX_0_390180644); /* -c3+c5 */ + tmp13 = MULTIPLY(tmp13, - FIX_1_961570560); /* -c3-c5 */ + tmp12 += z1; + tmp13 += z1; + + z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ + tmp0 = MULTIPLY(tmp0, FIX_1_501321110); /* c1+c3-c5-c7 */ + tmp3 = MULTIPLY(tmp3, FIX_0_298631336); /* -c1+c3+c5-c7 */ + tmp0 += z1 + tmp12; + tmp3 += z1 + tmp13; + + z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ + tmp1 = MULTIPLY(tmp1, FIX_3_072711026); /* c1+c3+c5-c7 */ + tmp2 = MULTIPLY(tmp2, FIX_2_053119869); /* c1+c3-c5+c7 */ + tmp1 += z1 + tmp13; + tmp2 += z1 + tmp12; + + dataptr[1] = (DCTELEM) RIGHT_SHIFT(tmp0, CONST_BITS-PASS1_BITS); + dataptr[3] = (DCTELEM) RIGHT_SHIFT(tmp1, CONST_BITS-PASS1_BITS); + dataptr[5] = (DCTELEM) RIGHT_SHIFT(tmp2, CONST_BITS-PASS1_BITS); + dataptr[7] = (DCTELEM) RIGHT_SHIFT(tmp3, CONST_BITS-PASS1_BITS); + + dataptr += DCTSIZE; /* advance pointer to next row */ + } + + /* Pass 2: process columns. + * We remove the PASS1_BITS scaling, but leave the results scaled up + * by an overall factor of 8. + * cK represents sqrt(2) * cos(K*pi/16). + */ + + dataptr = data; + for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { + /* Even part per LL&M figure 1 --- note that published figure is faulty; + * rotator "c1" should be "c6". + */ + + tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*7]; + tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*6]; + tmp2 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*5]; + tmp3 = dataptr[DCTSIZE*3] + dataptr[DCTSIZE*4]; + + /* Add fudge factor here for final descale. */ + tmp10 = tmp0 + tmp3 + (ONE << (PASS1_BITS-1)); + tmp12 = tmp0 - tmp3; + tmp11 = tmp1 + tmp2; + tmp13 = tmp1 - tmp2; + + tmp0 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*7]; + tmp1 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*6]; + tmp2 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*5]; + tmp3 = dataptr[DCTSIZE*3] - dataptr[DCTSIZE*4]; + + dataptr[DCTSIZE*0] = (DCTELEM) RIGHT_SHIFT(tmp10 + tmp11, PASS1_BITS); + dataptr[DCTSIZE*4] = (DCTELEM) RIGHT_SHIFT(tmp10 - tmp11, PASS1_BITS); + + z1 = MULTIPLY(tmp12 + tmp13, FIX_0_541196100); /* c6 */ + /* Add fudge factor here for final descale. */ + z1 += ONE << (CONST_BITS+PASS1_BITS-1); + + dataptr[DCTSIZE*2] = (DCTELEM) + RIGHT_SHIFT(z1 + MULTIPLY(tmp12, FIX_0_765366865), /* c2-c6 */ + CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*6] = (DCTELEM) + RIGHT_SHIFT(z1 - MULTIPLY(tmp13, FIX_1_847759065), /* c2+c6 */ + CONST_BITS+PASS1_BITS); + + /* Odd part per figure 8 --- note paper omits factor of sqrt(2). + * i0..i3 in the paper are tmp0..tmp3 here. + */ + + tmp12 = tmp0 + tmp2; + tmp13 = tmp1 + tmp3; + + z1 = MULTIPLY(tmp12 + tmp13, FIX_1_175875602); /* c3 */ + /* Add fudge factor here for final descale. */ + z1 += ONE << (CONST_BITS+PASS1_BITS-1); + + tmp12 = MULTIPLY(tmp12, - FIX_0_390180644); /* -c3+c5 */ + tmp13 = MULTIPLY(tmp13, - FIX_1_961570560); /* -c3-c5 */ + tmp12 += z1; + tmp13 += z1; + + z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ + tmp0 = MULTIPLY(tmp0, FIX_1_501321110); /* c1+c3-c5-c7 */ + tmp3 = MULTIPLY(tmp3, FIX_0_298631336); /* -c1+c3+c5-c7 */ + tmp0 += z1 + tmp12; + tmp3 += z1 + tmp13; + + z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ + tmp1 = MULTIPLY(tmp1, FIX_3_072711026); /* c1+c3+c5-c7 */ + tmp2 = MULTIPLY(tmp2, FIX_2_053119869); /* c1+c3-c5+c7 */ + tmp1 += z1 + tmp13; + tmp2 += z1 + tmp12; + + dataptr[DCTSIZE*1] = (DCTELEM) RIGHT_SHIFT(tmp0, CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*3] = (DCTELEM) RIGHT_SHIFT(tmp1, CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*5] = (DCTELEM) RIGHT_SHIFT(tmp2, CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*7] = (DCTELEM) RIGHT_SHIFT(tmp3, CONST_BITS+PASS1_BITS); + + dataptr++; /* advance pointer to next column */ + } +} + +#ifdef DCT_SCALING_SUPPORTED + + +/* + * Perform the forward DCT on a 7x7 sample block. + */ + +GLOBAL(void) +jpeg_fdct_7x7 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3; + INT32 tmp10, tmp11, tmp12; + INT32 z1, z2, z3; + DCTELEM *dataptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pre-zero output coefficient block. */ + MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT; + * furthermore, we scale the results by 2**PASS1_BITS. + * cK represents sqrt(2) * cos(K*pi/14). + */ + + dataptr = data; + for (ctr = 0; ctr < 7; ctr++) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[6]); + tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[5]); + tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[4]); + tmp3 = GETJSAMPLE(elemptr[3]); + + tmp10 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[6]); + tmp11 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[5]); + tmp12 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[4]); + + z1 = tmp0 + tmp2; + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) + ((z1 + tmp1 + tmp3 - 7 * CENTERJSAMPLE) << PASS1_BITS); + tmp3 += tmp3; + z1 -= tmp3; + z1 -= tmp3; + z1 = MULTIPLY(z1, FIX(0.353553391)); /* (c2+c6-c4)/2 */ + z2 = MULTIPLY(tmp0 - tmp2, FIX(0.920609002)); /* (c2+c4-c6)/2 */ + z3 = MULTIPLY(tmp1 - tmp2, FIX(0.314692123)); /* c6 */ + dataptr[2] = (DCTELEM) DESCALE(z1 + z2 + z3, CONST_BITS-PASS1_BITS); + z1 -= z2; + z2 = MULTIPLY(tmp0 - tmp1, FIX(0.881747734)); /* c4 */ + dataptr[4] = (DCTELEM) + DESCALE(z2 + z3 - MULTIPLY(tmp1 - tmp3, FIX(0.707106781)), /* c2+c6-c4 */ + CONST_BITS-PASS1_BITS); + dataptr[6] = (DCTELEM) DESCALE(z1 + z2, CONST_BITS-PASS1_BITS); + + /* Odd part */ + + tmp1 = MULTIPLY(tmp10 + tmp11, FIX(0.935414347)); /* (c3+c1-c5)/2 */ + tmp2 = MULTIPLY(tmp10 - tmp11, FIX(0.170262339)); /* (c3+c5-c1)/2 */ + tmp0 = tmp1 - tmp2; + tmp1 += tmp2; + tmp2 = MULTIPLY(tmp11 + tmp12, - FIX(1.378756276)); /* -c1 */ + tmp1 += tmp2; + tmp3 = MULTIPLY(tmp10 + tmp12, FIX(0.613604268)); /* c5 */ + tmp0 += tmp3; + tmp2 += tmp3 + MULTIPLY(tmp12, FIX(1.870828693)); /* c3+c1-c5 */ + + dataptr[1] = (DCTELEM) DESCALE(tmp0, CONST_BITS-PASS1_BITS); + dataptr[3] = (DCTELEM) DESCALE(tmp1, CONST_BITS-PASS1_BITS); + dataptr[5] = (DCTELEM) DESCALE(tmp2, CONST_BITS-PASS1_BITS); + + dataptr += DCTSIZE; /* advance pointer to next row */ + } + + /* Pass 2: process columns. + * We remove the PASS1_BITS scaling, but leave the results scaled up + * by an overall factor of 8. + * We must also scale the output by (8/7)**2 = 64/49, which we fold + * into the constant multipliers: + * cK now represents sqrt(2) * cos(K*pi/14) * 64/49. + */ + + dataptr = data; + for (ctr = 0; ctr < 7; ctr++) { + /* Even part */ + + tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*6]; + tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*5]; + tmp2 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*4]; + tmp3 = dataptr[DCTSIZE*3]; + + tmp10 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*6]; + tmp11 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*5]; + tmp12 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*4]; + + z1 = tmp0 + tmp2; + dataptr[DCTSIZE*0] = (DCTELEM) + DESCALE(MULTIPLY(z1 + tmp1 + tmp3, FIX(1.306122449)), /* 64/49 */ + CONST_BITS+PASS1_BITS); + tmp3 += tmp3; + z1 -= tmp3; + z1 -= tmp3; + z1 = MULTIPLY(z1, FIX(0.461784020)); /* (c2+c6-c4)/2 */ + z2 = MULTIPLY(tmp0 - tmp2, FIX(1.202428084)); /* (c2+c4-c6)/2 */ + z3 = MULTIPLY(tmp1 - tmp2, FIX(0.411026446)); /* c6 */ + dataptr[DCTSIZE*2] = (DCTELEM) DESCALE(z1 + z2 + z3, CONST_BITS+PASS1_BITS); + z1 -= z2; + z2 = MULTIPLY(tmp0 - tmp1, FIX(1.151670509)); /* c4 */ + dataptr[DCTSIZE*4] = (DCTELEM) + DESCALE(z2 + z3 - MULTIPLY(tmp1 - tmp3, FIX(0.923568041)), /* c2+c6-c4 */ + CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*6] = (DCTELEM) DESCALE(z1 + z2, CONST_BITS+PASS1_BITS); + + /* Odd part */ + + tmp1 = MULTIPLY(tmp10 + tmp11, FIX(1.221765677)); /* (c3+c1-c5)/2 */ + tmp2 = MULTIPLY(tmp10 - tmp11, FIX(0.222383464)); /* (c3+c5-c1)/2 */ + tmp0 = tmp1 - tmp2; + tmp1 += tmp2; + tmp2 = MULTIPLY(tmp11 + tmp12, - FIX(1.800824523)); /* -c1 */ + tmp1 += tmp2; + tmp3 = MULTIPLY(tmp10 + tmp12, FIX(0.801442310)); /* c5 */ + tmp0 += tmp3; + tmp2 += tmp3 + MULTIPLY(tmp12, FIX(2.443531355)); /* c3+c1-c5 */ + + dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp0, CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp1, CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp2, CONST_BITS+PASS1_BITS); + + dataptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 6x6 sample block. + */ + +GLOBAL(void) +jpeg_fdct_6x6 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2; + INT32 tmp10, tmp11, tmp12; + DCTELEM *dataptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pre-zero output coefficient block. */ + MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT; + * furthermore, we scale the results by 2**PASS1_BITS. + * cK represents sqrt(2) * cos(K*pi/12). + */ + + dataptr = data; + for (ctr = 0; ctr < 6; ctr++) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[5]); + tmp11 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[4]); + tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[3]); + + tmp10 = tmp0 + tmp2; + tmp12 = tmp0 - tmp2; + + tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[5]); + tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[4]); + tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[3]); + + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) + ((tmp10 + tmp11 - 6 * CENTERJSAMPLE) << PASS1_BITS); + dataptr[2] = (DCTELEM) + DESCALE(MULTIPLY(tmp12, FIX(1.224744871)), /* c2 */ + CONST_BITS-PASS1_BITS); + dataptr[4] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp11 - tmp11, FIX(0.707106781)), /* c4 */ + CONST_BITS-PASS1_BITS); + + /* Odd part */ + + tmp10 = DESCALE(MULTIPLY(tmp0 + tmp2, FIX(0.366025404)), /* c5 */ + CONST_BITS-PASS1_BITS); + + dataptr[1] = (DCTELEM) (tmp10 + ((tmp0 + tmp1) << PASS1_BITS)); + dataptr[3] = (DCTELEM) ((tmp0 - tmp1 - tmp2) << PASS1_BITS); + dataptr[5] = (DCTELEM) (tmp10 + ((tmp2 - tmp1) << PASS1_BITS)); + + dataptr += DCTSIZE; /* advance pointer to next row */ + } + + /* Pass 2: process columns. + * We remove the PASS1_BITS scaling, but leave the results scaled up + * by an overall factor of 8. + * We must also scale the output by (8/6)**2 = 16/9, which we fold + * into the constant multipliers: + * cK now represents sqrt(2) * cos(K*pi/12) * 16/9. + */ + + dataptr = data; + for (ctr = 0; ctr < 6; ctr++) { + /* Even part */ + + tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*5]; + tmp11 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*4]; + tmp2 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*3]; + + tmp10 = tmp0 + tmp2; + tmp12 = tmp0 - tmp2; + + tmp0 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*5]; + tmp1 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*4]; + tmp2 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*3]; + + dataptr[DCTSIZE*0] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 + tmp11, FIX(1.777777778)), /* 16/9 */ + CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*2] = (DCTELEM) + DESCALE(MULTIPLY(tmp12, FIX(2.177324216)), /* c2 */ + CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*4] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp11 - tmp11, FIX(1.257078722)), /* c4 */ + CONST_BITS+PASS1_BITS); + + /* Odd part */ + + tmp10 = MULTIPLY(tmp0 + tmp2, FIX(0.650711829)); /* c5 */ + + dataptr[DCTSIZE*1] = (DCTELEM) + DESCALE(tmp10 + MULTIPLY(tmp0 + tmp1, FIX(1.777777778)), /* 16/9 */ + CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*3] = (DCTELEM) + DESCALE(MULTIPLY(tmp0 - tmp1 - tmp2, FIX(1.777777778)), /* 16/9 */ + CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*5] = (DCTELEM) + DESCALE(tmp10 + MULTIPLY(tmp2 - tmp1, FIX(1.777777778)), /* 16/9 */ + CONST_BITS+PASS1_BITS); + + dataptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 5x5 sample block. + */ + +GLOBAL(void) +jpeg_fdct_5x5 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2; + INT32 tmp10, tmp11; + DCTELEM *dataptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pre-zero output coefficient block. */ + MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT; + * furthermore, we scale the results by 2**PASS1_BITS. + * We scale the results further by 2 as part of output adaption + * scaling for different DCT size. + * cK represents sqrt(2) * cos(K*pi/10). + */ + + dataptr = data; + for (ctr = 0; ctr < 5; ctr++) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[4]); + tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[3]); + tmp2 = GETJSAMPLE(elemptr[2]); + + tmp10 = tmp0 + tmp1; + tmp11 = tmp0 - tmp1; + + tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[4]); + tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[3]); + + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) + ((tmp10 + tmp2 - 5 * CENTERJSAMPLE) << (PASS1_BITS+1)); + tmp11 = MULTIPLY(tmp11, FIX(0.790569415)); /* (c2+c4)/2 */ + tmp10 -= tmp2 << 2; + tmp10 = MULTIPLY(tmp10, FIX(0.353553391)); /* (c2-c4)/2 */ + dataptr[2] = (DCTELEM) DESCALE(tmp11 + tmp10, CONST_BITS-PASS1_BITS-1); + dataptr[4] = (DCTELEM) DESCALE(tmp11 - tmp10, CONST_BITS-PASS1_BITS-1); + + /* Odd part */ + + tmp10 = MULTIPLY(tmp0 + tmp1, FIX(0.831253876)); /* c3 */ + + dataptr[1] = (DCTELEM) + DESCALE(tmp10 + MULTIPLY(tmp0, FIX(0.513743148)), /* c1-c3 */ + CONST_BITS-PASS1_BITS-1); + dataptr[3] = (DCTELEM) + DESCALE(tmp10 - MULTIPLY(tmp1, FIX(2.176250899)), /* c1+c3 */ + CONST_BITS-PASS1_BITS-1); + + dataptr += DCTSIZE; /* advance pointer to next row */ + } + + /* Pass 2: process columns. + * We remove the PASS1_BITS scaling, but leave the results scaled up + * by an overall factor of 8. + * We must also scale the output by (8/5)**2 = 64/25, which we partially + * fold into the constant multipliers (other part was done in pass 1): + * cK now represents sqrt(2) * cos(K*pi/10) * 32/25. + */ + + dataptr = data; + for (ctr = 0; ctr < 5; ctr++) { + /* Even part */ + + tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*4]; + tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*3]; + tmp2 = dataptr[DCTSIZE*2]; + + tmp10 = tmp0 + tmp1; + tmp11 = tmp0 - tmp1; + + tmp0 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*4]; + tmp1 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*3]; + + dataptr[DCTSIZE*0] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 + tmp2, FIX(1.28)), /* 32/25 */ + CONST_BITS+PASS1_BITS); + tmp11 = MULTIPLY(tmp11, FIX(1.011928851)); /* (c2+c4)/2 */ + tmp10 -= tmp2 << 2; + tmp10 = MULTIPLY(tmp10, FIX(0.452548340)); /* (c2-c4)/2 */ + dataptr[DCTSIZE*2] = (DCTELEM) DESCALE(tmp11 + tmp10, CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*4] = (DCTELEM) DESCALE(tmp11 - tmp10, CONST_BITS+PASS1_BITS); + + /* Odd part */ + + tmp10 = MULTIPLY(tmp0 + tmp1, FIX(1.064004961)); /* c3 */ + + dataptr[DCTSIZE*1] = (DCTELEM) + DESCALE(tmp10 + MULTIPLY(tmp0, FIX(0.657591230)), /* c1-c3 */ + CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*3] = (DCTELEM) + DESCALE(tmp10 - MULTIPLY(tmp1, FIX(2.785601151)), /* c1+c3 */ + CONST_BITS+PASS1_BITS); + + dataptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 4x4 sample block. + */ + +GLOBAL(void) +jpeg_fdct_4x4 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1; + INT32 tmp10, tmp11; + DCTELEM *dataptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pre-zero output coefficient block. */ + MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT; + * furthermore, we scale the results by 2**PASS1_BITS. + * We must also scale the output by (8/4)**2 = 2**2, which we add here. + * cK represents sqrt(2) * cos(K*pi/16) [refers to 8-point FDCT]. + */ + + dataptr = data; + for (ctr = 0; ctr < 4; ctr++) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[3]); + tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[2]); + + tmp10 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[3]); + tmp11 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[2]); + + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) + ((tmp0 + tmp1 - 4 * CENTERJSAMPLE) << (PASS1_BITS+2)); + dataptr[2] = (DCTELEM) ((tmp0 - tmp1) << (PASS1_BITS+2)); + + /* Odd part */ + + tmp0 = MULTIPLY(tmp10 + tmp11, FIX_0_541196100); /* c6 */ + /* Add fudge factor here for final descale. */ + tmp0 += ONE << (CONST_BITS-PASS1_BITS-3); + + dataptr[1] = (DCTELEM) + RIGHT_SHIFT(tmp0 + MULTIPLY(tmp10, FIX_0_765366865), /* c2-c6 */ + CONST_BITS-PASS1_BITS-2); + dataptr[3] = (DCTELEM) + RIGHT_SHIFT(tmp0 - MULTIPLY(tmp11, FIX_1_847759065), /* c2+c6 */ + CONST_BITS-PASS1_BITS-2); + + dataptr += DCTSIZE; /* advance pointer to next row */ + } + + /* Pass 2: process columns. + * We remove the PASS1_BITS scaling, but leave the results scaled up + * by an overall factor of 8. + * cK represents sqrt(2) * cos(K*pi/16) [refers to 8-point FDCT]. + */ + + dataptr = data; + for (ctr = 0; ctr < 4; ctr++) { + /* Even part */ + + /* Add fudge factor here for final descale. */ + tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*3] + (ONE << (PASS1_BITS-1)); + tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*2]; + + tmp10 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*3]; + tmp11 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*2]; + + dataptr[DCTSIZE*0] = (DCTELEM) RIGHT_SHIFT(tmp0 + tmp1, PASS1_BITS); + dataptr[DCTSIZE*2] = (DCTELEM) RIGHT_SHIFT(tmp0 - tmp1, PASS1_BITS); + + /* Odd part */ + + tmp0 = MULTIPLY(tmp10 + tmp11, FIX_0_541196100); /* c6 */ + /* Add fudge factor here for final descale. */ + tmp0 += ONE << (CONST_BITS+PASS1_BITS-1); + + dataptr[DCTSIZE*1] = (DCTELEM) + RIGHT_SHIFT(tmp0 + MULTIPLY(tmp10, FIX_0_765366865), /* c2-c6 */ + CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*3] = (DCTELEM) + RIGHT_SHIFT(tmp0 - MULTIPLY(tmp11, FIX_1_847759065), /* c2+c6 */ + CONST_BITS+PASS1_BITS); + + dataptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 3x3 sample block. + */ + +GLOBAL(void) +jpeg_fdct_3x3 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2; + DCTELEM *dataptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pre-zero output coefficient block. */ + MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT; + * furthermore, we scale the results by 2**PASS1_BITS. + * We scale the results further by 2**2 as part of output adaption + * scaling for different DCT size. + * cK represents sqrt(2) * cos(K*pi/6). + */ + + dataptr = data; + for (ctr = 0; ctr < 3; ctr++) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[2]); + tmp1 = GETJSAMPLE(elemptr[1]); + + tmp2 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[2]); + + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) + ((tmp0 + tmp1 - 3 * CENTERJSAMPLE) << (PASS1_BITS+2)); + dataptr[2] = (DCTELEM) + DESCALE(MULTIPLY(tmp0 - tmp1 - tmp1, FIX(0.707106781)), /* c2 */ + CONST_BITS-PASS1_BITS-2); + + /* Odd part */ + + dataptr[1] = (DCTELEM) + DESCALE(MULTIPLY(tmp2, FIX(1.224744871)), /* c1 */ + CONST_BITS-PASS1_BITS-2); + + dataptr += DCTSIZE; /* advance pointer to next row */ + } + + /* Pass 2: process columns. + * We remove the PASS1_BITS scaling, but leave the results scaled up + * by an overall factor of 8. + * We must also scale the output by (8/3)**2 = 64/9, which we partially + * fold into the constant multipliers (other part was done in pass 1): + * cK now represents sqrt(2) * cos(K*pi/6) * 16/9. + */ + + dataptr = data; + for (ctr = 0; ctr < 3; ctr++) { + /* Even part */ + + tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*2]; + tmp1 = dataptr[DCTSIZE*1]; + + tmp2 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*2]; + + dataptr[DCTSIZE*0] = (DCTELEM) + DESCALE(MULTIPLY(tmp0 + tmp1, FIX(1.777777778)), /* 16/9 */ + CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*2] = (DCTELEM) + DESCALE(MULTIPLY(tmp0 - tmp1 - tmp1, FIX(1.257078722)), /* c2 */ + CONST_BITS+PASS1_BITS); + + /* Odd part */ + + dataptr[DCTSIZE*1] = (DCTELEM) + DESCALE(MULTIPLY(tmp2, FIX(2.177324216)), /* c1 */ + CONST_BITS+PASS1_BITS); + + dataptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 2x2 sample block. + */ + +GLOBAL(void) +jpeg_fdct_2x2 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + DCTELEM tmp0, tmp1, tmp2, tmp3; + JSAMPROW elemptr; + + /* Pre-zero output coefficient block. */ + MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT. + */ + + /* Row 0 */ + elemptr = sample_data[0] + start_col; + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[1]); + tmp1 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[1]); + + /* Row 1 */ + elemptr = sample_data[1] + start_col; + + tmp2 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[1]); + tmp3 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[1]); + + /* Pass 2: process columns. + * We leave the results scaled up by an overall factor of 8. + * We must also scale the output by (8/2)**2 = 2**4. + */ + + /* Column 0 */ + /* Apply unsigned->signed conversion. */ + data[DCTSIZE*0] = (tmp0 + tmp2 - 4 * CENTERJSAMPLE) << 4; + data[DCTSIZE*1] = (tmp0 - tmp2) << 4; + + /* Column 1 */ + data[DCTSIZE*0+1] = (tmp1 + tmp3) << 4; + data[DCTSIZE*1+1] = (tmp1 - tmp3) << 4; +} + + +/* + * Perform the forward DCT on a 1x1 sample block. + */ + +GLOBAL(void) +jpeg_fdct_1x1 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + DCTELEM dcval; + + /* Pre-zero output coefficient block. */ + MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); + + dcval = GETJSAMPLE(sample_data[0][start_col]); + + /* We leave the result scaled up by an overall factor of 8. */ + /* We must also scale the output by (8/1)**2 = 2**6. */ + /* Apply unsigned->signed conversion. */ + data[0] = (dcval - CENTERJSAMPLE) << 6; +} + + +/* + * Perform the forward DCT on a 9x9 sample block. + */ + +GLOBAL(void) +jpeg_fdct_9x9 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3, tmp4; + INT32 tmp10, tmp11, tmp12, tmp13; + INT32 z1, z2; + DCTELEM workspace[8]; + DCTELEM *dataptr; + DCTELEM *wsptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT; + * we scale the results further by 2 as part of output adaption + * scaling for different DCT size. + * cK represents sqrt(2) * cos(K*pi/18). + */ + + dataptr = data; + ctr = 0; + for (;;) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[8]); + tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[7]); + tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[6]); + tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[5]); + tmp4 = GETJSAMPLE(elemptr[4]); + + tmp10 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[8]); + tmp11 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[7]); + tmp12 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[6]); + tmp13 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[5]); + + z1 = tmp0 + tmp2 + tmp3; + z2 = tmp1 + tmp4; + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) ((z1 + z2 - 9 * CENTERJSAMPLE) << 1); + dataptr[6] = (DCTELEM) + DESCALE(MULTIPLY(z1 - z2 - z2, FIX(0.707106781)), /* c6 */ + CONST_BITS-1); + z1 = MULTIPLY(tmp0 - tmp2, FIX(1.328926049)); /* c2 */ + z2 = MULTIPLY(tmp1 - tmp4 - tmp4, FIX(0.707106781)); /* c6 */ + dataptr[2] = (DCTELEM) + DESCALE(MULTIPLY(tmp2 - tmp3, FIX(1.083350441)) /* c4 */ + + z1 + z2, CONST_BITS-1); + dataptr[4] = (DCTELEM) + DESCALE(MULTIPLY(tmp3 - tmp0, FIX(0.245575608)) /* c8 */ + + z1 - z2, CONST_BITS-1); + + /* Odd part */ + + dataptr[3] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp12 - tmp13, FIX(1.224744871)), /* c3 */ + CONST_BITS-1); + + tmp11 = MULTIPLY(tmp11, FIX(1.224744871)); /* c3 */ + tmp0 = MULTIPLY(tmp10 + tmp12, FIX(0.909038955)); /* c5 */ + tmp1 = MULTIPLY(tmp10 + tmp13, FIX(0.483689525)); /* c7 */ + + dataptr[1] = (DCTELEM) DESCALE(tmp11 + tmp0 + tmp1, CONST_BITS-1); + + tmp2 = MULTIPLY(tmp12 - tmp13, FIX(1.392728481)); /* c1 */ + + dataptr[5] = (DCTELEM) DESCALE(tmp0 - tmp11 - tmp2, CONST_BITS-1); + dataptr[7] = (DCTELEM) DESCALE(tmp1 - tmp11 + tmp2, CONST_BITS-1); + + ctr++; + + if (ctr != DCTSIZE) { + if (ctr == 9) + break; /* Done. */ + dataptr += DCTSIZE; /* advance pointer to next row */ + } else + dataptr = workspace; /* switch pointer to extended workspace */ + } + + /* Pass 2: process columns. + * We leave the results scaled up by an overall factor of 8. + * We must also scale the output by (8/9)**2 = 64/81, which we partially + * fold into the constant multipliers and final/initial shifting: + * cK now represents sqrt(2) * cos(K*pi/18) * 128/81. + */ + + dataptr = data; + wsptr = workspace; + for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { + /* Even part */ + + tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*0]; + tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*7]; + tmp2 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*6]; + tmp3 = dataptr[DCTSIZE*3] + dataptr[DCTSIZE*5]; + tmp4 = dataptr[DCTSIZE*4]; + + tmp10 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*0]; + tmp11 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*7]; + tmp12 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*6]; + tmp13 = dataptr[DCTSIZE*3] - dataptr[DCTSIZE*5]; + + z1 = tmp0 + tmp2 + tmp3; + z2 = tmp1 + tmp4; + dataptr[DCTSIZE*0] = (DCTELEM) + DESCALE(MULTIPLY(z1 + z2, FIX(1.580246914)), /* 128/81 */ + CONST_BITS+2); + dataptr[DCTSIZE*6] = (DCTELEM) + DESCALE(MULTIPLY(z1 - z2 - z2, FIX(1.117403309)), /* c6 */ + CONST_BITS+2); + z1 = MULTIPLY(tmp0 - tmp2, FIX(2.100031287)); /* c2 */ + z2 = MULTIPLY(tmp1 - tmp4 - tmp4, FIX(1.117403309)); /* c6 */ + dataptr[DCTSIZE*2] = (DCTELEM) + DESCALE(MULTIPLY(tmp2 - tmp3, FIX(1.711961190)) /* c4 */ + + z1 + z2, CONST_BITS+2); + dataptr[DCTSIZE*4] = (DCTELEM) + DESCALE(MULTIPLY(tmp3 - tmp0, FIX(0.388070096)) /* c8 */ + + z1 - z2, CONST_BITS+2); + + /* Odd part */ + + dataptr[DCTSIZE*3] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp12 - tmp13, FIX(1.935399303)), /* c3 */ + CONST_BITS+2); + + tmp11 = MULTIPLY(tmp11, FIX(1.935399303)); /* c3 */ + tmp0 = MULTIPLY(tmp10 + tmp12, FIX(1.436506004)); /* c5 */ + tmp1 = MULTIPLY(tmp10 + tmp13, FIX(0.764348879)); /* c7 */ + + dataptr[DCTSIZE*1] = (DCTELEM) + DESCALE(tmp11 + tmp0 + tmp1, CONST_BITS+2); + + tmp2 = MULTIPLY(tmp12 - tmp13, FIX(2.200854883)); /* c1 */ + + dataptr[DCTSIZE*5] = (DCTELEM) + DESCALE(tmp0 - tmp11 - tmp2, CONST_BITS+2); + dataptr[DCTSIZE*7] = (DCTELEM) + DESCALE(tmp1 - tmp11 + tmp2, CONST_BITS+2); + + dataptr++; /* advance pointer to next column */ + wsptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 10x10 sample block. + */ + +GLOBAL(void) +jpeg_fdct_10x10 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3, tmp4; + INT32 tmp10, tmp11, tmp12, tmp13, tmp14; + DCTELEM workspace[8*2]; + DCTELEM *dataptr; + DCTELEM *wsptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT; + * we scale the results further by 2 as part of output adaption + * scaling for different DCT size. + * cK represents sqrt(2) * cos(K*pi/20). + */ + + dataptr = data; + ctr = 0; + for (;;) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[9]); + tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[8]); + tmp12 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[7]); + tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[6]); + tmp4 = GETJSAMPLE(elemptr[4]) + GETJSAMPLE(elemptr[5]); + + tmp10 = tmp0 + tmp4; + tmp13 = tmp0 - tmp4; + tmp11 = tmp1 + tmp3; + tmp14 = tmp1 - tmp3; + + tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[9]); + tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[8]); + tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[7]); + tmp3 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[6]); + tmp4 = GETJSAMPLE(elemptr[4]) - GETJSAMPLE(elemptr[5]); + + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) + ((tmp10 + tmp11 + tmp12 - 10 * CENTERJSAMPLE) << 1); + tmp12 += tmp12; + dataptr[4] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp12, FIX(1.144122806)) - /* c4 */ + MULTIPLY(tmp11 - tmp12, FIX(0.437016024)), /* c8 */ + CONST_BITS-1); + tmp10 = MULTIPLY(tmp13 + tmp14, FIX(0.831253876)); /* c6 */ + dataptr[2] = (DCTELEM) + DESCALE(tmp10 + MULTIPLY(tmp13, FIX(0.513743148)), /* c2-c6 */ + CONST_BITS-1); + dataptr[6] = (DCTELEM) + DESCALE(tmp10 - MULTIPLY(tmp14, FIX(2.176250899)), /* c2+c6 */ + CONST_BITS-1); + + /* Odd part */ + + tmp10 = tmp0 + tmp4; + tmp11 = tmp1 - tmp3; + dataptr[5] = (DCTELEM) ((tmp10 - tmp11 - tmp2) << 1); + tmp2 <<= CONST_BITS; + dataptr[1] = (DCTELEM) + DESCALE(MULTIPLY(tmp0, FIX(1.396802247)) + /* c1 */ + MULTIPLY(tmp1, FIX(1.260073511)) + tmp2 + /* c3 */ + MULTIPLY(tmp3, FIX(0.642039522)) + /* c7 */ + MULTIPLY(tmp4, FIX(0.221231742)), /* c9 */ + CONST_BITS-1); + tmp12 = MULTIPLY(tmp0 - tmp4, FIX(0.951056516)) - /* (c3+c7)/2 */ + MULTIPLY(tmp1 + tmp3, FIX(0.587785252)); /* (c1-c9)/2 */ + tmp13 = MULTIPLY(tmp10 + tmp11, FIX(0.309016994)) + /* (c3-c7)/2 */ + (tmp11 << (CONST_BITS - 1)) - tmp2; + dataptr[3] = (DCTELEM) DESCALE(tmp12 + tmp13, CONST_BITS-1); + dataptr[7] = (DCTELEM) DESCALE(tmp12 - tmp13, CONST_BITS-1); + + ctr++; + + if (ctr != DCTSIZE) { + if (ctr == 10) + break; /* Done. */ + dataptr += DCTSIZE; /* advance pointer to next row */ + } else + dataptr = workspace; /* switch pointer to extended workspace */ + } + + /* Pass 2: process columns. + * We leave the results scaled up by an overall factor of 8. + * We must also scale the output by (8/10)**2 = 16/25, which we partially + * fold into the constant multipliers and final/initial shifting: + * cK now represents sqrt(2) * cos(K*pi/20) * 32/25. + */ + + dataptr = data; + wsptr = workspace; + for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { + /* Even part */ + + tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*1]; + tmp1 = dataptr[DCTSIZE*1] + wsptr[DCTSIZE*0]; + tmp12 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*7]; + tmp3 = dataptr[DCTSIZE*3] + dataptr[DCTSIZE*6]; + tmp4 = dataptr[DCTSIZE*4] + dataptr[DCTSIZE*5]; + + tmp10 = tmp0 + tmp4; + tmp13 = tmp0 - tmp4; + tmp11 = tmp1 + tmp3; + tmp14 = tmp1 - tmp3; + + tmp0 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*1]; + tmp1 = dataptr[DCTSIZE*1] - wsptr[DCTSIZE*0]; + tmp2 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*7]; + tmp3 = dataptr[DCTSIZE*3] - dataptr[DCTSIZE*6]; + tmp4 = dataptr[DCTSIZE*4] - dataptr[DCTSIZE*5]; + + dataptr[DCTSIZE*0] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 + tmp11 + tmp12, FIX(1.28)), /* 32/25 */ + CONST_BITS+2); + tmp12 += tmp12; + dataptr[DCTSIZE*4] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp12, FIX(1.464477191)) - /* c4 */ + MULTIPLY(tmp11 - tmp12, FIX(0.559380511)), /* c8 */ + CONST_BITS+2); + tmp10 = MULTIPLY(tmp13 + tmp14, FIX(1.064004961)); /* c6 */ + dataptr[DCTSIZE*2] = (DCTELEM) + DESCALE(tmp10 + MULTIPLY(tmp13, FIX(0.657591230)), /* c2-c6 */ + CONST_BITS+2); + dataptr[DCTSIZE*6] = (DCTELEM) + DESCALE(tmp10 - MULTIPLY(tmp14, FIX(2.785601151)), /* c2+c6 */ + CONST_BITS+2); + + /* Odd part */ + + tmp10 = tmp0 + tmp4; + tmp11 = tmp1 - tmp3; + dataptr[DCTSIZE*5] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp11 - tmp2, FIX(1.28)), /* 32/25 */ + CONST_BITS+2); + tmp2 = MULTIPLY(tmp2, FIX(1.28)); /* 32/25 */ + dataptr[DCTSIZE*1] = (DCTELEM) + DESCALE(MULTIPLY(tmp0, FIX(1.787906876)) + /* c1 */ + MULTIPLY(tmp1, FIX(1.612894094)) + tmp2 + /* c3 */ + MULTIPLY(tmp3, FIX(0.821810588)) + /* c7 */ + MULTIPLY(tmp4, FIX(0.283176630)), /* c9 */ + CONST_BITS+2); + tmp12 = MULTIPLY(tmp0 - tmp4, FIX(1.217352341)) - /* (c3+c7)/2 */ + MULTIPLY(tmp1 + tmp3, FIX(0.752365123)); /* (c1-c9)/2 */ + tmp13 = MULTIPLY(tmp10 + tmp11, FIX(0.395541753)) + /* (c3-c7)/2 */ + MULTIPLY(tmp11, FIX(0.64)) - tmp2; /* 16/25 */ + dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp12 + tmp13, CONST_BITS+2); + dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp12 - tmp13, CONST_BITS+2); + + dataptr++; /* advance pointer to next column */ + wsptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on an 11x11 sample block. + */ + +GLOBAL(void) +jpeg_fdct_11x11 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5; + INT32 tmp10, tmp11, tmp12, tmp13, tmp14; + INT32 z1, z2, z3; + DCTELEM workspace[8*3]; + DCTELEM *dataptr; + DCTELEM *wsptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT; + * we scale the results further by 2 as part of output adaption + * scaling for different DCT size. + * cK represents sqrt(2) * cos(K*pi/22). + */ + + dataptr = data; + ctr = 0; + for (;;) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[10]); + tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[9]); + tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[8]); + tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[7]); + tmp4 = GETJSAMPLE(elemptr[4]) + GETJSAMPLE(elemptr[6]); + tmp5 = GETJSAMPLE(elemptr[5]); + + tmp10 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[10]); + tmp11 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[9]); + tmp12 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[8]); + tmp13 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[7]); + tmp14 = GETJSAMPLE(elemptr[4]) - GETJSAMPLE(elemptr[6]); + + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) + ((tmp0 + tmp1 + tmp2 + tmp3 + tmp4 + tmp5 - 11 * CENTERJSAMPLE) << 1); + tmp5 += tmp5; + tmp0 -= tmp5; + tmp1 -= tmp5; + tmp2 -= tmp5; + tmp3 -= tmp5; + tmp4 -= tmp5; + z1 = MULTIPLY(tmp0 + tmp3, FIX(1.356927976)) + /* c2 */ + MULTIPLY(tmp2 + tmp4, FIX(0.201263574)); /* c10 */ + z2 = MULTIPLY(tmp1 - tmp3, FIX(0.926112931)); /* c6 */ + z3 = MULTIPLY(tmp0 - tmp1, FIX(1.189712156)); /* c4 */ + dataptr[2] = (DCTELEM) + DESCALE(z1 + z2 - MULTIPLY(tmp3, FIX(1.018300590)) /* c2+c8-c6 */ + - MULTIPLY(tmp4, FIX(1.390975730)), /* c4+c10 */ + CONST_BITS-1); + dataptr[4] = (DCTELEM) + DESCALE(z2 + z3 + MULTIPLY(tmp1, FIX(0.062335650)) /* c4-c6-c10 */ + - MULTIPLY(tmp2, FIX(1.356927976)) /* c2 */ + + MULTIPLY(tmp4, FIX(0.587485545)), /* c8 */ + CONST_BITS-1); + dataptr[6] = (DCTELEM) + DESCALE(z1 + z3 - MULTIPLY(tmp0, FIX(1.620527200)) /* c2+c4-c6 */ + - MULTIPLY(tmp2, FIX(0.788749120)), /* c8+c10 */ + CONST_BITS-1); + + /* Odd part */ + + tmp1 = MULTIPLY(tmp10 + tmp11, FIX(1.286413905)); /* c3 */ + tmp2 = MULTIPLY(tmp10 + tmp12, FIX(1.068791298)); /* c5 */ + tmp3 = MULTIPLY(tmp10 + tmp13, FIX(0.764581576)); /* c7 */ + tmp0 = tmp1 + tmp2 + tmp3 - MULTIPLY(tmp10, FIX(1.719967871)) /* c7+c5+c3-c1 */ + + MULTIPLY(tmp14, FIX(0.398430003)); /* c9 */ + tmp4 = MULTIPLY(tmp11 + tmp12, - FIX(0.764581576)); /* -c7 */ + tmp5 = MULTIPLY(tmp11 + tmp13, - FIX(1.399818907)); /* -c1 */ + tmp1 += tmp4 + tmp5 + MULTIPLY(tmp11, FIX(1.276416582)) /* c9+c7+c1-c3 */ + - MULTIPLY(tmp14, FIX(1.068791298)); /* c5 */ + tmp10 = MULTIPLY(tmp12 + tmp13, FIX(0.398430003)); /* c9 */ + tmp2 += tmp4 + tmp10 - MULTIPLY(tmp12, FIX(1.989053629)) /* c9+c5+c3-c7 */ + + MULTIPLY(tmp14, FIX(1.399818907)); /* c1 */ + tmp3 += tmp5 + tmp10 + MULTIPLY(tmp13, FIX(1.305598626)) /* c1+c5-c9-c7 */ + - MULTIPLY(tmp14, FIX(1.286413905)); /* c3 */ + + dataptr[1] = (DCTELEM) DESCALE(tmp0, CONST_BITS-1); + dataptr[3] = (DCTELEM) DESCALE(tmp1, CONST_BITS-1); + dataptr[5] = (DCTELEM) DESCALE(tmp2, CONST_BITS-1); + dataptr[7] = (DCTELEM) DESCALE(tmp3, CONST_BITS-1); + + ctr++; + + if (ctr != DCTSIZE) { + if (ctr == 11) + break; /* Done. */ + dataptr += DCTSIZE; /* advance pointer to next row */ + } else + dataptr = workspace; /* switch pointer to extended workspace */ + } + + /* Pass 2: process columns. + * We leave the results scaled up by an overall factor of 8. + * We must also scale the output by (8/11)**2 = 64/121, which we partially + * fold into the constant multipliers and final/initial shifting: + * cK now represents sqrt(2) * cos(K*pi/22) * 128/121. + */ + + dataptr = data; + wsptr = workspace; + for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { + /* Even part */ + + tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*2]; + tmp1 = dataptr[DCTSIZE*1] + wsptr[DCTSIZE*1]; + tmp2 = dataptr[DCTSIZE*2] + wsptr[DCTSIZE*0]; + tmp3 = dataptr[DCTSIZE*3] + dataptr[DCTSIZE*7]; + tmp4 = dataptr[DCTSIZE*4] + dataptr[DCTSIZE*6]; + tmp5 = dataptr[DCTSIZE*5]; + + tmp10 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*2]; + tmp11 = dataptr[DCTSIZE*1] - wsptr[DCTSIZE*1]; + tmp12 = dataptr[DCTSIZE*2] - wsptr[DCTSIZE*0]; + tmp13 = dataptr[DCTSIZE*3] - dataptr[DCTSIZE*7]; + tmp14 = dataptr[DCTSIZE*4] - dataptr[DCTSIZE*6]; + + dataptr[DCTSIZE*0] = (DCTELEM) + DESCALE(MULTIPLY(tmp0 + tmp1 + tmp2 + tmp3 + tmp4 + tmp5, + FIX(1.057851240)), /* 128/121 */ + CONST_BITS+2); + tmp5 += tmp5; + tmp0 -= tmp5; + tmp1 -= tmp5; + tmp2 -= tmp5; + tmp3 -= tmp5; + tmp4 -= tmp5; + z1 = MULTIPLY(tmp0 + tmp3, FIX(1.435427942)) + /* c2 */ + MULTIPLY(tmp2 + tmp4, FIX(0.212906922)); /* c10 */ + z2 = MULTIPLY(tmp1 - tmp3, FIX(0.979689713)); /* c6 */ + z3 = MULTIPLY(tmp0 - tmp1, FIX(1.258538479)); /* c4 */ + dataptr[DCTSIZE*2] = (DCTELEM) + DESCALE(z1 + z2 - MULTIPLY(tmp3, FIX(1.077210542)) /* c2+c8-c6 */ + - MULTIPLY(tmp4, FIX(1.471445400)), /* c4+c10 */ + CONST_BITS+2); + dataptr[DCTSIZE*4] = (DCTELEM) + DESCALE(z2 + z3 + MULTIPLY(tmp1, FIX(0.065941844)) /* c4-c6-c10 */ + - MULTIPLY(tmp2, FIX(1.435427942)) /* c2 */ + + MULTIPLY(tmp4, FIX(0.621472312)), /* c8 */ + CONST_BITS+2); + dataptr[DCTSIZE*6] = (DCTELEM) + DESCALE(z1 + z3 - MULTIPLY(tmp0, FIX(1.714276708)) /* c2+c4-c6 */ + - MULTIPLY(tmp2, FIX(0.834379234)), /* c8+c10 */ + CONST_BITS+2); + + /* Odd part */ + + tmp1 = MULTIPLY(tmp10 + tmp11, FIX(1.360834544)); /* c3 */ + tmp2 = MULTIPLY(tmp10 + tmp12, FIX(1.130622199)); /* c5 */ + tmp3 = MULTIPLY(tmp10 + tmp13, FIX(0.808813568)); /* c7 */ + tmp0 = tmp1 + tmp2 + tmp3 - MULTIPLY(tmp10, FIX(1.819470145)) /* c7+c5+c3-c1 */ + + MULTIPLY(tmp14, FIX(0.421479672)); /* c9 */ + tmp4 = MULTIPLY(tmp11 + tmp12, - FIX(0.808813568)); /* -c7 */ + tmp5 = MULTIPLY(tmp11 + tmp13, - FIX(1.480800167)); /* -c1 */ + tmp1 += tmp4 + tmp5 + MULTIPLY(tmp11, FIX(1.350258864)) /* c9+c7+c1-c3 */ + - MULTIPLY(tmp14, FIX(1.130622199)); /* c5 */ + tmp10 = MULTIPLY(tmp12 + tmp13, FIX(0.421479672)); /* c9 */ + tmp2 += tmp4 + tmp10 - MULTIPLY(tmp12, FIX(2.104122847)) /* c9+c5+c3-c7 */ + + MULTIPLY(tmp14, FIX(1.480800167)); /* c1 */ + tmp3 += tmp5 + tmp10 + MULTIPLY(tmp13, FIX(1.381129125)) /* c1+c5-c9-c7 */ + - MULTIPLY(tmp14, FIX(1.360834544)); /* c3 */ + + dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp0, CONST_BITS+2); + dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp1, CONST_BITS+2); + dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp2, CONST_BITS+2); + dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp3, CONST_BITS+2); + + dataptr++; /* advance pointer to next column */ + wsptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 12x12 sample block. + */ + +GLOBAL(void) +jpeg_fdct_12x12 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5; + INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15; + DCTELEM workspace[8*4]; + DCTELEM *dataptr; + DCTELEM *wsptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT. + * cK represents sqrt(2) * cos(K*pi/24). + */ + + dataptr = data; + ctr = 0; + for (;;) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[11]); + tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[10]); + tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[9]); + tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[8]); + tmp4 = GETJSAMPLE(elemptr[4]) + GETJSAMPLE(elemptr[7]); + tmp5 = GETJSAMPLE(elemptr[5]) + GETJSAMPLE(elemptr[6]); + + tmp10 = tmp0 + tmp5; + tmp13 = tmp0 - tmp5; + tmp11 = tmp1 + tmp4; + tmp14 = tmp1 - tmp4; + tmp12 = tmp2 + tmp3; + tmp15 = tmp2 - tmp3; + + tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[11]); + tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[10]); + tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[9]); + tmp3 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[8]); + tmp4 = GETJSAMPLE(elemptr[4]) - GETJSAMPLE(elemptr[7]); + tmp5 = GETJSAMPLE(elemptr[5]) - GETJSAMPLE(elemptr[6]); + + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) (tmp10 + tmp11 + tmp12 - 12 * CENTERJSAMPLE); + dataptr[6] = (DCTELEM) (tmp13 - tmp14 - tmp15); + dataptr[4] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp12, FIX(1.224744871)), /* c4 */ + CONST_BITS); + dataptr[2] = (DCTELEM) + DESCALE(tmp14 - tmp15 + MULTIPLY(tmp13 + tmp15, FIX(1.366025404)), /* c2 */ + CONST_BITS); + + /* Odd part */ + + tmp10 = MULTIPLY(tmp1 + tmp4, FIX_0_541196100); /* c9 */ + tmp14 = tmp10 + MULTIPLY(tmp1, FIX_0_765366865); /* c3-c9 */ + tmp15 = tmp10 - MULTIPLY(tmp4, FIX_1_847759065); /* c3+c9 */ + tmp12 = MULTIPLY(tmp0 + tmp2, FIX(1.121971054)); /* c5 */ + tmp13 = MULTIPLY(tmp0 + tmp3, FIX(0.860918669)); /* c7 */ + tmp10 = tmp12 + tmp13 + tmp14 - MULTIPLY(tmp0, FIX(0.580774953)) /* c5+c7-c1 */ + + MULTIPLY(tmp5, FIX(0.184591911)); /* c11 */ + tmp11 = MULTIPLY(tmp2 + tmp3, - FIX(0.184591911)); /* -c11 */ + tmp12 += tmp11 - tmp15 - MULTIPLY(tmp2, FIX(2.339493912)) /* c1+c5-c11 */ + + MULTIPLY(tmp5, FIX(0.860918669)); /* c7 */ + tmp13 += tmp11 - tmp14 + MULTIPLY(tmp3, FIX(0.725788011)) /* c1+c11-c7 */ + - MULTIPLY(tmp5, FIX(1.121971054)); /* c5 */ + tmp11 = tmp15 + MULTIPLY(tmp0 - tmp3, FIX(1.306562965)) /* c3 */ + - MULTIPLY(tmp2 + tmp5, FIX_0_541196100); /* c9 */ + + dataptr[1] = (DCTELEM) DESCALE(tmp10, CONST_BITS); + dataptr[3] = (DCTELEM) DESCALE(tmp11, CONST_BITS); + dataptr[5] = (DCTELEM) DESCALE(tmp12, CONST_BITS); + dataptr[7] = (DCTELEM) DESCALE(tmp13, CONST_BITS); + + ctr++; + + if (ctr != DCTSIZE) { + if (ctr == 12) + break; /* Done. */ + dataptr += DCTSIZE; /* advance pointer to next row */ + } else + dataptr = workspace; /* switch pointer to extended workspace */ + } + + /* Pass 2: process columns. + * We leave the results scaled up by an overall factor of 8. + * We must also scale the output by (8/12)**2 = 4/9, which we partially + * fold into the constant multipliers and final shifting: + * cK now represents sqrt(2) * cos(K*pi/24) * 8/9. + */ + + dataptr = data; + wsptr = workspace; + for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { + /* Even part */ + + tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*3]; + tmp1 = dataptr[DCTSIZE*1] + wsptr[DCTSIZE*2]; + tmp2 = dataptr[DCTSIZE*2] + wsptr[DCTSIZE*1]; + tmp3 = dataptr[DCTSIZE*3] + wsptr[DCTSIZE*0]; + tmp4 = dataptr[DCTSIZE*4] + dataptr[DCTSIZE*7]; + tmp5 = dataptr[DCTSIZE*5] + dataptr[DCTSIZE*6]; + + tmp10 = tmp0 + tmp5; + tmp13 = tmp0 - tmp5; + tmp11 = tmp1 + tmp4; + tmp14 = tmp1 - tmp4; + tmp12 = tmp2 + tmp3; + tmp15 = tmp2 - tmp3; + + tmp0 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*3]; + tmp1 = dataptr[DCTSIZE*1] - wsptr[DCTSIZE*2]; + tmp2 = dataptr[DCTSIZE*2] - wsptr[DCTSIZE*1]; + tmp3 = dataptr[DCTSIZE*3] - wsptr[DCTSIZE*0]; + tmp4 = dataptr[DCTSIZE*4] - dataptr[DCTSIZE*7]; + tmp5 = dataptr[DCTSIZE*5] - dataptr[DCTSIZE*6]; + + dataptr[DCTSIZE*0] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 + tmp11 + tmp12, FIX(0.888888889)), /* 8/9 */ + CONST_BITS+1); + dataptr[DCTSIZE*6] = (DCTELEM) + DESCALE(MULTIPLY(tmp13 - tmp14 - tmp15, FIX(0.888888889)), /* 8/9 */ + CONST_BITS+1); + dataptr[DCTSIZE*4] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp12, FIX(1.088662108)), /* c4 */ + CONST_BITS+1); + dataptr[DCTSIZE*2] = (DCTELEM) + DESCALE(MULTIPLY(tmp14 - tmp15, FIX(0.888888889)) + /* 8/9 */ + MULTIPLY(tmp13 + tmp15, FIX(1.214244803)), /* c2 */ + CONST_BITS+1); + + /* Odd part */ + + tmp10 = MULTIPLY(tmp1 + tmp4, FIX(0.481063200)); /* c9 */ + tmp14 = tmp10 + MULTIPLY(tmp1, FIX(0.680326102)); /* c3-c9 */ + tmp15 = tmp10 - MULTIPLY(tmp4, FIX(1.642452502)); /* c3+c9 */ + tmp12 = MULTIPLY(tmp0 + tmp2, FIX(0.997307603)); /* c5 */ + tmp13 = MULTIPLY(tmp0 + tmp3, FIX(0.765261039)); /* c7 */ + tmp10 = tmp12 + tmp13 + tmp14 - MULTIPLY(tmp0, FIX(0.516244403)) /* c5+c7-c1 */ + + MULTIPLY(tmp5, FIX(0.164081699)); /* c11 */ + tmp11 = MULTIPLY(tmp2 + tmp3, - FIX(0.164081699)); /* -c11 */ + tmp12 += tmp11 - tmp15 - MULTIPLY(tmp2, FIX(2.079550144)) /* c1+c5-c11 */ + + MULTIPLY(tmp5, FIX(0.765261039)); /* c7 */ + tmp13 += tmp11 - tmp14 + MULTIPLY(tmp3, FIX(0.645144899)) /* c1+c11-c7 */ + - MULTIPLY(tmp5, FIX(0.997307603)); /* c5 */ + tmp11 = tmp15 + MULTIPLY(tmp0 - tmp3, FIX(1.161389302)) /* c3 */ + - MULTIPLY(tmp2 + tmp5, FIX(0.481063200)); /* c9 */ + + dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp10, CONST_BITS+1); + dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp11, CONST_BITS+1); + dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp12, CONST_BITS+1); + dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp13, CONST_BITS+1); + + dataptr++; /* advance pointer to next column */ + wsptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 13x13 sample block. + */ + +GLOBAL(void) +jpeg_fdct_13x13 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6; + INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15; + INT32 z1, z2; + DCTELEM workspace[8*5]; + DCTELEM *dataptr; + DCTELEM *wsptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT. + * cK represents sqrt(2) * cos(K*pi/26). + */ + + dataptr = data; + ctr = 0; + for (;;) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[12]); + tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[11]); + tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[10]); + tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[9]); + tmp4 = GETJSAMPLE(elemptr[4]) + GETJSAMPLE(elemptr[8]); + tmp5 = GETJSAMPLE(elemptr[5]) + GETJSAMPLE(elemptr[7]); + tmp6 = GETJSAMPLE(elemptr[6]); + + tmp10 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[12]); + tmp11 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[11]); + tmp12 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[10]); + tmp13 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[9]); + tmp14 = GETJSAMPLE(elemptr[4]) - GETJSAMPLE(elemptr[8]); + tmp15 = GETJSAMPLE(elemptr[5]) - GETJSAMPLE(elemptr[7]); + + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) + (tmp0 + tmp1 + tmp2 + tmp3 + tmp4 + tmp5 + tmp6 - 13 * CENTERJSAMPLE); + tmp6 += tmp6; + tmp0 -= tmp6; + tmp1 -= tmp6; + tmp2 -= tmp6; + tmp3 -= tmp6; + tmp4 -= tmp6; + tmp5 -= tmp6; + dataptr[2] = (DCTELEM) + DESCALE(MULTIPLY(tmp0, FIX(1.373119086)) + /* c2 */ + MULTIPLY(tmp1, FIX(1.058554052)) + /* c6 */ + MULTIPLY(tmp2, FIX(0.501487041)) - /* c10 */ + MULTIPLY(tmp3, FIX(0.170464608)) - /* c12 */ + MULTIPLY(tmp4, FIX(0.803364869)) - /* c8 */ + MULTIPLY(tmp5, FIX(1.252223920)), /* c4 */ + CONST_BITS); + z1 = MULTIPLY(tmp0 - tmp2, FIX(1.155388986)) - /* (c4+c6)/2 */ + MULTIPLY(tmp3 - tmp4, FIX(0.435816023)) - /* (c2-c10)/2 */ + MULTIPLY(tmp1 - tmp5, FIX(0.316450131)); /* (c8-c12)/2 */ + z2 = MULTIPLY(tmp0 + tmp2, FIX(0.096834934)) - /* (c4-c6)/2 */ + MULTIPLY(tmp3 + tmp4, FIX(0.937303064)) + /* (c2+c10)/2 */ + MULTIPLY(tmp1 + tmp5, FIX(0.486914739)); /* (c8+c12)/2 */ + + dataptr[4] = (DCTELEM) DESCALE(z1 + z2, CONST_BITS); + dataptr[6] = (DCTELEM) DESCALE(z1 - z2, CONST_BITS); + + /* Odd part */ + + tmp1 = MULTIPLY(tmp10 + tmp11, FIX(1.322312651)); /* c3 */ + tmp2 = MULTIPLY(tmp10 + tmp12, FIX(1.163874945)); /* c5 */ + tmp3 = MULTIPLY(tmp10 + tmp13, FIX(0.937797057)) + /* c7 */ + MULTIPLY(tmp14 + tmp15, FIX(0.338443458)); /* c11 */ + tmp0 = tmp1 + tmp2 + tmp3 - + MULTIPLY(tmp10, FIX(2.020082300)) + /* c3+c5+c7-c1 */ + MULTIPLY(tmp14, FIX(0.318774355)); /* c9-c11 */ + tmp4 = MULTIPLY(tmp14 - tmp15, FIX(0.937797057)) - /* c7 */ + MULTIPLY(tmp11 + tmp12, FIX(0.338443458)); /* c11 */ + tmp5 = MULTIPLY(tmp11 + tmp13, - FIX(1.163874945)); /* -c5 */ + tmp1 += tmp4 + tmp5 + + MULTIPLY(tmp11, FIX(0.837223564)) - /* c5+c9+c11-c3 */ + MULTIPLY(tmp14, FIX(2.341699410)); /* c1+c7 */ + tmp6 = MULTIPLY(tmp12 + tmp13, - FIX(0.657217813)); /* -c9 */ + tmp2 += tmp4 + tmp6 - + MULTIPLY(tmp12, FIX(1.572116027)) + /* c1+c5-c9-c11 */ + MULTIPLY(tmp15, FIX(2.260109708)); /* c3+c7 */ + tmp3 += tmp5 + tmp6 + + MULTIPLY(tmp13, FIX(2.205608352)) - /* c3+c5+c9-c7 */ + MULTIPLY(tmp15, FIX(1.742345811)); /* c1+c11 */ + + dataptr[1] = (DCTELEM) DESCALE(tmp0, CONST_BITS); + dataptr[3] = (DCTELEM) DESCALE(tmp1, CONST_BITS); + dataptr[5] = (DCTELEM) DESCALE(tmp2, CONST_BITS); + dataptr[7] = (DCTELEM) DESCALE(tmp3, CONST_BITS); + + ctr++; + + if (ctr != DCTSIZE) { + if (ctr == 13) + break; /* Done. */ + dataptr += DCTSIZE; /* advance pointer to next row */ + } else + dataptr = workspace; /* switch pointer to extended workspace */ + } + + /* Pass 2: process columns. + * We leave the results scaled up by an overall factor of 8. + * We must also scale the output by (8/13)**2 = 64/169, which we partially + * fold into the constant multipliers and final shifting: + * cK now represents sqrt(2) * cos(K*pi/26) * 128/169. + */ + + dataptr = data; + wsptr = workspace; + for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { + /* Even part */ + + tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*4]; + tmp1 = dataptr[DCTSIZE*1] + wsptr[DCTSIZE*3]; + tmp2 = dataptr[DCTSIZE*2] + wsptr[DCTSIZE*2]; + tmp3 = dataptr[DCTSIZE*3] + wsptr[DCTSIZE*1]; + tmp4 = dataptr[DCTSIZE*4] + wsptr[DCTSIZE*0]; + tmp5 = dataptr[DCTSIZE*5] + dataptr[DCTSIZE*7]; + tmp6 = dataptr[DCTSIZE*6]; + + tmp10 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*4]; + tmp11 = dataptr[DCTSIZE*1] - wsptr[DCTSIZE*3]; + tmp12 = dataptr[DCTSIZE*2] - wsptr[DCTSIZE*2]; + tmp13 = dataptr[DCTSIZE*3] - wsptr[DCTSIZE*1]; + tmp14 = dataptr[DCTSIZE*4] - wsptr[DCTSIZE*0]; + tmp15 = dataptr[DCTSIZE*5] - dataptr[DCTSIZE*7]; + + dataptr[DCTSIZE*0] = (DCTELEM) + DESCALE(MULTIPLY(tmp0 + tmp1 + tmp2 + tmp3 + tmp4 + tmp5 + tmp6, + FIX(0.757396450)), /* 128/169 */ + CONST_BITS+1); + tmp6 += tmp6; + tmp0 -= tmp6; + tmp1 -= tmp6; + tmp2 -= tmp6; + tmp3 -= tmp6; + tmp4 -= tmp6; + tmp5 -= tmp6; + dataptr[DCTSIZE*2] = (DCTELEM) + DESCALE(MULTIPLY(tmp0, FIX(1.039995521)) + /* c2 */ + MULTIPLY(tmp1, FIX(0.801745081)) + /* c6 */ + MULTIPLY(tmp2, FIX(0.379824504)) - /* c10 */ + MULTIPLY(tmp3, FIX(0.129109289)) - /* c12 */ + MULTIPLY(tmp4, FIX(0.608465700)) - /* c8 */ + MULTIPLY(tmp5, FIX(0.948429952)), /* c4 */ + CONST_BITS+1); + z1 = MULTIPLY(tmp0 - tmp2, FIX(0.875087516)) - /* (c4+c6)/2 */ + MULTIPLY(tmp3 - tmp4, FIX(0.330085509)) - /* (c2-c10)/2 */ + MULTIPLY(tmp1 - tmp5, FIX(0.239678205)); /* (c8-c12)/2 */ + z2 = MULTIPLY(tmp0 + tmp2, FIX(0.073342435)) - /* (c4-c6)/2 */ + MULTIPLY(tmp3 + tmp4, FIX(0.709910013)) + /* (c2+c10)/2 */ + MULTIPLY(tmp1 + tmp5, FIX(0.368787494)); /* (c8+c12)/2 */ + + dataptr[DCTSIZE*4] = (DCTELEM) DESCALE(z1 + z2, CONST_BITS+1); + dataptr[DCTSIZE*6] = (DCTELEM) DESCALE(z1 - z2, CONST_BITS+1); + + /* Odd part */ + + tmp1 = MULTIPLY(tmp10 + tmp11, FIX(1.001514908)); /* c3 */ + tmp2 = MULTIPLY(tmp10 + tmp12, FIX(0.881514751)); /* c5 */ + tmp3 = MULTIPLY(tmp10 + tmp13, FIX(0.710284161)) + /* c7 */ + MULTIPLY(tmp14 + tmp15, FIX(0.256335874)); /* c11 */ + tmp0 = tmp1 + tmp2 + tmp3 - + MULTIPLY(tmp10, FIX(1.530003162)) + /* c3+c5+c7-c1 */ + MULTIPLY(tmp14, FIX(0.241438564)); /* c9-c11 */ + tmp4 = MULTIPLY(tmp14 - tmp15, FIX(0.710284161)) - /* c7 */ + MULTIPLY(tmp11 + tmp12, FIX(0.256335874)); /* c11 */ + tmp5 = MULTIPLY(tmp11 + tmp13, - FIX(0.881514751)); /* -c5 */ + tmp1 += tmp4 + tmp5 + + MULTIPLY(tmp11, FIX(0.634110155)) - /* c5+c9+c11-c3 */ + MULTIPLY(tmp14, FIX(1.773594819)); /* c1+c7 */ + tmp6 = MULTIPLY(tmp12 + tmp13, - FIX(0.497774438)); /* -c9 */ + tmp2 += tmp4 + tmp6 - + MULTIPLY(tmp12, FIX(1.190715098)) + /* c1+c5-c9-c11 */ + MULTIPLY(tmp15, FIX(1.711799069)); /* c3+c7 */ + tmp3 += tmp5 + tmp6 + + MULTIPLY(tmp13, FIX(1.670519935)) - /* c3+c5+c9-c7 */ + MULTIPLY(tmp15, FIX(1.319646532)); /* c1+c11 */ + + dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp0, CONST_BITS+1); + dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp1, CONST_BITS+1); + dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp2, CONST_BITS+1); + dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp3, CONST_BITS+1); + + dataptr++; /* advance pointer to next column */ + wsptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 14x14 sample block. + */ + +GLOBAL(void) +jpeg_fdct_14x14 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6; + INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15, tmp16; + DCTELEM workspace[8*6]; + DCTELEM *dataptr; + DCTELEM *wsptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT. + * cK represents sqrt(2) * cos(K*pi/28). + */ + + dataptr = data; + ctr = 0; + for (;;) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[13]); + tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[12]); + tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[11]); + tmp13 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[10]); + tmp4 = GETJSAMPLE(elemptr[4]) + GETJSAMPLE(elemptr[9]); + tmp5 = GETJSAMPLE(elemptr[5]) + GETJSAMPLE(elemptr[8]); + tmp6 = GETJSAMPLE(elemptr[6]) + GETJSAMPLE(elemptr[7]); + + tmp10 = tmp0 + tmp6; + tmp14 = tmp0 - tmp6; + tmp11 = tmp1 + tmp5; + tmp15 = tmp1 - tmp5; + tmp12 = tmp2 + tmp4; + tmp16 = tmp2 - tmp4; + + tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[13]); + tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[12]); + tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[11]); + tmp3 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[10]); + tmp4 = GETJSAMPLE(elemptr[4]) - GETJSAMPLE(elemptr[9]); + tmp5 = GETJSAMPLE(elemptr[5]) - GETJSAMPLE(elemptr[8]); + tmp6 = GETJSAMPLE(elemptr[6]) - GETJSAMPLE(elemptr[7]); + + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) + (tmp10 + tmp11 + tmp12 + tmp13 - 14 * CENTERJSAMPLE); + tmp13 += tmp13; + dataptr[4] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp13, FIX(1.274162392)) + /* c4 */ + MULTIPLY(tmp11 - tmp13, FIX(0.314692123)) - /* c12 */ + MULTIPLY(tmp12 - tmp13, FIX(0.881747734)), /* c8 */ + CONST_BITS); + + tmp10 = MULTIPLY(tmp14 + tmp15, FIX(1.105676686)); /* c6 */ + + dataptr[2] = (DCTELEM) + DESCALE(tmp10 + MULTIPLY(tmp14, FIX(0.273079590)) /* c2-c6 */ + + MULTIPLY(tmp16, FIX(0.613604268)), /* c10 */ + CONST_BITS); + dataptr[6] = (DCTELEM) + DESCALE(tmp10 - MULTIPLY(tmp15, FIX(1.719280954)) /* c6+c10 */ + - MULTIPLY(tmp16, FIX(1.378756276)), /* c2 */ + CONST_BITS); + + /* Odd part */ + + tmp10 = tmp1 + tmp2; + tmp11 = tmp5 - tmp4; + dataptr[7] = (DCTELEM) (tmp0 - tmp10 + tmp3 - tmp11 - tmp6); + tmp3 <<= CONST_BITS; + tmp10 = MULTIPLY(tmp10, - FIX(0.158341681)); /* -c13 */ + tmp11 = MULTIPLY(tmp11, FIX(1.405321284)); /* c1 */ + tmp10 += tmp11 - tmp3; + tmp11 = MULTIPLY(tmp0 + tmp2, FIX(1.197448846)) + /* c5 */ + MULTIPLY(tmp4 + tmp6, FIX(0.752406978)); /* c9 */ + dataptr[5] = (DCTELEM) + DESCALE(tmp10 + tmp11 - MULTIPLY(tmp2, FIX(2.373959773)) /* c3+c5-c13 */ + + MULTIPLY(tmp4, FIX(1.119999435)), /* c1+c11-c9 */ + CONST_BITS); + tmp12 = MULTIPLY(tmp0 + tmp1, FIX(1.334852607)) + /* c3 */ + MULTIPLY(tmp5 - tmp6, FIX(0.467085129)); /* c11 */ + dataptr[3] = (DCTELEM) + DESCALE(tmp10 + tmp12 - MULTIPLY(tmp1, FIX(0.424103948)) /* c3-c9-c13 */ + - MULTIPLY(tmp5, FIX(3.069855259)), /* c1+c5+c11 */ + CONST_BITS); + dataptr[1] = (DCTELEM) + DESCALE(tmp11 + tmp12 + tmp3 + tmp6 - + MULTIPLY(tmp0 + tmp6, FIX(1.126980169)), /* c3+c5-c1 */ + CONST_BITS); + + ctr++; + + if (ctr != DCTSIZE) { + if (ctr == 14) + break; /* Done. */ + dataptr += DCTSIZE; /* advance pointer to next row */ + } else + dataptr = workspace; /* switch pointer to extended workspace */ + } + + /* Pass 2: process columns. + * We leave the results scaled up by an overall factor of 8. + * We must also scale the output by (8/14)**2 = 16/49, which we partially + * fold into the constant multipliers and final shifting: + * cK now represents sqrt(2) * cos(K*pi/28) * 32/49. + */ + + dataptr = data; + wsptr = workspace; + for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { + /* Even part */ + + tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*5]; + tmp1 = dataptr[DCTSIZE*1] + wsptr[DCTSIZE*4]; + tmp2 = dataptr[DCTSIZE*2] + wsptr[DCTSIZE*3]; + tmp13 = dataptr[DCTSIZE*3] + wsptr[DCTSIZE*2]; + tmp4 = dataptr[DCTSIZE*4] + wsptr[DCTSIZE*1]; + tmp5 = dataptr[DCTSIZE*5] + wsptr[DCTSIZE*0]; + tmp6 = dataptr[DCTSIZE*6] + dataptr[DCTSIZE*7]; + + tmp10 = tmp0 + tmp6; + tmp14 = tmp0 - tmp6; + tmp11 = tmp1 + tmp5; + tmp15 = tmp1 - tmp5; + tmp12 = tmp2 + tmp4; + tmp16 = tmp2 - tmp4; + + tmp0 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*5]; + tmp1 = dataptr[DCTSIZE*1] - wsptr[DCTSIZE*4]; + tmp2 = dataptr[DCTSIZE*2] - wsptr[DCTSIZE*3]; + tmp3 = dataptr[DCTSIZE*3] - wsptr[DCTSIZE*2]; + tmp4 = dataptr[DCTSIZE*4] - wsptr[DCTSIZE*1]; + tmp5 = dataptr[DCTSIZE*5] - wsptr[DCTSIZE*0]; + tmp6 = dataptr[DCTSIZE*6] - dataptr[DCTSIZE*7]; + + dataptr[DCTSIZE*0] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 + tmp11 + tmp12 + tmp13, + FIX(0.653061224)), /* 32/49 */ + CONST_BITS+1); + tmp13 += tmp13; + dataptr[DCTSIZE*4] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp13, FIX(0.832106052)) + /* c4 */ + MULTIPLY(tmp11 - tmp13, FIX(0.205513223)) - /* c12 */ + MULTIPLY(tmp12 - tmp13, FIX(0.575835255)), /* c8 */ + CONST_BITS+1); + + tmp10 = MULTIPLY(tmp14 + tmp15, FIX(0.722074570)); /* c6 */ + + dataptr[DCTSIZE*2] = (DCTELEM) + DESCALE(tmp10 + MULTIPLY(tmp14, FIX(0.178337691)) /* c2-c6 */ + + MULTIPLY(tmp16, FIX(0.400721155)), /* c10 */ + CONST_BITS+1); + dataptr[DCTSIZE*6] = (DCTELEM) + DESCALE(tmp10 - MULTIPLY(tmp15, FIX(1.122795725)) /* c6+c10 */ + - MULTIPLY(tmp16, FIX(0.900412262)), /* c2 */ + CONST_BITS+1); + + /* Odd part */ + + tmp10 = tmp1 + tmp2; + tmp11 = tmp5 - tmp4; + dataptr[DCTSIZE*7] = (DCTELEM) + DESCALE(MULTIPLY(tmp0 - tmp10 + tmp3 - tmp11 - tmp6, + FIX(0.653061224)), /* 32/49 */ + CONST_BITS+1); + tmp3 = MULTIPLY(tmp3 , FIX(0.653061224)); /* 32/49 */ + tmp10 = MULTIPLY(tmp10, - FIX(0.103406812)); /* -c13 */ + tmp11 = MULTIPLY(tmp11, FIX(0.917760839)); /* c1 */ + tmp10 += tmp11 - tmp3; + tmp11 = MULTIPLY(tmp0 + tmp2, FIX(0.782007410)) + /* c5 */ + MULTIPLY(tmp4 + tmp6, FIX(0.491367823)); /* c9 */ + dataptr[DCTSIZE*5] = (DCTELEM) + DESCALE(tmp10 + tmp11 - MULTIPLY(tmp2, FIX(1.550341076)) /* c3+c5-c13 */ + + MULTIPLY(tmp4, FIX(0.731428202)), /* c1+c11-c9 */ + CONST_BITS+1); + tmp12 = MULTIPLY(tmp0 + tmp1, FIX(0.871740478)) + /* c3 */ + MULTIPLY(tmp5 - tmp6, FIX(0.305035186)); /* c11 */ + dataptr[DCTSIZE*3] = (DCTELEM) + DESCALE(tmp10 + tmp12 - MULTIPLY(tmp1, FIX(0.276965844)) /* c3-c9-c13 */ + - MULTIPLY(tmp5, FIX(2.004803435)), /* c1+c5+c11 */ + CONST_BITS+1); + dataptr[DCTSIZE*1] = (DCTELEM) + DESCALE(tmp11 + tmp12 + tmp3 + - MULTIPLY(tmp0, FIX(0.735987049)) /* c3+c5-c1 */ + - MULTIPLY(tmp6, FIX(0.082925825)), /* c9-c11-c13 */ + CONST_BITS+1); + + dataptr++; /* advance pointer to next column */ + wsptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 15x15 sample block. + */ + +GLOBAL(void) +jpeg_fdct_15x15 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7; + INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15, tmp16; + INT32 z1, z2, z3; + DCTELEM workspace[8*7]; + DCTELEM *dataptr; + DCTELEM *wsptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT. + * cK represents sqrt(2) * cos(K*pi/30). + */ + + dataptr = data; + ctr = 0; + for (;;) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[14]); + tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[13]); + tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[12]); + tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[11]); + tmp4 = GETJSAMPLE(elemptr[4]) + GETJSAMPLE(elemptr[10]); + tmp5 = GETJSAMPLE(elemptr[5]) + GETJSAMPLE(elemptr[9]); + tmp6 = GETJSAMPLE(elemptr[6]) + GETJSAMPLE(elemptr[8]); + tmp7 = GETJSAMPLE(elemptr[7]); + + tmp10 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[14]); + tmp11 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[13]); + tmp12 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[12]); + tmp13 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[11]); + tmp14 = GETJSAMPLE(elemptr[4]) - GETJSAMPLE(elemptr[10]); + tmp15 = GETJSAMPLE(elemptr[5]) - GETJSAMPLE(elemptr[9]); + tmp16 = GETJSAMPLE(elemptr[6]) - GETJSAMPLE(elemptr[8]); + + z1 = tmp0 + tmp4 + tmp5; + z2 = tmp1 + tmp3 + tmp6; + z3 = tmp2 + tmp7; + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) (z1 + z2 + z3 - 15 * CENTERJSAMPLE); + z3 += z3; + dataptr[6] = (DCTELEM) + DESCALE(MULTIPLY(z1 - z3, FIX(1.144122806)) - /* c6 */ + MULTIPLY(z2 - z3, FIX(0.437016024)), /* c12 */ + CONST_BITS); + tmp2 += ((tmp1 + tmp4) >> 1) - tmp7 - tmp7; + z1 = MULTIPLY(tmp3 - tmp2, FIX(1.531135173)) - /* c2+c14 */ + MULTIPLY(tmp6 - tmp2, FIX(2.238241955)); /* c4+c8 */ + z2 = MULTIPLY(tmp5 - tmp2, FIX(0.798468008)) - /* c8-c14 */ + MULTIPLY(tmp0 - tmp2, FIX(0.091361227)); /* c2-c4 */ + z3 = MULTIPLY(tmp0 - tmp3, FIX(1.383309603)) + /* c2 */ + MULTIPLY(tmp6 - tmp5, FIX(0.946293579)) + /* c8 */ + MULTIPLY(tmp1 - tmp4, FIX(0.790569415)); /* (c6+c12)/2 */ + + dataptr[2] = (DCTELEM) DESCALE(z1 + z3, CONST_BITS); + dataptr[4] = (DCTELEM) DESCALE(z2 + z3, CONST_BITS); + + /* Odd part */ + + tmp2 = MULTIPLY(tmp10 - tmp12 - tmp13 + tmp15 + tmp16, + FIX(1.224744871)); /* c5 */ + tmp1 = MULTIPLY(tmp10 - tmp14 - tmp15, FIX(1.344997024)) + /* c3 */ + MULTIPLY(tmp11 - tmp13 - tmp16, FIX(0.831253876)); /* c9 */ + tmp12 = MULTIPLY(tmp12, FIX(1.224744871)); /* c5 */ + tmp4 = MULTIPLY(tmp10 - tmp16, FIX(1.406466353)) + /* c1 */ + MULTIPLY(tmp11 + tmp14, FIX(1.344997024)) + /* c3 */ + MULTIPLY(tmp13 + tmp15, FIX(0.575212477)); /* c11 */ + tmp0 = MULTIPLY(tmp13, FIX(0.475753014)) - /* c7-c11 */ + MULTIPLY(tmp14, FIX(0.513743148)) + /* c3-c9 */ + MULTIPLY(tmp16, FIX(1.700497885)) + tmp4 + tmp12; /* c1+c13 */ + tmp3 = MULTIPLY(tmp10, - FIX(0.355500862)) - /* -(c1-c7) */ + MULTIPLY(tmp11, FIX(2.176250899)) - /* c3+c9 */ + MULTIPLY(tmp15, FIX(0.869244010)) + tmp4 - tmp12; /* c11+c13 */ + + dataptr[1] = (DCTELEM) DESCALE(tmp0, CONST_BITS); + dataptr[3] = (DCTELEM) DESCALE(tmp1, CONST_BITS); + dataptr[5] = (DCTELEM) DESCALE(tmp2, CONST_BITS); + dataptr[7] = (DCTELEM) DESCALE(tmp3, CONST_BITS); + + ctr++; + + if (ctr != DCTSIZE) { + if (ctr == 15) + break; /* Done. */ + dataptr += DCTSIZE; /* advance pointer to next row */ + } else + dataptr = workspace; /* switch pointer to extended workspace */ + } + + /* Pass 2: process columns. + * We leave the results scaled up by an overall factor of 8. + * We must also scale the output by (8/15)**2 = 64/225, which we partially + * fold into the constant multipliers and final shifting: + * cK now represents sqrt(2) * cos(K*pi/30) * 256/225. + */ + + dataptr = data; + wsptr = workspace; + for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { + /* Even part */ + + tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*6]; + tmp1 = dataptr[DCTSIZE*1] + wsptr[DCTSIZE*5]; + tmp2 = dataptr[DCTSIZE*2] + wsptr[DCTSIZE*4]; + tmp3 = dataptr[DCTSIZE*3] + wsptr[DCTSIZE*3]; + tmp4 = dataptr[DCTSIZE*4] + wsptr[DCTSIZE*2]; + tmp5 = dataptr[DCTSIZE*5] + wsptr[DCTSIZE*1]; + tmp6 = dataptr[DCTSIZE*6] + wsptr[DCTSIZE*0]; + tmp7 = dataptr[DCTSIZE*7]; + + tmp10 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*6]; + tmp11 = dataptr[DCTSIZE*1] - wsptr[DCTSIZE*5]; + tmp12 = dataptr[DCTSIZE*2] - wsptr[DCTSIZE*4]; + tmp13 = dataptr[DCTSIZE*3] - wsptr[DCTSIZE*3]; + tmp14 = dataptr[DCTSIZE*4] - wsptr[DCTSIZE*2]; + tmp15 = dataptr[DCTSIZE*5] - wsptr[DCTSIZE*1]; + tmp16 = dataptr[DCTSIZE*6] - wsptr[DCTSIZE*0]; + + z1 = tmp0 + tmp4 + tmp5; + z2 = tmp1 + tmp3 + tmp6; + z3 = tmp2 + tmp7; + dataptr[DCTSIZE*0] = (DCTELEM) + DESCALE(MULTIPLY(z1 + z2 + z3, FIX(1.137777778)), /* 256/225 */ + CONST_BITS+2); + z3 += z3; + dataptr[DCTSIZE*6] = (DCTELEM) + DESCALE(MULTIPLY(z1 - z3, FIX(1.301757503)) - /* c6 */ + MULTIPLY(z2 - z3, FIX(0.497227121)), /* c12 */ + CONST_BITS+2); + tmp2 += ((tmp1 + tmp4) >> 1) - tmp7 - tmp7; + z1 = MULTIPLY(tmp3 - tmp2, FIX(1.742091575)) - /* c2+c14 */ + MULTIPLY(tmp6 - tmp2, FIX(2.546621957)); /* c4+c8 */ + z2 = MULTIPLY(tmp5 - tmp2, FIX(0.908479156)) - /* c8-c14 */ + MULTIPLY(tmp0 - tmp2, FIX(0.103948774)); /* c2-c4 */ + z3 = MULTIPLY(tmp0 - tmp3, FIX(1.573898926)) + /* c2 */ + MULTIPLY(tmp6 - tmp5, FIX(1.076671805)) + /* c8 */ + MULTIPLY(tmp1 - tmp4, FIX(0.899492312)); /* (c6+c12)/2 */ + + dataptr[DCTSIZE*2] = (DCTELEM) DESCALE(z1 + z3, CONST_BITS+2); + dataptr[DCTSIZE*4] = (DCTELEM) DESCALE(z2 + z3, CONST_BITS+2); + + /* Odd part */ + + tmp2 = MULTIPLY(tmp10 - tmp12 - tmp13 + tmp15 + tmp16, + FIX(1.393487498)); /* c5 */ + tmp1 = MULTIPLY(tmp10 - tmp14 - tmp15, FIX(1.530307725)) + /* c3 */ + MULTIPLY(tmp11 - tmp13 - tmp16, FIX(0.945782187)); /* c9 */ + tmp12 = MULTIPLY(tmp12, FIX(1.393487498)); /* c5 */ + tmp4 = MULTIPLY(tmp10 - tmp16, FIX(1.600246161)) + /* c1 */ + MULTIPLY(tmp11 + tmp14, FIX(1.530307725)) + /* c3 */ + MULTIPLY(tmp13 + tmp15, FIX(0.654463974)); /* c11 */ + tmp0 = MULTIPLY(tmp13, FIX(0.541301207)) - /* c7-c11 */ + MULTIPLY(tmp14, FIX(0.584525538)) + /* c3-c9 */ + MULTIPLY(tmp16, FIX(1.934788705)) + tmp4 + tmp12; /* c1+c13 */ + tmp3 = MULTIPLY(tmp10, - FIX(0.404480980)) - /* -(c1-c7) */ + MULTIPLY(tmp11, FIX(2.476089912)) - /* c3+c9 */ + MULTIPLY(tmp15, FIX(0.989006518)) + tmp4 - tmp12; /* c11+c13 */ + + dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp0, CONST_BITS+2); + dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp1, CONST_BITS+2); + dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp2, CONST_BITS+2); + dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp3, CONST_BITS+2); + + dataptr++; /* advance pointer to next column */ + wsptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 16x16 sample block. + */ + +GLOBAL(void) +jpeg_fdct_16x16 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7; + INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15, tmp16, tmp17; + DCTELEM workspace[DCTSIZE2]; + DCTELEM *dataptr; + DCTELEM *wsptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT; + * furthermore, we scale the results by 2**PASS1_BITS. + * cK represents sqrt(2) * cos(K*pi/32). + */ + + dataptr = data; + ctr = 0; + for (;;) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[15]); + tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[14]); + tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[13]); + tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[12]); + tmp4 = GETJSAMPLE(elemptr[4]) + GETJSAMPLE(elemptr[11]); + tmp5 = GETJSAMPLE(elemptr[5]) + GETJSAMPLE(elemptr[10]); + tmp6 = GETJSAMPLE(elemptr[6]) + GETJSAMPLE(elemptr[9]); + tmp7 = GETJSAMPLE(elemptr[7]) + GETJSAMPLE(elemptr[8]); + + tmp10 = tmp0 + tmp7; + tmp14 = tmp0 - tmp7; + tmp11 = tmp1 + tmp6; + tmp15 = tmp1 - tmp6; + tmp12 = tmp2 + tmp5; + tmp16 = tmp2 - tmp5; + tmp13 = tmp3 + tmp4; + tmp17 = tmp3 - tmp4; + + tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[15]); + tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[14]); + tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[13]); + tmp3 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[12]); + tmp4 = GETJSAMPLE(elemptr[4]) - GETJSAMPLE(elemptr[11]); + tmp5 = GETJSAMPLE(elemptr[5]) - GETJSAMPLE(elemptr[10]); + tmp6 = GETJSAMPLE(elemptr[6]) - GETJSAMPLE(elemptr[9]); + tmp7 = GETJSAMPLE(elemptr[7]) - GETJSAMPLE(elemptr[8]); + + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) + ((tmp10 + tmp11 + tmp12 + tmp13 - 16 * CENTERJSAMPLE) << PASS1_BITS); + dataptr[4] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp13, FIX(1.306562965)) + /* c4[16] = c2[8] */ + MULTIPLY(tmp11 - tmp12, FIX_0_541196100), /* c12[16] = c6[8] */ + CONST_BITS-PASS1_BITS); + + tmp10 = MULTIPLY(tmp17 - tmp15, FIX(0.275899379)) + /* c14[16] = c7[8] */ + MULTIPLY(tmp14 - tmp16, FIX(1.387039845)); /* c2[16] = c1[8] */ + + dataptr[2] = (DCTELEM) + DESCALE(tmp10 + MULTIPLY(tmp15, FIX(1.451774982)) /* c6+c14 */ + + MULTIPLY(tmp16, FIX(2.172734804)), /* c2+c10 */ + CONST_BITS-PASS1_BITS); + dataptr[6] = (DCTELEM) + DESCALE(tmp10 - MULTIPLY(tmp14, FIX(0.211164243)) /* c2-c6 */ + - MULTIPLY(tmp17, FIX(1.061594338)), /* c10+c14 */ + CONST_BITS-PASS1_BITS); + + /* Odd part */ + + tmp11 = MULTIPLY(tmp0 + tmp1, FIX(1.353318001)) + /* c3 */ + MULTIPLY(tmp6 - tmp7, FIX(0.410524528)); /* c13 */ + tmp12 = MULTIPLY(tmp0 + tmp2, FIX(1.247225013)) + /* c5 */ + MULTIPLY(tmp5 + tmp7, FIX(0.666655658)); /* c11 */ + tmp13 = MULTIPLY(tmp0 + tmp3, FIX(1.093201867)) + /* c7 */ + MULTIPLY(tmp4 - tmp7, FIX(0.897167586)); /* c9 */ + tmp14 = MULTIPLY(tmp1 + tmp2, FIX(0.138617169)) + /* c15 */ + MULTIPLY(tmp6 - tmp5, FIX(1.407403738)); /* c1 */ + tmp15 = MULTIPLY(tmp1 + tmp3, - FIX(0.666655658)) + /* -c11 */ + MULTIPLY(tmp4 + tmp6, - FIX(1.247225013)); /* -c5 */ + tmp16 = MULTIPLY(tmp2 + tmp3, - FIX(1.353318001)) + /* -c3 */ + MULTIPLY(tmp5 - tmp4, FIX(0.410524528)); /* c13 */ + tmp10 = tmp11 + tmp12 + tmp13 - + MULTIPLY(tmp0, FIX(2.286341144)) + /* c7+c5+c3-c1 */ + MULTIPLY(tmp7, FIX(0.779653625)); /* c15+c13-c11+c9 */ + tmp11 += tmp14 + tmp15 + MULTIPLY(tmp1, FIX(0.071888074)) /* c9-c3-c15+c11 */ + - MULTIPLY(tmp6, FIX(1.663905119)); /* c7+c13+c1-c5 */ + tmp12 += tmp14 + tmp16 - MULTIPLY(tmp2, FIX(1.125726048)) /* c7+c5+c15-c3 */ + + MULTIPLY(tmp5, FIX(1.227391138)); /* c9-c11+c1-c13 */ + tmp13 += tmp15 + tmp16 + MULTIPLY(tmp3, FIX(1.065388962)) /* c15+c3+c11-c7 */ + + MULTIPLY(tmp4, FIX(2.167985692)); /* c1+c13+c5-c9 */ + + dataptr[1] = (DCTELEM) DESCALE(tmp10, CONST_BITS-PASS1_BITS); + dataptr[3] = (DCTELEM) DESCALE(tmp11, CONST_BITS-PASS1_BITS); + dataptr[5] = (DCTELEM) DESCALE(tmp12, CONST_BITS-PASS1_BITS); + dataptr[7] = (DCTELEM) DESCALE(tmp13, CONST_BITS-PASS1_BITS); + + ctr++; + + if (ctr != DCTSIZE) { + if (ctr == DCTSIZE * 2) + break; /* Done. */ + dataptr += DCTSIZE; /* advance pointer to next row */ + } else + dataptr = workspace; /* switch pointer to extended workspace */ + } + + /* Pass 2: process columns. + * We remove the PASS1_BITS scaling, but leave the results scaled up + * by an overall factor of 8. + * We must also scale the output by (8/16)**2 = 1/2**2. + * cK represents sqrt(2) * cos(K*pi/32). + */ + + dataptr = data; + wsptr = workspace; + for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { + /* Even part */ + + tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*7]; + tmp1 = dataptr[DCTSIZE*1] + wsptr[DCTSIZE*6]; + tmp2 = dataptr[DCTSIZE*2] + wsptr[DCTSIZE*5]; + tmp3 = dataptr[DCTSIZE*3] + wsptr[DCTSIZE*4]; + tmp4 = dataptr[DCTSIZE*4] + wsptr[DCTSIZE*3]; + tmp5 = dataptr[DCTSIZE*5] + wsptr[DCTSIZE*2]; + tmp6 = dataptr[DCTSIZE*6] + wsptr[DCTSIZE*1]; + tmp7 = dataptr[DCTSIZE*7] + wsptr[DCTSIZE*0]; + + tmp10 = tmp0 + tmp7; + tmp14 = tmp0 - tmp7; + tmp11 = tmp1 + tmp6; + tmp15 = tmp1 - tmp6; + tmp12 = tmp2 + tmp5; + tmp16 = tmp2 - tmp5; + tmp13 = tmp3 + tmp4; + tmp17 = tmp3 - tmp4; + + tmp0 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*7]; + tmp1 = dataptr[DCTSIZE*1] - wsptr[DCTSIZE*6]; + tmp2 = dataptr[DCTSIZE*2] - wsptr[DCTSIZE*5]; + tmp3 = dataptr[DCTSIZE*3] - wsptr[DCTSIZE*4]; + tmp4 = dataptr[DCTSIZE*4] - wsptr[DCTSIZE*3]; + tmp5 = dataptr[DCTSIZE*5] - wsptr[DCTSIZE*2]; + tmp6 = dataptr[DCTSIZE*6] - wsptr[DCTSIZE*1]; + tmp7 = dataptr[DCTSIZE*7] - wsptr[DCTSIZE*0]; + + dataptr[DCTSIZE*0] = (DCTELEM) + DESCALE(tmp10 + tmp11 + tmp12 + tmp13, PASS1_BITS+2); + dataptr[DCTSIZE*4] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp13, FIX(1.306562965)) + /* c4[16] = c2[8] */ + MULTIPLY(tmp11 - tmp12, FIX_0_541196100), /* c12[16] = c6[8] */ + CONST_BITS+PASS1_BITS+2); + + tmp10 = MULTIPLY(tmp17 - tmp15, FIX(0.275899379)) + /* c14[16] = c7[8] */ + MULTIPLY(tmp14 - tmp16, FIX(1.387039845)); /* c2[16] = c1[8] */ + + dataptr[DCTSIZE*2] = (DCTELEM) + DESCALE(tmp10 + MULTIPLY(tmp15, FIX(1.451774982)) /* c6+c14 */ + + MULTIPLY(tmp16, FIX(2.172734804)), /* c2+10 */ + CONST_BITS+PASS1_BITS+2); + dataptr[DCTSIZE*6] = (DCTELEM) + DESCALE(tmp10 - MULTIPLY(tmp14, FIX(0.211164243)) /* c2-c6 */ + - MULTIPLY(tmp17, FIX(1.061594338)), /* c10+c14 */ + CONST_BITS+PASS1_BITS+2); + + /* Odd part */ + + tmp11 = MULTIPLY(tmp0 + tmp1, FIX(1.353318001)) + /* c3 */ + MULTIPLY(tmp6 - tmp7, FIX(0.410524528)); /* c13 */ + tmp12 = MULTIPLY(tmp0 + tmp2, FIX(1.247225013)) + /* c5 */ + MULTIPLY(tmp5 + tmp7, FIX(0.666655658)); /* c11 */ + tmp13 = MULTIPLY(tmp0 + tmp3, FIX(1.093201867)) + /* c7 */ + MULTIPLY(tmp4 - tmp7, FIX(0.897167586)); /* c9 */ + tmp14 = MULTIPLY(tmp1 + tmp2, FIX(0.138617169)) + /* c15 */ + MULTIPLY(tmp6 - tmp5, FIX(1.407403738)); /* c1 */ + tmp15 = MULTIPLY(tmp1 + tmp3, - FIX(0.666655658)) + /* -c11 */ + MULTIPLY(tmp4 + tmp6, - FIX(1.247225013)); /* -c5 */ + tmp16 = MULTIPLY(tmp2 + tmp3, - FIX(1.353318001)) + /* -c3 */ + MULTIPLY(tmp5 - tmp4, FIX(0.410524528)); /* c13 */ + tmp10 = tmp11 + tmp12 + tmp13 - + MULTIPLY(tmp0, FIX(2.286341144)) + /* c7+c5+c3-c1 */ + MULTIPLY(tmp7, FIX(0.779653625)); /* c15+c13-c11+c9 */ + tmp11 += tmp14 + tmp15 + MULTIPLY(tmp1, FIX(0.071888074)) /* c9-c3-c15+c11 */ + - MULTIPLY(tmp6, FIX(1.663905119)); /* c7+c13+c1-c5 */ + tmp12 += tmp14 + tmp16 - MULTIPLY(tmp2, FIX(1.125726048)) /* c7+c5+c15-c3 */ + + MULTIPLY(tmp5, FIX(1.227391138)); /* c9-c11+c1-c13 */ + tmp13 += tmp15 + tmp16 + MULTIPLY(tmp3, FIX(1.065388962)) /* c15+c3+c11-c7 */ + + MULTIPLY(tmp4, FIX(2.167985692)); /* c1+c13+c5-c9 */ + + dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp10, CONST_BITS+PASS1_BITS+2); + dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp11, CONST_BITS+PASS1_BITS+2); + dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp12, CONST_BITS+PASS1_BITS+2); + dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp13, CONST_BITS+PASS1_BITS+2); + + dataptr++; /* advance pointer to next column */ + wsptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 16x8 sample block. + * + * 16-point FDCT in pass 1 (rows), 8-point in pass 2 (columns). + */ + +GLOBAL(void) +jpeg_fdct_16x8 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7; + INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15, tmp16, tmp17; + INT32 z1; + DCTELEM *dataptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT; + * furthermore, we scale the results by 2**PASS1_BITS. + * 16-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/32). + */ + + dataptr = data; + ctr = 0; + for (ctr = 0; ctr < DCTSIZE; ctr++) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[15]); + tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[14]); + tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[13]); + tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[12]); + tmp4 = GETJSAMPLE(elemptr[4]) + GETJSAMPLE(elemptr[11]); + tmp5 = GETJSAMPLE(elemptr[5]) + GETJSAMPLE(elemptr[10]); + tmp6 = GETJSAMPLE(elemptr[6]) + GETJSAMPLE(elemptr[9]); + tmp7 = GETJSAMPLE(elemptr[7]) + GETJSAMPLE(elemptr[8]); + + tmp10 = tmp0 + tmp7; + tmp14 = tmp0 - tmp7; + tmp11 = tmp1 + tmp6; + tmp15 = tmp1 - tmp6; + tmp12 = tmp2 + tmp5; + tmp16 = tmp2 - tmp5; + tmp13 = tmp3 + tmp4; + tmp17 = tmp3 - tmp4; + + tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[15]); + tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[14]); + tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[13]); + tmp3 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[12]); + tmp4 = GETJSAMPLE(elemptr[4]) - GETJSAMPLE(elemptr[11]); + tmp5 = GETJSAMPLE(elemptr[5]) - GETJSAMPLE(elemptr[10]); + tmp6 = GETJSAMPLE(elemptr[6]) - GETJSAMPLE(elemptr[9]); + tmp7 = GETJSAMPLE(elemptr[7]) - GETJSAMPLE(elemptr[8]); + + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) + ((tmp10 + tmp11 + tmp12 + tmp13 - 16 * CENTERJSAMPLE) << PASS1_BITS); + dataptr[4] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp13, FIX(1.306562965)) + /* c4[16] = c2[8] */ + MULTIPLY(tmp11 - tmp12, FIX_0_541196100), /* c12[16] = c6[8] */ + CONST_BITS-PASS1_BITS); + + tmp10 = MULTIPLY(tmp17 - tmp15, FIX(0.275899379)) + /* c14[16] = c7[8] */ + MULTIPLY(tmp14 - tmp16, FIX(1.387039845)); /* c2[16] = c1[8] */ + + dataptr[2] = (DCTELEM) + DESCALE(tmp10 + MULTIPLY(tmp15, FIX(1.451774982)) /* c6+c14 */ + + MULTIPLY(tmp16, FIX(2.172734804)), /* c2+c10 */ + CONST_BITS-PASS1_BITS); + dataptr[6] = (DCTELEM) + DESCALE(tmp10 - MULTIPLY(tmp14, FIX(0.211164243)) /* c2-c6 */ + - MULTIPLY(tmp17, FIX(1.061594338)), /* c10+c14 */ + CONST_BITS-PASS1_BITS); + + /* Odd part */ + + tmp11 = MULTIPLY(tmp0 + tmp1, FIX(1.353318001)) + /* c3 */ + MULTIPLY(tmp6 - tmp7, FIX(0.410524528)); /* c13 */ + tmp12 = MULTIPLY(tmp0 + tmp2, FIX(1.247225013)) + /* c5 */ + MULTIPLY(tmp5 + tmp7, FIX(0.666655658)); /* c11 */ + tmp13 = MULTIPLY(tmp0 + tmp3, FIX(1.093201867)) + /* c7 */ + MULTIPLY(tmp4 - tmp7, FIX(0.897167586)); /* c9 */ + tmp14 = MULTIPLY(tmp1 + tmp2, FIX(0.138617169)) + /* c15 */ + MULTIPLY(tmp6 - tmp5, FIX(1.407403738)); /* c1 */ + tmp15 = MULTIPLY(tmp1 + tmp3, - FIX(0.666655658)) + /* -c11 */ + MULTIPLY(tmp4 + tmp6, - FIX(1.247225013)); /* -c5 */ + tmp16 = MULTIPLY(tmp2 + tmp3, - FIX(1.353318001)) + /* -c3 */ + MULTIPLY(tmp5 - tmp4, FIX(0.410524528)); /* c13 */ + tmp10 = tmp11 + tmp12 + tmp13 - + MULTIPLY(tmp0, FIX(2.286341144)) + /* c7+c5+c3-c1 */ + MULTIPLY(tmp7, FIX(0.779653625)); /* c15+c13-c11+c9 */ + tmp11 += tmp14 + tmp15 + MULTIPLY(tmp1, FIX(0.071888074)) /* c9-c3-c15+c11 */ + - MULTIPLY(tmp6, FIX(1.663905119)); /* c7+c13+c1-c5 */ + tmp12 += tmp14 + tmp16 - MULTIPLY(tmp2, FIX(1.125726048)) /* c7+c5+c15-c3 */ + + MULTIPLY(tmp5, FIX(1.227391138)); /* c9-c11+c1-c13 */ + tmp13 += tmp15 + tmp16 + MULTIPLY(tmp3, FIX(1.065388962)) /* c15+c3+c11-c7 */ + + MULTIPLY(tmp4, FIX(2.167985692)); /* c1+c13+c5-c9 */ + + dataptr[1] = (DCTELEM) DESCALE(tmp10, CONST_BITS-PASS1_BITS); + dataptr[3] = (DCTELEM) DESCALE(tmp11, CONST_BITS-PASS1_BITS); + dataptr[5] = (DCTELEM) DESCALE(tmp12, CONST_BITS-PASS1_BITS); + dataptr[7] = (DCTELEM) DESCALE(tmp13, CONST_BITS-PASS1_BITS); + + dataptr += DCTSIZE; /* advance pointer to next row */ + } + + /* Pass 2: process columns. + * We remove the PASS1_BITS scaling, but leave the results scaled up + * by an overall factor of 8. + * We must also scale the output by 8/16 = 1/2. + * 8-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/16). + */ + + dataptr = data; + for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { + /* Even part per LL&M figure 1 --- note that published figure is faulty; + * rotator "c1" should be "c6". + */ + + tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*7]; + tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*6]; + tmp2 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*5]; + tmp3 = dataptr[DCTSIZE*3] + dataptr[DCTSIZE*4]; + + tmp10 = tmp0 + tmp3; + tmp12 = tmp0 - tmp3; + tmp11 = tmp1 + tmp2; + tmp13 = tmp1 - tmp2; + + tmp0 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*7]; + tmp1 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*6]; + tmp2 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*5]; + tmp3 = dataptr[DCTSIZE*3] - dataptr[DCTSIZE*4]; + + dataptr[DCTSIZE*0] = (DCTELEM) DESCALE(tmp10 + tmp11, PASS1_BITS+1); + dataptr[DCTSIZE*4] = (DCTELEM) DESCALE(tmp10 - tmp11, PASS1_BITS+1); + + z1 = MULTIPLY(tmp12 + tmp13, FIX_0_541196100); /* c6 */ + dataptr[DCTSIZE*2] = (DCTELEM) + DESCALE(z1 + MULTIPLY(tmp12, FIX_0_765366865), /* c2-c6 */ + CONST_BITS+PASS1_BITS+1); + dataptr[DCTSIZE*6] = (DCTELEM) + DESCALE(z1 - MULTIPLY(tmp13, FIX_1_847759065), /* c2+c6 */ + CONST_BITS+PASS1_BITS+1); + + /* Odd part per figure 8 --- note paper omits factor of sqrt(2). + * i0..i3 in the paper are tmp0..tmp3 here. + */ + + tmp12 = tmp0 + tmp2; + tmp13 = tmp1 + tmp3; + + z1 = MULTIPLY(tmp12 + tmp13, FIX_1_175875602); /* c3 */ + tmp12 = MULTIPLY(tmp12, - FIX_0_390180644); /* -c3+c5 */ + tmp13 = MULTIPLY(tmp13, - FIX_1_961570560); /* -c3-c5 */ + tmp12 += z1; + tmp13 += z1; + + z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ + tmp0 = MULTIPLY(tmp0, FIX_1_501321110); /* c1+c3-c5-c7 */ + tmp3 = MULTIPLY(tmp3, FIX_0_298631336); /* -c1+c3+c5-c7 */ + tmp0 += z1 + tmp12; + tmp3 += z1 + tmp13; + + z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ + tmp1 = MULTIPLY(tmp1, FIX_3_072711026); /* c1+c3+c5-c7 */ + tmp2 = MULTIPLY(tmp2, FIX_2_053119869); /* c1+c3-c5+c7 */ + tmp1 += z1 + tmp13; + tmp2 += z1 + tmp12; + + dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp0, CONST_BITS+PASS1_BITS+1); + dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp1, CONST_BITS+PASS1_BITS+1); + dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp2, CONST_BITS+PASS1_BITS+1); + dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp3, CONST_BITS+PASS1_BITS+1); + + dataptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 14x7 sample block. + * + * 14-point FDCT in pass 1 (rows), 7-point in pass 2 (columns). + */ + +GLOBAL(void) +jpeg_fdct_14x7 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6; + INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15, tmp16; + INT32 z1, z2, z3; + DCTELEM *dataptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Zero bottom row of output coefficient block. */ + MEMZERO(&data[DCTSIZE*7], SIZEOF(DCTELEM) * DCTSIZE); + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT; + * furthermore, we scale the results by 2**PASS1_BITS. + * 14-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/28). + */ + + dataptr = data; + for (ctr = 0; ctr < 7; ctr++) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[13]); + tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[12]); + tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[11]); + tmp13 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[10]); + tmp4 = GETJSAMPLE(elemptr[4]) + GETJSAMPLE(elemptr[9]); + tmp5 = GETJSAMPLE(elemptr[5]) + GETJSAMPLE(elemptr[8]); + tmp6 = GETJSAMPLE(elemptr[6]) + GETJSAMPLE(elemptr[7]); + + tmp10 = tmp0 + tmp6; + tmp14 = tmp0 - tmp6; + tmp11 = tmp1 + tmp5; + tmp15 = tmp1 - tmp5; + tmp12 = tmp2 + tmp4; + tmp16 = tmp2 - tmp4; + + tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[13]); + tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[12]); + tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[11]); + tmp3 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[10]); + tmp4 = GETJSAMPLE(elemptr[4]) - GETJSAMPLE(elemptr[9]); + tmp5 = GETJSAMPLE(elemptr[5]) - GETJSAMPLE(elemptr[8]); + tmp6 = GETJSAMPLE(elemptr[6]) - GETJSAMPLE(elemptr[7]); + + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) + ((tmp10 + tmp11 + tmp12 + tmp13 - 14 * CENTERJSAMPLE) << PASS1_BITS); + tmp13 += tmp13; + dataptr[4] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp13, FIX(1.274162392)) + /* c4 */ + MULTIPLY(tmp11 - tmp13, FIX(0.314692123)) - /* c12 */ + MULTIPLY(tmp12 - tmp13, FIX(0.881747734)), /* c8 */ + CONST_BITS-PASS1_BITS); + + tmp10 = MULTIPLY(tmp14 + tmp15, FIX(1.105676686)); /* c6 */ + + dataptr[2] = (DCTELEM) + DESCALE(tmp10 + MULTIPLY(tmp14, FIX(0.273079590)) /* c2-c6 */ + + MULTIPLY(tmp16, FIX(0.613604268)), /* c10 */ + CONST_BITS-PASS1_BITS); + dataptr[6] = (DCTELEM) + DESCALE(tmp10 - MULTIPLY(tmp15, FIX(1.719280954)) /* c6+c10 */ + - MULTIPLY(tmp16, FIX(1.378756276)), /* c2 */ + CONST_BITS-PASS1_BITS); + + /* Odd part */ + + tmp10 = tmp1 + tmp2; + tmp11 = tmp5 - tmp4; + dataptr[7] = (DCTELEM) ((tmp0 - tmp10 + tmp3 - tmp11 - tmp6) << PASS1_BITS); + tmp3 <<= CONST_BITS; + tmp10 = MULTIPLY(tmp10, - FIX(0.158341681)); /* -c13 */ + tmp11 = MULTIPLY(tmp11, FIX(1.405321284)); /* c1 */ + tmp10 += tmp11 - tmp3; + tmp11 = MULTIPLY(tmp0 + tmp2, FIX(1.197448846)) + /* c5 */ + MULTIPLY(tmp4 + tmp6, FIX(0.752406978)); /* c9 */ + dataptr[5] = (DCTELEM) + DESCALE(tmp10 + tmp11 - MULTIPLY(tmp2, FIX(2.373959773)) /* c3+c5-c13 */ + + MULTIPLY(tmp4, FIX(1.119999435)), /* c1+c11-c9 */ + CONST_BITS-PASS1_BITS); + tmp12 = MULTIPLY(tmp0 + tmp1, FIX(1.334852607)) + /* c3 */ + MULTIPLY(tmp5 - tmp6, FIX(0.467085129)); /* c11 */ + dataptr[3] = (DCTELEM) + DESCALE(tmp10 + tmp12 - MULTIPLY(tmp1, FIX(0.424103948)) /* c3-c9-c13 */ + - MULTIPLY(tmp5, FIX(3.069855259)), /* c1+c5+c11 */ + CONST_BITS-PASS1_BITS); + dataptr[1] = (DCTELEM) + DESCALE(tmp11 + tmp12 + tmp3 + tmp6 - + MULTIPLY(tmp0 + tmp6, FIX(1.126980169)), /* c3+c5-c1 */ + CONST_BITS-PASS1_BITS); + + dataptr += DCTSIZE; /* advance pointer to next row */ + } + + /* Pass 2: process columns. + * We remove the PASS1_BITS scaling, but leave the results scaled up + * by an overall factor of 8. + * We must also scale the output by (8/14)*(8/7) = 32/49, which we + * partially fold into the constant multipliers and final shifting: + * 7-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/14) * 64/49. + */ + + dataptr = data; + for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { + /* Even part */ + + tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*6]; + tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*5]; + tmp2 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*4]; + tmp3 = dataptr[DCTSIZE*3]; + + tmp10 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*6]; + tmp11 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*5]; + tmp12 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*4]; + + z1 = tmp0 + tmp2; + dataptr[DCTSIZE*0] = (DCTELEM) + DESCALE(MULTIPLY(z1 + tmp1 + tmp3, FIX(1.306122449)), /* 64/49 */ + CONST_BITS+PASS1_BITS+1); + tmp3 += tmp3; + z1 -= tmp3; + z1 -= tmp3; + z1 = MULTIPLY(z1, FIX(0.461784020)); /* (c2+c6-c4)/2 */ + z2 = MULTIPLY(tmp0 - tmp2, FIX(1.202428084)); /* (c2+c4-c6)/2 */ + z3 = MULTIPLY(tmp1 - tmp2, FIX(0.411026446)); /* c6 */ + dataptr[DCTSIZE*2] = (DCTELEM) DESCALE(z1 + z2 + z3, CONST_BITS+PASS1_BITS+1); + z1 -= z2; + z2 = MULTIPLY(tmp0 - tmp1, FIX(1.151670509)); /* c4 */ + dataptr[DCTSIZE*4] = (DCTELEM) + DESCALE(z2 + z3 - MULTIPLY(tmp1 - tmp3, FIX(0.923568041)), /* c2+c6-c4 */ + CONST_BITS+PASS1_BITS+1); + dataptr[DCTSIZE*6] = (DCTELEM) DESCALE(z1 + z2, CONST_BITS+PASS1_BITS+1); + + /* Odd part */ + + tmp1 = MULTIPLY(tmp10 + tmp11, FIX(1.221765677)); /* (c3+c1-c5)/2 */ + tmp2 = MULTIPLY(tmp10 - tmp11, FIX(0.222383464)); /* (c3+c5-c1)/2 */ + tmp0 = tmp1 - tmp2; + tmp1 += tmp2; + tmp2 = MULTIPLY(tmp11 + tmp12, - FIX(1.800824523)); /* -c1 */ + tmp1 += tmp2; + tmp3 = MULTIPLY(tmp10 + tmp12, FIX(0.801442310)); /* c5 */ + tmp0 += tmp3; + tmp2 += tmp3 + MULTIPLY(tmp12, FIX(2.443531355)); /* c3+c1-c5 */ + + dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp0, CONST_BITS+PASS1_BITS+1); + dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp1, CONST_BITS+PASS1_BITS+1); + dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp2, CONST_BITS+PASS1_BITS+1); + + dataptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 12x6 sample block. + * + * 12-point FDCT in pass 1 (rows), 6-point in pass 2 (columns). + */ + +GLOBAL(void) +jpeg_fdct_12x6 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5; + INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15; + DCTELEM *dataptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Zero 2 bottom rows of output coefficient block. */ + MEMZERO(&data[DCTSIZE*6], SIZEOF(DCTELEM) * DCTSIZE * 2); + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT; + * furthermore, we scale the results by 2**PASS1_BITS. + * 12-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/24). + */ + + dataptr = data; + for (ctr = 0; ctr < 6; ctr++) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[11]); + tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[10]); + tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[9]); + tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[8]); + tmp4 = GETJSAMPLE(elemptr[4]) + GETJSAMPLE(elemptr[7]); + tmp5 = GETJSAMPLE(elemptr[5]) + GETJSAMPLE(elemptr[6]); + + tmp10 = tmp0 + tmp5; + tmp13 = tmp0 - tmp5; + tmp11 = tmp1 + tmp4; + tmp14 = tmp1 - tmp4; + tmp12 = tmp2 + tmp3; + tmp15 = tmp2 - tmp3; + + tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[11]); + tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[10]); + tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[9]); + tmp3 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[8]); + tmp4 = GETJSAMPLE(elemptr[4]) - GETJSAMPLE(elemptr[7]); + tmp5 = GETJSAMPLE(elemptr[5]) - GETJSAMPLE(elemptr[6]); + + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) + ((tmp10 + tmp11 + tmp12 - 12 * CENTERJSAMPLE) << PASS1_BITS); + dataptr[6] = (DCTELEM) ((tmp13 - tmp14 - tmp15) << PASS1_BITS); + dataptr[4] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp12, FIX(1.224744871)), /* c4 */ + CONST_BITS-PASS1_BITS); + dataptr[2] = (DCTELEM) + DESCALE(tmp14 - tmp15 + MULTIPLY(tmp13 + tmp15, FIX(1.366025404)), /* c2 */ + CONST_BITS-PASS1_BITS); + + /* Odd part */ + + tmp10 = MULTIPLY(tmp1 + tmp4, FIX_0_541196100); /* c9 */ + tmp14 = tmp10 + MULTIPLY(tmp1, FIX_0_765366865); /* c3-c9 */ + tmp15 = tmp10 - MULTIPLY(tmp4, FIX_1_847759065); /* c3+c9 */ + tmp12 = MULTIPLY(tmp0 + tmp2, FIX(1.121971054)); /* c5 */ + tmp13 = MULTIPLY(tmp0 + tmp3, FIX(0.860918669)); /* c7 */ + tmp10 = tmp12 + tmp13 + tmp14 - MULTIPLY(tmp0, FIX(0.580774953)) /* c5+c7-c1 */ + + MULTIPLY(tmp5, FIX(0.184591911)); /* c11 */ + tmp11 = MULTIPLY(tmp2 + tmp3, - FIX(0.184591911)); /* -c11 */ + tmp12 += tmp11 - tmp15 - MULTIPLY(tmp2, FIX(2.339493912)) /* c1+c5-c11 */ + + MULTIPLY(tmp5, FIX(0.860918669)); /* c7 */ + tmp13 += tmp11 - tmp14 + MULTIPLY(tmp3, FIX(0.725788011)) /* c1+c11-c7 */ + - MULTIPLY(tmp5, FIX(1.121971054)); /* c5 */ + tmp11 = tmp15 + MULTIPLY(tmp0 - tmp3, FIX(1.306562965)) /* c3 */ + - MULTIPLY(tmp2 + tmp5, FIX_0_541196100); /* c9 */ + + dataptr[1] = (DCTELEM) DESCALE(tmp10, CONST_BITS-PASS1_BITS); + dataptr[3] = (DCTELEM) DESCALE(tmp11, CONST_BITS-PASS1_BITS); + dataptr[5] = (DCTELEM) DESCALE(tmp12, CONST_BITS-PASS1_BITS); + dataptr[7] = (DCTELEM) DESCALE(tmp13, CONST_BITS-PASS1_BITS); + + dataptr += DCTSIZE; /* advance pointer to next row */ + } + + /* Pass 2: process columns. + * We remove the PASS1_BITS scaling, but leave the results scaled up + * by an overall factor of 8. + * We must also scale the output by (8/12)*(8/6) = 8/9, which we + * partially fold into the constant multipliers and final shifting: + * 6-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/12) * 16/9. + */ + + dataptr = data; + for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { + /* Even part */ + + tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*5]; + tmp11 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*4]; + tmp2 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*3]; + + tmp10 = tmp0 + tmp2; + tmp12 = tmp0 - tmp2; + + tmp0 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*5]; + tmp1 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*4]; + tmp2 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*3]; + + dataptr[DCTSIZE*0] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 + tmp11, FIX(1.777777778)), /* 16/9 */ + CONST_BITS+PASS1_BITS+1); + dataptr[DCTSIZE*2] = (DCTELEM) + DESCALE(MULTIPLY(tmp12, FIX(2.177324216)), /* c2 */ + CONST_BITS+PASS1_BITS+1); + dataptr[DCTSIZE*4] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp11 - tmp11, FIX(1.257078722)), /* c4 */ + CONST_BITS+PASS1_BITS+1); + + /* Odd part */ + + tmp10 = MULTIPLY(tmp0 + tmp2, FIX(0.650711829)); /* c5 */ + + dataptr[DCTSIZE*1] = (DCTELEM) + DESCALE(tmp10 + MULTIPLY(tmp0 + tmp1, FIX(1.777777778)), /* 16/9 */ + CONST_BITS+PASS1_BITS+1); + dataptr[DCTSIZE*3] = (DCTELEM) + DESCALE(MULTIPLY(tmp0 - tmp1 - tmp2, FIX(1.777777778)), /* 16/9 */ + CONST_BITS+PASS1_BITS+1); + dataptr[DCTSIZE*5] = (DCTELEM) + DESCALE(tmp10 + MULTIPLY(tmp2 - tmp1, FIX(1.777777778)), /* 16/9 */ + CONST_BITS+PASS1_BITS+1); + + dataptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 10x5 sample block. + * + * 10-point FDCT in pass 1 (rows), 5-point in pass 2 (columns). + */ + +GLOBAL(void) +jpeg_fdct_10x5 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3, tmp4; + INT32 tmp10, tmp11, tmp12, tmp13, tmp14; + DCTELEM *dataptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Zero 3 bottom rows of output coefficient block. */ + MEMZERO(&data[DCTSIZE*5], SIZEOF(DCTELEM) * DCTSIZE * 3); + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT; + * furthermore, we scale the results by 2**PASS1_BITS. + * 10-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/20). + */ + + dataptr = data; + for (ctr = 0; ctr < 5; ctr++) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[9]); + tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[8]); + tmp12 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[7]); + tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[6]); + tmp4 = GETJSAMPLE(elemptr[4]) + GETJSAMPLE(elemptr[5]); + + tmp10 = tmp0 + tmp4; + tmp13 = tmp0 - tmp4; + tmp11 = tmp1 + tmp3; + tmp14 = tmp1 - tmp3; + + tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[9]); + tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[8]); + tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[7]); + tmp3 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[6]); + tmp4 = GETJSAMPLE(elemptr[4]) - GETJSAMPLE(elemptr[5]); + + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) + ((tmp10 + tmp11 + tmp12 - 10 * CENTERJSAMPLE) << PASS1_BITS); + tmp12 += tmp12; + dataptr[4] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp12, FIX(1.144122806)) - /* c4 */ + MULTIPLY(tmp11 - tmp12, FIX(0.437016024)), /* c8 */ + CONST_BITS-PASS1_BITS); + tmp10 = MULTIPLY(tmp13 + tmp14, FIX(0.831253876)); /* c6 */ + dataptr[2] = (DCTELEM) + DESCALE(tmp10 + MULTIPLY(tmp13, FIX(0.513743148)), /* c2-c6 */ + CONST_BITS-PASS1_BITS); + dataptr[6] = (DCTELEM) + DESCALE(tmp10 - MULTIPLY(tmp14, FIX(2.176250899)), /* c2+c6 */ + CONST_BITS-PASS1_BITS); + + /* Odd part */ + + tmp10 = tmp0 + tmp4; + tmp11 = tmp1 - tmp3; + dataptr[5] = (DCTELEM) ((tmp10 - tmp11 - tmp2) << PASS1_BITS); + tmp2 <<= CONST_BITS; + dataptr[1] = (DCTELEM) + DESCALE(MULTIPLY(tmp0, FIX(1.396802247)) + /* c1 */ + MULTIPLY(tmp1, FIX(1.260073511)) + tmp2 + /* c3 */ + MULTIPLY(tmp3, FIX(0.642039522)) + /* c7 */ + MULTIPLY(tmp4, FIX(0.221231742)), /* c9 */ + CONST_BITS-PASS1_BITS); + tmp12 = MULTIPLY(tmp0 - tmp4, FIX(0.951056516)) - /* (c3+c7)/2 */ + MULTIPLY(tmp1 + tmp3, FIX(0.587785252)); /* (c1-c9)/2 */ + tmp13 = MULTIPLY(tmp10 + tmp11, FIX(0.309016994)) + /* (c3-c7)/2 */ + (tmp11 << (CONST_BITS - 1)) - tmp2; + dataptr[3] = (DCTELEM) DESCALE(tmp12 + tmp13, CONST_BITS-PASS1_BITS); + dataptr[7] = (DCTELEM) DESCALE(tmp12 - tmp13, CONST_BITS-PASS1_BITS); + + dataptr += DCTSIZE; /* advance pointer to next row */ + } + + /* Pass 2: process columns. + * We remove the PASS1_BITS scaling, but leave the results scaled up + * by an overall factor of 8. + * We must also scale the output by (8/10)*(8/5) = 32/25, which we + * fold into the constant multipliers: + * 5-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/10) * 32/25. + */ + + dataptr = data; + for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { + /* Even part */ + + tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*4]; + tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*3]; + tmp2 = dataptr[DCTSIZE*2]; + + tmp10 = tmp0 + tmp1; + tmp11 = tmp0 - tmp1; + + tmp0 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*4]; + tmp1 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*3]; + + dataptr[DCTSIZE*0] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 + tmp2, FIX(1.28)), /* 32/25 */ + CONST_BITS+PASS1_BITS); + tmp11 = MULTIPLY(tmp11, FIX(1.011928851)); /* (c2+c4)/2 */ + tmp10 -= tmp2 << 2; + tmp10 = MULTIPLY(tmp10, FIX(0.452548340)); /* (c2-c4)/2 */ + dataptr[DCTSIZE*2] = (DCTELEM) DESCALE(tmp11 + tmp10, CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*4] = (DCTELEM) DESCALE(tmp11 - tmp10, CONST_BITS+PASS1_BITS); + + /* Odd part */ + + tmp10 = MULTIPLY(tmp0 + tmp1, FIX(1.064004961)); /* c3 */ + + dataptr[DCTSIZE*1] = (DCTELEM) + DESCALE(tmp10 + MULTIPLY(tmp0, FIX(0.657591230)), /* c1-c3 */ + CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*3] = (DCTELEM) + DESCALE(tmp10 - MULTIPLY(tmp1, FIX(2.785601151)), /* c1+c3 */ + CONST_BITS+PASS1_BITS); + + dataptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on an 8x4 sample block. + * + * 8-point FDCT in pass 1 (rows), 4-point in pass 2 (columns). + */ + +GLOBAL(void) +jpeg_fdct_8x4 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3; + INT32 tmp10, tmp11, tmp12, tmp13; + INT32 z1; + DCTELEM *dataptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Zero 4 bottom rows of output coefficient block. */ + MEMZERO(&data[DCTSIZE*4], SIZEOF(DCTELEM) * DCTSIZE * 4); + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT; + * furthermore, we scale the results by 2**PASS1_BITS. + * We must also scale the output by 8/4 = 2, which we add here. + * 8-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/16). + */ + + dataptr = data; + for (ctr = 0; ctr < 4; ctr++) { + elemptr = sample_data[ctr] + start_col; + + /* Even part per LL&M figure 1 --- note that published figure is faulty; + * rotator "c1" should be "c6". + */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[7]); + tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[6]); + tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[5]); + tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[4]); + + tmp10 = tmp0 + tmp3; + tmp12 = tmp0 - tmp3; + tmp11 = tmp1 + tmp2; + tmp13 = tmp1 - tmp2; + + tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[7]); + tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[6]); + tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[5]); + tmp3 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[4]); + + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) + ((tmp10 + tmp11 - 8 * CENTERJSAMPLE) << (PASS1_BITS+1)); + dataptr[4] = (DCTELEM) ((tmp10 - tmp11) << (PASS1_BITS+1)); + + z1 = MULTIPLY(tmp12 + tmp13, FIX_0_541196100); /* c6 */ + /* Add fudge factor here for final descale. */ + z1 += ONE << (CONST_BITS-PASS1_BITS-2); + + dataptr[2] = (DCTELEM) + RIGHT_SHIFT(z1 + MULTIPLY(tmp12, FIX_0_765366865), /* c2-c6 */ + CONST_BITS-PASS1_BITS-1); + dataptr[6] = (DCTELEM) + RIGHT_SHIFT(z1 - MULTIPLY(tmp13, FIX_1_847759065), /* c2+c6 */ + CONST_BITS-PASS1_BITS-1); + + /* Odd part per figure 8 --- note paper omits factor of sqrt(2). + * i0..i3 in the paper are tmp0..tmp3 here. + */ + + tmp12 = tmp0 + tmp2; + tmp13 = tmp1 + tmp3; + + z1 = MULTIPLY(tmp12 + tmp13, FIX_1_175875602); /* c3 */ + /* Add fudge factor here for final descale. */ + z1 += ONE << (CONST_BITS-PASS1_BITS-2); + + tmp12 = MULTIPLY(tmp12, - FIX_0_390180644); /* -c3+c5 */ + tmp13 = MULTIPLY(tmp13, - FIX_1_961570560); /* -c3-c5 */ + tmp12 += z1; + tmp13 += z1; + + z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ + tmp0 = MULTIPLY(tmp0, FIX_1_501321110); /* c1+c3-c5-c7 */ + tmp3 = MULTIPLY(tmp3, FIX_0_298631336); /* -c1+c3+c5-c7 */ + tmp0 += z1 + tmp12; + tmp3 += z1 + tmp13; + + z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ + tmp1 = MULTIPLY(tmp1, FIX_3_072711026); /* c1+c3+c5-c7 */ + tmp2 = MULTIPLY(tmp2, FIX_2_053119869); /* c1+c3-c5+c7 */ + tmp1 += z1 + tmp13; + tmp2 += z1 + tmp12; + + dataptr[1] = (DCTELEM) RIGHT_SHIFT(tmp0, CONST_BITS-PASS1_BITS-1); + dataptr[3] = (DCTELEM) RIGHT_SHIFT(tmp1, CONST_BITS-PASS1_BITS-1); + dataptr[5] = (DCTELEM) RIGHT_SHIFT(tmp2, CONST_BITS-PASS1_BITS-1); + dataptr[7] = (DCTELEM) RIGHT_SHIFT(tmp3, CONST_BITS-PASS1_BITS-1); + + dataptr += DCTSIZE; /* advance pointer to next row */ + } + + /* Pass 2: process columns. + * We remove the PASS1_BITS scaling, but leave the results scaled up + * by an overall factor of 8. + * 4-point FDCT kernel, + * cK represents sqrt(2) * cos(K*pi/16) [refers to 8-point FDCT]. + */ + + dataptr = data; + for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { + /* Even part */ + + /* Add fudge factor here for final descale. */ + tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*3] + (ONE << (PASS1_BITS-1)); + tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*2]; + + tmp10 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*3]; + tmp11 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*2]; + + dataptr[DCTSIZE*0] = (DCTELEM) RIGHT_SHIFT(tmp0 + tmp1, PASS1_BITS); + dataptr[DCTSIZE*2] = (DCTELEM) RIGHT_SHIFT(tmp0 - tmp1, PASS1_BITS); + + /* Odd part */ + + tmp0 = MULTIPLY(tmp10 + tmp11, FIX_0_541196100); /* c6 */ + /* Add fudge factor here for final descale. */ + tmp0 += ONE << (CONST_BITS+PASS1_BITS-1); + + dataptr[DCTSIZE*1] = (DCTELEM) + RIGHT_SHIFT(tmp0 + MULTIPLY(tmp10, FIX_0_765366865), /* c2-c6 */ + CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*3] = (DCTELEM) + RIGHT_SHIFT(tmp0 - MULTIPLY(tmp11, FIX_1_847759065), /* c2+c6 */ + CONST_BITS+PASS1_BITS); + + dataptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 6x3 sample block. + * + * 6-point FDCT in pass 1 (rows), 3-point in pass 2 (columns). + */ + +GLOBAL(void) +jpeg_fdct_6x3 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2; + INT32 tmp10, tmp11, tmp12; + DCTELEM *dataptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pre-zero output coefficient block. */ + MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT; + * furthermore, we scale the results by 2**PASS1_BITS. + * We scale the results further by 2 as part of output adaption + * scaling for different DCT size. + * 6-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/12). + */ + + dataptr = data; + for (ctr = 0; ctr < 3; ctr++) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[5]); + tmp11 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[4]); + tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[3]); + + tmp10 = tmp0 + tmp2; + tmp12 = tmp0 - tmp2; + + tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[5]); + tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[4]); + tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[3]); + + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) + ((tmp10 + tmp11 - 6 * CENTERJSAMPLE) << (PASS1_BITS+1)); + dataptr[2] = (DCTELEM) + DESCALE(MULTIPLY(tmp12, FIX(1.224744871)), /* c2 */ + CONST_BITS-PASS1_BITS-1); + dataptr[4] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp11 - tmp11, FIX(0.707106781)), /* c4 */ + CONST_BITS-PASS1_BITS-1); + + /* Odd part */ + + tmp10 = DESCALE(MULTIPLY(tmp0 + tmp2, FIX(0.366025404)), /* c5 */ + CONST_BITS-PASS1_BITS-1); + + dataptr[1] = (DCTELEM) (tmp10 + ((tmp0 + tmp1) << (PASS1_BITS+1))); + dataptr[3] = (DCTELEM) ((tmp0 - tmp1 - tmp2) << (PASS1_BITS+1)); + dataptr[5] = (DCTELEM) (tmp10 + ((tmp2 - tmp1) << (PASS1_BITS+1))); + + dataptr += DCTSIZE; /* advance pointer to next row */ + } + + /* Pass 2: process columns. + * We remove the PASS1_BITS scaling, but leave the results scaled up + * by an overall factor of 8. + * We must also scale the output by (8/6)*(8/3) = 32/9, which we partially + * fold into the constant multipliers (other part was done in pass 1): + * 3-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/6) * 16/9. + */ + + dataptr = data; + for (ctr = 0; ctr < 6; ctr++) { + /* Even part */ + + tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*2]; + tmp1 = dataptr[DCTSIZE*1]; + + tmp2 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*2]; + + dataptr[DCTSIZE*0] = (DCTELEM) + DESCALE(MULTIPLY(tmp0 + tmp1, FIX(1.777777778)), /* 16/9 */ + CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*2] = (DCTELEM) + DESCALE(MULTIPLY(tmp0 - tmp1 - tmp1, FIX(1.257078722)), /* c2 */ + CONST_BITS+PASS1_BITS); + + /* Odd part */ + + dataptr[DCTSIZE*1] = (DCTELEM) + DESCALE(MULTIPLY(tmp2, FIX(2.177324216)), /* c1 */ + CONST_BITS+PASS1_BITS); + + dataptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 4x2 sample block. + * + * 4-point FDCT in pass 1 (rows), 2-point in pass 2 (columns). + */ + +GLOBAL(void) +jpeg_fdct_4x2 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + DCTELEM tmp0, tmp2, tmp10, tmp12, tmp4, tmp5; + INT32 tmp1, tmp3, tmp11, tmp13; + INT32 z1, z2, z3; + JSAMPROW elemptr; + SHIFT_TEMPS + + /* Pre-zero output coefficient block. */ + MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT. + * 4-point FDCT kernel, + * cK represents sqrt(2) * cos(K*pi/16) [refers to 8-point FDCT]. + */ + + /* Row 0 */ + elemptr = sample_data[0] + start_col; + + /* Even part */ + + tmp4 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[3]); + tmp5 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[2]); + + tmp0 = tmp4 + tmp5; + tmp2 = tmp4 - tmp5; + + /* Odd part */ + + z2 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[3]); + z3 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[2]); + + z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ + /* Add fudge factor here for final descale. */ + z1 += ONE << (CONST_BITS-3-1); + tmp1 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ + tmp3 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ + + /* Row 1 */ + elemptr = sample_data[1] + start_col; + + /* Even part */ + + tmp4 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[3]); + tmp5 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[2]); + + tmp10 = tmp4 + tmp5; + tmp12 = tmp4 - tmp5; + + /* Odd part */ + + z2 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[3]); + z3 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[2]); + + z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ + tmp11 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ + tmp13 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ + + /* Pass 2: process columns. + * We leave the results scaled up by an overall factor of 8. + * We must also scale the output by (8/4)*(8/2) = 2**3. + */ + + /* Column 0 */ + /* Apply unsigned->signed conversion. */ + data[DCTSIZE*0] = (tmp0 + tmp10 - 8 * CENTERJSAMPLE) << 3; + data[DCTSIZE*1] = (tmp0 - tmp10) << 3; + + /* Column 1 */ + data[DCTSIZE*0+1] = (DCTELEM) RIGHT_SHIFT(tmp1 + tmp11, CONST_BITS-3); + data[DCTSIZE*1+1] = (DCTELEM) RIGHT_SHIFT(tmp1 - tmp11, CONST_BITS-3); + + /* Column 2 */ + data[DCTSIZE*0+2] = (tmp2 + tmp12) << 3; + data[DCTSIZE*1+2] = (tmp2 - tmp12) << 3; + + /* Column 3 */ + data[DCTSIZE*0+3] = (DCTELEM) RIGHT_SHIFT(tmp3 + tmp13, CONST_BITS-3); + data[DCTSIZE*1+3] = (DCTELEM) RIGHT_SHIFT(tmp3 - tmp13, CONST_BITS-3); +} + + +/* + * Perform the forward DCT on a 2x1 sample block. + * + * 2-point FDCT in pass 1 (rows), 1-point in pass 2 (columns). + */ + +GLOBAL(void) +jpeg_fdct_2x1 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + DCTELEM tmp0, tmp1; + JSAMPROW elemptr; + + /* Pre-zero output coefficient block. */ + MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); + + elemptr = sample_data[0] + start_col; + + tmp0 = GETJSAMPLE(elemptr[0]); + tmp1 = GETJSAMPLE(elemptr[1]); + + /* We leave the results scaled up by an overall factor of 8. + * We must also scale the output by (8/2)*(8/1) = 2**5. + */ + + /* Even part */ + + /* Apply unsigned->signed conversion. */ + data[0] = (tmp0 + tmp1 - 2 * CENTERJSAMPLE) << 5; + + /* Odd part */ + + data[1] = (tmp0 - tmp1) << 5; +} + + +/* + * Perform the forward DCT on an 8x16 sample block. + * + * 8-point FDCT in pass 1 (rows), 16-point in pass 2 (columns). + */ + +GLOBAL(void) +jpeg_fdct_8x16 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7; + INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15, tmp16, tmp17; + INT32 z1; + DCTELEM workspace[DCTSIZE2]; + DCTELEM *dataptr; + DCTELEM *wsptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT; + * furthermore, we scale the results by 2**PASS1_BITS. + * 8-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/16). + */ + + dataptr = data; + ctr = 0; + for (;;) { + elemptr = sample_data[ctr] + start_col; + + /* Even part per LL&M figure 1 --- note that published figure is faulty; + * rotator "c1" should be "c6". + */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[7]); + tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[6]); + tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[5]); + tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[4]); + + tmp10 = tmp0 + tmp3; + tmp12 = tmp0 - tmp3; + tmp11 = tmp1 + tmp2; + tmp13 = tmp1 - tmp2; + + tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[7]); + tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[6]); + tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[5]); + tmp3 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[4]); + + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) ((tmp10 + tmp11 - 8 * CENTERJSAMPLE) << PASS1_BITS); + dataptr[4] = (DCTELEM) ((tmp10 - tmp11) << PASS1_BITS); + + z1 = MULTIPLY(tmp12 + tmp13, FIX_0_541196100); /* c6 */ + dataptr[2] = (DCTELEM) + DESCALE(z1 + MULTIPLY(tmp12, FIX_0_765366865), /* c2-c6 */ + CONST_BITS-PASS1_BITS); + dataptr[6] = (DCTELEM) + DESCALE(z1 - MULTIPLY(tmp13, FIX_1_847759065), /* c2+c6 */ + CONST_BITS-PASS1_BITS); + + /* Odd part per figure 8 --- note paper omits factor of sqrt(2). + * i0..i3 in the paper are tmp0..tmp3 here. + */ + + tmp12 = tmp0 + tmp2; + tmp13 = tmp1 + tmp3; + + z1 = MULTIPLY(tmp12 + tmp13, FIX_1_175875602); /* c3 */ + tmp12 = MULTIPLY(tmp12, - FIX_0_390180644); /* -c3+c5 */ + tmp13 = MULTIPLY(tmp13, - FIX_1_961570560); /* -c3-c5 */ + tmp12 += z1; + tmp13 += z1; + + z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ + tmp0 = MULTIPLY(tmp0, FIX_1_501321110); /* c1+c3-c5-c7 */ + tmp3 = MULTIPLY(tmp3, FIX_0_298631336); /* -c1+c3+c5-c7 */ + tmp0 += z1 + tmp12; + tmp3 += z1 + tmp13; + + z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ + tmp1 = MULTIPLY(tmp1, FIX_3_072711026); /* c1+c3+c5-c7 */ + tmp2 = MULTIPLY(tmp2, FIX_2_053119869); /* c1+c3-c5+c7 */ + tmp1 += z1 + tmp13; + tmp2 += z1 + tmp12; + + dataptr[1] = (DCTELEM) DESCALE(tmp0, CONST_BITS-PASS1_BITS); + dataptr[3] = (DCTELEM) DESCALE(tmp1, CONST_BITS-PASS1_BITS); + dataptr[5] = (DCTELEM) DESCALE(tmp2, CONST_BITS-PASS1_BITS); + dataptr[7] = (DCTELEM) DESCALE(tmp3, CONST_BITS-PASS1_BITS); + + ctr++; + + if (ctr != DCTSIZE) { + if (ctr == DCTSIZE * 2) + break; /* Done. */ + dataptr += DCTSIZE; /* advance pointer to next row */ + } else + dataptr = workspace; /* switch pointer to extended workspace */ + } + + /* Pass 2: process columns. + * We remove the PASS1_BITS scaling, but leave the results scaled up + * by an overall factor of 8. + * We must also scale the output by 8/16 = 1/2. + * 16-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/32). + */ + + dataptr = data; + wsptr = workspace; + for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { + /* Even part */ + + tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*7]; + tmp1 = dataptr[DCTSIZE*1] + wsptr[DCTSIZE*6]; + tmp2 = dataptr[DCTSIZE*2] + wsptr[DCTSIZE*5]; + tmp3 = dataptr[DCTSIZE*3] + wsptr[DCTSIZE*4]; + tmp4 = dataptr[DCTSIZE*4] + wsptr[DCTSIZE*3]; + tmp5 = dataptr[DCTSIZE*5] + wsptr[DCTSIZE*2]; + tmp6 = dataptr[DCTSIZE*6] + wsptr[DCTSIZE*1]; + tmp7 = dataptr[DCTSIZE*7] + wsptr[DCTSIZE*0]; + + tmp10 = tmp0 + tmp7; + tmp14 = tmp0 - tmp7; + tmp11 = tmp1 + tmp6; + tmp15 = tmp1 - tmp6; + tmp12 = tmp2 + tmp5; + tmp16 = tmp2 - tmp5; + tmp13 = tmp3 + tmp4; + tmp17 = tmp3 - tmp4; + + tmp0 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*7]; + tmp1 = dataptr[DCTSIZE*1] - wsptr[DCTSIZE*6]; + tmp2 = dataptr[DCTSIZE*2] - wsptr[DCTSIZE*5]; + tmp3 = dataptr[DCTSIZE*3] - wsptr[DCTSIZE*4]; + tmp4 = dataptr[DCTSIZE*4] - wsptr[DCTSIZE*3]; + tmp5 = dataptr[DCTSIZE*5] - wsptr[DCTSIZE*2]; + tmp6 = dataptr[DCTSIZE*6] - wsptr[DCTSIZE*1]; + tmp7 = dataptr[DCTSIZE*7] - wsptr[DCTSIZE*0]; + + dataptr[DCTSIZE*0] = (DCTELEM) + DESCALE(tmp10 + tmp11 + tmp12 + tmp13, PASS1_BITS+1); + dataptr[DCTSIZE*4] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp13, FIX(1.306562965)) + /* c4[16] = c2[8] */ + MULTIPLY(tmp11 - tmp12, FIX_0_541196100), /* c12[16] = c6[8] */ + CONST_BITS+PASS1_BITS+1); + + tmp10 = MULTIPLY(tmp17 - tmp15, FIX(0.275899379)) + /* c14[16] = c7[8] */ + MULTIPLY(tmp14 - tmp16, FIX(1.387039845)); /* c2[16] = c1[8] */ + + dataptr[DCTSIZE*2] = (DCTELEM) + DESCALE(tmp10 + MULTIPLY(tmp15, FIX(1.451774982)) /* c6+c14 */ + + MULTIPLY(tmp16, FIX(2.172734804)), /* c2+c10 */ + CONST_BITS+PASS1_BITS+1); + dataptr[DCTSIZE*6] = (DCTELEM) + DESCALE(tmp10 - MULTIPLY(tmp14, FIX(0.211164243)) /* c2-c6 */ + - MULTIPLY(tmp17, FIX(1.061594338)), /* c10+c14 */ + CONST_BITS+PASS1_BITS+1); + + /* Odd part */ + + tmp11 = MULTIPLY(tmp0 + tmp1, FIX(1.353318001)) + /* c3 */ + MULTIPLY(tmp6 - tmp7, FIX(0.410524528)); /* c13 */ + tmp12 = MULTIPLY(tmp0 + tmp2, FIX(1.247225013)) + /* c5 */ + MULTIPLY(tmp5 + tmp7, FIX(0.666655658)); /* c11 */ + tmp13 = MULTIPLY(tmp0 + tmp3, FIX(1.093201867)) + /* c7 */ + MULTIPLY(tmp4 - tmp7, FIX(0.897167586)); /* c9 */ + tmp14 = MULTIPLY(tmp1 + tmp2, FIX(0.138617169)) + /* c15 */ + MULTIPLY(tmp6 - tmp5, FIX(1.407403738)); /* c1 */ + tmp15 = MULTIPLY(tmp1 + tmp3, - FIX(0.666655658)) + /* -c11 */ + MULTIPLY(tmp4 + tmp6, - FIX(1.247225013)); /* -c5 */ + tmp16 = MULTIPLY(tmp2 + tmp3, - FIX(1.353318001)) + /* -c3 */ + MULTIPLY(tmp5 - tmp4, FIX(0.410524528)); /* c13 */ + tmp10 = tmp11 + tmp12 + tmp13 - + MULTIPLY(tmp0, FIX(2.286341144)) + /* c7+c5+c3-c1 */ + MULTIPLY(tmp7, FIX(0.779653625)); /* c15+c13-c11+c9 */ + tmp11 += tmp14 + tmp15 + MULTIPLY(tmp1, FIX(0.071888074)) /* c9-c3-c15+c11 */ + - MULTIPLY(tmp6, FIX(1.663905119)); /* c7+c13+c1-c5 */ + tmp12 += tmp14 + tmp16 - MULTIPLY(tmp2, FIX(1.125726048)) /* c7+c5+c15-c3 */ + + MULTIPLY(tmp5, FIX(1.227391138)); /* c9-c11+c1-c13 */ + tmp13 += tmp15 + tmp16 + MULTIPLY(tmp3, FIX(1.065388962)) /* c15+c3+c11-c7 */ + + MULTIPLY(tmp4, FIX(2.167985692)); /* c1+c13+c5-c9 */ + + dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp10, CONST_BITS+PASS1_BITS+1); + dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp11, CONST_BITS+PASS1_BITS+1); + dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp12, CONST_BITS+PASS1_BITS+1); + dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp13, CONST_BITS+PASS1_BITS+1); + + dataptr++; /* advance pointer to next column */ + wsptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 7x14 sample block. + * + * 7-point FDCT in pass 1 (rows), 14-point in pass 2 (columns). + */ + +GLOBAL(void) +jpeg_fdct_7x14 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6; + INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15, tmp16; + INT32 z1, z2, z3; + DCTELEM workspace[8*6]; + DCTELEM *dataptr; + DCTELEM *wsptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pre-zero output coefficient block. */ + MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT; + * furthermore, we scale the results by 2**PASS1_BITS. + * 7-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/14). + */ + + dataptr = data; + ctr = 0; + for (;;) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[6]); + tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[5]); + tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[4]); + tmp3 = GETJSAMPLE(elemptr[3]); + + tmp10 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[6]); + tmp11 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[5]); + tmp12 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[4]); + + z1 = tmp0 + tmp2; + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) + ((z1 + tmp1 + tmp3 - 7 * CENTERJSAMPLE) << PASS1_BITS); + tmp3 += tmp3; + z1 -= tmp3; + z1 -= tmp3; + z1 = MULTIPLY(z1, FIX(0.353553391)); /* (c2+c6-c4)/2 */ + z2 = MULTIPLY(tmp0 - tmp2, FIX(0.920609002)); /* (c2+c4-c6)/2 */ + z3 = MULTIPLY(tmp1 - tmp2, FIX(0.314692123)); /* c6 */ + dataptr[2] = (DCTELEM) DESCALE(z1 + z2 + z3, CONST_BITS-PASS1_BITS); + z1 -= z2; + z2 = MULTIPLY(tmp0 - tmp1, FIX(0.881747734)); /* c4 */ + dataptr[4] = (DCTELEM) + DESCALE(z2 + z3 - MULTIPLY(tmp1 - tmp3, FIX(0.707106781)), /* c2+c6-c4 */ + CONST_BITS-PASS1_BITS); + dataptr[6] = (DCTELEM) DESCALE(z1 + z2, CONST_BITS-PASS1_BITS); + + /* Odd part */ + + tmp1 = MULTIPLY(tmp10 + tmp11, FIX(0.935414347)); /* (c3+c1-c5)/2 */ + tmp2 = MULTIPLY(tmp10 - tmp11, FIX(0.170262339)); /* (c3+c5-c1)/2 */ + tmp0 = tmp1 - tmp2; + tmp1 += tmp2; + tmp2 = MULTIPLY(tmp11 + tmp12, - FIX(1.378756276)); /* -c1 */ + tmp1 += tmp2; + tmp3 = MULTIPLY(tmp10 + tmp12, FIX(0.613604268)); /* c5 */ + tmp0 += tmp3; + tmp2 += tmp3 + MULTIPLY(tmp12, FIX(1.870828693)); /* c3+c1-c5 */ + + dataptr[1] = (DCTELEM) DESCALE(tmp0, CONST_BITS-PASS1_BITS); + dataptr[3] = (DCTELEM) DESCALE(tmp1, CONST_BITS-PASS1_BITS); + dataptr[5] = (DCTELEM) DESCALE(tmp2, CONST_BITS-PASS1_BITS); + + ctr++; + + if (ctr != DCTSIZE) { + if (ctr == 14) + break; /* Done. */ + dataptr += DCTSIZE; /* advance pointer to next row */ + } else + dataptr = workspace; /* switch pointer to extended workspace */ + } + + /* Pass 2: process columns. + * We remove the PASS1_BITS scaling, but leave the results scaled up + * by an overall factor of 8. + * We must also scale the output by (8/7)*(8/14) = 32/49, which we + * fold into the constant multipliers: + * 14-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/28) * 32/49. + */ + + dataptr = data; + wsptr = workspace; + for (ctr = 0; ctr < 7; ctr++) { + /* Even part */ + + tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*5]; + tmp1 = dataptr[DCTSIZE*1] + wsptr[DCTSIZE*4]; + tmp2 = dataptr[DCTSIZE*2] + wsptr[DCTSIZE*3]; + tmp13 = dataptr[DCTSIZE*3] + wsptr[DCTSIZE*2]; + tmp4 = dataptr[DCTSIZE*4] + wsptr[DCTSIZE*1]; + tmp5 = dataptr[DCTSIZE*5] + wsptr[DCTSIZE*0]; + tmp6 = dataptr[DCTSIZE*6] + dataptr[DCTSIZE*7]; + + tmp10 = tmp0 + tmp6; + tmp14 = tmp0 - tmp6; + tmp11 = tmp1 + tmp5; + tmp15 = tmp1 - tmp5; + tmp12 = tmp2 + tmp4; + tmp16 = tmp2 - tmp4; + + tmp0 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*5]; + tmp1 = dataptr[DCTSIZE*1] - wsptr[DCTSIZE*4]; + tmp2 = dataptr[DCTSIZE*2] - wsptr[DCTSIZE*3]; + tmp3 = dataptr[DCTSIZE*3] - wsptr[DCTSIZE*2]; + tmp4 = dataptr[DCTSIZE*4] - wsptr[DCTSIZE*1]; + tmp5 = dataptr[DCTSIZE*5] - wsptr[DCTSIZE*0]; + tmp6 = dataptr[DCTSIZE*6] - dataptr[DCTSIZE*7]; + + dataptr[DCTSIZE*0] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 + tmp11 + tmp12 + tmp13, + FIX(0.653061224)), /* 32/49 */ + CONST_BITS+PASS1_BITS); + tmp13 += tmp13; + dataptr[DCTSIZE*4] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp13, FIX(0.832106052)) + /* c4 */ + MULTIPLY(tmp11 - tmp13, FIX(0.205513223)) - /* c12 */ + MULTIPLY(tmp12 - tmp13, FIX(0.575835255)), /* c8 */ + CONST_BITS+PASS1_BITS); + + tmp10 = MULTIPLY(tmp14 + tmp15, FIX(0.722074570)); /* c6 */ + + dataptr[DCTSIZE*2] = (DCTELEM) + DESCALE(tmp10 + MULTIPLY(tmp14, FIX(0.178337691)) /* c2-c6 */ + + MULTIPLY(tmp16, FIX(0.400721155)), /* c10 */ + CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*6] = (DCTELEM) + DESCALE(tmp10 - MULTIPLY(tmp15, FIX(1.122795725)) /* c6+c10 */ + - MULTIPLY(tmp16, FIX(0.900412262)), /* c2 */ + CONST_BITS+PASS1_BITS); + + /* Odd part */ + + tmp10 = tmp1 + tmp2; + tmp11 = tmp5 - tmp4; + dataptr[DCTSIZE*7] = (DCTELEM) + DESCALE(MULTIPLY(tmp0 - tmp10 + tmp3 - tmp11 - tmp6, + FIX(0.653061224)), /* 32/49 */ + CONST_BITS+PASS1_BITS); + tmp3 = MULTIPLY(tmp3 , FIX(0.653061224)); /* 32/49 */ + tmp10 = MULTIPLY(tmp10, - FIX(0.103406812)); /* -c13 */ + tmp11 = MULTIPLY(tmp11, FIX(0.917760839)); /* c1 */ + tmp10 += tmp11 - tmp3; + tmp11 = MULTIPLY(tmp0 + tmp2, FIX(0.782007410)) + /* c5 */ + MULTIPLY(tmp4 + tmp6, FIX(0.491367823)); /* c9 */ + dataptr[DCTSIZE*5] = (DCTELEM) + DESCALE(tmp10 + tmp11 - MULTIPLY(tmp2, FIX(1.550341076)) /* c3+c5-c13 */ + + MULTIPLY(tmp4, FIX(0.731428202)), /* c1+c11-c9 */ + CONST_BITS+PASS1_BITS); + tmp12 = MULTIPLY(tmp0 + tmp1, FIX(0.871740478)) + /* c3 */ + MULTIPLY(tmp5 - tmp6, FIX(0.305035186)); /* c11 */ + dataptr[DCTSIZE*3] = (DCTELEM) + DESCALE(tmp10 + tmp12 - MULTIPLY(tmp1, FIX(0.276965844)) /* c3-c9-c13 */ + - MULTIPLY(tmp5, FIX(2.004803435)), /* c1+c5+c11 */ + CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*1] = (DCTELEM) + DESCALE(tmp11 + tmp12 + tmp3 + - MULTIPLY(tmp0, FIX(0.735987049)) /* c3+c5-c1 */ + - MULTIPLY(tmp6, FIX(0.082925825)), /* c9-c11-c13 */ + CONST_BITS+PASS1_BITS); + + dataptr++; /* advance pointer to next column */ + wsptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 6x12 sample block. + * + * 6-point FDCT in pass 1 (rows), 12-point in pass 2 (columns). + */ + +GLOBAL(void) +jpeg_fdct_6x12 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5; + INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15; + DCTELEM workspace[8*4]; + DCTELEM *dataptr; + DCTELEM *wsptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pre-zero output coefficient block. */ + MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT; + * furthermore, we scale the results by 2**PASS1_BITS. + * 6-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/12). + */ + + dataptr = data; + ctr = 0; + for (;;) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[5]); + tmp11 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[4]); + tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[3]); + + tmp10 = tmp0 + tmp2; + tmp12 = tmp0 - tmp2; + + tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[5]); + tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[4]); + tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[3]); + + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) + ((tmp10 + tmp11 - 6 * CENTERJSAMPLE) << PASS1_BITS); + dataptr[2] = (DCTELEM) + DESCALE(MULTIPLY(tmp12, FIX(1.224744871)), /* c2 */ + CONST_BITS-PASS1_BITS); + dataptr[4] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp11 - tmp11, FIX(0.707106781)), /* c4 */ + CONST_BITS-PASS1_BITS); + + /* Odd part */ + + tmp10 = DESCALE(MULTIPLY(tmp0 + tmp2, FIX(0.366025404)), /* c5 */ + CONST_BITS-PASS1_BITS); + + dataptr[1] = (DCTELEM) (tmp10 + ((tmp0 + tmp1) << PASS1_BITS)); + dataptr[3] = (DCTELEM) ((tmp0 - tmp1 - tmp2) << PASS1_BITS); + dataptr[5] = (DCTELEM) (tmp10 + ((tmp2 - tmp1) << PASS1_BITS)); + + ctr++; + + if (ctr != DCTSIZE) { + if (ctr == 12) + break; /* Done. */ + dataptr += DCTSIZE; /* advance pointer to next row */ + } else + dataptr = workspace; /* switch pointer to extended workspace */ + } + + /* Pass 2: process columns. + * We remove the PASS1_BITS scaling, but leave the results scaled up + * by an overall factor of 8. + * We must also scale the output by (8/6)*(8/12) = 8/9, which we + * fold into the constant multipliers: + * 12-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/24) * 8/9. + */ + + dataptr = data; + wsptr = workspace; + for (ctr = 0; ctr < 6; ctr++) { + /* Even part */ + + tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*3]; + tmp1 = dataptr[DCTSIZE*1] + wsptr[DCTSIZE*2]; + tmp2 = dataptr[DCTSIZE*2] + wsptr[DCTSIZE*1]; + tmp3 = dataptr[DCTSIZE*3] + wsptr[DCTSIZE*0]; + tmp4 = dataptr[DCTSIZE*4] + dataptr[DCTSIZE*7]; + tmp5 = dataptr[DCTSIZE*5] + dataptr[DCTSIZE*6]; + + tmp10 = tmp0 + tmp5; + tmp13 = tmp0 - tmp5; + tmp11 = tmp1 + tmp4; + tmp14 = tmp1 - tmp4; + tmp12 = tmp2 + tmp3; + tmp15 = tmp2 - tmp3; + + tmp0 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*3]; + tmp1 = dataptr[DCTSIZE*1] - wsptr[DCTSIZE*2]; + tmp2 = dataptr[DCTSIZE*2] - wsptr[DCTSIZE*1]; + tmp3 = dataptr[DCTSIZE*3] - wsptr[DCTSIZE*0]; + tmp4 = dataptr[DCTSIZE*4] - dataptr[DCTSIZE*7]; + tmp5 = dataptr[DCTSIZE*5] - dataptr[DCTSIZE*6]; + + dataptr[DCTSIZE*0] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 + tmp11 + tmp12, FIX(0.888888889)), /* 8/9 */ + CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*6] = (DCTELEM) + DESCALE(MULTIPLY(tmp13 - tmp14 - tmp15, FIX(0.888888889)), /* 8/9 */ + CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*4] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp12, FIX(1.088662108)), /* c4 */ + CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*2] = (DCTELEM) + DESCALE(MULTIPLY(tmp14 - tmp15, FIX(0.888888889)) + /* 8/9 */ + MULTIPLY(tmp13 + tmp15, FIX(1.214244803)), /* c2 */ + CONST_BITS+PASS1_BITS); + + /* Odd part */ + + tmp10 = MULTIPLY(tmp1 + tmp4, FIX(0.481063200)); /* c9 */ + tmp14 = tmp10 + MULTIPLY(tmp1, FIX(0.680326102)); /* c3-c9 */ + tmp15 = tmp10 - MULTIPLY(tmp4, FIX(1.642452502)); /* c3+c9 */ + tmp12 = MULTIPLY(tmp0 + tmp2, FIX(0.997307603)); /* c5 */ + tmp13 = MULTIPLY(tmp0 + tmp3, FIX(0.765261039)); /* c7 */ + tmp10 = tmp12 + tmp13 + tmp14 - MULTIPLY(tmp0, FIX(0.516244403)) /* c5+c7-c1 */ + + MULTIPLY(tmp5, FIX(0.164081699)); /* c11 */ + tmp11 = MULTIPLY(tmp2 + tmp3, - FIX(0.164081699)); /* -c11 */ + tmp12 += tmp11 - tmp15 - MULTIPLY(tmp2, FIX(2.079550144)) /* c1+c5-c11 */ + + MULTIPLY(tmp5, FIX(0.765261039)); /* c7 */ + tmp13 += tmp11 - tmp14 + MULTIPLY(tmp3, FIX(0.645144899)) /* c1+c11-c7 */ + - MULTIPLY(tmp5, FIX(0.997307603)); /* c5 */ + tmp11 = tmp15 + MULTIPLY(tmp0 - tmp3, FIX(1.161389302)) /* c3 */ + - MULTIPLY(tmp2 + tmp5, FIX(0.481063200)); /* c9 */ + + dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp10, CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp11, CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp12, CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp13, CONST_BITS+PASS1_BITS); + + dataptr++; /* advance pointer to next column */ + wsptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 5x10 sample block. + * + * 5-point FDCT in pass 1 (rows), 10-point in pass 2 (columns). + */ + +GLOBAL(void) +jpeg_fdct_5x10 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3, tmp4; + INT32 tmp10, tmp11, tmp12, tmp13, tmp14; + DCTELEM workspace[8*2]; + DCTELEM *dataptr; + DCTELEM *wsptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pre-zero output coefficient block. */ + MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT; + * furthermore, we scale the results by 2**PASS1_BITS. + * 5-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/10). + */ + + dataptr = data; + ctr = 0; + for (;;) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[4]); + tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[3]); + tmp2 = GETJSAMPLE(elemptr[2]); + + tmp10 = tmp0 + tmp1; + tmp11 = tmp0 - tmp1; + + tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[4]); + tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[3]); + + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) + ((tmp10 + tmp2 - 5 * CENTERJSAMPLE) << PASS1_BITS); + tmp11 = MULTIPLY(tmp11, FIX(0.790569415)); /* (c2+c4)/2 */ + tmp10 -= tmp2 << 2; + tmp10 = MULTIPLY(tmp10, FIX(0.353553391)); /* (c2-c4)/2 */ + dataptr[2] = (DCTELEM) DESCALE(tmp11 + tmp10, CONST_BITS-PASS1_BITS); + dataptr[4] = (DCTELEM) DESCALE(tmp11 - tmp10, CONST_BITS-PASS1_BITS); + + /* Odd part */ + + tmp10 = MULTIPLY(tmp0 + tmp1, FIX(0.831253876)); /* c3 */ + + dataptr[1] = (DCTELEM) + DESCALE(tmp10 + MULTIPLY(tmp0, FIX(0.513743148)), /* c1-c3 */ + CONST_BITS-PASS1_BITS); + dataptr[3] = (DCTELEM) + DESCALE(tmp10 - MULTIPLY(tmp1, FIX(2.176250899)), /* c1+c3 */ + CONST_BITS-PASS1_BITS); + + ctr++; + + if (ctr != DCTSIZE) { + if (ctr == 10) + break; /* Done. */ + dataptr += DCTSIZE; /* advance pointer to next row */ + } else + dataptr = workspace; /* switch pointer to extended workspace */ + } + + /* Pass 2: process columns. + * We remove the PASS1_BITS scaling, but leave the results scaled up + * by an overall factor of 8. + * We must also scale the output by (8/5)*(8/10) = 32/25, which we + * fold into the constant multipliers: + * 10-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/20) * 32/25. + */ + + dataptr = data; + wsptr = workspace; + for (ctr = 0; ctr < 5; ctr++) { + /* Even part */ + + tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*1]; + tmp1 = dataptr[DCTSIZE*1] + wsptr[DCTSIZE*0]; + tmp12 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*7]; + tmp3 = dataptr[DCTSIZE*3] + dataptr[DCTSIZE*6]; + tmp4 = dataptr[DCTSIZE*4] + dataptr[DCTSIZE*5]; + + tmp10 = tmp0 + tmp4; + tmp13 = tmp0 - tmp4; + tmp11 = tmp1 + tmp3; + tmp14 = tmp1 - tmp3; + + tmp0 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*1]; + tmp1 = dataptr[DCTSIZE*1] - wsptr[DCTSIZE*0]; + tmp2 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*7]; + tmp3 = dataptr[DCTSIZE*3] - dataptr[DCTSIZE*6]; + tmp4 = dataptr[DCTSIZE*4] - dataptr[DCTSIZE*5]; + + dataptr[DCTSIZE*0] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 + tmp11 + tmp12, FIX(1.28)), /* 32/25 */ + CONST_BITS+PASS1_BITS); + tmp12 += tmp12; + dataptr[DCTSIZE*4] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp12, FIX(1.464477191)) - /* c4 */ + MULTIPLY(tmp11 - tmp12, FIX(0.559380511)), /* c8 */ + CONST_BITS+PASS1_BITS); + tmp10 = MULTIPLY(tmp13 + tmp14, FIX(1.064004961)); /* c6 */ + dataptr[DCTSIZE*2] = (DCTELEM) + DESCALE(tmp10 + MULTIPLY(tmp13, FIX(0.657591230)), /* c2-c6 */ + CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*6] = (DCTELEM) + DESCALE(tmp10 - MULTIPLY(tmp14, FIX(2.785601151)), /* c2+c6 */ + CONST_BITS+PASS1_BITS); + + /* Odd part */ + + tmp10 = tmp0 + tmp4; + tmp11 = tmp1 - tmp3; + dataptr[DCTSIZE*5] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp11 - tmp2, FIX(1.28)), /* 32/25 */ + CONST_BITS+PASS1_BITS); + tmp2 = MULTIPLY(tmp2, FIX(1.28)); /* 32/25 */ + dataptr[DCTSIZE*1] = (DCTELEM) + DESCALE(MULTIPLY(tmp0, FIX(1.787906876)) + /* c1 */ + MULTIPLY(tmp1, FIX(1.612894094)) + tmp2 + /* c3 */ + MULTIPLY(tmp3, FIX(0.821810588)) + /* c7 */ + MULTIPLY(tmp4, FIX(0.283176630)), /* c9 */ + CONST_BITS+PASS1_BITS); + tmp12 = MULTIPLY(tmp0 - tmp4, FIX(1.217352341)) - /* (c3+c7)/2 */ + MULTIPLY(tmp1 + tmp3, FIX(0.752365123)); /* (c1-c9)/2 */ + tmp13 = MULTIPLY(tmp10 + tmp11, FIX(0.395541753)) + /* (c3-c7)/2 */ + MULTIPLY(tmp11, FIX(0.64)) - tmp2; /* 16/25 */ + dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp12 + tmp13, CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp12 - tmp13, CONST_BITS+PASS1_BITS); + + dataptr++; /* advance pointer to next column */ + wsptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 4x8 sample block. + * + * 4-point FDCT in pass 1 (rows), 8-point in pass 2 (columns). + */ + +GLOBAL(void) +jpeg_fdct_4x8 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3; + INT32 tmp10, tmp11, tmp12, tmp13; + INT32 z1; + DCTELEM *dataptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pre-zero output coefficient block. */ + MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT; + * furthermore, we scale the results by 2**PASS1_BITS. + * We must also scale the output by 8/4 = 2, which we add here. + * 4-point FDCT kernel, + * cK represents sqrt(2) * cos(K*pi/16) [refers to 8-point FDCT]. + */ + + dataptr = data; + for (ctr = 0; ctr < DCTSIZE; ctr++) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[3]); + tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[2]); + + tmp10 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[3]); + tmp11 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[2]); + + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) + ((tmp0 + tmp1 - 4 * CENTERJSAMPLE) << (PASS1_BITS+1)); + dataptr[2] = (DCTELEM) ((tmp0 - tmp1) << (PASS1_BITS+1)); + + /* Odd part */ + + tmp0 = MULTIPLY(tmp10 + tmp11, FIX_0_541196100); /* c6 */ + /* Add fudge factor here for final descale. */ + tmp0 += ONE << (CONST_BITS-PASS1_BITS-2); + + dataptr[1] = (DCTELEM) + RIGHT_SHIFT(tmp0 + MULTIPLY(tmp10, FIX_0_765366865), /* c2-c6 */ + CONST_BITS-PASS1_BITS-1); + dataptr[3] = (DCTELEM) + RIGHT_SHIFT(tmp0 - MULTIPLY(tmp11, FIX_1_847759065), /* c2+c6 */ + CONST_BITS-PASS1_BITS-1); + + dataptr += DCTSIZE; /* advance pointer to next row */ + } + + /* Pass 2: process columns. + * We remove the PASS1_BITS scaling, but leave the results scaled up + * by an overall factor of 8. + * 8-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/16). + */ + + dataptr = data; + for (ctr = 0; ctr < 4; ctr++) { + /* Even part per LL&M figure 1 --- note that published figure is faulty; + * rotator "c1" should be "c6". + */ + + tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*7]; + tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*6]; + tmp2 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*5]; + tmp3 = dataptr[DCTSIZE*3] + dataptr[DCTSIZE*4]; + + /* Add fudge factor here for final descale. */ + tmp10 = tmp0 + tmp3 + (ONE << (PASS1_BITS-1)); + tmp12 = tmp0 - tmp3; + tmp11 = tmp1 + tmp2; + tmp13 = tmp1 - tmp2; + + tmp0 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*7]; + tmp1 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*6]; + tmp2 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*5]; + tmp3 = dataptr[DCTSIZE*3] - dataptr[DCTSIZE*4]; + + dataptr[DCTSIZE*0] = (DCTELEM) RIGHT_SHIFT(tmp10 + tmp11, PASS1_BITS); + dataptr[DCTSIZE*4] = (DCTELEM) RIGHT_SHIFT(tmp10 - tmp11, PASS1_BITS); + + z1 = MULTIPLY(tmp12 + tmp13, FIX_0_541196100); /* c6 */ + /* Add fudge factor here for final descale. */ + z1 += ONE << (CONST_BITS+PASS1_BITS-1); + + dataptr[DCTSIZE*2] = (DCTELEM) + RIGHT_SHIFT(z1 + MULTIPLY(tmp12, FIX_0_765366865), /* c2-c6 */ + CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*6] = (DCTELEM) + RIGHT_SHIFT(z1 - MULTIPLY(tmp13, FIX_1_847759065), /* c2+c6 */ + CONST_BITS+PASS1_BITS); + + /* Odd part per figure 8 --- note paper omits factor of sqrt(2). + * i0..i3 in the paper are tmp0..tmp3 here. + */ + + tmp12 = tmp0 + tmp2; + tmp13 = tmp1 + tmp3; + + z1 = MULTIPLY(tmp12 + tmp13, FIX_1_175875602); /* c3 */ + /* Add fudge factor here for final descale. */ + z1 += ONE << (CONST_BITS+PASS1_BITS-1); + + tmp12 = MULTIPLY(tmp12, - FIX_0_390180644); /* -c3+c5 */ + tmp13 = MULTIPLY(tmp13, - FIX_1_961570560); /* -c3-c5 */ + tmp12 += z1; + tmp13 += z1; + + z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ + tmp0 = MULTIPLY(tmp0, FIX_1_501321110); /* c1+c3-c5-c7 */ + tmp3 = MULTIPLY(tmp3, FIX_0_298631336); /* -c1+c3+c5-c7 */ + tmp0 += z1 + tmp12; + tmp3 += z1 + tmp13; + + z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ + tmp1 = MULTIPLY(tmp1, FIX_3_072711026); /* c1+c3+c5-c7 */ + tmp2 = MULTIPLY(tmp2, FIX_2_053119869); /* c1+c3-c5+c7 */ + tmp1 += z1 + tmp13; + tmp2 += z1 + tmp12; + + dataptr[DCTSIZE*1] = (DCTELEM) RIGHT_SHIFT(tmp0, CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*3] = (DCTELEM) RIGHT_SHIFT(tmp1, CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*5] = (DCTELEM) RIGHT_SHIFT(tmp2, CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*7] = (DCTELEM) RIGHT_SHIFT(tmp3, CONST_BITS+PASS1_BITS); + + dataptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 3x6 sample block. + * + * 3-point FDCT in pass 1 (rows), 6-point in pass 2 (columns). + */ + +GLOBAL(void) +jpeg_fdct_3x6 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1, tmp2; + INT32 tmp10, tmp11, tmp12; + DCTELEM *dataptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pre-zero output coefficient block. */ + MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT; + * furthermore, we scale the results by 2**PASS1_BITS. + * We scale the results further by 2 as part of output adaption + * scaling for different DCT size. + * 3-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/6). + */ + + dataptr = data; + for (ctr = 0; ctr < 6; ctr++) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[2]); + tmp1 = GETJSAMPLE(elemptr[1]); + + tmp2 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[2]); + + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) + ((tmp0 + tmp1 - 3 * CENTERJSAMPLE) << (PASS1_BITS+1)); + dataptr[2] = (DCTELEM) + DESCALE(MULTIPLY(tmp0 - tmp1 - tmp1, FIX(0.707106781)), /* c2 */ + CONST_BITS-PASS1_BITS-1); + + /* Odd part */ + + dataptr[1] = (DCTELEM) + DESCALE(MULTIPLY(tmp2, FIX(1.224744871)), /* c1 */ + CONST_BITS-PASS1_BITS-1); + + dataptr += DCTSIZE; /* advance pointer to next row */ + } + + /* Pass 2: process columns. + * We remove the PASS1_BITS scaling, but leave the results scaled up + * by an overall factor of 8. + * We must also scale the output by (8/6)*(8/3) = 32/9, which we partially + * fold into the constant multipliers (other part was done in pass 1): + * 6-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/12) * 16/9. + */ + + dataptr = data; + for (ctr = 0; ctr < 3; ctr++) { + /* Even part */ + + tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*5]; + tmp11 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*4]; + tmp2 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*3]; + + tmp10 = tmp0 + tmp2; + tmp12 = tmp0 - tmp2; + + tmp0 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*5]; + tmp1 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*4]; + tmp2 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*3]; + + dataptr[DCTSIZE*0] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 + tmp11, FIX(1.777777778)), /* 16/9 */ + CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*2] = (DCTELEM) + DESCALE(MULTIPLY(tmp12, FIX(2.177324216)), /* c2 */ + CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*4] = (DCTELEM) + DESCALE(MULTIPLY(tmp10 - tmp11 - tmp11, FIX(1.257078722)), /* c4 */ + CONST_BITS+PASS1_BITS); + + /* Odd part */ + + tmp10 = MULTIPLY(tmp0 + tmp2, FIX(0.650711829)); /* c5 */ + + dataptr[DCTSIZE*1] = (DCTELEM) + DESCALE(tmp10 + MULTIPLY(tmp0 + tmp1, FIX(1.777777778)), /* 16/9 */ + CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*3] = (DCTELEM) + DESCALE(MULTIPLY(tmp0 - tmp1 - tmp2, FIX(1.777777778)), /* 16/9 */ + CONST_BITS+PASS1_BITS); + dataptr[DCTSIZE*5] = (DCTELEM) + DESCALE(tmp10 + MULTIPLY(tmp2 - tmp1, FIX(1.777777778)), /* 16/9 */ + CONST_BITS+PASS1_BITS); + + dataptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 2x4 sample block. + * + * 2-point FDCT in pass 1 (rows), 4-point in pass 2 (columns). + */ + +GLOBAL(void) +jpeg_fdct_2x4 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + INT32 tmp0, tmp1; + INT32 tmp10, tmp11; + DCTELEM *dataptr; + JSAMPROW elemptr; + int ctr; + SHIFT_TEMPS + + /* Pre-zero output coefficient block. */ + MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); + + /* Pass 1: process rows. + * Note results are scaled up by sqrt(8) compared to a true DCT. + */ + + dataptr = data; + for (ctr = 0; ctr < 4; ctr++) { + elemptr = sample_data[ctr] + start_col; + + /* Even part */ + + tmp0 = GETJSAMPLE(elemptr[0]); + tmp1 = GETJSAMPLE(elemptr[1]); + + /* Apply unsigned->signed conversion. */ + dataptr[0] = (DCTELEM) (tmp0 + tmp1 - 2 * CENTERJSAMPLE); + + /* Odd part */ + + dataptr[1] = (DCTELEM) (tmp0 - tmp1); + + dataptr += DCTSIZE; /* advance pointer to next row */ + } + + /* Pass 2: process columns. + * We leave the results scaled up by an overall factor of 8. + * We must also scale the output by (8/2)*(8/4) = 2**3. + * 4-point FDCT kernel, + * cK represents sqrt(2) * cos(K*pi/16) [refers to 8-point FDCT]. + */ + + dataptr = data; + for (ctr = 0; ctr < 2; ctr++) { + /* Even part */ + + tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*3]; + tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*2]; + + tmp10 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*3]; + tmp11 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*2]; + + dataptr[DCTSIZE*0] = (DCTELEM) ((tmp0 + tmp1) << 3); + dataptr[DCTSIZE*2] = (DCTELEM) ((tmp0 - tmp1) << 3); + + /* Odd part */ + + tmp0 = MULTIPLY(tmp10 + tmp11, FIX_0_541196100); /* c6 */ + /* Add fudge factor here for final descale. */ + tmp0 += ONE << (CONST_BITS-3-1); + + dataptr[DCTSIZE*1] = (DCTELEM) + RIGHT_SHIFT(tmp0 + MULTIPLY(tmp10, FIX_0_765366865), /* c2-c6 */ + CONST_BITS-3); + dataptr[DCTSIZE*3] = (DCTELEM) + RIGHT_SHIFT(tmp0 - MULTIPLY(tmp11, FIX_1_847759065), /* c2+c6 */ + CONST_BITS-3); + + dataptr++; /* advance pointer to next column */ + } +} + + +/* + * Perform the forward DCT on a 1x2 sample block. + * + * 1-point FDCT in pass 1 (rows), 2-point in pass 2 (columns). + */ + +GLOBAL(void) +jpeg_fdct_1x2 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) +{ + DCTELEM tmp0, tmp1; + + /* Pre-zero output coefficient block. */ + MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); + + /* Pass 1: empty. */ + + /* Pass 2: process columns. + * We leave the results scaled up by an overall factor of 8. + * We must also scale the output by (8/1)*(8/2) = 2**5. + */ + + /* Even part */ + + tmp0 = GETJSAMPLE(sample_data[0][start_col]); + tmp1 = GETJSAMPLE(sample_data[1][start_col]); + + /* Apply unsigned->signed conversion. */ + data[DCTSIZE*0] = (tmp0 + tmp1 - 2 * CENTERJSAMPLE) << 5; + + /* Odd part */ + + data[DCTSIZE*1] = (tmp0 - tmp1) << 5; +} + +#endif /* DCT_SCALING_SUPPORTED */ +#endif /* DCT_ISLOW_SUPPORTED */ diff --git a/cmake/externals/libjpeg/jidctflt.c b/cmake/externals/libjpeg/jidctflt.c new file mode 100644 index 000000000..e33a2b5e4 --- /dev/null +++ b/cmake/externals/libjpeg/jidctflt.c @@ -0,0 +1,238 @@ +/* + * jidctflt.c + * + * Copyright (C) 1994-1998, Thomas G. Lane. + * Modified 2010-2017 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains a floating-point implementation of the + * inverse DCT (Discrete Cosine Transform). In the IJG code, this routine + * must also perform dequantization of the input coefficients. + * + * This implementation should be more accurate than either of the integer + * IDCT implementations. However, it may not give the same results on all + * machines because of differences in roundoff behavior. Speed will depend + * on the hardware's floating point capacity. + * + * A 2-D IDCT can be done by 1-D IDCT on each column followed by 1-D IDCT + * on each row (or vice versa, but it's more convenient to emit a row at + * a time). Direct algorithms are also available, but they are much more + * complex and seem not to be any faster when reduced to code. + * + * This implementation is based on Arai, Agui, and Nakajima's algorithm for + * scaled DCT. Their original paper (Trans. IEICE E-71(11):1095) is in + * Japanese, but the algorithm is described in the Pennebaker & Mitchell + * JPEG textbook (see REFERENCES section in file README). The following code + * is based directly on figure 4-8 in P&M. + * While an 8-point DCT cannot be done in less than 11 multiplies, it is + * possible to arrange the computation so that many of the multiplies are + * simple scalings of the final outputs. These multiplies can then be + * folded into the multiplications or divisions by the JPEG quantization + * table entries. The AA&N method leaves only 5 multiplies and 29 adds + * to be done in the DCT itself. + * The primary disadvantage of this method is that with a fixed-point + * implementation, accuracy is lost due to imprecise representation of the + * scaled quantization values. However, that problem does not arise if + * we use floating point arithmetic. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" +#include "jdct.h" /* Private declarations for DCT subsystem */ + +#ifdef DCT_FLOAT_SUPPORTED + + +/* + * This module is specialized to the case DCTSIZE = 8. + */ + +#if DCTSIZE != 8 + Sorry, this code only copes with 8x8 DCT blocks. /* deliberate syntax err */ +#endif + + +/* Dequantize a coefficient by multiplying it by the multiplier-table + * entry; produce a float result. + */ + +#define DEQUANTIZE(coef,quantval) (((FAST_FLOAT) (coef)) * (quantval)) + + +/* + * Perform dequantization and inverse DCT on one block of coefficients. + * + * cK represents cos(K*pi/16). + */ + +GLOBAL(void) +jpeg_idct_float (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + FAST_FLOAT tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7; + FAST_FLOAT tmp10, tmp11, tmp12, tmp13; + FAST_FLOAT z5, z10, z11, z12, z13; + JCOEFPTR inptr; + FLOAT_MULT_TYPE * quantptr; + FAST_FLOAT * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + FAST_FLOAT workspace[DCTSIZE2]; /* buffers data between passes */ + + /* Pass 1: process columns from input, store into work array. */ + + inptr = coef_block; + quantptr = (FLOAT_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = DCTSIZE; ctr > 0; ctr--) { + /* Due to quantization, we will usually find that many of the input + * coefficients are zero, especially the AC terms. We can exploit this + * by short-circuiting the IDCT calculation for any column in which all + * the AC terms are zero. In that case each output is equal to the + * DC coefficient (with scale factor as needed). + * With typical images and quantization tables, half or more of the + * column DCT calculations can be simplified this way. + */ + + if (inptr[DCTSIZE*1] == 0 && inptr[DCTSIZE*2] == 0 && + inptr[DCTSIZE*3] == 0 && inptr[DCTSIZE*4] == 0 && + inptr[DCTSIZE*5] == 0 && inptr[DCTSIZE*6] == 0 && + inptr[DCTSIZE*7] == 0) { + /* AC terms all zero */ + FAST_FLOAT dcval = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + + wsptr[DCTSIZE*0] = dcval; + wsptr[DCTSIZE*1] = dcval; + wsptr[DCTSIZE*2] = dcval; + wsptr[DCTSIZE*3] = dcval; + wsptr[DCTSIZE*4] = dcval; + wsptr[DCTSIZE*5] = dcval; + wsptr[DCTSIZE*6] = dcval; + wsptr[DCTSIZE*7] = dcval; + + inptr++; /* advance pointers to next column */ + quantptr++; + wsptr++; + continue; + } + + /* Even part */ + + tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + tmp1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + tmp2 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + tmp3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); + + tmp10 = tmp0 + tmp2; /* phase 3 */ + tmp11 = tmp0 - tmp2; + + tmp13 = tmp1 + tmp3; /* phases 5-3 */ + tmp12 = (tmp1 - tmp3) * ((FAST_FLOAT) 1.414213562) - tmp13; /* 2*c4 */ + + tmp0 = tmp10 + tmp13; /* phase 2 */ + tmp3 = tmp10 - tmp13; + tmp1 = tmp11 + tmp12; + tmp2 = tmp11 - tmp12; + + /* Odd part */ + + tmp4 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + tmp5 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + tmp6 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); + tmp7 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); + + z13 = tmp6 + tmp5; /* phase 6 */ + z10 = tmp6 - tmp5; + z11 = tmp4 + tmp7; + z12 = tmp4 - tmp7; + + tmp7 = z11 + z13; /* phase 5 */ + tmp11 = (z11 - z13) * ((FAST_FLOAT) 1.414213562); /* 2*c4 */ + + z5 = (z10 + z12) * ((FAST_FLOAT) 1.847759065); /* 2*c2 */ + tmp10 = z5 - z12 * ((FAST_FLOAT) 1.082392200); /* 2*(c2-c6) */ + tmp12 = z5 - z10 * ((FAST_FLOAT) 2.613125930); /* 2*(c2+c6) */ + + tmp6 = tmp12 - tmp7; /* phase 2 */ + tmp5 = tmp11 - tmp6; + tmp4 = tmp10 - tmp5; + + wsptr[DCTSIZE*0] = tmp0 + tmp7; + wsptr[DCTSIZE*7] = tmp0 - tmp7; + wsptr[DCTSIZE*1] = tmp1 + tmp6; + wsptr[DCTSIZE*6] = tmp1 - tmp6; + wsptr[DCTSIZE*2] = tmp2 + tmp5; + wsptr[DCTSIZE*5] = tmp2 - tmp5; + wsptr[DCTSIZE*3] = tmp3 + tmp4; + wsptr[DCTSIZE*4] = tmp3 - tmp4; + + inptr++; /* advance pointers to next column */ + quantptr++; + wsptr++; + } + + /* Pass 2: process rows from work array, store into output array. */ + + wsptr = workspace; + for (ctr = 0; ctr < DCTSIZE; ctr++) { + outptr = output_buf[ctr] + output_col; + /* Rows of zeroes can be exploited in the same way as we did with columns. + * However, the column calculation has created many nonzero AC terms, so + * the simplification applies less often (typically 5% to 10% of the time). + * And testing floats for zero is relatively expensive, so we don't bother. + */ + + /* Even part */ + + /* Prepare range-limit and float->int conversion */ + z5 = wsptr[0] + (((FAST_FLOAT) RANGE_CENTER) + ((FAST_FLOAT) 0.5)); + tmp10 = z5 + wsptr[4]; + tmp11 = z5 - wsptr[4]; + + tmp13 = wsptr[2] + wsptr[6]; + tmp12 = (wsptr[2] - wsptr[6]) * + ((FAST_FLOAT) 1.414213562) - tmp13; /* 2*c4 */ + + tmp0 = tmp10 + tmp13; + tmp3 = tmp10 - tmp13; + tmp1 = tmp11 + tmp12; + tmp2 = tmp11 - tmp12; + + /* Odd part */ + + z13 = wsptr[5] + wsptr[3]; + z10 = wsptr[5] - wsptr[3]; + z11 = wsptr[1] + wsptr[7]; + z12 = wsptr[1] - wsptr[7]; + + tmp7 = z11 + z13; /* phase 5 */ + tmp11 = (z11 - z13) * ((FAST_FLOAT) 1.414213562); /* 2*c4 */ + + z5 = (z10 + z12) * ((FAST_FLOAT) 1.847759065); /* 2*c2 */ + tmp10 = z5 - z12 * ((FAST_FLOAT) 1.082392200); /* 2*(c2-c6) */ + tmp12 = z5 - z10 * ((FAST_FLOAT) 2.613125930); /* 2*(c2+c6) */ + + tmp6 = tmp12 - tmp7; /* phase 2 */ + tmp5 = tmp11 - tmp6; + tmp4 = tmp10 - tmp5; + + /* Final output stage: float->int conversion and range-limit */ + + outptr[0] = range_limit[(int) (tmp0 + tmp7) & RANGE_MASK]; + outptr[7] = range_limit[(int) (tmp0 - tmp7) & RANGE_MASK]; + outptr[1] = range_limit[(int) (tmp1 + tmp6) & RANGE_MASK]; + outptr[6] = range_limit[(int) (tmp1 - tmp6) & RANGE_MASK]; + outptr[2] = range_limit[(int) (tmp2 + tmp5) & RANGE_MASK]; + outptr[5] = range_limit[(int) (tmp2 - tmp5) & RANGE_MASK]; + outptr[3] = range_limit[(int) (tmp3 + tmp4) & RANGE_MASK]; + outptr[4] = range_limit[(int) (tmp3 - tmp4) & RANGE_MASK]; + + wsptr += DCTSIZE; /* advance pointer to next row */ + } +} + +#endif /* DCT_FLOAT_SUPPORTED */ diff --git a/cmake/externals/libjpeg/jidctfst.c b/cmake/externals/libjpeg/jidctfst.c new file mode 100644 index 000000000..1ac3e39cb --- /dev/null +++ b/cmake/externals/libjpeg/jidctfst.c @@ -0,0 +1,351 @@ +/* + * jidctfst.c + * + * Copyright (C) 1994-1998, Thomas G. Lane. + * Modified 2015-2017 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains a fast, not so accurate integer implementation of the + * inverse DCT (Discrete Cosine Transform). In the IJG code, this routine + * must also perform dequantization of the input coefficients. + * + * A 2-D IDCT can be done by 1-D IDCT on each column followed by 1-D IDCT + * on each row (or vice versa, but it's more convenient to emit a row at + * a time). Direct algorithms are also available, but they are much more + * complex and seem not to be any faster when reduced to code. + * + * This implementation is based on Arai, Agui, and Nakajima's algorithm for + * scaled DCT. Their original paper (Trans. IEICE E-71(11):1095) is in + * Japanese, but the algorithm is described in the Pennebaker & Mitchell + * JPEG textbook (see REFERENCES section in file README). The following code + * is based directly on figure 4-8 in P&M. + * While an 8-point DCT cannot be done in less than 11 multiplies, it is + * possible to arrange the computation so that many of the multiplies are + * simple scalings of the final outputs. These multiplies can then be + * folded into the multiplications or divisions by the JPEG quantization + * table entries. The AA&N method leaves only 5 multiplies and 29 adds + * to be done in the DCT itself. + * The primary disadvantage of this method is that with fixed-point math, + * accuracy is lost due to imprecise representation of the scaled + * quantization values. The smaller the quantization table entry, the less + * precise the scaled value, so this implementation does worse with high- + * quality-setting files than with low-quality ones. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" +#include "jdct.h" /* Private declarations for DCT subsystem */ + +#ifdef DCT_IFAST_SUPPORTED + + +/* + * This module is specialized to the case DCTSIZE = 8. + */ + +#if DCTSIZE != 8 + Sorry, this code only copes with 8x8 DCT blocks. /* deliberate syntax err */ +#endif + + +/* Scaling decisions are generally the same as in the LL&M algorithm; + * see jidctint.c for more details. However, we choose to descale + * (right shift) multiplication products as soon as they are formed, + * rather than carrying additional fractional bits into subsequent additions. + * This compromises accuracy slightly, but it lets us save a few shifts. + * More importantly, 16-bit arithmetic is then adequate (for 8-bit samples) + * everywhere except in the multiplications proper; this saves a good deal + * of work on 16-bit-int machines. + * + * The dequantized coefficients are not integers because the AA&N scaling + * factors have been incorporated. We represent them scaled up by PASS1_BITS, + * so that the first and second IDCT rounds have the same input scaling. + * For 8-bit JSAMPLEs, we choose IFAST_SCALE_BITS = PASS1_BITS so as to + * avoid a descaling shift; this compromises accuracy rather drastically + * for small quantization table entries, but it saves a lot of shifts. + * For 12-bit JSAMPLEs, there's no hope of using 16x16 multiplies anyway, + * so we use a much larger scaling factor to preserve accuracy. + * + * A final compromise is to represent the multiplicative constants to only + * 8 fractional bits, rather than 13. This saves some shifting work on some + * machines, and may also reduce the cost of multiplication (since there + * are fewer one-bits in the constants). + */ + +#if BITS_IN_JSAMPLE == 8 +#define CONST_BITS 8 +#define PASS1_BITS 2 +#else +#define CONST_BITS 8 +#define PASS1_BITS 1 /* lose a little precision to avoid overflow */ +#endif + +/* Some C compilers fail to reduce "FIX(constant)" at compile time, thus + * causing a lot of useless floating-point operations at run time. + * To get around this we use the following pre-calculated constants. + * If you change CONST_BITS you may want to add appropriate values. + * (With a reasonable C compiler, you can just rely on the FIX() macro...) + */ + +#if CONST_BITS == 8 +#define FIX_1_082392200 ((INT32) 277) /* FIX(1.082392200) */ +#define FIX_1_414213562 ((INT32) 362) /* FIX(1.414213562) */ +#define FIX_1_847759065 ((INT32) 473) /* FIX(1.847759065) */ +#define FIX_2_613125930 ((INT32) 669) /* FIX(2.613125930) */ +#else +#define FIX_1_082392200 FIX(1.082392200) +#define FIX_1_414213562 FIX(1.414213562) +#define FIX_1_847759065 FIX(1.847759065) +#define FIX_2_613125930 FIX(2.613125930) +#endif + + +/* We can gain a little more speed, with a further compromise in accuracy, + * by omitting the addition in a descaling shift. This yields an incorrectly + * rounded result half the time... + */ + +#ifndef USE_ACCURATE_ROUNDING +#undef DESCALE +#define DESCALE(x,n) RIGHT_SHIFT(x, n) +#endif + + +/* Multiply a DCTELEM variable by an INT32 constant, and immediately + * descale to yield a DCTELEM result. + */ + +#define MULTIPLY(var,const) ((DCTELEM) DESCALE((var) * (const), CONST_BITS)) + + +/* Dequantize a coefficient by multiplying it by the multiplier-table + * entry; produce a DCTELEM result. For 8-bit data a 16x16->16 + * multiplication will do. For 12-bit data, the multiplier table is + * declared INT32, so a 32-bit multiply will be used. + */ + +#if BITS_IN_JSAMPLE == 8 +#define DEQUANTIZE(coef,quantval) (((IFAST_MULT_TYPE) (coef)) * (quantval)) +#else +#define DEQUANTIZE(coef,quantval) \ + DESCALE((coef)*(quantval), IFAST_SCALE_BITS-PASS1_BITS) +#endif + + +/* + * Perform dequantization and inverse DCT on one block of coefficients. + * + * cK represents cos(K*pi/16). + */ + +GLOBAL(void) +jpeg_idct_ifast (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + DCTELEM tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7; + DCTELEM tmp10, tmp11, tmp12, tmp13; + DCTELEM z5, z10, z11, z12, z13; + JCOEFPTR inptr; + IFAST_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[DCTSIZE2]; /* buffers data between passes */ + SHIFT_TEMPS /* for DESCALE */ + ISHIFT_TEMPS /* for IRIGHT_SHIFT */ + + /* Pass 1: process columns from input, store into work array. */ + + inptr = coef_block; + quantptr = (IFAST_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = DCTSIZE; ctr > 0; ctr--) { + /* Due to quantization, we will usually find that many of the input + * coefficients are zero, especially the AC terms. We can exploit this + * by short-circuiting the IDCT calculation for any column in which all + * the AC terms are zero. In that case each output is equal to the + * DC coefficient (with scale factor as needed). + * With typical images and quantization tables, half or more of the + * column DCT calculations can be simplified this way. + */ + + if (inptr[DCTSIZE*1] == 0 && inptr[DCTSIZE*2] == 0 && + inptr[DCTSIZE*3] == 0 && inptr[DCTSIZE*4] == 0 && + inptr[DCTSIZE*5] == 0 && inptr[DCTSIZE*6] == 0 && + inptr[DCTSIZE*7] == 0) { + /* AC terms all zero */ + int dcval = (int) DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + + wsptr[DCTSIZE*0] = dcval; + wsptr[DCTSIZE*1] = dcval; + wsptr[DCTSIZE*2] = dcval; + wsptr[DCTSIZE*3] = dcval; + wsptr[DCTSIZE*4] = dcval; + wsptr[DCTSIZE*5] = dcval; + wsptr[DCTSIZE*6] = dcval; + wsptr[DCTSIZE*7] = dcval; + + inptr++; /* advance pointers to next column */ + quantptr++; + wsptr++; + continue; + } + + /* Even part */ + + tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + tmp1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + tmp2 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + tmp3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); + + tmp10 = tmp0 + tmp2; /* phase 3 */ + tmp11 = tmp0 - tmp2; + + tmp13 = tmp1 + tmp3; /* phases 5-3 */ + tmp12 = MULTIPLY(tmp1 - tmp3, FIX_1_414213562) - tmp13; /* 2*c4 */ + + tmp0 = tmp10 + tmp13; /* phase 2 */ + tmp3 = tmp10 - tmp13; + tmp1 = tmp11 + tmp12; + tmp2 = tmp11 - tmp12; + + /* Odd part */ + + tmp4 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + tmp5 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + tmp6 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); + tmp7 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); + + z13 = tmp6 + tmp5; /* phase 6 */ + z10 = tmp6 - tmp5; + z11 = tmp4 + tmp7; + z12 = tmp4 - tmp7; + + tmp7 = z11 + z13; /* phase 5 */ + tmp11 = MULTIPLY(z11 - z13, FIX_1_414213562); /* 2*c4 */ + + z5 = MULTIPLY(z10 + z12, FIX_1_847759065); /* 2*c2 */ + tmp10 = z5 - MULTIPLY(z12, FIX_1_082392200); /* 2*(c2-c6) */ + tmp12 = z5 - MULTIPLY(z10, FIX_2_613125930); /* 2*(c2+c6) */ + + tmp6 = tmp12 - tmp7; /* phase 2 */ + tmp5 = tmp11 - tmp6; + tmp4 = tmp10 - tmp5; + + wsptr[DCTSIZE*0] = (int) (tmp0 + tmp7); + wsptr[DCTSIZE*7] = (int) (tmp0 - tmp7); + wsptr[DCTSIZE*1] = (int) (tmp1 + tmp6); + wsptr[DCTSIZE*6] = (int) (tmp1 - tmp6); + wsptr[DCTSIZE*2] = (int) (tmp2 + tmp5); + wsptr[DCTSIZE*5] = (int) (tmp2 - tmp5); + wsptr[DCTSIZE*3] = (int) (tmp3 + tmp4); + wsptr[DCTSIZE*4] = (int) (tmp3 - tmp4); + + inptr++; /* advance pointers to next column */ + quantptr++; + wsptr++; + } + + /* Pass 2: process rows from work array, store into output array. + * Note that we must descale the results by a factor of 8 == 2**3, + * and also undo the PASS1_BITS scaling. + */ + + wsptr = workspace; + for (ctr = 0; ctr < DCTSIZE; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Add range center and fudge factor for final descale and range-limit. */ + z5 = (DCTELEM) wsptr[0] + + ((((DCTELEM) RANGE_CENTER) << (PASS1_BITS+3)) + + (1 << (PASS1_BITS+2))); + + /* Rows of zeroes can be exploited in the same way as we did with columns. + * However, the column calculation has created many nonzero AC terms, so + * the simplification applies less often (typically 5% to 10% of the time). + * On machines with very fast multiplication, it's possible that the + * test takes more time than it's worth. In that case this section + * may be commented out. + */ + +#ifndef NO_ZERO_ROW_TEST + if (wsptr[1] == 0 && wsptr[2] == 0 && wsptr[3] == 0 && wsptr[4] == 0 && + wsptr[5] == 0 && wsptr[6] == 0 && wsptr[7] == 0) { + /* AC terms all zero */ + JSAMPLE dcval = range_limit[(int) IRIGHT_SHIFT(z5, PASS1_BITS+3) + & RANGE_MASK]; + + outptr[0] = dcval; + outptr[1] = dcval; + outptr[2] = dcval; + outptr[3] = dcval; + outptr[4] = dcval; + outptr[5] = dcval; + outptr[6] = dcval; + outptr[7] = dcval; + + wsptr += DCTSIZE; /* advance pointer to next row */ + continue; + } +#endif + + /* Even part */ + + tmp10 = z5 + (DCTELEM) wsptr[4]; + tmp11 = z5 - (DCTELEM) wsptr[4]; + + tmp13 = (DCTELEM) wsptr[2] + (DCTELEM) wsptr[6]; + tmp12 = MULTIPLY((DCTELEM) wsptr[2] - (DCTELEM) wsptr[6], + FIX_1_414213562) - tmp13; /* 2*c4 */ + + tmp0 = tmp10 + tmp13; + tmp3 = tmp10 - tmp13; + tmp1 = tmp11 + tmp12; + tmp2 = tmp11 - tmp12; + + /* Odd part */ + + z13 = (DCTELEM) wsptr[5] + (DCTELEM) wsptr[3]; + z10 = (DCTELEM) wsptr[5] - (DCTELEM) wsptr[3]; + z11 = (DCTELEM) wsptr[1] + (DCTELEM) wsptr[7]; + z12 = (DCTELEM) wsptr[1] - (DCTELEM) wsptr[7]; + + tmp7 = z11 + z13; /* phase 5 */ + tmp11 = MULTIPLY(z11 - z13, FIX_1_414213562); /* 2*c4 */ + + z5 = MULTIPLY(z10 + z12, FIX_1_847759065); /* 2*c2 */ + tmp10 = z5 - MULTIPLY(z12, FIX_1_082392200); /* 2*(c2-c6) */ + tmp12 = z5 - MULTIPLY(z10, FIX_2_613125930); /* 2*(c2+c6) */ + + tmp6 = tmp12 - tmp7; /* phase 2 */ + tmp5 = tmp11 - tmp6; + tmp4 = tmp10 - tmp5; + + /* Final output stage: scale down by a factor of 8 and range-limit */ + + outptr[0] = range_limit[(int) IRIGHT_SHIFT(tmp0 + tmp7, PASS1_BITS+3) + & RANGE_MASK]; + outptr[7] = range_limit[(int) IRIGHT_SHIFT(tmp0 - tmp7, PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) IRIGHT_SHIFT(tmp1 + tmp6, PASS1_BITS+3) + & RANGE_MASK]; + outptr[6] = range_limit[(int) IRIGHT_SHIFT(tmp1 - tmp6, PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) IRIGHT_SHIFT(tmp2 + tmp5, PASS1_BITS+3) + & RANGE_MASK]; + outptr[5] = range_limit[(int) IRIGHT_SHIFT(tmp2 - tmp5, PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) IRIGHT_SHIFT(tmp3 + tmp4, PASS1_BITS+3) + & RANGE_MASK]; + outptr[4] = range_limit[(int) IRIGHT_SHIFT(tmp3 - tmp4, PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += DCTSIZE; /* advance pointer to next row */ + } +} + +#endif /* DCT_IFAST_SUPPORTED */ diff --git a/cmake/externals/libjpeg/jidctint.c b/cmake/externals/libjpeg/jidctint.c new file mode 100644 index 000000000..e30ec8c8a --- /dev/null +++ b/cmake/externals/libjpeg/jidctint.c @@ -0,0 +1,5240 @@ +/* + * jidctint.c + * + * Copyright (C) 1991-1998, Thomas G. Lane. + * Modification developed 2002-2018 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains a slow-but-accurate integer implementation of the + * inverse DCT (Discrete Cosine Transform). In the IJG code, this routine + * must also perform dequantization of the input coefficients. + * + * A 2-D IDCT can be done by 1-D IDCT on each column followed by 1-D IDCT + * on each row (or vice versa, but it's more convenient to emit a row at + * a time). Direct algorithms are also available, but they are much more + * complex and seem not to be any faster when reduced to code. + * + * This implementation is based on an algorithm described in + * C. Loeffler, A. Ligtenberg and G. Moschytz, "Practical Fast 1-D DCT + * Algorithms with 11 Multiplications", Proc. Int'l. Conf. on Acoustics, + * Speech, and Signal Processing 1989 (ICASSP '89), pp. 988-991. + * The primary algorithm described there uses 11 multiplies and 29 adds. + * We use their alternate method with 12 multiplies and 32 adds. + * The advantage of this method is that no data path contains more than one + * multiplication; this allows a very simple and accurate implementation in + * scaled fixed-point arithmetic, with a minimal number of shifts. + * + * We also provide IDCT routines with various output sample block sizes for + * direct resolution reduction or enlargement and for direct resolving the + * common 2x1 and 1x2 subsampling cases without additional resampling: NxN + * (N=1...16), 2NxN, and Nx2N (N=1...8) pixels for one 8x8 input DCT block. + * + * For N<8 we simply take the corresponding low-frequency coefficients of + * the 8x8 input DCT block and apply an NxN point IDCT on the sub-block + * to yield the downscaled outputs. + * This can be seen as direct low-pass downsampling from the DCT domain + * point of view rather than the usual spatial domain point of view, + * yielding significant computational savings and results at least + * as good as common bilinear (averaging) spatial downsampling. + * + * For N>8 we apply a partial NxN IDCT on the 8 input coefficients as + * lower frequencies and higher frequencies assumed to be zero. + * It turns out that the computational effort is similar to the 8x8 IDCT + * regarding the output size. + * Furthermore, the scaling and descaling is the same for all IDCT sizes. + * + * CAUTION: We rely on the FIX() macro except for the N=1,2,4,8 cases + * since there would be too many additional constants to pre-calculate. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" +#include "jdct.h" /* Private declarations for DCT subsystem */ + +#ifdef DCT_ISLOW_SUPPORTED + + +/* + * This module is specialized to the case DCTSIZE = 8. + */ + +#if DCTSIZE != 8 + Sorry, this code only copes with 8x8 DCT blocks. /* deliberate syntax err */ +#endif + + +/* + * The poop on this scaling stuff is as follows: + * + * Each 1-D IDCT step produces outputs which are a factor of sqrt(N) + * larger than the true IDCT outputs. The final outputs are therefore + * a factor of N larger than desired; since N=8 this can be cured by + * a simple right shift at the end of the algorithm. The advantage of + * this arrangement is that we save two multiplications per 1-D IDCT, + * because the y0 and y4 inputs need not be divided by sqrt(N). + * + * We have to do addition and subtraction of the integer inputs, which + * is no problem, and multiplication by fractional constants, which is + * a problem to do in integer arithmetic. We multiply all the constants + * by CONST_SCALE and convert them to integer constants (thus retaining + * CONST_BITS bits of precision in the constants). After doing a + * multiplication we have to divide the product by CONST_SCALE, with proper + * rounding, to produce the correct output. This division can be done + * cheaply as a right shift of CONST_BITS bits. We postpone shifting + * as long as possible so that partial sums can be added together with + * full fractional precision. + * + * The outputs of the first pass are scaled up by PASS1_BITS bits so that + * they are represented to better-than-integral precision. These outputs + * require BITS_IN_JSAMPLE + PASS1_BITS + 3 bits; this fits in a 16-bit word + * with the recommended scaling. (To scale up 12-bit sample data further, an + * intermediate INT32 array would be needed.) + * + * To avoid overflow of the 32-bit intermediate results in pass 2, we must + * have BITS_IN_JSAMPLE + CONST_BITS + PASS1_BITS <= 26. Error analysis + * shows that the values given below are the most effective. + */ + +#if BITS_IN_JSAMPLE == 8 +#define CONST_BITS 13 +#define PASS1_BITS 2 +#else +#define CONST_BITS 13 +#define PASS1_BITS 1 /* lose a little precision to avoid overflow */ +#endif + +/* Some C compilers fail to reduce "FIX(constant)" at compile time, thus + * causing a lot of useless floating-point operations at run time. + * To get around this we use the following pre-calculated constants. + * If you change CONST_BITS you may want to add appropriate values. + * (With a reasonable C compiler, you can just rely on the FIX() macro...) + */ + +#if CONST_BITS == 13 +#define FIX_0_298631336 ((INT32) 2446) /* FIX(0.298631336) */ +#define FIX_0_390180644 ((INT32) 3196) /* FIX(0.390180644) */ +#define FIX_0_541196100 ((INT32) 4433) /* FIX(0.541196100) */ +#define FIX_0_765366865 ((INT32) 6270) /* FIX(0.765366865) */ +#define FIX_0_899976223 ((INT32) 7373) /* FIX(0.899976223) */ +#define FIX_1_175875602 ((INT32) 9633) /* FIX(1.175875602) */ +#define FIX_1_501321110 ((INT32) 12299) /* FIX(1.501321110) */ +#define FIX_1_847759065 ((INT32) 15137) /* FIX(1.847759065) */ +#define FIX_1_961570560 ((INT32) 16069) /* FIX(1.961570560) */ +#define FIX_2_053119869 ((INT32) 16819) /* FIX(2.053119869) */ +#define FIX_2_562915447 ((INT32) 20995) /* FIX(2.562915447) */ +#define FIX_3_072711026 ((INT32) 25172) /* FIX(3.072711026) */ +#else +#define FIX_0_298631336 FIX(0.298631336) +#define FIX_0_390180644 FIX(0.390180644) +#define FIX_0_541196100 FIX(0.541196100) +#define FIX_0_765366865 FIX(0.765366865) +#define FIX_0_899976223 FIX(0.899976223) +#define FIX_1_175875602 FIX(1.175875602) +#define FIX_1_501321110 FIX(1.501321110) +#define FIX_1_847759065 FIX(1.847759065) +#define FIX_1_961570560 FIX(1.961570560) +#define FIX_2_053119869 FIX(2.053119869) +#define FIX_2_562915447 FIX(2.562915447) +#define FIX_3_072711026 FIX(3.072711026) +#endif + + +/* Multiply an INT32 variable by an INT32 constant to yield an INT32 result. + * For 8-bit samples with the recommended scaling, all the variable + * and constant values involved are no more than 16 bits wide, so a + * 16x16->32 bit multiply can be used instead of a full 32x32 multiply. + * For 12-bit samples, a full 32-bit multiplication will be needed. + */ + +#if BITS_IN_JSAMPLE == 8 +#define MULTIPLY(var,const) MULTIPLY16C16(var,const) +#else +#define MULTIPLY(var,const) ((var) * (const)) +#endif + + +/* Dequantize a coefficient by multiplying it by the multiplier-table + * entry; produce an int result. In this module, both inputs and result + * are 16 bits or less, so either int or short multiply will work. + */ + +#define DEQUANTIZE(coef,quantval) (((ISLOW_MULT_TYPE) (coef)) * (quantval)) + + +/* + * Perform dequantization and inverse DCT on one block of coefficients. + * + * Optimized algorithm with 12 multiplications in the 1-D kernel. + * cK represents sqrt(2) * cos(K*pi/16). + */ + +GLOBAL(void) +jpeg_idct_islow (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3; + INT32 tmp10, tmp11, tmp12, tmp13; + INT32 z1, z2, z3; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[DCTSIZE2]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. + * Note results are scaled up by sqrt(8) compared to a true IDCT; + * furthermore, we scale the results by 2**PASS1_BITS. + */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = DCTSIZE; ctr > 0; ctr--) { + /* Due to quantization, we will usually find that many of the input + * coefficients are zero, especially the AC terms. We can exploit this + * by short-circuiting the IDCT calculation for any column in which all + * the AC terms are zero. In that case each output is equal to the + * DC coefficient (with scale factor as needed). + * With typical images and quantization tables, half or more of the + * column DCT calculations can be simplified this way. + */ + + if (inptr[DCTSIZE*1] == 0 && inptr[DCTSIZE*2] == 0 && + inptr[DCTSIZE*3] == 0 && inptr[DCTSIZE*4] == 0 && + inptr[DCTSIZE*5] == 0 && inptr[DCTSIZE*6] == 0 && + inptr[DCTSIZE*7] == 0) { + /* AC terms all zero */ + int dcval = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]) << PASS1_BITS; + + wsptr[DCTSIZE*0] = dcval; + wsptr[DCTSIZE*1] = dcval; + wsptr[DCTSIZE*2] = dcval; + wsptr[DCTSIZE*3] = dcval; + wsptr[DCTSIZE*4] = dcval; + wsptr[DCTSIZE*5] = dcval; + wsptr[DCTSIZE*6] = dcval; + wsptr[DCTSIZE*7] = dcval; + + inptr++; /* advance pointers to next column */ + quantptr++; + wsptr++; + continue; + } + + /* Even part: reverse the even part of the forward DCT. + * The rotator is c(-6). + */ + + z2 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + z3 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + z2 <<= CONST_BITS; + z3 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + z2 += ONE << (CONST_BITS-PASS1_BITS-1); + + tmp0 = z2 + z3; + tmp1 = z2 - z3; + + z2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + z3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); + + z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ + tmp2 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ + tmp3 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ + + tmp10 = tmp0 + tmp2; + tmp13 = tmp0 - tmp2; + tmp11 = tmp1 + tmp3; + tmp12 = tmp1 - tmp3; + + /* Odd part per figure 8; the matrix is unitary and hence its + * transpose is its inverse. i0..i3 are y7,y5,y3,y1 respectively. + */ + + tmp0 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); + tmp1 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); + tmp2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + tmp3 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + + z2 = tmp0 + tmp2; + z3 = tmp1 + tmp3; + + z1 = MULTIPLY(z2 + z3, FIX_1_175875602); /* c3 */ + z2 = MULTIPLY(z2, - FIX_1_961570560); /* -c3-c5 */ + z3 = MULTIPLY(z3, - FIX_0_390180644); /* -c3+c5 */ + z2 += z1; + z3 += z1; + + z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ + tmp0 = MULTIPLY(tmp0, FIX_0_298631336); /* -c1+c3+c5-c7 */ + tmp3 = MULTIPLY(tmp3, FIX_1_501321110); /* c1+c3-c5-c7 */ + tmp0 += z1 + z2; + tmp3 += z1 + z3; + + z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ + tmp1 = MULTIPLY(tmp1, FIX_2_053119869); /* c1+c3-c5+c7 */ + tmp2 = MULTIPLY(tmp2, FIX_3_072711026); /* c1+c3+c5-c7 */ + tmp1 += z1 + z3; + tmp2 += z1 + z2; + + /* Final output stage: inputs are tmp10..tmp13, tmp0..tmp3 */ + + wsptr[DCTSIZE*0] = (int) RIGHT_SHIFT(tmp10 + tmp3, CONST_BITS-PASS1_BITS); + wsptr[DCTSIZE*7] = (int) RIGHT_SHIFT(tmp10 - tmp3, CONST_BITS-PASS1_BITS); + wsptr[DCTSIZE*1] = (int) RIGHT_SHIFT(tmp11 + tmp2, CONST_BITS-PASS1_BITS); + wsptr[DCTSIZE*6] = (int) RIGHT_SHIFT(tmp11 - tmp2, CONST_BITS-PASS1_BITS); + wsptr[DCTSIZE*2] = (int) RIGHT_SHIFT(tmp12 + tmp1, CONST_BITS-PASS1_BITS); + wsptr[DCTSIZE*5] = (int) RIGHT_SHIFT(tmp12 - tmp1, CONST_BITS-PASS1_BITS); + wsptr[DCTSIZE*3] = (int) RIGHT_SHIFT(tmp13 + tmp0, CONST_BITS-PASS1_BITS); + wsptr[DCTSIZE*4] = (int) RIGHT_SHIFT(tmp13 - tmp0, CONST_BITS-PASS1_BITS); + + inptr++; /* advance pointers to next column */ + quantptr++; + wsptr++; + } + + /* Pass 2: process rows from work array, store into output array. + * Note that we must descale the results by a factor of 8 == 2**3, + * and also undo the PASS1_BITS scaling. + */ + + wsptr = workspace; + for (ctr = 0; ctr < DCTSIZE; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Add range center and fudge factor for final descale and range-limit. */ + z2 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + + /* Rows of zeroes can be exploited in the same way as we did with columns. + * However, the column calculation has created many nonzero AC terms, so + * the simplification applies less often (typically 5% to 10% of the time). + * On machines with very fast multiplication, it's possible that the + * test takes more time than it's worth. In that case this section + * may be commented out. + */ + +#ifndef NO_ZERO_ROW_TEST + if (wsptr[1] == 0 && wsptr[2] == 0 && wsptr[3] == 0 && wsptr[4] == 0 && + wsptr[5] == 0 && wsptr[6] == 0 && wsptr[7] == 0) { + /* AC terms all zero */ + JSAMPLE dcval = range_limit[(int) RIGHT_SHIFT(z2, PASS1_BITS+3) + & RANGE_MASK]; + + outptr[0] = dcval; + outptr[1] = dcval; + outptr[2] = dcval; + outptr[3] = dcval; + outptr[4] = dcval; + outptr[5] = dcval; + outptr[6] = dcval; + outptr[7] = dcval; + + wsptr += DCTSIZE; /* advance pointer to next row */ + continue; + } +#endif + + /* Even part: reverse the even part of the forward DCT. + * The rotator is c(-6). + */ + + z3 = (INT32) wsptr[4]; + + tmp0 = (z2 + z3) << CONST_BITS; + tmp1 = (z2 - z3) << CONST_BITS; + + z2 = (INT32) wsptr[2]; + z3 = (INT32) wsptr[6]; + + z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ + tmp2 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ + tmp3 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ + + tmp10 = tmp0 + tmp2; + tmp13 = tmp0 - tmp2; + tmp11 = tmp1 + tmp3; + tmp12 = tmp1 - tmp3; + + /* Odd part per figure 8; the matrix is unitary and hence its + * transpose is its inverse. i0..i3 are y7,y5,y3,y1 respectively. + */ + + tmp0 = (INT32) wsptr[7]; + tmp1 = (INT32) wsptr[5]; + tmp2 = (INT32) wsptr[3]; + tmp3 = (INT32) wsptr[1]; + + z2 = tmp0 + tmp2; + z3 = tmp1 + tmp3; + + z1 = MULTIPLY(z2 + z3, FIX_1_175875602); /* c3 */ + z2 = MULTIPLY(z2, - FIX_1_961570560); /* -c3-c5 */ + z3 = MULTIPLY(z3, - FIX_0_390180644); /* -c3+c5 */ + z2 += z1; + z3 += z1; + + z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ + tmp0 = MULTIPLY(tmp0, FIX_0_298631336); /* -c1+c3+c5-c7 */ + tmp3 = MULTIPLY(tmp3, FIX_1_501321110); /* c1+c3-c5-c7 */ + tmp0 += z1 + z2; + tmp3 += z1 + z3; + + z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ + tmp1 = MULTIPLY(tmp1, FIX_2_053119869); /* c1+c3-c5+c7 */ + tmp2 = MULTIPLY(tmp2, FIX_3_072711026); /* c1+c3+c5-c7 */ + tmp1 += z1 + z3; + tmp2 += z1 + z2; + + /* Final output stage: inputs are tmp10..tmp13, tmp0..tmp3 */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp3, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp3, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp11 + tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp11 - tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12 + tmp1, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp12 - tmp1, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp13 + tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp13 - tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += DCTSIZE; /* advance pointer to next row */ + } +} + +#ifdef IDCT_SCALING_SUPPORTED + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a reduced-size 7x7 output block. + * + * Optimized algorithm with 12 multiplications in the 1-D kernel. + * cK represents sqrt(2) * cos(K*pi/14). + */ + +GLOBAL(void) +jpeg_idct_7x7 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp0, tmp1, tmp2, tmp10, tmp11, tmp12, tmp13; + INT32 z1, z2, z3; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[7*7]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 7; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + tmp13 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + tmp13 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + tmp13 += ONE << (CONST_BITS-PASS1_BITS-1); + + z1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + z2 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + z3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); + + tmp10 = MULTIPLY(z2 - z3, FIX(0.881747734)); /* c4 */ + tmp12 = MULTIPLY(z1 - z2, FIX(0.314692123)); /* c6 */ + tmp11 = tmp10 + tmp12 + tmp13 - MULTIPLY(z2, FIX(1.841218003)); /* c2+c4-c6 */ + tmp0 = z1 + z3; + z2 -= tmp0; + tmp0 = MULTIPLY(tmp0, FIX(1.274162392)) + tmp13; /* c2 */ + tmp10 += tmp0 - MULTIPLY(z3, FIX(0.077722536)); /* c2-c4-c6 */ + tmp12 += tmp0 - MULTIPLY(z1, FIX(2.470602249)); /* c2+c4+c6 */ + tmp13 += MULTIPLY(z2, FIX(1.414213562)); /* c0 */ + + /* Odd part */ + + z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); + + tmp1 = MULTIPLY(z1 + z2, FIX(0.935414347)); /* (c3+c1-c5)/2 */ + tmp2 = MULTIPLY(z1 - z2, FIX(0.170262339)); /* (c3+c5-c1)/2 */ + tmp0 = tmp1 - tmp2; + tmp1 += tmp2; + tmp2 = MULTIPLY(z2 + z3, - FIX(1.378756276)); /* -c1 */ + tmp1 += tmp2; + z2 = MULTIPLY(z1 + z3, FIX(0.613604268)); /* c5 */ + tmp0 += z2; + tmp2 += z2 + MULTIPLY(z3, FIX(1.870828693)); /* c3+c1-c5 */ + + /* Final output stage */ + + wsptr[7*0] = (int) RIGHT_SHIFT(tmp10 + tmp0, CONST_BITS-PASS1_BITS); + wsptr[7*6] = (int) RIGHT_SHIFT(tmp10 - tmp0, CONST_BITS-PASS1_BITS); + wsptr[7*1] = (int) RIGHT_SHIFT(tmp11 + tmp1, CONST_BITS-PASS1_BITS); + wsptr[7*5] = (int) RIGHT_SHIFT(tmp11 - tmp1, CONST_BITS-PASS1_BITS); + wsptr[7*2] = (int) RIGHT_SHIFT(tmp12 + tmp2, CONST_BITS-PASS1_BITS); + wsptr[7*4] = (int) RIGHT_SHIFT(tmp12 - tmp2, CONST_BITS-PASS1_BITS); + wsptr[7*3] = (int) RIGHT_SHIFT(tmp13, CONST_BITS-PASS1_BITS); + } + + /* Pass 2: process 7 rows from work array, store into output array. */ + + wsptr = workspace; + for (ctr = 0; ctr < 7; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + tmp13 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + tmp13 <<= CONST_BITS; + + z1 = (INT32) wsptr[2]; + z2 = (INT32) wsptr[4]; + z3 = (INT32) wsptr[6]; + + tmp10 = MULTIPLY(z2 - z3, FIX(0.881747734)); /* c4 */ + tmp12 = MULTIPLY(z1 - z2, FIX(0.314692123)); /* c6 */ + tmp11 = tmp10 + tmp12 + tmp13 - MULTIPLY(z2, FIX(1.841218003)); /* c2+c4-c6 */ + tmp0 = z1 + z3; + z2 -= tmp0; + tmp0 = MULTIPLY(tmp0, FIX(1.274162392)) + tmp13; /* c2 */ + tmp10 += tmp0 - MULTIPLY(z3, FIX(0.077722536)); /* c2-c4-c6 */ + tmp12 += tmp0 - MULTIPLY(z1, FIX(2.470602249)); /* c2+c4+c6 */ + tmp13 += MULTIPLY(z2, FIX(1.414213562)); /* c0 */ + + /* Odd part */ + + z1 = (INT32) wsptr[1]; + z2 = (INT32) wsptr[3]; + z3 = (INT32) wsptr[5]; + + tmp1 = MULTIPLY(z1 + z2, FIX(0.935414347)); /* (c3+c1-c5)/2 */ + tmp2 = MULTIPLY(z1 - z2, FIX(0.170262339)); /* (c3+c5-c1)/2 */ + tmp0 = tmp1 - tmp2; + tmp1 += tmp2; + tmp2 = MULTIPLY(z2 + z3, - FIX(1.378756276)); /* -c1 */ + tmp1 += tmp2; + z2 = MULTIPLY(z1 + z3, FIX(0.613604268)); /* c5 */ + tmp0 += z2; + tmp2 += z2 + MULTIPLY(z3, FIX(1.870828693)); /* c3+c1-c5 */ + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp11 + tmp1, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp11 - tmp1, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12 + tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp12 - tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += 7; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a reduced-size 6x6 output block. + * + * Optimized algorithm with 3 multiplications in the 1-D kernel. + * cK represents sqrt(2) * cos(K*pi/12). + */ + +GLOBAL(void) +jpeg_idct_6x6 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp0, tmp1, tmp2, tmp10, tmp11, tmp12; + INT32 z1, z2, z3; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[6*6]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 6; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + tmp0 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + tmp0 += ONE << (CONST_BITS-PASS1_BITS-1); + tmp2 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + tmp10 = MULTIPLY(tmp2, FIX(0.707106781)); /* c4 */ + tmp1 = tmp0 + tmp10; + tmp11 = RIGHT_SHIFT(tmp0 - tmp10 - tmp10, CONST_BITS-PASS1_BITS); + tmp10 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + tmp0 = MULTIPLY(tmp10, FIX(1.224744871)); /* c2 */ + tmp10 = tmp1 + tmp0; + tmp12 = tmp1 - tmp0; + + /* Odd part */ + + z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); + tmp1 = MULTIPLY(z1 + z3, FIX(0.366025404)); /* c5 */ + tmp0 = tmp1 + ((z1 + z2) << CONST_BITS); + tmp2 = tmp1 + ((z3 - z2) << CONST_BITS); + tmp1 = (z1 - z2 - z3) << PASS1_BITS; + + /* Final output stage */ + + wsptr[6*0] = (int) RIGHT_SHIFT(tmp10 + tmp0, CONST_BITS-PASS1_BITS); + wsptr[6*5] = (int) RIGHT_SHIFT(tmp10 - tmp0, CONST_BITS-PASS1_BITS); + wsptr[6*1] = (int) (tmp11 + tmp1); + wsptr[6*4] = (int) (tmp11 - tmp1); + wsptr[6*2] = (int) RIGHT_SHIFT(tmp12 + tmp2, CONST_BITS-PASS1_BITS); + wsptr[6*3] = (int) RIGHT_SHIFT(tmp12 - tmp2, CONST_BITS-PASS1_BITS); + } + + /* Pass 2: process 6 rows from work array, store into output array. */ + + wsptr = workspace; + for (ctr = 0; ctr < 6; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + tmp0 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + tmp0 <<= CONST_BITS; + tmp2 = (INT32) wsptr[4]; + tmp10 = MULTIPLY(tmp2, FIX(0.707106781)); /* c4 */ + tmp1 = tmp0 + tmp10; + tmp11 = tmp0 - tmp10 - tmp10; + tmp10 = (INT32) wsptr[2]; + tmp0 = MULTIPLY(tmp10, FIX(1.224744871)); /* c2 */ + tmp10 = tmp1 + tmp0; + tmp12 = tmp1 - tmp0; + + /* Odd part */ + + z1 = (INT32) wsptr[1]; + z2 = (INT32) wsptr[3]; + z3 = (INT32) wsptr[5]; + tmp1 = MULTIPLY(z1 + z3, FIX(0.366025404)); /* c5 */ + tmp0 = tmp1 + ((z1 + z2) << CONST_BITS); + tmp2 = tmp1 + ((z3 - z2) << CONST_BITS); + tmp1 = (z1 - z2 - z3) << CONST_BITS; + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp11 + tmp1, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp11 - tmp1, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12 + tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp12 - tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += 6; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a reduced-size 5x5 output block. + * + * Optimized algorithm with 5 multiplications in the 1-D kernel. + * cK represents sqrt(2) * cos(K*pi/10). + */ + +GLOBAL(void) +jpeg_idct_5x5 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp0, tmp1, tmp10, tmp11, tmp12; + INT32 z1, z2, z3; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[5*5]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 5; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + tmp12 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + tmp12 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + tmp12 += ONE << (CONST_BITS-PASS1_BITS-1); + tmp0 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + tmp1 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + z1 = MULTIPLY(tmp0 + tmp1, FIX(0.790569415)); /* (c2+c4)/2 */ + z2 = MULTIPLY(tmp0 - tmp1, FIX(0.353553391)); /* (c2-c4)/2 */ + z3 = tmp12 + z2; + tmp10 = z3 + z1; + tmp11 = z3 - z1; + tmp12 -= z2 << 2; + + /* Odd part */ + + z2 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + z3 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + + z1 = MULTIPLY(z2 + z3, FIX(0.831253876)); /* c3 */ + tmp0 = z1 + MULTIPLY(z2, FIX(0.513743148)); /* c1-c3 */ + tmp1 = z1 - MULTIPLY(z3, FIX(2.176250899)); /* c1+c3 */ + + /* Final output stage */ + + wsptr[5*0] = (int) RIGHT_SHIFT(tmp10 + tmp0, CONST_BITS-PASS1_BITS); + wsptr[5*4] = (int) RIGHT_SHIFT(tmp10 - tmp0, CONST_BITS-PASS1_BITS); + wsptr[5*1] = (int) RIGHT_SHIFT(tmp11 + tmp1, CONST_BITS-PASS1_BITS); + wsptr[5*3] = (int) RIGHT_SHIFT(tmp11 - tmp1, CONST_BITS-PASS1_BITS); + wsptr[5*2] = (int) RIGHT_SHIFT(tmp12, CONST_BITS-PASS1_BITS); + } + + /* Pass 2: process 5 rows from work array, store into output array. */ + + wsptr = workspace; + for (ctr = 0; ctr < 5; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + tmp12 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + tmp12 <<= CONST_BITS; + tmp0 = (INT32) wsptr[2]; + tmp1 = (INT32) wsptr[4]; + z1 = MULTIPLY(tmp0 + tmp1, FIX(0.790569415)); /* (c2+c4)/2 */ + z2 = MULTIPLY(tmp0 - tmp1, FIX(0.353553391)); /* (c2-c4)/2 */ + z3 = tmp12 + z2; + tmp10 = z3 + z1; + tmp11 = z3 - z1; + tmp12 -= z2 << 2; + + /* Odd part */ + + z2 = (INT32) wsptr[1]; + z3 = (INT32) wsptr[3]; + + z1 = MULTIPLY(z2 + z3, FIX(0.831253876)); /* c3 */ + tmp0 = z1 + MULTIPLY(z2, FIX(0.513743148)); /* c1-c3 */ + tmp1 = z1 - MULTIPLY(z3, FIX(2.176250899)); /* c1+c3 */ + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp11 + tmp1, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp11 - tmp1, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += 5; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a reduced-size 4x4 output block. + * + * Optimized algorithm with 3 multiplications in the 1-D kernel. + * cK represents sqrt(2) * cos(K*pi/16) [refers to 8-point IDCT]. + */ + +GLOBAL(void) +jpeg_idct_4x4 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp0, tmp2, tmp10, tmp12; + INT32 z1, z2, z3; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[4*4]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 4; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + tmp2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + + tmp10 = (tmp0 + tmp2) << PASS1_BITS; + tmp12 = (tmp0 - tmp2) << PASS1_BITS; + + /* Odd part */ + /* Same rotation as in the even part of the 8x8 LL&M IDCT */ + + z2 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + z3 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + + z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ + /* Add fudge factor here for final descale. */ + z1 += ONE << (CONST_BITS-PASS1_BITS-1); + tmp0 = RIGHT_SHIFT(z1 + MULTIPLY(z2, FIX_0_765366865), /* c2-c6 */ + CONST_BITS-PASS1_BITS); + tmp2 = RIGHT_SHIFT(z1 - MULTIPLY(z3, FIX_1_847759065), /* c2+c6 */ + CONST_BITS-PASS1_BITS); + + /* Final output stage */ + + wsptr[4*0] = (int) (tmp10 + tmp0); + wsptr[4*3] = (int) (tmp10 - tmp0); + wsptr[4*1] = (int) (tmp12 + tmp2); + wsptr[4*2] = (int) (tmp12 - tmp2); + } + + /* Pass 2: process 4 rows from work array, store into output array. */ + + wsptr = workspace; + for (ctr = 0; ctr < 4; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + tmp0 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + tmp2 = (INT32) wsptr[2]; + + tmp10 = (tmp0 + tmp2) << CONST_BITS; + tmp12 = (tmp0 - tmp2) << CONST_BITS; + + /* Odd part */ + /* Same rotation as in the even part of the 8x8 LL&M IDCT */ + + z2 = (INT32) wsptr[1]; + z3 = (INT32) wsptr[3]; + + z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ + tmp0 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ + tmp2 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp12 + tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12 - tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += 4; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a reduced-size 3x3 output block. + * + * Optimized algorithm with 2 multiplications in the 1-D kernel. + * cK represents sqrt(2) * cos(K*pi/6). + */ + +GLOBAL(void) +jpeg_idct_3x3 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp0, tmp2, tmp10, tmp12; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[3*3]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 3; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + tmp0 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + tmp0 += ONE << (CONST_BITS-PASS1_BITS-1); + tmp2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + tmp12 = MULTIPLY(tmp2, FIX(0.707106781)); /* c2 */ + tmp10 = tmp0 + tmp12; + tmp2 = tmp0 - tmp12 - tmp12; + + /* Odd part */ + + tmp12 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + tmp0 = MULTIPLY(tmp12, FIX(1.224744871)); /* c1 */ + + /* Final output stage */ + + wsptr[3*0] = (int) RIGHT_SHIFT(tmp10 + tmp0, CONST_BITS-PASS1_BITS); + wsptr[3*2] = (int) RIGHT_SHIFT(tmp10 - tmp0, CONST_BITS-PASS1_BITS); + wsptr[3*1] = (int) RIGHT_SHIFT(tmp2, CONST_BITS-PASS1_BITS); + } + + /* Pass 2: process 3 rows from work array, store into output array. */ + + wsptr = workspace; + for (ctr = 0; ctr < 3; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + tmp0 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + tmp0 <<= CONST_BITS; + tmp2 = (INT32) wsptr[2]; + tmp12 = MULTIPLY(tmp2, FIX(0.707106781)); /* c2 */ + tmp10 = tmp0 + tmp12; + tmp2 = tmp0 - tmp12 - tmp12; + + /* Odd part */ + + tmp12 = (INT32) wsptr[1]; + tmp0 = MULTIPLY(tmp12, FIX(1.224744871)); /* c1 */ + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += 3; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a reduced-size 2x2 output block. + * + * Multiplication-less algorithm. + */ + +GLOBAL(void) +jpeg_idct_2x2 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + DCTELEM tmp0, tmp1, tmp2, tmp3, tmp4, tmp5; + ISLOW_MULT_TYPE * quantptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + ISHIFT_TEMPS + + /* Pass 1: process columns from input. */ + + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + + /* Column 0 */ + tmp4 = DEQUANTIZE(coef_block[DCTSIZE*0], quantptr[DCTSIZE*0]); + tmp5 = DEQUANTIZE(coef_block[DCTSIZE*1], quantptr[DCTSIZE*1]); + /* Add range center and fudge factor for final descale and range-limit. */ + tmp4 += (((DCTELEM) RANGE_CENTER) << 3) + (1 << 2); + + tmp0 = tmp4 + tmp5; + tmp2 = tmp4 - tmp5; + + /* Column 1 */ + tmp4 = DEQUANTIZE(coef_block[DCTSIZE*0+1], quantptr[DCTSIZE*0+1]); + tmp5 = DEQUANTIZE(coef_block[DCTSIZE*1+1], quantptr[DCTSIZE*1+1]); + + tmp1 = tmp4 + tmp5; + tmp3 = tmp4 - tmp5; + + /* Pass 2: process 2 rows, store into output array. */ + + /* Row 0 */ + outptr = output_buf[0] + output_col; + + outptr[0] = range_limit[(int) IRIGHT_SHIFT(tmp0 + tmp1, 3) & RANGE_MASK]; + outptr[1] = range_limit[(int) IRIGHT_SHIFT(tmp0 - tmp1, 3) & RANGE_MASK]; + + /* Row 1 */ + outptr = output_buf[1] + output_col; + + outptr[0] = range_limit[(int) IRIGHT_SHIFT(tmp2 + tmp3, 3) & RANGE_MASK]; + outptr[1] = range_limit[(int) IRIGHT_SHIFT(tmp2 - tmp3, 3) & RANGE_MASK]; +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a reduced-size 1x1 output block. + * + * We hardly need an inverse DCT routine for this: just take the + * average pixel value, which is one-eighth of the DC coefficient. + */ + +GLOBAL(void) +jpeg_idct_1x1 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + DCTELEM dcval; + ISLOW_MULT_TYPE * quantptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + ISHIFT_TEMPS + + /* 1x1 is trivial: just take the DC coefficient divided by 8. */ + + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + + dcval = DEQUANTIZE(coef_block[0], quantptr[0]); + /* Add range center and fudge factor for descale and range-limit. */ + dcval += (((DCTELEM) RANGE_CENTER) << 3) + (1 << 2); + + output_buf[0][output_col] = + range_limit[(int) IRIGHT_SHIFT(dcval, 3) & RANGE_MASK]; +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a 9x9 output block. + * + * Optimized algorithm with 10 multiplications in the 1-D kernel. + * cK represents sqrt(2) * cos(K*pi/18). + */ + +GLOBAL(void) +jpeg_idct_9x9 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3, tmp10, tmp11, tmp12, tmp13, tmp14; + INT32 z1, z2, z3, z4; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[8*9]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + tmp0 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + tmp0 += ONE << (CONST_BITS-PASS1_BITS-1); + + z1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + z2 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + z3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); + + tmp3 = MULTIPLY(z3, FIX(0.707106781)); /* c6 */ + tmp1 = tmp0 + tmp3; + tmp2 = tmp0 - tmp3 - tmp3; + + tmp0 = MULTIPLY(z1 - z2, FIX(0.707106781)); /* c6 */ + tmp11 = tmp2 + tmp0; + tmp14 = tmp2 - tmp0 - tmp0; + + tmp0 = MULTIPLY(z1 + z2, FIX(1.328926049)); /* c2 */ + tmp2 = MULTIPLY(z1, FIX(1.083350441)); /* c4 */ + tmp3 = MULTIPLY(z2, FIX(0.245575608)); /* c8 */ + + tmp10 = tmp1 + tmp0 - tmp3; + tmp12 = tmp1 - tmp0 + tmp2; + tmp13 = tmp1 - tmp2 + tmp3; + + /* Odd part */ + + z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); + z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); + + z2 = MULTIPLY(z2, - FIX(1.224744871)); /* -c3 */ + + tmp2 = MULTIPLY(z1 + z3, FIX(0.909038955)); /* c5 */ + tmp3 = MULTIPLY(z1 + z4, FIX(0.483689525)); /* c7 */ + tmp0 = tmp2 + tmp3 - z2; + tmp1 = MULTIPLY(z3 - z4, FIX(1.392728481)); /* c1 */ + tmp2 += z2 - tmp1; + tmp3 += z2 + tmp1; + tmp1 = MULTIPLY(z1 - z3 - z4, FIX(1.224744871)); /* c3 */ + + /* Final output stage */ + + wsptr[8*0] = (int) RIGHT_SHIFT(tmp10 + tmp0, CONST_BITS-PASS1_BITS); + wsptr[8*8] = (int) RIGHT_SHIFT(tmp10 - tmp0, CONST_BITS-PASS1_BITS); + wsptr[8*1] = (int) RIGHT_SHIFT(tmp11 + tmp1, CONST_BITS-PASS1_BITS); + wsptr[8*7] = (int) RIGHT_SHIFT(tmp11 - tmp1, CONST_BITS-PASS1_BITS); + wsptr[8*2] = (int) RIGHT_SHIFT(tmp12 + tmp2, CONST_BITS-PASS1_BITS); + wsptr[8*6] = (int) RIGHT_SHIFT(tmp12 - tmp2, CONST_BITS-PASS1_BITS); + wsptr[8*3] = (int) RIGHT_SHIFT(tmp13 + tmp3, CONST_BITS-PASS1_BITS); + wsptr[8*5] = (int) RIGHT_SHIFT(tmp13 - tmp3, CONST_BITS-PASS1_BITS); + wsptr[8*4] = (int) RIGHT_SHIFT(tmp14, CONST_BITS-PASS1_BITS); + } + + /* Pass 2: process 9 rows from work array, store into output array. */ + + wsptr = workspace; + for (ctr = 0; ctr < 9; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + tmp0 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + tmp0 <<= CONST_BITS; + + z1 = (INT32) wsptr[2]; + z2 = (INT32) wsptr[4]; + z3 = (INT32) wsptr[6]; + + tmp3 = MULTIPLY(z3, FIX(0.707106781)); /* c6 */ + tmp1 = tmp0 + tmp3; + tmp2 = tmp0 - tmp3 - tmp3; + + tmp0 = MULTIPLY(z1 - z2, FIX(0.707106781)); /* c6 */ + tmp11 = tmp2 + tmp0; + tmp14 = tmp2 - tmp0 - tmp0; + + tmp0 = MULTIPLY(z1 + z2, FIX(1.328926049)); /* c2 */ + tmp2 = MULTIPLY(z1, FIX(1.083350441)); /* c4 */ + tmp3 = MULTIPLY(z2, FIX(0.245575608)); /* c8 */ + + tmp10 = tmp1 + tmp0 - tmp3; + tmp12 = tmp1 - tmp0 + tmp2; + tmp13 = tmp1 - tmp2 + tmp3; + + /* Odd part */ + + z1 = (INT32) wsptr[1]; + z2 = (INT32) wsptr[3]; + z3 = (INT32) wsptr[5]; + z4 = (INT32) wsptr[7]; + + z2 = MULTIPLY(z2, - FIX(1.224744871)); /* -c3 */ + + tmp2 = MULTIPLY(z1 + z3, FIX(0.909038955)); /* c5 */ + tmp3 = MULTIPLY(z1 + z4, FIX(0.483689525)); /* c7 */ + tmp0 = tmp2 + tmp3 - z2; + tmp1 = MULTIPLY(z3 - z4, FIX(1.392728481)); /* c1 */ + tmp2 += z2 - tmp1; + tmp3 += z2 + tmp1; + tmp1 = MULTIPLY(z1 - z3 - z4, FIX(1.224744871)); /* c3 */ + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp11 + tmp1, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp11 - tmp1, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12 + tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp12 - tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp13 + tmp3, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp13 - tmp3, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp14, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += 8; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a 10x10 output block. + * + * Optimized algorithm with 12 multiplications in the 1-D kernel. + * cK represents sqrt(2) * cos(K*pi/20). + */ + +GLOBAL(void) +jpeg_idct_10x10 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp10, tmp11, tmp12, tmp13, tmp14; + INT32 tmp20, tmp21, tmp22, tmp23, tmp24; + INT32 z1, z2, z3, z4, z5; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[8*10]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + z3 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + z3 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + z3 += ONE << (CONST_BITS-PASS1_BITS-1); + z4 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + z1 = MULTIPLY(z4, FIX(1.144122806)); /* c4 */ + z2 = MULTIPLY(z4, FIX(0.437016024)); /* c8 */ + tmp10 = z3 + z1; + tmp11 = z3 - z2; + + tmp22 = RIGHT_SHIFT(z3 - ((z1 - z2) << 1), /* c0 = (c4-c8)*2 */ + CONST_BITS-PASS1_BITS); + + z2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + z3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); + + z1 = MULTIPLY(z2 + z3, FIX(0.831253876)); /* c6 */ + tmp12 = z1 + MULTIPLY(z2, FIX(0.513743148)); /* c2-c6 */ + tmp13 = z1 - MULTIPLY(z3, FIX(2.176250899)); /* c2+c6 */ + + tmp20 = tmp10 + tmp12; + tmp24 = tmp10 - tmp12; + tmp21 = tmp11 + tmp13; + tmp23 = tmp11 - tmp13; + + /* Odd part */ + + z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); + z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); + + tmp11 = z2 + z4; + tmp13 = z2 - z4; + + tmp12 = MULTIPLY(tmp13, FIX(0.309016994)); /* (c3-c7)/2 */ + z5 = z3 << CONST_BITS; + + z2 = MULTIPLY(tmp11, FIX(0.951056516)); /* (c3+c7)/2 */ + z4 = z5 + tmp12; + + tmp10 = MULTIPLY(z1, FIX(1.396802247)) + z2 + z4; /* c1 */ + tmp14 = MULTIPLY(z1, FIX(0.221231742)) - z2 + z4; /* c9 */ + + z2 = MULTIPLY(tmp11, FIX(0.587785252)); /* (c1-c9)/2 */ + z4 = z5 - tmp12 - (tmp13 << (CONST_BITS - 1)); + + tmp12 = (z1 - tmp13 - z3) << PASS1_BITS; + + tmp11 = MULTIPLY(z1, FIX(1.260073511)) - z2 - z4; /* c3 */ + tmp13 = MULTIPLY(z1, FIX(0.642039522)) - z2 + z4; /* c7 */ + + /* Final output stage */ + + wsptr[8*0] = (int) RIGHT_SHIFT(tmp20 + tmp10, CONST_BITS-PASS1_BITS); + wsptr[8*9] = (int) RIGHT_SHIFT(tmp20 - tmp10, CONST_BITS-PASS1_BITS); + wsptr[8*1] = (int) RIGHT_SHIFT(tmp21 + tmp11, CONST_BITS-PASS1_BITS); + wsptr[8*8] = (int) RIGHT_SHIFT(tmp21 - tmp11, CONST_BITS-PASS1_BITS); + wsptr[8*2] = (int) (tmp22 + tmp12); + wsptr[8*7] = (int) (tmp22 - tmp12); + wsptr[8*3] = (int) RIGHT_SHIFT(tmp23 + tmp13, CONST_BITS-PASS1_BITS); + wsptr[8*6] = (int) RIGHT_SHIFT(tmp23 - tmp13, CONST_BITS-PASS1_BITS); + wsptr[8*4] = (int) RIGHT_SHIFT(tmp24 + tmp14, CONST_BITS-PASS1_BITS); + wsptr[8*5] = (int) RIGHT_SHIFT(tmp24 - tmp14, CONST_BITS-PASS1_BITS); + } + + /* Pass 2: process 10 rows from work array, store into output array. */ + + wsptr = workspace; + for (ctr = 0; ctr < 10; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + z3 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + z3 <<= CONST_BITS; + z4 = (INT32) wsptr[4]; + z1 = MULTIPLY(z4, FIX(1.144122806)); /* c4 */ + z2 = MULTIPLY(z4, FIX(0.437016024)); /* c8 */ + tmp10 = z3 + z1; + tmp11 = z3 - z2; + + tmp22 = z3 - ((z1 - z2) << 1); /* c0 = (c4-c8)*2 */ + + z2 = (INT32) wsptr[2]; + z3 = (INT32) wsptr[6]; + + z1 = MULTIPLY(z2 + z3, FIX(0.831253876)); /* c6 */ + tmp12 = z1 + MULTIPLY(z2, FIX(0.513743148)); /* c2-c6 */ + tmp13 = z1 - MULTIPLY(z3, FIX(2.176250899)); /* c2+c6 */ + + tmp20 = tmp10 + tmp12; + tmp24 = tmp10 - tmp12; + tmp21 = tmp11 + tmp13; + tmp23 = tmp11 - tmp13; + + /* Odd part */ + + z1 = (INT32) wsptr[1]; + z2 = (INT32) wsptr[3]; + z3 = (INT32) wsptr[5]; + z3 <<= CONST_BITS; + z4 = (INT32) wsptr[7]; + + tmp11 = z2 + z4; + tmp13 = z2 - z4; + + tmp12 = MULTIPLY(tmp13, FIX(0.309016994)); /* (c3-c7)/2 */ + + z2 = MULTIPLY(tmp11, FIX(0.951056516)); /* (c3+c7)/2 */ + z4 = z3 + tmp12; + + tmp10 = MULTIPLY(z1, FIX(1.396802247)) + z2 + z4; /* c1 */ + tmp14 = MULTIPLY(z1, FIX(0.221231742)) - z2 + z4; /* c9 */ + + z2 = MULTIPLY(tmp11, FIX(0.587785252)); /* (c1-c9)/2 */ + z4 = z3 - tmp12 - (tmp13 << (CONST_BITS - 1)); + + tmp12 = ((z1 - tmp13) << CONST_BITS) - z3; + + tmp11 = MULTIPLY(z1, FIX(1.260073511)) - z2 - z4; /* c3 */ + tmp13 = MULTIPLY(z1, FIX(0.642039522)) - z2 + z4; /* c7 */ + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[9] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23 + tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp23 - tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp24 + tmp14, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp24 - tmp14, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += 8; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing an 11x11 output block. + * + * Optimized algorithm with 24 multiplications in the 1-D kernel. + * cK represents sqrt(2) * cos(K*pi/22). + */ + +GLOBAL(void) +jpeg_idct_11x11 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp10, tmp11, tmp12, tmp13, tmp14; + INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25; + INT32 z1, z2, z3, z4; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[8*11]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + tmp10 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + tmp10 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + tmp10 += ONE << (CONST_BITS-PASS1_BITS-1); + + z1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + z2 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + z3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); + + tmp20 = MULTIPLY(z2 - z3, FIX(2.546640132)); /* c2+c4 */ + tmp23 = MULTIPLY(z2 - z1, FIX(0.430815045)); /* c2-c6 */ + z4 = z1 + z3; + tmp24 = MULTIPLY(z4, - FIX(1.155664402)); /* -(c2-c10) */ + z4 -= z2; + tmp25 = tmp10 + MULTIPLY(z4, FIX(1.356927976)); /* c2 */ + tmp21 = tmp20 + tmp23 + tmp25 - + MULTIPLY(z2, FIX(1.821790775)); /* c2+c4+c10-c6 */ + tmp20 += tmp25 + MULTIPLY(z3, FIX(2.115825087)); /* c4+c6 */ + tmp23 += tmp25 - MULTIPLY(z1, FIX(1.513598477)); /* c6+c8 */ + tmp24 += tmp25; + tmp22 = tmp24 - MULTIPLY(z3, FIX(0.788749120)); /* c8+c10 */ + tmp24 += MULTIPLY(z2, FIX(1.944413522)) - /* c2+c8 */ + MULTIPLY(z1, FIX(1.390975730)); /* c4+c10 */ + tmp25 = tmp10 - MULTIPLY(z4, FIX(1.414213562)); /* c0 */ + + /* Odd part */ + + z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); + z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); + + tmp11 = z1 + z2; + tmp14 = MULTIPLY(tmp11 + z3 + z4, FIX(0.398430003)); /* c9 */ + tmp11 = MULTIPLY(tmp11, FIX(0.887983902)); /* c3-c9 */ + tmp12 = MULTIPLY(z1 + z3, FIX(0.670361295)); /* c5-c9 */ + tmp13 = tmp14 + MULTIPLY(z1 + z4, FIX(0.366151574)); /* c7-c9 */ + tmp10 = tmp11 + tmp12 + tmp13 - + MULTIPLY(z1, FIX(0.923107866)); /* c7+c5+c3-c1-2*c9 */ + z1 = tmp14 - MULTIPLY(z2 + z3, FIX(1.163011579)); /* c7+c9 */ + tmp11 += z1 + MULTIPLY(z2, FIX(2.073276588)); /* c1+c7+3*c9-c3 */ + tmp12 += z1 - MULTIPLY(z3, FIX(1.192193623)); /* c3+c5-c7-c9 */ + z1 = MULTIPLY(z2 + z4, - FIX(1.798248910)); /* -(c1+c9) */ + tmp11 += z1; + tmp13 += z1 + MULTIPLY(z4, FIX(2.102458632)); /* c1+c5+c9-c7 */ + tmp14 += MULTIPLY(z2, - FIX(1.467221301)) + /* -(c5+c9) */ + MULTIPLY(z3, FIX(1.001388905)) - /* c1-c9 */ + MULTIPLY(z4, FIX(1.684843907)); /* c3+c9 */ + + /* Final output stage */ + + wsptr[8*0] = (int) RIGHT_SHIFT(tmp20 + tmp10, CONST_BITS-PASS1_BITS); + wsptr[8*10] = (int) RIGHT_SHIFT(tmp20 - tmp10, CONST_BITS-PASS1_BITS); + wsptr[8*1] = (int) RIGHT_SHIFT(tmp21 + tmp11, CONST_BITS-PASS1_BITS); + wsptr[8*9] = (int) RIGHT_SHIFT(tmp21 - tmp11, CONST_BITS-PASS1_BITS); + wsptr[8*2] = (int) RIGHT_SHIFT(tmp22 + tmp12, CONST_BITS-PASS1_BITS); + wsptr[8*8] = (int) RIGHT_SHIFT(tmp22 - tmp12, CONST_BITS-PASS1_BITS); + wsptr[8*3] = (int) RIGHT_SHIFT(tmp23 + tmp13, CONST_BITS-PASS1_BITS); + wsptr[8*7] = (int) RIGHT_SHIFT(tmp23 - tmp13, CONST_BITS-PASS1_BITS); + wsptr[8*4] = (int) RIGHT_SHIFT(tmp24 + tmp14, CONST_BITS-PASS1_BITS); + wsptr[8*6] = (int) RIGHT_SHIFT(tmp24 - tmp14, CONST_BITS-PASS1_BITS); + wsptr[8*5] = (int) RIGHT_SHIFT(tmp25, CONST_BITS-PASS1_BITS); + } + + /* Pass 2: process 11 rows from work array, store into output array. */ + + wsptr = workspace; + for (ctr = 0; ctr < 11; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + tmp10 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + tmp10 <<= CONST_BITS; + + z1 = (INT32) wsptr[2]; + z2 = (INT32) wsptr[4]; + z3 = (INT32) wsptr[6]; + + tmp20 = MULTIPLY(z2 - z3, FIX(2.546640132)); /* c2+c4 */ + tmp23 = MULTIPLY(z2 - z1, FIX(0.430815045)); /* c2-c6 */ + z4 = z1 + z3; + tmp24 = MULTIPLY(z4, - FIX(1.155664402)); /* -(c2-c10) */ + z4 -= z2; + tmp25 = tmp10 + MULTIPLY(z4, FIX(1.356927976)); /* c2 */ + tmp21 = tmp20 + tmp23 + tmp25 - + MULTIPLY(z2, FIX(1.821790775)); /* c2+c4+c10-c6 */ + tmp20 += tmp25 + MULTIPLY(z3, FIX(2.115825087)); /* c4+c6 */ + tmp23 += tmp25 - MULTIPLY(z1, FIX(1.513598477)); /* c6+c8 */ + tmp24 += tmp25; + tmp22 = tmp24 - MULTIPLY(z3, FIX(0.788749120)); /* c8+c10 */ + tmp24 += MULTIPLY(z2, FIX(1.944413522)) - /* c2+c8 */ + MULTIPLY(z1, FIX(1.390975730)); /* c4+c10 */ + tmp25 = tmp10 - MULTIPLY(z4, FIX(1.414213562)); /* c0 */ + + /* Odd part */ + + z1 = (INT32) wsptr[1]; + z2 = (INT32) wsptr[3]; + z3 = (INT32) wsptr[5]; + z4 = (INT32) wsptr[7]; + + tmp11 = z1 + z2; + tmp14 = MULTIPLY(tmp11 + z3 + z4, FIX(0.398430003)); /* c9 */ + tmp11 = MULTIPLY(tmp11, FIX(0.887983902)); /* c3-c9 */ + tmp12 = MULTIPLY(z1 + z3, FIX(0.670361295)); /* c5-c9 */ + tmp13 = tmp14 + MULTIPLY(z1 + z4, FIX(0.366151574)); /* c7-c9 */ + tmp10 = tmp11 + tmp12 + tmp13 - + MULTIPLY(z1, FIX(0.923107866)); /* c7+c5+c3-c1-2*c9 */ + z1 = tmp14 - MULTIPLY(z2 + z3, FIX(1.163011579)); /* c7+c9 */ + tmp11 += z1 + MULTIPLY(z2, FIX(2.073276588)); /* c1+c7+3*c9-c3 */ + tmp12 += z1 - MULTIPLY(z3, FIX(1.192193623)); /* c3+c5-c7-c9 */ + z1 = MULTIPLY(z2 + z4, - FIX(1.798248910)); /* -(c1+c9) */ + tmp11 += z1; + tmp13 += z1 + MULTIPLY(z4, FIX(2.102458632)); /* c1+c5+c9-c7 */ + tmp14 += MULTIPLY(z2, - FIX(1.467221301)) + /* -(c5+c9) */ + MULTIPLY(z3, FIX(1.001388905)) - /* c1-c9 */ + MULTIPLY(z4, FIX(1.684843907)); /* c3+c9 */ + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[10] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[9] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23 + tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp23 - tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp24 + tmp14, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp24 - tmp14, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp25, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += 8; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a 12x12 output block. + * + * Optimized algorithm with 15 multiplications in the 1-D kernel. + * cK represents sqrt(2) * cos(K*pi/24). + */ + +GLOBAL(void) +jpeg_idct_12x12 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15; + INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25; + INT32 z1, z2, z3, z4; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[8*12]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + z3 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + z3 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + z3 += ONE << (CONST_BITS-PASS1_BITS-1); + + z4 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + z4 = MULTIPLY(z4, FIX(1.224744871)); /* c4 */ + + tmp10 = z3 + z4; + tmp11 = z3 - z4; + + z1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + z4 = MULTIPLY(z1, FIX(1.366025404)); /* c2 */ + z1 <<= CONST_BITS; + z2 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); + z2 <<= CONST_BITS; + + tmp12 = z1 - z2; + + tmp21 = z3 + tmp12; + tmp24 = z3 - tmp12; + + tmp12 = z4 + z2; + + tmp20 = tmp10 + tmp12; + tmp25 = tmp10 - tmp12; + + tmp12 = z4 - z1 - z2; + + tmp22 = tmp11 + tmp12; + tmp23 = tmp11 - tmp12; + + /* Odd part */ + + z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); + z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); + + tmp11 = MULTIPLY(z2, FIX(1.306562965)); /* c3 */ + tmp14 = MULTIPLY(z2, - FIX_0_541196100); /* -c9 */ + + tmp10 = z1 + z3; + tmp15 = MULTIPLY(tmp10 + z4, FIX(0.860918669)); /* c7 */ + tmp12 = tmp15 + MULTIPLY(tmp10, FIX(0.261052384)); /* c5-c7 */ + tmp10 = tmp12 + tmp11 + MULTIPLY(z1, FIX(0.280143716)); /* c1-c5 */ + tmp13 = MULTIPLY(z3 + z4, - FIX(1.045510580)); /* -(c7+c11) */ + tmp12 += tmp13 + tmp14 - MULTIPLY(z3, FIX(1.478575242)); /* c1+c5-c7-c11 */ + tmp13 += tmp15 - tmp11 + MULTIPLY(z4, FIX(1.586706681)); /* c1+c11 */ + tmp15 += tmp14 - MULTIPLY(z1, FIX(0.676326758)) - /* c7-c11 */ + MULTIPLY(z4, FIX(1.982889723)); /* c5+c7 */ + + z1 -= z4; + z2 -= z3; + z3 = MULTIPLY(z1 + z2, FIX_0_541196100); /* c9 */ + tmp11 = z3 + MULTIPLY(z1, FIX_0_765366865); /* c3-c9 */ + tmp14 = z3 - MULTIPLY(z2, FIX_1_847759065); /* c3+c9 */ + + /* Final output stage */ + + wsptr[8*0] = (int) RIGHT_SHIFT(tmp20 + tmp10, CONST_BITS-PASS1_BITS); + wsptr[8*11] = (int) RIGHT_SHIFT(tmp20 - tmp10, CONST_BITS-PASS1_BITS); + wsptr[8*1] = (int) RIGHT_SHIFT(tmp21 + tmp11, CONST_BITS-PASS1_BITS); + wsptr[8*10] = (int) RIGHT_SHIFT(tmp21 - tmp11, CONST_BITS-PASS1_BITS); + wsptr[8*2] = (int) RIGHT_SHIFT(tmp22 + tmp12, CONST_BITS-PASS1_BITS); + wsptr[8*9] = (int) RIGHT_SHIFT(tmp22 - tmp12, CONST_BITS-PASS1_BITS); + wsptr[8*3] = (int) RIGHT_SHIFT(tmp23 + tmp13, CONST_BITS-PASS1_BITS); + wsptr[8*8] = (int) RIGHT_SHIFT(tmp23 - tmp13, CONST_BITS-PASS1_BITS); + wsptr[8*4] = (int) RIGHT_SHIFT(tmp24 + tmp14, CONST_BITS-PASS1_BITS); + wsptr[8*7] = (int) RIGHT_SHIFT(tmp24 - tmp14, CONST_BITS-PASS1_BITS); + wsptr[8*5] = (int) RIGHT_SHIFT(tmp25 + tmp15, CONST_BITS-PASS1_BITS); + wsptr[8*6] = (int) RIGHT_SHIFT(tmp25 - tmp15, CONST_BITS-PASS1_BITS); + } + + /* Pass 2: process 12 rows from work array, store into output array. */ + + wsptr = workspace; + for (ctr = 0; ctr < 12; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + z3 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + z3 <<= CONST_BITS; + + z4 = (INT32) wsptr[4]; + z4 = MULTIPLY(z4, FIX(1.224744871)); /* c4 */ + + tmp10 = z3 + z4; + tmp11 = z3 - z4; + + z1 = (INT32) wsptr[2]; + z4 = MULTIPLY(z1, FIX(1.366025404)); /* c2 */ + z1 <<= CONST_BITS; + z2 = (INT32) wsptr[6]; + z2 <<= CONST_BITS; + + tmp12 = z1 - z2; + + tmp21 = z3 + tmp12; + tmp24 = z3 - tmp12; + + tmp12 = z4 + z2; + + tmp20 = tmp10 + tmp12; + tmp25 = tmp10 - tmp12; + + tmp12 = z4 - z1 - z2; + + tmp22 = tmp11 + tmp12; + tmp23 = tmp11 - tmp12; + + /* Odd part */ + + z1 = (INT32) wsptr[1]; + z2 = (INT32) wsptr[3]; + z3 = (INT32) wsptr[5]; + z4 = (INT32) wsptr[7]; + + tmp11 = MULTIPLY(z2, FIX(1.306562965)); /* c3 */ + tmp14 = MULTIPLY(z2, - FIX_0_541196100); /* -c9 */ + + tmp10 = z1 + z3; + tmp15 = MULTIPLY(tmp10 + z4, FIX(0.860918669)); /* c7 */ + tmp12 = tmp15 + MULTIPLY(tmp10, FIX(0.261052384)); /* c5-c7 */ + tmp10 = tmp12 + tmp11 + MULTIPLY(z1, FIX(0.280143716)); /* c1-c5 */ + tmp13 = MULTIPLY(z3 + z4, - FIX(1.045510580)); /* -(c7+c11) */ + tmp12 += tmp13 + tmp14 - MULTIPLY(z3, FIX(1.478575242)); /* c1+c5-c7-c11 */ + tmp13 += tmp15 - tmp11 + MULTIPLY(z4, FIX(1.586706681)); /* c1+c11 */ + tmp15 += tmp14 - MULTIPLY(z1, FIX(0.676326758)) - /* c7-c11 */ + MULTIPLY(z4, FIX(1.982889723)); /* c5+c7 */ + + z1 -= z4; + z2 -= z3; + z3 = MULTIPLY(z1 + z2, FIX_0_541196100); /* c9 */ + tmp11 = z3 + MULTIPLY(z1, FIX_0_765366865); /* c3-c9 */ + tmp14 = z3 - MULTIPLY(z2, FIX_1_847759065); /* c3+c9 */ + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[11] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[10] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[9] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23 + tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp23 - tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp24 + tmp14, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp24 - tmp14, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp25 + tmp15, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp25 - tmp15, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += 8; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a 13x13 output block. + * + * Optimized algorithm with 29 multiplications in the 1-D kernel. + * cK represents sqrt(2) * cos(K*pi/26). + */ + +GLOBAL(void) +jpeg_idct_13x13 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15; + INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25, tmp26; + INT32 z1, z2, z3, z4; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[8*13]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + z1 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + z1 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + z1 += ONE << (CONST_BITS-PASS1_BITS-1); + + z2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + z3 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + z4 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); + + tmp10 = z3 + z4; + tmp11 = z3 - z4; + + tmp12 = MULTIPLY(tmp10, FIX(1.155388986)); /* (c4+c6)/2 */ + tmp13 = MULTIPLY(tmp11, FIX(0.096834934)) + z1; /* (c4-c6)/2 */ + + tmp20 = MULTIPLY(z2, FIX(1.373119086)) + tmp12 + tmp13; /* c2 */ + tmp22 = MULTIPLY(z2, FIX(0.501487041)) - tmp12 + tmp13; /* c10 */ + + tmp12 = MULTIPLY(tmp10, FIX(0.316450131)); /* (c8-c12)/2 */ + tmp13 = MULTIPLY(tmp11, FIX(0.486914739)) + z1; /* (c8+c12)/2 */ + + tmp21 = MULTIPLY(z2, FIX(1.058554052)) - tmp12 + tmp13; /* c6 */ + tmp25 = MULTIPLY(z2, - FIX(1.252223920)) + tmp12 + tmp13; /* c4 */ + + tmp12 = MULTIPLY(tmp10, FIX(0.435816023)); /* (c2-c10)/2 */ + tmp13 = MULTIPLY(tmp11, FIX(0.937303064)) - z1; /* (c2+c10)/2 */ + + tmp23 = MULTIPLY(z2, - FIX(0.170464608)) - tmp12 - tmp13; /* c12 */ + tmp24 = MULTIPLY(z2, - FIX(0.803364869)) + tmp12 - tmp13; /* c8 */ + + tmp26 = MULTIPLY(tmp11 - z2, FIX(1.414213562)) + z1; /* c0 */ + + /* Odd part */ + + z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); + z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); + + tmp11 = MULTIPLY(z1 + z2, FIX(1.322312651)); /* c3 */ + tmp12 = MULTIPLY(z1 + z3, FIX(1.163874945)); /* c5 */ + tmp15 = z1 + z4; + tmp13 = MULTIPLY(tmp15, FIX(0.937797057)); /* c7 */ + tmp10 = tmp11 + tmp12 + tmp13 - + MULTIPLY(z1, FIX(2.020082300)); /* c7+c5+c3-c1 */ + tmp14 = MULTIPLY(z2 + z3, - FIX(0.338443458)); /* -c11 */ + tmp11 += tmp14 + MULTIPLY(z2, FIX(0.837223564)); /* c5+c9+c11-c3 */ + tmp12 += tmp14 - MULTIPLY(z3, FIX(1.572116027)); /* c1+c5-c9-c11 */ + tmp14 = MULTIPLY(z2 + z4, - FIX(1.163874945)); /* -c5 */ + tmp11 += tmp14; + tmp13 += tmp14 + MULTIPLY(z4, FIX(2.205608352)); /* c3+c5+c9-c7 */ + tmp14 = MULTIPLY(z3 + z4, - FIX(0.657217813)); /* -c9 */ + tmp12 += tmp14; + tmp13 += tmp14; + tmp15 = MULTIPLY(tmp15, FIX(0.338443458)); /* c11 */ + tmp14 = tmp15 + MULTIPLY(z1, FIX(0.318774355)) - /* c9-c11 */ + MULTIPLY(z2, FIX(0.466105296)); /* c1-c7 */ + z1 = MULTIPLY(z3 - z2, FIX(0.937797057)); /* c7 */ + tmp14 += z1; + tmp15 += z1 + MULTIPLY(z3, FIX(0.384515595)) - /* c3-c7 */ + MULTIPLY(z4, FIX(1.742345811)); /* c1+c11 */ + + /* Final output stage */ + + wsptr[8*0] = (int) RIGHT_SHIFT(tmp20 + tmp10, CONST_BITS-PASS1_BITS); + wsptr[8*12] = (int) RIGHT_SHIFT(tmp20 - tmp10, CONST_BITS-PASS1_BITS); + wsptr[8*1] = (int) RIGHT_SHIFT(tmp21 + tmp11, CONST_BITS-PASS1_BITS); + wsptr[8*11] = (int) RIGHT_SHIFT(tmp21 - tmp11, CONST_BITS-PASS1_BITS); + wsptr[8*2] = (int) RIGHT_SHIFT(tmp22 + tmp12, CONST_BITS-PASS1_BITS); + wsptr[8*10] = (int) RIGHT_SHIFT(tmp22 - tmp12, CONST_BITS-PASS1_BITS); + wsptr[8*3] = (int) RIGHT_SHIFT(tmp23 + tmp13, CONST_BITS-PASS1_BITS); + wsptr[8*9] = (int) RIGHT_SHIFT(tmp23 - tmp13, CONST_BITS-PASS1_BITS); + wsptr[8*4] = (int) RIGHT_SHIFT(tmp24 + tmp14, CONST_BITS-PASS1_BITS); + wsptr[8*8] = (int) RIGHT_SHIFT(tmp24 - tmp14, CONST_BITS-PASS1_BITS); + wsptr[8*5] = (int) RIGHT_SHIFT(tmp25 + tmp15, CONST_BITS-PASS1_BITS); + wsptr[8*7] = (int) RIGHT_SHIFT(tmp25 - tmp15, CONST_BITS-PASS1_BITS); + wsptr[8*6] = (int) RIGHT_SHIFT(tmp26, CONST_BITS-PASS1_BITS); + } + + /* Pass 2: process 13 rows from work array, store into output array. */ + + wsptr = workspace; + for (ctr = 0; ctr < 13; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + z1 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + z1 <<= CONST_BITS; + + z2 = (INT32) wsptr[2]; + z3 = (INT32) wsptr[4]; + z4 = (INT32) wsptr[6]; + + tmp10 = z3 + z4; + tmp11 = z3 - z4; + + tmp12 = MULTIPLY(tmp10, FIX(1.155388986)); /* (c4+c6)/2 */ + tmp13 = MULTIPLY(tmp11, FIX(0.096834934)) + z1; /* (c4-c6)/2 */ + + tmp20 = MULTIPLY(z2, FIX(1.373119086)) + tmp12 + tmp13; /* c2 */ + tmp22 = MULTIPLY(z2, FIX(0.501487041)) - tmp12 + tmp13; /* c10 */ + + tmp12 = MULTIPLY(tmp10, FIX(0.316450131)); /* (c8-c12)/2 */ + tmp13 = MULTIPLY(tmp11, FIX(0.486914739)) + z1; /* (c8+c12)/2 */ + + tmp21 = MULTIPLY(z2, FIX(1.058554052)) - tmp12 + tmp13; /* c6 */ + tmp25 = MULTIPLY(z2, - FIX(1.252223920)) + tmp12 + tmp13; /* c4 */ + + tmp12 = MULTIPLY(tmp10, FIX(0.435816023)); /* (c2-c10)/2 */ + tmp13 = MULTIPLY(tmp11, FIX(0.937303064)) - z1; /* (c2+c10)/2 */ + + tmp23 = MULTIPLY(z2, - FIX(0.170464608)) - tmp12 - tmp13; /* c12 */ + tmp24 = MULTIPLY(z2, - FIX(0.803364869)) + tmp12 - tmp13; /* c8 */ + + tmp26 = MULTIPLY(tmp11 - z2, FIX(1.414213562)) + z1; /* c0 */ + + /* Odd part */ + + z1 = (INT32) wsptr[1]; + z2 = (INT32) wsptr[3]; + z3 = (INT32) wsptr[5]; + z4 = (INT32) wsptr[7]; + + tmp11 = MULTIPLY(z1 + z2, FIX(1.322312651)); /* c3 */ + tmp12 = MULTIPLY(z1 + z3, FIX(1.163874945)); /* c5 */ + tmp15 = z1 + z4; + tmp13 = MULTIPLY(tmp15, FIX(0.937797057)); /* c7 */ + tmp10 = tmp11 + tmp12 + tmp13 - + MULTIPLY(z1, FIX(2.020082300)); /* c7+c5+c3-c1 */ + tmp14 = MULTIPLY(z2 + z3, - FIX(0.338443458)); /* -c11 */ + tmp11 += tmp14 + MULTIPLY(z2, FIX(0.837223564)); /* c5+c9+c11-c3 */ + tmp12 += tmp14 - MULTIPLY(z3, FIX(1.572116027)); /* c1+c5-c9-c11 */ + tmp14 = MULTIPLY(z2 + z4, - FIX(1.163874945)); /* -c5 */ + tmp11 += tmp14; + tmp13 += tmp14 + MULTIPLY(z4, FIX(2.205608352)); /* c3+c5+c9-c7 */ + tmp14 = MULTIPLY(z3 + z4, - FIX(0.657217813)); /* -c9 */ + tmp12 += tmp14; + tmp13 += tmp14; + tmp15 = MULTIPLY(tmp15, FIX(0.338443458)); /* c11 */ + tmp14 = tmp15 + MULTIPLY(z1, FIX(0.318774355)) - /* c9-c11 */ + MULTIPLY(z2, FIX(0.466105296)); /* c1-c7 */ + z1 = MULTIPLY(z3 - z2, FIX(0.937797057)); /* c7 */ + tmp14 += z1; + tmp15 += z1 + MULTIPLY(z3, FIX(0.384515595)) - /* c3-c7 */ + MULTIPLY(z4, FIX(1.742345811)); /* c1+c11 */ + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[12] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[11] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[10] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23 + tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[9] = range_limit[(int) RIGHT_SHIFT(tmp23 - tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp24 + tmp14, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp24 - tmp14, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp25 + tmp15, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp25 - tmp15, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp26, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += 8; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a 14x14 output block. + * + * Optimized algorithm with 20 multiplications in the 1-D kernel. + * cK represents sqrt(2) * cos(K*pi/28). + */ + +GLOBAL(void) +jpeg_idct_14x14 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15, tmp16; + INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25, tmp26; + INT32 z1, z2, z3, z4; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[8*14]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + z1 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + z1 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + z1 += ONE << (CONST_BITS-PASS1_BITS-1); + z4 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + z2 = MULTIPLY(z4, FIX(1.274162392)); /* c4 */ + z3 = MULTIPLY(z4, FIX(0.314692123)); /* c12 */ + z4 = MULTIPLY(z4, FIX(0.881747734)); /* c8 */ + + tmp10 = z1 + z2; + tmp11 = z1 + z3; + tmp12 = z1 - z4; + + tmp23 = RIGHT_SHIFT(z1 - ((z2 + z3 - z4) << 1), /* c0 = (c4+c12-c8)*2 */ + CONST_BITS-PASS1_BITS); + + z1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + z2 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); + + z3 = MULTIPLY(z1 + z2, FIX(1.105676686)); /* c6 */ + + tmp13 = z3 + MULTIPLY(z1, FIX(0.273079590)); /* c2-c6 */ + tmp14 = z3 - MULTIPLY(z2, FIX(1.719280954)); /* c6+c10 */ + tmp15 = MULTIPLY(z1, FIX(0.613604268)) - /* c10 */ + MULTIPLY(z2, FIX(1.378756276)); /* c2 */ + + tmp20 = tmp10 + tmp13; + tmp26 = tmp10 - tmp13; + tmp21 = tmp11 + tmp14; + tmp25 = tmp11 - tmp14; + tmp22 = tmp12 + tmp15; + tmp24 = tmp12 - tmp15; + + /* Odd part */ + + z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); + z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); + tmp13 = z4 << CONST_BITS; + + tmp14 = z1 + z3; + tmp11 = MULTIPLY(z1 + z2, FIX(1.334852607)); /* c3 */ + tmp12 = MULTIPLY(tmp14, FIX(1.197448846)); /* c5 */ + tmp10 = tmp11 + tmp12 + tmp13 - MULTIPLY(z1, FIX(1.126980169)); /* c3+c5-c1 */ + tmp14 = MULTIPLY(tmp14, FIX(0.752406978)); /* c9 */ + tmp16 = tmp14 - MULTIPLY(z1, FIX(1.061150426)); /* c9+c11-c13 */ + z1 -= z2; + tmp15 = MULTIPLY(z1, FIX(0.467085129)) - tmp13; /* c11 */ + tmp16 += tmp15; + z1 += z4; + z4 = MULTIPLY(z2 + z3, - FIX(0.158341681)) - tmp13; /* -c13 */ + tmp11 += z4 - MULTIPLY(z2, FIX(0.424103948)); /* c3-c9-c13 */ + tmp12 += z4 - MULTIPLY(z3, FIX(2.373959773)); /* c3+c5-c13 */ + z4 = MULTIPLY(z3 - z2, FIX(1.405321284)); /* c1 */ + tmp14 += z4 + tmp13 - MULTIPLY(z3, FIX(1.6906431334)); /* c1+c9-c11 */ + tmp15 += z4 + MULTIPLY(z2, FIX(0.674957567)); /* c1+c11-c5 */ + + tmp13 = (z1 - z3) << PASS1_BITS; + + /* Final output stage */ + + wsptr[8*0] = (int) RIGHT_SHIFT(tmp20 + tmp10, CONST_BITS-PASS1_BITS); + wsptr[8*13] = (int) RIGHT_SHIFT(tmp20 - tmp10, CONST_BITS-PASS1_BITS); + wsptr[8*1] = (int) RIGHT_SHIFT(tmp21 + tmp11, CONST_BITS-PASS1_BITS); + wsptr[8*12] = (int) RIGHT_SHIFT(tmp21 - tmp11, CONST_BITS-PASS1_BITS); + wsptr[8*2] = (int) RIGHT_SHIFT(tmp22 + tmp12, CONST_BITS-PASS1_BITS); + wsptr[8*11] = (int) RIGHT_SHIFT(tmp22 - tmp12, CONST_BITS-PASS1_BITS); + wsptr[8*3] = (int) (tmp23 + tmp13); + wsptr[8*10] = (int) (tmp23 - tmp13); + wsptr[8*4] = (int) RIGHT_SHIFT(tmp24 + tmp14, CONST_BITS-PASS1_BITS); + wsptr[8*9] = (int) RIGHT_SHIFT(tmp24 - tmp14, CONST_BITS-PASS1_BITS); + wsptr[8*5] = (int) RIGHT_SHIFT(tmp25 + tmp15, CONST_BITS-PASS1_BITS); + wsptr[8*8] = (int) RIGHT_SHIFT(tmp25 - tmp15, CONST_BITS-PASS1_BITS); + wsptr[8*6] = (int) RIGHT_SHIFT(tmp26 + tmp16, CONST_BITS-PASS1_BITS); + wsptr[8*7] = (int) RIGHT_SHIFT(tmp26 - tmp16, CONST_BITS-PASS1_BITS); + } + + /* Pass 2: process 14 rows from work array, store into output array. */ + + wsptr = workspace; + for (ctr = 0; ctr < 14; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + z1 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + z1 <<= CONST_BITS; + z4 = (INT32) wsptr[4]; + z2 = MULTIPLY(z4, FIX(1.274162392)); /* c4 */ + z3 = MULTIPLY(z4, FIX(0.314692123)); /* c12 */ + z4 = MULTIPLY(z4, FIX(0.881747734)); /* c8 */ + + tmp10 = z1 + z2; + tmp11 = z1 + z3; + tmp12 = z1 - z4; + + tmp23 = z1 - ((z2 + z3 - z4) << 1); /* c0 = (c4+c12-c8)*2 */ + + z1 = (INT32) wsptr[2]; + z2 = (INT32) wsptr[6]; + + z3 = MULTIPLY(z1 + z2, FIX(1.105676686)); /* c6 */ + + tmp13 = z3 + MULTIPLY(z1, FIX(0.273079590)); /* c2-c6 */ + tmp14 = z3 - MULTIPLY(z2, FIX(1.719280954)); /* c6+c10 */ + tmp15 = MULTIPLY(z1, FIX(0.613604268)) - /* c10 */ + MULTIPLY(z2, FIX(1.378756276)); /* c2 */ + + tmp20 = tmp10 + tmp13; + tmp26 = tmp10 - tmp13; + tmp21 = tmp11 + tmp14; + tmp25 = tmp11 - tmp14; + tmp22 = tmp12 + tmp15; + tmp24 = tmp12 - tmp15; + + /* Odd part */ + + z1 = (INT32) wsptr[1]; + z2 = (INT32) wsptr[3]; + z3 = (INT32) wsptr[5]; + z4 = (INT32) wsptr[7]; + z4 <<= CONST_BITS; + + tmp14 = z1 + z3; + tmp11 = MULTIPLY(z1 + z2, FIX(1.334852607)); /* c3 */ + tmp12 = MULTIPLY(tmp14, FIX(1.197448846)); /* c5 */ + tmp10 = tmp11 + tmp12 + z4 - MULTIPLY(z1, FIX(1.126980169)); /* c3+c5-c1 */ + tmp14 = MULTIPLY(tmp14, FIX(0.752406978)); /* c9 */ + tmp16 = tmp14 - MULTIPLY(z1, FIX(1.061150426)); /* c9+c11-c13 */ + z1 -= z2; + tmp15 = MULTIPLY(z1, FIX(0.467085129)) - z4; /* c11 */ + tmp16 += tmp15; + tmp13 = MULTIPLY(z2 + z3, - FIX(0.158341681)) - z4; /* -c13 */ + tmp11 += tmp13 - MULTIPLY(z2, FIX(0.424103948)); /* c3-c9-c13 */ + tmp12 += tmp13 - MULTIPLY(z3, FIX(2.373959773)); /* c3+c5-c13 */ + tmp13 = MULTIPLY(z3 - z2, FIX(1.405321284)); /* c1 */ + tmp14 += tmp13 + z4 - MULTIPLY(z3, FIX(1.6906431334)); /* c1+c9-c11 */ + tmp15 += tmp13 + MULTIPLY(z2, FIX(0.674957567)); /* c1+c11-c5 */ + + tmp13 = ((z1 - z3) << CONST_BITS) + z4; + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[13] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[12] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[11] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23 + tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[10] = range_limit[(int) RIGHT_SHIFT(tmp23 - tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp24 + tmp14, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[9] = range_limit[(int) RIGHT_SHIFT(tmp24 - tmp14, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp25 + tmp15, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp25 - tmp15, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp26 + tmp16, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp26 - tmp16, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += 8; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a 15x15 output block. + * + * Optimized algorithm with 22 multiplications in the 1-D kernel. + * cK represents sqrt(2) * cos(K*pi/30). + */ + +GLOBAL(void) +jpeg_idct_15x15 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15, tmp16; + INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25, tmp26, tmp27; + INT32 z1, z2, z3, z4; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[8*15]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + z1 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + z1 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + z1 += ONE << (CONST_BITS-PASS1_BITS-1); + + z2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + z3 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + z4 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); + + tmp10 = MULTIPLY(z4, FIX(0.437016024)); /* c12 */ + tmp11 = MULTIPLY(z4, FIX(1.144122806)); /* c6 */ + + tmp12 = z1 - tmp10; + tmp13 = z1 + tmp11; + z1 -= (tmp11 - tmp10) << 1; /* c0 = (c6-c12)*2 */ + + z4 = z2 - z3; + z3 += z2; + tmp10 = MULTIPLY(z3, FIX(1.337628990)); /* (c2+c4)/2 */ + tmp11 = MULTIPLY(z4, FIX(0.045680613)); /* (c2-c4)/2 */ + z2 = MULTIPLY(z2, FIX(1.439773946)); /* c4+c14 */ + + tmp20 = tmp13 + tmp10 + tmp11; + tmp23 = tmp12 - tmp10 + tmp11 + z2; + + tmp10 = MULTIPLY(z3, FIX(0.547059574)); /* (c8+c14)/2 */ + tmp11 = MULTIPLY(z4, FIX(0.399234004)); /* (c8-c14)/2 */ + + tmp25 = tmp13 - tmp10 - tmp11; + tmp26 = tmp12 + tmp10 - tmp11 - z2; + + tmp10 = MULTIPLY(z3, FIX(0.790569415)); /* (c6+c12)/2 */ + tmp11 = MULTIPLY(z4, FIX(0.353553391)); /* (c6-c12)/2 */ + + tmp21 = tmp12 + tmp10 + tmp11; + tmp24 = tmp13 - tmp10 + tmp11; + tmp11 += tmp11; + tmp22 = z1 + tmp11; /* c10 = c6-c12 */ + tmp27 = z1 - tmp11 - tmp11; /* c0 = (c6-c12)*2 */ + + /* Odd part */ + + z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + z4 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); + z3 = MULTIPLY(z4, FIX(1.224744871)); /* c5 */ + z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); + + tmp13 = z2 - z4; + tmp15 = MULTIPLY(z1 + tmp13, FIX(0.831253876)); /* c9 */ + tmp11 = tmp15 + MULTIPLY(z1, FIX(0.513743148)); /* c3-c9 */ + tmp14 = tmp15 - MULTIPLY(tmp13, FIX(2.176250899)); /* c3+c9 */ + + tmp13 = MULTIPLY(z2, - FIX(0.831253876)); /* -c9 */ + tmp15 = MULTIPLY(z2, - FIX(1.344997024)); /* -c3 */ + z2 = z1 - z4; + tmp12 = z3 + MULTIPLY(z2, FIX(1.406466353)); /* c1 */ + + tmp10 = tmp12 + MULTIPLY(z4, FIX(2.457431844)) - tmp15; /* c1+c7 */ + tmp16 = tmp12 - MULTIPLY(z1, FIX(1.112434820)) + tmp13; /* c1-c13 */ + tmp12 = MULTIPLY(z2, FIX(1.224744871)) - z3; /* c5 */ + z2 = MULTIPLY(z1 + z4, FIX(0.575212477)); /* c11 */ + tmp13 += z2 + MULTIPLY(z1, FIX(0.475753014)) - z3; /* c7-c11 */ + tmp15 += z2 - MULTIPLY(z4, FIX(0.869244010)) + z3; /* c11+c13 */ + + /* Final output stage */ + + wsptr[8*0] = (int) RIGHT_SHIFT(tmp20 + tmp10, CONST_BITS-PASS1_BITS); + wsptr[8*14] = (int) RIGHT_SHIFT(tmp20 - tmp10, CONST_BITS-PASS1_BITS); + wsptr[8*1] = (int) RIGHT_SHIFT(tmp21 + tmp11, CONST_BITS-PASS1_BITS); + wsptr[8*13] = (int) RIGHT_SHIFT(tmp21 - tmp11, CONST_BITS-PASS1_BITS); + wsptr[8*2] = (int) RIGHT_SHIFT(tmp22 + tmp12, CONST_BITS-PASS1_BITS); + wsptr[8*12] = (int) RIGHT_SHIFT(tmp22 - tmp12, CONST_BITS-PASS1_BITS); + wsptr[8*3] = (int) RIGHT_SHIFT(tmp23 + tmp13, CONST_BITS-PASS1_BITS); + wsptr[8*11] = (int) RIGHT_SHIFT(tmp23 - tmp13, CONST_BITS-PASS1_BITS); + wsptr[8*4] = (int) RIGHT_SHIFT(tmp24 + tmp14, CONST_BITS-PASS1_BITS); + wsptr[8*10] = (int) RIGHT_SHIFT(tmp24 - tmp14, CONST_BITS-PASS1_BITS); + wsptr[8*5] = (int) RIGHT_SHIFT(tmp25 + tmp15, CONST_BITS-PASS1_BITS); + wsptr[8*9] = (int) RIGHT_SHIFT(tmp25 - tmp15, CONST_BITS-PASS1_BITS); + wsptr[8*6] = (int) RIGHT_SHIFT(tmp26 + tmp16, CONST_BITS-PASS1_BITS); + wsptr[8*8] = (int) RIGHT_SHIFT(tmp26 - tmp16, CONST_BITS-PASS1_BITS); + wsptr[8*7] = (int) RIGHT_SHIFT(tmp27, CONST_BITS-PASS1_BITS); + } + + /* Pass 2: process 15 rows from work array, store into output array. */ + + wsptr = workspace; + for (ctr = 0; ctr < 15; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + z1 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + z1 <<= CONST_BITS; + + z2 = (INT32) wsptr[2]; + z3 = (INT32) wsptr[4]; + z4 = (INT32) wsptr[6]; + + tmp10 = MULTIPLY(z4, FIX(0.437016024)); /* c12 */ + tmp11 = MULTIPLY(z4, FIX(1.144122806)); /* c6 */ + + tmp12 = z1 - tmp10; + tmp13 = z1 + tmp11; + z1 -= (tmp11 - tmp10) << 1; /* c0 = (c6-c12)*2 */ + + z4 = z2 - z3; + z3 += z2; + tmp10 = MULTIPLY(z3, FIX(1.337628990)); /* (c2+c4)/2 */ + tmp11 = MULTIPLY(z4, FIX(0.045680613)); /* (c2-c4)/2 */ + z2 = MULTIPLY(z2, FIX(1.439773946)); /* c4+c14 */ + + tmp20 = tmp13 + tmp10 + tmp11; + tmp23 = tmp12 - tmp10 + tmp11 + z2; + + tmp10 = MULTIPLY(z3, FIX(0.547059574)); /* (c8+c14)/2 */ + tmp11 = MULTIPLY(z4, FIX(0.399234004)); /* (c8-c14)/2 */ + + tmp25 = tmp13 - tmp10 - tmp11; + tmp26 = tmp12 + tmp10 - tmp11 - z2; + + tmp10 = MULTIPLY(z3, FIX(0.790569415)); /* (c6+c12)/2 */ + tmp11 = MULTIPLY(z4, FIX(0.353553391)); /* (c6-c12)/2 */ + + tmp21 = tmp12 + tmp10 + tmp11; + tmp24 = tmp13 - tmp10 + tmp11; + tmp11 += tmp11; + tmp22 = z1 + tmp11; /* c10 = c6-c12 */ + tmp27 = z1 - tmp11 - tmp11; /* c0 = (c6-c12)*2 */ + + /* Odd part */ + + z1 = (INT32) wsptr[1]; + z2 = (INT32) wsptr[3]; + z4 = (INT32) wsptr[5]; + z3 = MULTIPLY(z4, FIX(1.224744871)); /* c5 */ + z4 = (INT32) wsptr[7]; + + tmp13 = z2 - z4; + tmp15 = MULTIPLY(z1 + tmp13, FIX(0.831253876)); /* c9 */ + tmp11 = tmp15 + MULTIPLY(z1, FIX(0.513743148)); /* c3-c9 */ + tmp14 = tmp15 - MULTIPLY(tmp13, FIX(2.176250899)); /* c3+c9 */ + + tmp13 = MULTIPLY(z2, - FIX(0.831253876)); /* -c9 */ + tmp15 = MULTIPLY(z2, - FIX(1.344997024)); /* -c3 */ + z2 = z1 - z4; + tmp12 = z3 + MULTIPLY(z2, FIX(1.406466353)); /* c1 */ + + tmp10 = tmp12 + MULTIPLY(z4, FIX(2.457431844)) - tmp15; /* c1+c7 */ + tmp16 = tmp12 - MULTIPLY(z1, FIX(1.112434820)) + tmp13; /* c1-c13 */ + tmp12 = MULTIPLY(z2, FIX(1.224744871)) - z3; /* c5 */ + z2 = MULTIPLY(z1 + z4, FIX(0.575212477)); /* c11 */ + tmp13 += z2 + MULTIPLY(z1, FIX(0.475753014)) - z3; /* c7-c11 */ + tmp15 += z2 - MULTIPLY(z4, FIX(0.869244010)) + z3; /* c11+c13 */ + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[14] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[13] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[12] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23 + tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[11] = range_limit[(int) RIGHT_SHIFT(tmp23 - tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp24 + tmp14, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[10] = range_limit[(int) RIGHT_SHIFT(tmp24 - tmp14, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp25 + tmp15, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[9] = range_limit[(int) RIGHT_SHIFT(tmp25 - tmp15, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp26 + tmp16, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp26 - tmp16, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp27, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += 8; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a 16x16 output block. + * + * Optimized algorithm with 28 multiplications in the 1-D kernel. + * cK represents sqrt(2) * cos(K*pi/32). + */ + +GLOBAL(void) +jpeg_idct_16x16 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3, tmp10, tmp11, tmp12, tmp13; + INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25, tmp26, tmp27; + INT32 z1, z2, z3, z4; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[8*16]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + tmp0 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + tmp0 += ONE << (CONST_BITS-PASS1_BITS-1); + + z1 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + tmp1 = MULTIPLY(z1, FIX(1.306562965)); /* c4[16] = c2[8] */ + tmp2 = MULTIPLY(z1, FIX_0_541196100); /* c12[16] = c6[8] */ + + tmp10 = tmp0 + tmp1; + tmp11 = tmp0 - tmp1; + tmp12 = tmp0 + tmp2; + tmp13 = tmp0 - tmp2; + + z1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + z2 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); + z3 = z1 - z2; + z4 = MULTIPLY(z3, FIX(0.275899379)); /* c14[16] = c7[8] */ + z3 = MULTIPLY(z3, FIX(1.387039845)); /* c2[16] = c1[8] */ + + tmp0 = z3 + MULTIPLY(z2, FIX_2_562915447); /* (c6+c2)[16] = (c3+c1)[8] */ + tmp1 = z4 + MULTIPLY(z1, FIX_0_899976223); /* (c6-c14)[16] = (c3-c7)[8] */ + tmp2 = z3 - MULTIPLY(z1, FIX(0.601344887)); /* (c2-c10)[16] = (c1-c5)[8] */ + tmp3 = z4 - MULTIPLY(z2, FIX(0.509795579)); /* (c10-c14)[16] = (c5-c7)[8] */ + + tmp20 = tmp10 + tmp0; + tmp27 = tmp10 - tmp0; + tmp21 = tmp12 + tmp1; + tmp26 = tmp12 - tmp1; + tmp22 = tmp13 + tmp2; + tmp25 = tmp13 - tmp2; + tmp23 = tmp11 + tmp3; + tmp24 = tmp11 - tmp3; + + /* Odd part */ + + z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); + z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); + + tmp11 = z1 + z3; + + tmp1 = MULTIPLY(z1 + z2, FIX(1.353318001)); /* c3 */ + tmp2 = MULTIPLY(tmp11, FIX(1.247225013)); /* c5 */ + tmp3 = MULTIPLY(z1 + z4, FIX(1.093201867)); /* c7 */ + tmp10 = MULTIPLY(z1 - z4, FIX(0.897167586)); /* c9 */ + tmp11 = MULTIPLY(tmp11, FIX(0.666655658)); /* c11 */ + tmp12 = MULTIPLY(z1 - z2, FIX(0.410524528)); /* c13 */ + tmp0 = tmp1 + tmp2 + tmp3 - + MULTIPLY(z1, FIX(2.286341144)); /* c7+c5+c3-c1 */ + tmp13 = tmp10 + tmp11 + tmp12 - + MULTIPLY(z1, FIX(1.835730603)); /* c9+c11+c13-c15 */ + z1 = MULTIPLY(z2 + z3, FIX(0.138617169)); /* c15 */ + tmp1 += z1 + MULTIPLY(z2, FIX(0.071888074)); /* c9+c11-c3-c15 */ + tmp2 += z1 - MULTIPLY(z3, FIX(1.125726048)); /* c5+c7+c15-c3 */ + z1 = MULTIPLY(z3 - z2, FIX(1.407403738)); /* c1 */ + tmp11 += z1 - MULTIPLY(z3, FIX(0.766367282)); /* c1+c11-c9-c13 */ + tmp12 += z1 + MULTIPLY(z2, FIX(1.971951411)); /* c1+c5+c13-c7 */ + z2 += z4; + z1 = MULTIPLY(z2, - FIX(0.666655658)); /* -c11 */ + tmp1 += z1; + tmp3 += z1 + MULTIPLY(z4, FIX(1.065388962)); /* c3+c11+c15-c7 */ + z2 = MULTIPLY(z2, - FIX(1.247225013)); /* -c5 */ + tmp10 += z2 + MULTIPLY(z4, FIX(3.141271809)); /* c1+c5+c9-c13 */ + tmp12 += z2; + z2 = MULTIPLY(z3 + z4, - FIX(1.353318001)); /* -c3 */ + tmp2 += z2; + tmp3 += z2; + z2 = MULTIPLY(z4 - z3, FIX(0.410524528)); /* c13 */ + tmp10 += z2; + tmp11 += z2; + + /* Final output stage */ + + wsptr[8*0] = (int) RIGHT_SHIFT(tmp20 + tmp0, CONST_BITS-PASS1_BITS); + wsptr[8*15] = (int) RIGHT_SHIFT(tmp20 - tmp0, CONST_BITS-PASS1_BITS); + wsptr[8*1] = (int) RIGHT_SHIFT(tmp21 + tmp1, CONST_BITS-PASS1_BITS); + wsptr[8*14] = (int) RIGHT_SHIFT(tmp21 - tmp1, CONST_BITS-PASS1_BITS); + wsptr[8*2] = (int) RIGHT_SHIFT(tmp22 + tmp2, CONST_BITS-PASS1_BITS); + wsptr[8*13] = (int) RIGHT_SHIFT(tmp22 - tmp2, CONST_BITS-PASS1_BITS); + wsptr[8*3] = (int) RIGHT_SHIFT(tmp23 + tmp3, CONST_BITS-PASS1_BITS); + wsptr[8*12] = (int) RIGHT_SHIFT(tmp23 - tmp3, CONST_BITS-PASS1_BITS); + wsptr[8*4] = (int) RIGHT_SHIFT(tmp24 + tmp10, CONST_BITS-PASS1_BITS); + wsptr[8*11] = (int) RIGHT_SHIFT(tmp24 - tmp10, CONST_BITS-PASS1_BITS); + wsptr[8*5] = (int) RIGHT_SHIFT(tmp25 + tmp11, CONST_BITS-PASS1_BITS); + wsptr[8*10] = (int) RIGHT_SHIFT(tmp25 - tmp11, CONST_BITS-PASS1_BITS); + wsptr[8*6] = (int) RIGHT_SHIFT(tmp26 + tmp12, CONST_BITS-PASS1_BITS); + wsptr[8*9] = (int) RIGHT_SHIFT(tmp26 - tmp12, CONST_BITS-PASS1_BITS); + wsptr[8*7] = (int) RIGHT_SHIFT(tmp27 + tmp13, CONST_BITS-PASS1_BITS); + wsptr[8*8] = (int) RIGHT_SHIFT(tmp27 - tmp13, CONST_BITS-PASS1_BITS); + } + + /* Pass 2: process 16 rows from work array, store into output array. */ + + wsptr = workspace; + for (ctr = 0; ctr < 16; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + tmp0 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + tmp0 <<= CONST_BITS; + + z1 = (INT32) wsptr[4]; + tmp1 = MULTIPLY(z1, FIX(1.306562965)); /* c4[16] = c2[8] */ + tmp2 = MULTIPLY(z1, FIX_0_541196100); /* c12[16] = c6[8] */ + + tmp10 = tmp0 + tmp1; + tmp11 = tmp0 - tmp1; + tmp12 = tmp0 + tmp2; + tmp13 = tmp0 - tmp2; + + z1 = (INT32) wsptr[2]; + z2 = (INT32) wsptr[6]; + z3 = z1 - z2; + z4 = MULTIPLY(z3, FIX(0.275899379)); /* c14[16] = c7[8] */ + z3 = MULTIPLY(z3, FIX(1.387039845)); /* c2[16] = c1[8] */ + + tmp0 = z3 + MULTIPLY(z2, FIX_2_562915447); /* (c6+c2)[16] = (c3+c1)[8] */ + tmp1 = z4 + MULTIPLY(z1, FIX_0_899976223); /* (c6-c14)[16] = (c3-c7)[8] */ + tmp2 = z3 - MULTIPLY(z1, FIX(0.601344887)); /* (c2-c10)[16] = (c1-c5)[8] */ + tmp3 = z4 - MULTIPLY(z2, FIX(0.509795579)); /* (c10-c14)[16] = (c5-c7)[8] */ + + tmp20 = tmp10 + tmp0; + tmp27 = tmp10 - tmp0; + tmp21 = tmp12 + tmp1; + tmp26 = tmp12 - tmp1; + tmp22 = tmp13 + tmp2; + tmp25 = tmp13 - tmp2; + tmp23 = tmp11 + tmp3; + tmp24 = tmp11 - tmp3; + + /* Odd part */ + + z1 = (INT32) wsptr[1]; + z2 = (INT32) wsptr[3]; + z3 = (INT32) wsptr[5]; + z4 = (INT32) wsptr[7]; + + tmp11 = z1 + z3; + + tmp1 = MULTIPLY(z1 + z2, FIX(1.353318001)); /* c3 */ + tmp2 = MULTIPLY(tmp11, FIX(1.247225013)); /* c5 */ + tmp3 = MULTIPLY(z1 + z4, FIX(1.093201867)); /* c7 */ + tmp10 = MULTIPLY(z1 - z4, FIX(0.897167586)); /* c9 */ + tmp11 = MULTIPLY(tmp11, FIX(0.666655658)); /* c11 */ + tmp12 = MULTIPLY(z1 - z2, FIX(0.410524528)); /* c13 */ + tmp0 = tmp1 + tmp2 + tmp3 - + MULTIPLY(z1, FIX(2.286341144)); /* c7+c5+c3-c1 */ + tmp13 = tmp10 + tmp11 + tmp12 - + MULTIPLY(z1, FIX(1.835730603)); /* c9+c11+c13-c15 */ + z1 = MULTIPLY(z2 + z3, FIX(0.138617169)); /* c15 */ + tmp1 += z1 + MULTIPLY(z2, FIX(0.071888074)); /* c9+c11-c3-c15 */ + tmp2 += z1 - MULTIPLY(z3, FIX(1.125726048)); /* c5+c7+c15-c3 */ + z1 = MULTIPLY(z3 - z2, FIX(1.407403738)); /* c1 */ + tmp11 += z1 - MULTIPLY(z3, FIX(0.766367282)); /* c1+c11-c9-c13 */ + tmp12 += z1 + MULTIPLY(z2, FIX(1.971951411)); /* c1+c5+c13-c7 */ + z2 += z4; + z1 = MULTIPLY(z2, - FIX(0.666655658)); /* -c11 */ + tmp1 += z1; + tmp3 += z1 + MULTIPLY(z4, FIX(1.065388962)); /* c3+c11+c15-c7 */ + z2 = MULTIPLY(z2, - FIX(1.247225013)); /* -c5 */ + tmp10 += z2 + MULTIPLY(z4, FIX(3.141271809)); /* c1+c5+c9-c13 */ + tmp12 += z2; + z2 = MULTIPLY(z3 + z4, - FIX(1.353318001)); /* -c3 */ + tmp2 += z2; + tmp3 += z2; + z2 = MULTIPLY(z4 - z3, FIX(0.410524528)); /* c13 */ + tmp10 += z2; + tmp11 += z2; + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[15] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp1, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[14] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp1, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[13] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23 + tmp3, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[12] = range_limit[(int) RIGHT_SHIFT(tmp23 - tmp3, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp24 + tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[11] = range_limit[(int) RIGHT_SHIFT(tmp24 - tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp25 + tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[10] = range_limit[(int) RIGHT_SHIFT(tmp25 - tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp26 + tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[9] = range_limit[(int) RIGHT_SHIFT(tmp26 - tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp27 + tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp27 - tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += 8; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a 16x8 output block. + * + * 8-point IDCT in pass 1 (columns), 16-point in pass 2 (rows). + */ + +GLOBAL(void) +jpeg_idct_16x8 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3, tmp10, tmp11, tmp12, tmp13; + INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25, tmp26, tmp27; + INT32 z1, z2, z3, z4; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[8*8]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. + * Note results are scaled up by sqrt(8) compared to a true IDCT; + * furthermore, we scale the results by 2**PASS1_BITS. + * 8-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/16). + */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = DCTSIZE; ctr > 0; ctr--) { + /* Due to quantization, we will usually find that many of the input + * coefficients are zero, especially the AC terms. We can exploit this + * by short-circuiting the IDCT calculation for any column in which all + * the AC terms are zero. In that case each output is equal to the + * DC coefficient (with scale factor as needed). + * With typical images and quantization tables, half or more of the + * column DCT calculations can be simplified this way. + */ + + if (inptr[DCTSIZE*1] == 0 && inptr[DCTSIZE*2] == 0 && + inptr[DCTSIZE*3] == 0 && inptr[DCTSIZE*4] == 0 && + inptr[DCTSIZE*5] == 0 && inptr[DCTSIZE*6] == 0 && + inptr[DCTSIZE*7] == 0) { + /* AC terms all zero */ + int dcval = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]) << PASS1_BITS; + + wsptr[DCTSIZE*0] = dcval; + wsptr[DCTSIZE*1] = dcval; + wsptr[DCTSIZE*2] = dcval; + wsptr[DCTSIZE*3] = dcval; + wsptr[DCTSIZE*4] = dcval; + wsptr[DCTSIZE*5] = dcval; + wsptr[DCTSIZE*6] = dcval; + wsptr[DCTSIZE*7] = dcval; + + inptr++; /* advance pointers to next column */ + quantptr++; + wsptr++; + continue; + } + + /* Even part: reverse the even part of the forward DCT. + * The rotator is c(-6). + */ + + z2 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + z3 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + z2 <<= CONST_BITS; + z3 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + z2 += ONE << (CONST_BITS-PASS1_BITS-1); + + tmp0 = z2 + z3; + tmp1 = z2 - z3; + + z2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + z3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); + + z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ + tmp2 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ + tmp3 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ + + tmp10 = tmp0 + tmp2; + tmp13 = tmp0 - tmp2; + tmp11 = tmp1 + tmp3; + tmp12 = tmp1 - tmp3; + + /* Odd part per figure 8; the matrix is unitary and hence its + * transpose is its inverse. i0..i3 are y7,y5,y3,y1 respectively. + */ + + tmp0 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); + tmp1 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); + tmp2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + tmp3 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + + z2 = tmp0 + tmp2; + z3 = tmp1 + tmp3; + + z1 = MULTIPLY(z2 + z3, FIX_1_175875602); /* c3 */ + z2 = MULTIPLY(z2, - FIX_1_961570560); /* -c3-c5 */ + z3 = MULTIPLY(z3, - FIX_0_390180644); /* -c3+c5 */ + z2 += z1; + z3 += z1; + + z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ + tmp0 = MULTIPLY(tmp0, FIX_0_298631336); /* -c1+c3+c5-c7 */ + tmp3 = MULTIPLY(tmp3, FIX_1_501321110); /* c1+c3-c5-c7 */ + tmp0 += z1 + z2; + tmp3 += z1 + z3; + + z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ + tmp1 = MULTIPLY(tmp1, FIX_2_053119869); /* c1+c3-c5+c7 */ + tmp2 = MULTIPLY(tmp2, FIX_3_072711026); /* c1+c3+c5-c7 */ + tmp1 += z1 + z3; + tmp2 += z1 + z2; + + /* Final output stage: inputs are tmp10..tmp13, tmp0..tmp3 */ + + wsptr[DCTSIZE*0] = (int) RIGHT_SHIFT(tmp10 + tmp3, CONST_BITS-PASS1_BITS); + wsptr[DCTSIZE*7] = (int) RIGHT_SHIFT(tmp10 - tmp3, CONST_BITS-PASS1_BITS); + wsptr[DCTSIZE*1] = (int) RIGHT_SHIFT(tmp11 + tmp2, CONST_BITS-PASS1_BITS); + wsptr[DCTSIZE*6] = (int) RIGHT_SHIFT(tmp11 - tmp2, CONST_BITS-PASS1_BITS); + wsptr[DCTSIZE*2] = (int) RIGHT_SHIFT(tmp12 + tmp1, CONST_BITS-PASS1_BITS); + wsptr[DCTSIZE*5] = (int) RIGHT_SHIFT(tmp12 - tmp1, CONST_BITS-PASS1_BITS); + wsptr[DCTSIZE*3] = (int) RIGHT_SHIFT(tmp13 + tmp0, CONST_BITS-PASS1_BITS); + wsptr[DCTSIZE*4] = (int) RIGHT_SHIFT(tmp13 - tmp0, CONST_BITS-PASS1_BITS); + + inptr++; /* advance pointers to next column */ + quantptr++; + wsptr++; + } + + /* Pass 2: process 8 rows from work array, store into output array. + * 16-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/32). + */ + + wsptr = workspace; + for (ctr = 0; ctr < 8; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + tmp0 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + tmp0 <<= CONST_BITS; + + z1 = (INT32) wsptr[4]; + tmp1 = MULTIPLY(z1, FIX(1.306562965)); /* c4[16] = c2[8] */ + tmp2 = MULTIPLY(z1, FIX_0_541196100); /* c12[16] = c6[8] */ + + tmp10 = tmp0 + tmp1; + tmp11 = tmp0 - tmp1; + tmp12 = tmp0 + tmp2; + tmp13 = tmp0 - tmp2; + + z1 = (INT32) wsptr[2]; + z2 = (INT32) wsptr[6]; + z3 = z1 - z2; + z4 = MULTIPLY(z3, FIX(0.275899379)); /* c14[16] = c7[8] */ + z3 = MULTIPLY(z3, FIX(1.387039845)); /* c2[16] = c1[8] */ + + tmp0 = z3 + MULTIPLY(z2, FIX_2_562915447); /* (c6+c2)[16] = (c3+c1)[8] */ + tmp1 = z4 + MULTIPLY(z1, FIX_0_899976223); /* (c6-c14)[16] = (c3-c7)[8] */ + tmp2 = z3 - MULTIPLY(z1, FIX(0.601344887)); /* (c2-c10)[16] = (c1-c5)[8] */ + tmp3 = z4 - MULTIPLY(z2, FIX(0.509795579)); /* (c10-c14)[16] = (c5-c7)[8] */ + + tmp20 = tmp10 + tmp0; + tmp27 = tmp10 - tmp0; + tmp21 = tmp12 + tmp1; + tmp26 = tmp12 - tmp1; + tmp22 = tmp13 + tmp2; + tmp25 = tmp13 - tmp2; + tmp23 = tmp11 + tmp3; + tmp24 = tmp11 - tmp3; + + /* Odd part */ + + z1 = (INT32) wsptr[1]; + z2 = (INT32) wsptr[3]; + z3 = (INT32) wsptr[5]; + z4 = (INT32) wsptr[7]; + + tmp11 = z1 + z3; + + tmp1 = MULTIPLY(z1 + z2, FIX(1.353318001)); /* c3 */ + tmp2 = MULTIPLY(tmp11, FIX(1.247225013)); /* c5 */ + tmp3 = MULTIPLY(z1 + z4, FIX(1.093201867)); /* c7 */ + tmp10 = MULTIPLY(z1 - z4, FIX(0.897167586)); /* c9 */ + tmp11 = MULTIPLY(tmp11, FIX(0.666655658)); /* c11 */ + tmp12 = MULTIPLY(z1 - z2, FIX(0.410524528)); /* c13 */ + tmp0 = tmp1 + tmp2 + tmp3 - + MULTIPLY(z1, FIX(2.286341144)); /* c7+c5+c3-c1 */ + tmp13 = tmp10 + tmp11 + tmp12 - + MULTIPLY(z1, FIX(1.835730603)); /* c9+c11+c13-c15 */ + z1 = MULTIPLY(z2 + z3, FIX(0.138617169)); /* c15 */ + tmp1 += z1 + MULTIPLY(z2, FIX(0.071888074)); /* c9+c11-c3-c15 */ + tmp2 += z1 - MULTIPLY(z3, FIX(1.125726048)); /* c5+c7+c15-c3 */ + z1 = MULTIPLY(z3 - z2, FIX(1.407403738)); /* c1 */ + tmp11 += z1 - MULTIPLY(z3, FIX(0.766367282)); /* c1+c11-c9-c13 */ + tmp12 += z1 + MULTIPLY(z2, FIX(1.971951411)); /* c1+c5+c13-c7 */ + z2 += z4; + z1 = MULTIPLY(z2, - FIX(0.666655658)); /* -c11 */ + tmp1 += z1; + tmp3 += z1 + MULTIPLY(z4, FIX(1.065388962)); /* c3+c11+c15-c7 */ + z2 = MULTIPLY(z2, - FIX(1.247225013)); /* -c5 */ + tmp10 += z2 + MULTIPLY(z4, FIX(3.141271809)); /* c1+c5+c9-c13 */ + tmp12 += z2; + z2 = MULTIPLY(z3 + z4, - FIX(1.353318001)); /* -c3 */ + tmp2 += z2; + tmp3 += z2; + z2 = MULTIPLY(z4 - z3, FIX(0.410524528)); /* c13 */ + tmp10 += z2; + tmp11 += z2; + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[15] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp1, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[14] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp1, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[13] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23 + tmp3, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[12] = range_limit[(int) RIGHT_SHIFT(tmp23 - tmp3, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp24 + tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[11] = range_limit[(int) RIGHT_SHIFT(tmp24 - tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp25 + tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[10] = range_limit[(int) RIGHT_SHIFT(tmp25 - tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp26 + tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[9] = range_limit[(int) RIGHT_SHIFT(tmp26 - tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp27 + tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp27 - tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += 8; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a 14x7 output block. + * + * 7-point IDCT in pass 1 (columns), 14-point in pass 2 (rows). + */ + +GLOBAL(void) +jpeg_idct_14x7 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15, tmp16; + INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25, tmp26; + INT32 z1, z2, z3, z4; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[8*7]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. + * 7-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/14). + */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + tmp23 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + tmp23 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + tmp23 += ONE << (CONST_BITS-PASS1_BITS-1); + + z1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + z2 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + z3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); + + tmp20 = MULTIPLY(z2 - z3, FIX(0.881747734)); /* c4 */ + tmp22 = MULTIPLY(z1 - z2, FIX(0.314692123)); /* c6 */ + tmp21 = tmp20 + tmp22 + tmp23 - MULTIPLY(z2, FIX(1.841218003)); /* c2+c4-c6 */ + tmp10 = z1 + z3; + z2 -= tmp10; + tmp10 = MULTIPLY(tmp10, FIX(1.274162392)) + tmp23; /* c2 */ + tmp20 += tmp10 - MULTIPLY(z3, FIX(0.077722536)); /* c2-c4-c6 */ + tmp22 += tmp10 - MULTIPLY(z1, FIX(2.470602249)); /* c2+c4+c6 */ + tmp23 += MULTIPLY(z2, FIX(1.414213562)); /* c0 */ + + /* Odd part */ + + z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); + + tmp11 = MULTIPLY(z1 + z2, FIX(0.935414347)); /* (c3+c1-c5)/2 */ + tmp12 = MULTIPLY(z1 - z2, FIX(0.170262339)); /* (c3+c5-c1)/2 */ + tmp10 = tmp11 - tmp12; + tmp11 += tmp12; + tmp12 = MULTIPLY(z2 + z3, - FIX(1.378756276)); /* -c1 */ + tmp11 += tmp12; + z2 = MULTIPLY(z1 + z3, FIX(0.613604268)); /* c5 */ + tmp10 += z2; + tmp12 += z2 + MULTIPLY(z3, FIX(1.870828693)); /* c3+c1-c5 */ + + /* Final output stage */ + + wsptr[8*0] = (int) RIGHT_SHIFT(tmp20 + tmp10, CONST_BITS-PASS1_BITS); + wsptr[8*6] = (int) RIGHT_SHIFT(tmp20 - tmp10, CONST_BITS-PASS1_BITS); + wsptr[8*1] = (int) RIGHT_SHIFT(tmp21 + tmp11, CONST_BITS-PASS1_BITS); + wsptr[8*5] = (int) RIGHT_SHIFT(tmp21 - tmp11, CONST_BITS-PASS1_BITS); + wsptr[8*2] = (int) RIGHT_SHIFT(tmp22 + tmp12, CONST_BITS-PASS1_BITS); + wsptr[8*4] = (int) RIGHT_SHIFT(tmp22 - tmp12, CONST_BITS-PASS1_BITS); + wsptr[8*3] = (int) RIGHT_SHIFT(tmp23, CONST_BITS-PASS1_BITS); + } + + /* Pass 2: process 7 rows from work array, store into output array. + * 14-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/28). + */ + + wsptr = workspace; + for (ctr = 0; ctr < 7; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + z1 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + z1 <<= CONST_BITS; + z4 = (INT32) wsptr[4]; + z2 = MULTIPLY(z4, FIX(1.274162392)); /* c4 */ + z3 = MULTIPLY(z4, FIX(0.314692123)); /* c12 */ + z4 = MULTIPLY(z4, FIX(0.881747734)); /* c8 */ + + tmp10 = z1 + z2; + tmp11 = z1 + z3; + tmp12 = z1 - z4; + + tmp23 = z1 - ((z2 + z3 - z4) << 1); /* c0 = (c4+c12-c8)*2 */ + + z1 = (INT32) wsptr[2]; + z2 = (INT32) wsptr[6]; + + z3 = MULTIPLY(z1 + z2, FIX(1.105676686)); /* c6 */ + + tmp13 = z3 + MULTIPLY(z1, FIX(0.273079590)); /* c2-c6 */ + tmp14 = z3 - MULTIPLY(z2, FIX(1.719280954)); /* c6+c10 */ + tmp15 = MULTIPLY(z1, FIX(0.613604268)) - /* c10 */ + MULTIPLY(z2, FIX(1.378756276)); /* c2 */ + + tmp20 = tmp10 + tmp13; + tmp26 = tmp10 - tmp13; + tmp21 = tmp11 + tmp14; + tmp25 = tmp11 - tmp14; + tmp22 = tmp12 + tmp15; + tmp24 = tmp12 - tmp15; + + /* Odd part */ + + z1 = (INT32) wsptr[1]; + z2 = (INT32) wsptr[3]; + z3 = (INT32) wsptr[5]; + z4 = (INT32) wsptr[7]; + z4 <<= CONST_BITS; + + tmp14 = z1 + z3; + tmp11 = MULTIPLY(z1 + z2, FIX(1.334852607)); /* c3 */ + tmp12 = MULTIPLY(tmp14, FIX(1.197448846)); /* c5 */ + tmp10 = tmp11 + tmp12 + z4 - MULTIPLY(z1, FIX(1.126980169)); /* c3+c5-c1 */ + tmp14 = MULTIPLY(tmp14, FIX(0.752406978)); /* c9 */ + tmp16 = tmp14 - MULTIPLY(z1, FIX(1.061150426)); /* c9+c11-c13 */ + z1 -= z2; + tmp15 = MULTIPLY(z1, FIX(0.467085129)) - z4; /* c11 */ + tmp16 += tmp15; + tmp13 = MULTIPLY(z2 + z3, - FIX(0.158341681)) - z4; /* -c13 */ + tmp11 += tmp13 - MULTIPLY(z2, FIX(0.424103948)); /* c3-c9-c13 */ + tmp12 += tmp13 - MULTIPLY(z3, FIX(2.373959773)); /* c3+c5-c13 */ + tmp13 = MULTIPLY(z3 - z2, FIX(1.405321284)); /* c1 */ + tmp14 += tmp13 + z4 - MULTIPLY(z3, FIX(1.6906431334)); /* c1+c9-c11 */ + tmp15 += tmp13 + MULTIPLY(z2, FIX(0.674957567)); /* c1+c11-c5 */ + + tmp13 = ((z1 - z3) << CONST_BITS) + z4; + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[13] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[12] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[11] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23 + tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[10] = range_limit[(int) RIGHT_SHIFT(tmp23 - tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp24 + tmp14, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[9] = range_limit[(int) RIGHT_SHIFT(tmp24 - tmp14, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp25 + tmp15, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp25 - tmp15, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp26 + tmp16, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp26 - tmp16, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += 8; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a 12x6 output block. + * + * 6-point IDCT in pass 1 (columns), 12-point in pass 2 (rows). + */ + +GLOBAL(void) +jpeg_idct_12x6 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15; + INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25; + INT32 z1, z2, z3, z4; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[8*6]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. + * 6-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/12). + */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + tmp10 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + tmp10 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + tmp10 += ONE << (CONST_BITS-PASS1_BITS-1); + tmp12 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + tmp20 = MULTIPLY(tmp12, FIX(0.707106781)); /* c4 */ + tmp11 = tmp10 + tmp20; + tmp21 = RIGHT_SHIFT(tmp10 - tmp20 - tmp20, CONST_BITS-PASS1_BITS); + tmp20 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + tmp10 = MULTIPLY(tmp20, FIX(1.224744871)); /* c2 */ + tmp20 = tmp11 + tmp10; + tmp22 = tmp11 - tmp10; + + /* Odd part */ + + z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); + tmp11 = MULTIPLY(z1 + z3, FIX(0.366025404)); /* c5 */ + tmp10 = tmp11 + ((z1 + z2) << CONST_BITS); + tmp12 = tmp11 + ((z3 - z2) << CONST_BITS); + tmp11 = (z1 - z2 - z3) << PASS1_BITS; + + /* Final output stage */ + + wsptr[8*0] = (int) RIGHT_SHIFT(tmp20 + tmp10, CONST_BITS-PASS1_BITS); + wsptr[8*5] = (int) RIGHT_SHIFT(tmp20 - tmp10, CONST_BITS-PASS1_BITS); + wsptr[8*1] = (int) (tmp21 + tmp11); + wsptr[8*4] = (int) (tmp21 - tmp11); + wsptr[8*2] = (int) RIGHT_SHIFT(tmp22 + tmp12, CONST_BITS-PASS1_BITS); + wsptr[8*3] = (int) RIGHT_SHIFT(tmp22 - tmp12, CONST_BITS-PASS1_BITS); + } + + /* Pass 2: process 6 rows from work array, store into output array. + * 12-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/24). + */ + + wsptr = workspace; + for (ctr = 0; ctr < 6; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + z3 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + z3 <<= CONST_BITS; + + z4 = (INT32) wsptr[4]; + z4 = MULTIPLY(z4, FIX(1.224744871)); /* c4 */ + + tmp10 = z3 + z4; + tmp11 = z3 - z4; + + z1 = (INT32) wsptr[2]; + z4 = MULTIPLY(z1, FIX(1.366025404)); /* c2 */ + z1 <<= CONST_BITS; + z2 = (INT32) wsptr[6]; + z2 <<= CONST_BITS; + + tmp12 = z1 - z2; + + tmp21 = z3 + tmp12; + tmp24 = z3 - tmp12; + + tmp12 = z4 + z2; + + tmp20 = tmp10 + tmp12; + tmp25 = tmp10 - tmp12; + + tmp12 = z4 - z1 - z2; + + tmp22 = tmp11 + tmp12; + tmp23 = tmp11 - tmp12; + + /* Odd part */ + + z1 = (INT32) wsptr[1]; + z2 = (INT32) wsptr[3]; + z3 = (INT32) wsptr[5]; + z4 = (INT32) wsptr[7]; + + tmp11 = MULTIPLY(z2, FIX(1.306562965)); /* c3 */ + tmp14 = MULTIPLY(z2, - FIX_0_541196100); /* -c9 */ + + tmp10 = z1 + z3; + tmp15 = MULTIPLY(tmp10 + z4, FIX(0.860918669)); /* c7 */ + tmp12 = tmp15 + MULTIPLY(tmp10, FIX(0.261052384)); /* c5-c7 */ + tmp10 = tmp12 + tmp11 + MULTIPLY(z1, FIX(0.280143716)); /* c1-c5 */ + tmp13 = MULTIPLY(z3 + z4, - FIX(1.045510580)); /* -(c7+c11) */ + tmp12 += tmp13 + tmp14 - MULTIPLY(z3, FIX(1.478575242)); /* c1+c5-c7-c11 */ + tmp13 += tmp15 - tmp11 + MULTIPLY(z4, FIX(1.586706681)); /* c1+c11 */ + tmp15 += tmp14 - MULTIPLY(z1, FIX(0.676326758)) - /* c7-c11 */ + MULTIPLY(z4, FIX(1.982889723)); /* c5+c7 */ + + z1 -= z4; + z2 -= z3; + z3 = MULTIPLY(z1 + z2, FIX_0_541196100); /* c9 */ + tmp11 = z3 + MULTIPLY(z1, FIX_0_765366865); /* c3-c9 */ + tmp14 = z3 - MULTIPLY(z2, FIX_1_847759065); /* c3+c9 */ + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[11] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[10] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[9] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23 + tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp23 - tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp24 + tmp14, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp24 - tmp14, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp25 + tmp15, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp25 - tmp15, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += 8; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a 10x5 output block. + * + * 5-point IDCT in pass 1 (columns), 10-point in pass 2 (rows). + */ + +GLOBAL(void) +jpeg_idct_10x5 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp10, tmp11, tmp12, tmp13, tmp14; + INT32 tmp20, tmp21, tmp22, tmp23, tmp24; + INT32 z1, z2, z3, z4; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[8*5]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. + * 5-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/10). + */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + tmp12 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + tmp12 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + tmp12 += ONE << (CONST_BITS-PASS1_BITS-1); + tmp13 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + tmp14 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + z1 = MULTIPLY(tmp13 + tmp14, FIX(0.790569415)); /* (c2+c4)/2 */ + z2 = MULTIPLY(tmp13 - tmp14, FIX(0.353553391)); /* (c2-c4)/2 */ + z3 = tmp12 + z2; + tmp10 = z3 + z1; + tmp11 = z3 - z1; + tmp12 -= z2 << 2; + + /* Odd part */ + + z2 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + z3 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + + z1 = MULTIPLY(z2 + z3, FIX(0.831253876)); /* c3 */ + tmp13 = z1 + MULTIPLY(z2, FIX(0.513743148)); /* c1-c3 */ + tmp14 = z1 - MULTIPLY(z3, FIX(2.176250899)); /* c1+c3 */ + + /* Final output stage */ + + wsptr[8*0] = (int) RIGHT_SHIFT(tmp10 + tmp13, CONST_BITS-PASS1_BITS); + wsptr[8*4] = (int) RIGHT_SHIFT(tmp10 - tmp13, CONST_BITS-PASS1_BITS); + wsptr[8*1] = (int) RIGHT_SHIFT(tmp11 + tmp14, CONST_BITS-PASS1_BITS); + wsptr[8*3] = (int) RIGHT_SHIFT(tmp11 - tmp14, CONST_BITS-PASS1_BITS); + wsptr[8*2] = (int) RIGHT_SHIFT(tmp12, CONST_BITS-PASS1_BITS); + } + + /* Pass 2: process 5 rows from work array, store into output array. + * 10-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/20). + */ + + wsptr = workspace; + for (ctr = 0; ctr < 5; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + z3 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + z3 <<= CONST_BITS; + z4 = (INT32) wsptr[4]; + z1 = MULTIPLY(z4, FIX(1.144122806)); /* c4 */ + z2 = MULTIPLY(z4, FIX(0.437016024)); /* c8 */ + tmp10 = z3 + z1; + tmp11 = z3 - z2; + + tmp22 = z3 - ((z1 - z2) << 1); /* c0 = (c4-c8)*2 */ + + z2 = (INT32) wsptr[2]; + z3 = (INT32) wsptr[6]; + + z1 = MULTIPLY(z2 + z3, FIX(0.831253876)); /* c6 */ + tmp12 = z1 + MULTIPLY(z2, FIX(0.513743148)); /* c2-c6 */ + tmp13 = z1 - MULTIPLY(z3, FIX(2.176250899)); /* c2+c6 */ + + tmp20 = tmp10 + tmp12; + tmp24 = tmp10 - tmp12; + tmp21 = tmp11 + tmp13; + tmp23 = tmp11 - tmp13; + + /* Odd part */ + + z1 = (INT32) wsptr[1]; + z2 = (INT32) wsptr[3]; + z3 = (INT32) wsptr[5]; + z3 <<= CONST_BITS; + z4 = (INT32) wsptr[7]; + + tmp11 = z2 + z4; + tmp13 = z2 - z4; + + tmp12 = MULTIPLY(tmp13, FIX(0.309016994)); /* (c3-c7)/2 */ + + z2 = MULTIPLY(tmp11, FIX(0.951056516)); /* (c3+c7)/2 */ + z4 = z3 + tmp12; + + tmp10 = MULTIPLY(z1, FIX(1.396802247)) + z2 + z4; /* c1 */ + tmp14 = MULTIPLY(z1, FIX(0.221231742)) - z2 + z4; /* c9 */ + + z2 = MULTIPLY(tmp11, FIX(0.587785252)); /* (c1-c9)/2 */ + z4 = z3 - tmp12 - (tmp13 << (CONST_BITS - 1)); + + tmp12 = ((z1 - tmp13) << CONST_BITS) - z3; + + tmp11 = MULTIPLY(z1, FIX(1.260073511)) - z2 - z4; /* c3 */ + tmp13 = MULTIPLY(z1, FIX(0.642039522)) - z2 + z4; /* c7 */ + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[9] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23 + tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp23 - tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp24 + tmp14, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp24 - tmp14, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += 8; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing an 8x4 output block. + * + * 4-point IDCT in pass 1 (columns), 8-point in pass 2 (rows). + */ + +GLOBAL(void) +jpeg_idct_8x4 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3; + INT32 tmp10, tmp11, tmp12, tmp13; + INT32 z1, z2, z3; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[8*4]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. + * 4-point IDCT kernel, + * cK represents sqrt(2) * cos(K*pi/16) [refers to 8-point IDCT]. + */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + tmp2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + + tmp10 = (tmp0 + tmp2) << PASS1_BITS; + tmp12 = (tmp0 - tmp2) << PASS1_BITS; + + /* Odd part */ + /* Same rotation as in the even part of the 8x8 LL&M IDCT */ + + z2 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + z3 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + + z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ + /* Add fudge factor here for final descale. */ + z1 += ONE << (CONST_BITS-PASS1_BITS-1); + tmp0 = RIGHT_SHIFT(z1 + MULTIPLY(z2, FIX_0_765366865), /* c2-c6 */ + CONST_BITS-PASS1_BITS); + tmp2 = RIGHT_SHIFT(z1 - MULTIPLY(z3, FIX_1_847759065), /* c2+c6 */ + CONST_BITS-PASS1_BITS); + + /* Final output stage */ + + wsptr[8*0] = (int) (tmp10 + tmp0); + wsptr[8*3] = (int) (tmp10 - tmp0); + wsptr[8*1] = (int) (tmp12 + tmp2); + wsptr[8*2] = (int) (tmp12 - tmp2); + } + + /* Pass 2: process rows from work array, store into output array. + * Note that we must descale the results by a factor of 8 == 2**3, + * and also undo the PASS1_BITS scaling. + * 8-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/16). + */ + + wsptr = workspace; + for (ctr = 0; ctr < 4; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part: reverse the even part of the forward DCT. + * The rotator is c(-6). + */ + + /* Add range center and fudge factor for final descale and range-limit. */ + z2 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + z3 = (INT32) wsptr[4]; + + tmp0 = (z2 + z3) << CONST_BITS; + tmp1 = (z2 - z3) << CONST_BITS; + + z2 = (INT32) wsptr[2]; + z3 = (INT32) wsptr[6]; + + z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ + tmp2 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ + tmp3 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ + + tmp10 = tmp0 + tmp2; + tmp13 = tmp0 - tmp2; + tmp11 = tmp1 + tmp3; + tmp12 = tmp1 - tmp3; + + /* Odd part per figure 8; the matrix is unitary and hence its + * transpose is its inverse. i0..i3 are y7,y5,y3,y1 respectively. + */ + + tmp0 = (INT32) wsptr[7]; + tmp1 = (INT32) wsptr[5]; + tmp2 = (INT32) wsptr[3]; + tmp3 = (INT32) wsptr[1]; + + z2 = tmp0 + tmp2; + z3 = tmp1 + tmp3; + + z1 = MULTIPLY(z2 + z3, FIX_1_175875602); /* c3 */ + z2 = MULTIPLY(z2, - FIX_1_961570560); /* -c3-c5 */ + z3 = MULTIPLY(z3, - FIX_0_390180644); /* -c3+c5 */ + z2 += z1; + z3 += z1; + + z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ + tmp0 = MULTIPLY(tmp0, FIX_0_298631336); /* -c1+c3+c5-c7 */ + tmp3 = MULTIPLY(tmp3, FIX_1_501321110); /* c1+c3-c5-c7 */ + tmp0 += z1 + z2; + tmp3 += z1 + z3; + + z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ + tmp1 = MULTIPLY(tmp1, FIX_2_053119869); /* c1+c3-c5+c7 */ + tmp2 = MULTIPLY(tmp2, FIX_3_072711026); /* c1+c3+c5-c7 */ + tmp1 += z1 + z3; + tmp2 += z1 + z2; + + /* Final output stage: inputs are tmp10..tmp13, tmp0..tmp3 */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp3, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp3, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp11 + tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp11 - tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12 + tmp1, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp12 - tmp1, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp13 + tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp13 - tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += DCTSIZE; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a 6x3 output block. + * + * 3-point IDCT in pass 1 (columns), 6-point in pass 2 (rows). + */ + +GLOBAL(void) +jpeg_idct_6x3 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp0, tmp1, tmp2, tmp10, tmp11, tmp12; + INT32 z1, z2, z3; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[6*3]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. + * 3-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/6). + */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 6; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + tmp0 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + tmp0 += ONE << (CONST_BITS-PASS1_BITS-1); + tmp2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + tmp12 = MULTIPLY(tmp2, FIX(0.707106781)); /* c2 */ + tmp10 = tmp0 + tmp12; + tmp2 = tmp0 - tmp12 - tmp12; + + /* Odd part */ + + tmp12 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + tmp0 = MULTIPLY(tmp12, FIX(1.224744871)); /* c1 */ + + /* Final output stage */ + + wsptr[6*0] = (int) RIGHT_SHIFT(tmp10 + tmp0, CONST_BITS-PASS1_BITS); + wsptr[6*2] = (int) RIGHT_SHIFT(tmp10 - tmp0, CONST_BITS-PASS1_BITS); + wsptr[6*1] = (int) RIGHT_SHIFT(tmp2, CONST_BITS-PASS1_BITS); + } + + /* Pass 2: process 3 rows from work array, store into output array. + * 6-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/12). + */ + + wsptr = workspace; + for (ctr = 0; ctr < 3; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + tmp0 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + tmp0 <<= CONST_BITS; + tmp2 = (INT32) wsptr[4]; + tmp10 = MULTIPLY(tmp2, FIX(0.707106781)); /* c4 */ + tmp1 = tmp0 + tmp10; + tmp11 = tmp0 - tmp10 - tmp10; + tmp10 = (INT32) wsptr[2]; + tmp0 = MULTIPLY(tmp10, FIX(1.224744871)); /* c2 */ + tmp10 = tmp1 + tmp0; + tmp12 = tmp1 - tmp0; + + /* Odd part */ + + z1 = (INT32) wsptr[1]; + z2 = (INT32) wsptr[3]; + z3 = (INT32) wsptr[5]; + tmp1 = MULTIPLY(z1 + z3, FIX(0.366025404)); /* c5 */ + tmp0 = tmp1 + ((z1 + z2) << CONST_BITS); + tmp2 = tmp1 + ((z3 - z2) << CONST_BITS); + tmp1 = (z1 - z2 - z3) << CONST_BITS; + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp11 + tmp1, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp11 - tmp1, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12 + tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp12 - tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += 6; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a 4x2 output block. + * + * 2-point IDCT in pass 1 (columns), 4-point in pass 2 (rows). + */ + +GLOBAL(void) +jpeg_idct_4x2 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp0, tmp2, tmp10, tmp12; + INT32 z1, z2, z3; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + INT32 * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + INT32 workspace[4*2]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 4; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + tmp10 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + + /* Odd part */ + + tmp0 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + + /* Final output stage */ + + wsptr[4*0] = tmp10 + tmp0; + wsptr[4*1] = tmp10 - tmp0; + } + + /* Pass 2: process 2 rows from work array, store into output array. + * 4-point IDCT kernel, + * cK represents sqrt(2) * cos(K*pi/16) [refers to 8-point IDCT]. + */ + + wsptr = workspace; + for (ctr = 0; ctr < 2; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + tmp0 = wsptr[0] + ((((INT32) RANGE_CENTER) << 3) + (ONE << 2)); + tmp2 = wsptr[2]; + + tmp10 = (tmp0 + tmp2) << CONST_BITS; + tmp12 = (tmp0 - tmp2) << CONST_BITS; + + /* Odd part */ + /* Same rotation as in the even part of the 8x8 LL&M IDCT */ + + z2 = wsptr[1]; + z3 = wsptr[3]; + + z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ + tmp0 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ + tmp2 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp0, + CONST_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp0, + CONST_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp12 + tmp2, + CONST_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12 - tmp2, + CONST_BITS+3) + & RANGE_MASK]; + + wsptr += 4; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a 2x1 output block. + * + * 1-point IDCT in pass 1 (columns), 2-point in pass 2 (rows). + */ + +GLOBAL(void) +jpeg_idct_2x1 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + DCTELEM tmp0, tmp1; + ISLOW_MULT_TYPE * quantptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + ISHIFT_TEMPS + + /* Pass 1: empty. */ + + /* Pass 2: process 1 row from input, store into output array. */ + + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + outptr = output_buf[0] + output_col; + + /* Even part */ + + tmp0 = DEQUANTIZE(coef_block[0], quantptr[0]); + /* Add range center and fudge factor for final descale and range-limit. */ + tmp0 += (((DCTELEM) RANGE_CENTER) << 3) + (1 << 2); + + /* Odd part */ + + tmp1 = DEQUANTIZE(coef_block[1], quantptr[1]); + + /* Final output stage */ + + outptr[0] = range_limit[(int) IRIGHT_SHIFT(tmp0 + tmp1, 3) & RANGE_MASK]; + outptr[1] = range_limit[(int) IRIGHT_SHIFT(tmp0 - tmp1, 3) & RANGE_MASK]; +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing an 8x16 output block. + * + * 16-point IDCT in pass 1 (columns), 8-point in pass 2 (rows). + */ + +GLOBAL(void) +jpeg_idct_8x16 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3, tmp10, tmp11, tmp12, tmp13; + INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25, tmp26, tmp27; + INT32 z1, z2, z3, z4; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[8*16]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. + * 16-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/32). + */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + tmp0 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + tmp0 += ONE << (CONST_BITS-PASS1_BITS-1); + + z1 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + tmp1 = MULTIPLY(z1, FIX(1.306562965)); /* c4[16] = c2[8] */ + tmp2 = MULTIPLY(z1, FIX_0_541196100); /* c12[16] = c6[8] */ + + tmp10 = tmp0 + tmp1; + tmp11 = tmp0 - tmp1; + tmp12 = tmp0 + tmp2; + tmp13 = tmp0 - tmp2; + + z1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + z2 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); + z3 = z1 - z2; + z4 = MULTIPLY(z3, FIX(0.275899379)); /* c14[16] = c7[8] */ + z3 = MULTIPLY(z3, FIX(1.387039845)); /* c2[16] = c1[8] */ + + tmp0 = z3 + MULTIPLY(z2, FIX_2_562915447); /* (c6+c2)[16] = (c3+c1)[8] */ + tmp1 = z4 + MULTIPLY(z1, FIX_0_899976223); /* (c6-c14)[16] = (c3-c7)[8] */ + tmp2 = z3 - MULTIPLY(z1, FIX(0.601344887)); /* (c2-c10)[16] = (c1-c5)[8] */ + tmp3 = z4 - MULTIPLY(z2, FIX(0.509795579)); /* (c10-c14)[16] = (c5-c7)[8] */ + + tmp20 = tmp10 + tmp0; + tmp27 = tmp10 - tmp0; + tmp21 = tmp12 + tmp1; + tmp26 = tmp12 - tmp1; + tmp22 = tmp13 + tmp2; + tmp25 = tmp13 - tmp2; + tmp23 = tmp11 + tmp3; + tmp24 = tmp11 - tmp3; + + /* Odd part */ + + z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); + z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); + + tmp11 = z1 + z3; + + tmp1 = MULTIPLY(z1 + z2, FIX(1.353318001)); /* c3 */ + tmp2 = MULTIPLY(tmp11, FIX(1.247225013)); /* c5 */ + tmp3 = MULTIPLY(z1 + z4, FIX(1.093201867)); /* c7 */ + tmp10 = MULTIPLY(z1 - z4, FIX(0.897167586)); /* c9 */ + tmp11 = MULTIPLY(tmp11, FIX(0.666655658)); /* c11 */ + tmp12 = MULTIPLY(z1 - z2, FIX(0.410524528)); /* c13 */ + tmp0 = tmp1 + tmp2 + tmp3 - + MULTIPLY(z1, FIX(2.286341144)); /* c7+c5+c3-c1 */ + tmp13 = tmp10 + tmp11 + tmp12 - + MULTIPLY(z1, FIX(1.835730603)); /* c9+c11+c13-c15 */ + z1 = MULTIPLY(z2 + z3, FIX(0.138617169)); /* c15 */ + tmp1 += z1 + MULTIPLY(z2, FIX(0.071888074)); /* c9+c11-c3-c15 */ + tmp2 += z1 - MULTIPLY(z3, FIX(1.125726048)); /* c5+c7+c15-c3 */ + z1 = MULTIPLY(z3 - z2, FIX(1.407403738)); /* c1 */ + tmp11 += z1 - MULTIPLY(z3, FIX(0.766367282)); /* c1+c11-c9-c13 */ + tmp12 += z1 + MULTIPLY(z2, FIX(1.971951411)); /* c1+c5+c13-c7 */ + z2 += z4; + z1 = MULTIPLY(z2, - FIX(0.666655658)); /* -c11 */ + tmp1 += z1; + tmp3 += z1 + MULTIPLY(z4, FIX(1.065388962)); /* c3+c11+c15-c7 */ + z2 = MULTIPLY(z2, - FIX(1.247225013)); /* -c5 */ + tmp10 += z2 + MULTIPLY(z4, FIX(3.141271809)); /* c1+c5+c9-c13 */ + tmp12 += z2; + z2 = MULTIPLY(z3 + z4, - FIX(1.353318001)); /* -c3 */ + tmp2 += z2; + tmp3 += z2; + z2 = MULTIPLY(z4 - z3, FIX(0.410524528)); /* c13 */ + tmp10 += z2; + tmp11 += z2; + + /* Final output stage */ + + wsptr[8*0] = (int) RIGHT_SHIFT(tmp20 + tmp0, CONST_BITS-PASS1_BITS); + wsptr[8*15] = (int) RIGHT_SHIFT(tmp20 - tmp0, CONST_BITS-PASS1_BITS); + wsptr[8*1] = (int) RIGHT_SHIFT(tmp21 + tmp1, CONST_BITS-PASS1_BITS); + wsptr[8*14] = (int) RIGHT_SHIFT(tmp21 - tmp1, CONST_BITS-PASS1_BITS); + wsptr[8*2] = (int) RIGHT_SHIFT(tmp22 + tmp2, CONST_BITS-PASS1_BITS); + wsptr[8*13] = (int) RIGHT_SHIFT(tmp22 - tmp2, CONST_BITS-PASS1_BITS); + wsptr[8*3] = (int) RIGHT_SHIFT(tmp23 + tmp3, CONST_BITS-PASS1_BITS); + wsptr[8*12] = (int) RIGHT_SHIFT(tmp23 - tmp3, CONST_BITS-PASS1_BITS); + wsptr[8*4] = (int) RIGHT_SHIFT(tmp24 + tmp10, CONST_BITS-PASS1_BITS); + wsptr[8*11] = (int) RIGHT_SHIFT(tmp24 - tmp10, CONST_BITS-PASS1_BITS); + wsptr[8*5] = (int) RIGHT_SHIFT(tmp25 + tmp11, CONST_BITS-PASS1_BITS); + wsptr[8*10] = (int) RIGHT_SHIFT(tmp25 - tmp11, CONST_BITS-PASS1_BITS); + wsptr[8*6] = (int) RIGHT_SHIFT(tmp26 + tmp12, CONST_BITS-PASS1_BITS); + wsptr[8*9] = (int) RIGHT_SHIFT(tmp26 - tmp12, CONST_BITS-PASS1_BITS); + wsptr[8*7] = (int) RIGHT_SHIFT(tmp27 + tmp13, CONST_BITS-PASS1_BITS); + wsptr[8*8] = (int) RIGHT_SHIFT(tmp27 - tmp13, CONST_BITS-PASS1_BITS); + } + + /* Pass 2: process rows from work array, store into output array. + * Note that we must descale the results by a factor of 8 == 2**3, + * and also undo the PASS1_BITS scaling. + * 8-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/16). + */ + + wsptr = workspace; + for (ctr = 0; ctr < 16; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part: reverse the even part of the forward DCT. + * The rotator is c(-6). + */ + + /* Add range center and fudge factor for final descale and range-limit. */ + z2 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + z3 = (INT32) wsptr[4]; + + tmp0 = (z2 + z3) << CONST_BITS; + tmp1 = (z2 - z3) << CONST_BITS; + + z2 = (INT32) wsptr[2]; + z3 = (INT32) wsptr[6]; + + z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ + tmp2 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ + tmp3 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ + + tmp10 = tmp0 + tmp2; + tmp13 = tmp0 - tmp2; + tmp11 = tmp1 + tmp3; + tmp12 = tmp1 - tmp3; + + /* Odd part per figure 8; the matrix is unitary and hence its + * transpose is its inverse. i0..i3 are y7,y5,y3,y1 respectively. + */ + + tmp0 = (INT32) wsptr[7]; + tmp1 = (INT32) wsptr[5]; + tmp2 = (INT32) wsptr[3]; + tmp3 = (INT32) wsptr[1]; + + z2 = tmp0 + tmp2; + z3 = tmp1 + tmp3; + + z1 = MULTIPLY(z2 + z3, FIX_1_175875602); /* c3 */ + z2 = MULTIPLY(z2, - FIX_1_961570560); /* -c3-c5 */ + z3 = MULTIPLY(z3, - FIX_0_390180644); /* -c3+c5 */ + z2 += z1; + z3 += z1; + + z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ + tmp0 = MULTIPLY(tmp0, FIX_0_298631336); /* -c1+c3+c5-c7 */ + tmp3 = MULTIPLY(tmp3, FIX_1_501321110); /* c1+c3-c5-c7 */ + tmp0 += z1 + z2; + tmp3 += z1 + z3; + + z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ + tmp1 = MULTIPLY(tmp1, FIX_2_053119869); /* c1+c3-c5+c7 */ + tmp2 = MULTIPLY(tmp2, FIX_3_072711026); /* c1+c3+c5-c7 */ + tmp1 += z1 + z3; + tmp2 += z1 + z2; + + /* Final output stage: inputs are tmp10..tmp13, tmp0..tmp3 */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp3, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp3, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp11 + tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp11 - tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12 + tmp1, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp12 - tmp1, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp13 + tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp13 - tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += DCTSIZE; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a 7x14 output block. + * + * 14-point IDCT in pass 1 (columns), 7-point in pass 2 (rows). + */ + +GLOBAL(void) +jpeg_idct_7x14 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15, tmp16; + INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25, tmp26; + INT32 z1, z2, z3, z4; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[7*14]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. + * 14-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/28). + */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 7; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + z1 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + z1 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + z1 += ONE << (CONST_BITS-PASS1_BITS-1); + z4 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + z2 = MULTIPLY(z4, FIX(1.274162392)); /* c4 */ + z3 = MULTIPLY(z4, FIX(0.314692123)); /* c12 */ + z4 = MULTIPLY(z4, FIX(0.881747734)); /* c8 */ + + tmp10 = z1 + z2; + tmp11 = z1 + z3; + tmp12 = z1 - z4; + + tmp23 = RIGHT_SHIFT(z1 - ((z2 + z3 - z4) << 1), /* c0 = (c4+c12-c8)*2 */ + CONST_BITS-PASS1_BITS); + + z1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + z2 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); + + z3 = MULTIPLY(z1 + z2, FIX(1.105676686)); /* c6 */ + + tmp13 = z3 + MULTIPLY(z1, FIX(0.273079590)); /* c2-c6 */ + tmp14 = z3 - MULTIPLY(z2, FIX(1.719280954)); /* c6+c10 */ + tmp15 = MULTIPLY(z1, FIX(0.613604268)) - /* c10 */ + MULTIPLY(z2, FIX(1.378756276)); /* c2 */ + + tmp20 = tmp10 + tmp13; + tmp26 = tmp10 - tmp13; + tmp21 = tmp11 + tmp14; + tmp25 = tmp11 - tmp14; + tmp22 = tmp12 + tmp15; + tmp24 = tmp12 - tmp15; + + /* Odd part */ + + z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); + z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); + tmp13 = z4 << CONST_BITS; + + tmp14 = z1 + z3; + tmp11 = MULTIPLY(z1 + z2, FIX(1.334852607)); /* c3 */ + tmp12 = MULTIPLY(tmp14, FIX(1.197448846)); /* c5 */ + tmp10 = tmp11 + tmp12 + tmp13 - MULTIPLY(z1, FIX(1.126980169)); /* c3+c5-c1 */ + tmp14 = MULTIPLY(tmp14, FIX(0.752406978)); /* c9 */ + tmp16 = tmp14 - MULTIPLY(z1, FIX(1.061150426)); /* c9+c11-c13 */ + z1 -= z2; + tmp15 = MULTIPLY(z1, FIX(0.467085129)) - tmp13; /* c11 */ + tmp16 += tmp15; + z1 += z4; + z4 = MULTIPLY(z2 + z3, - FIX(0.158341681)) - tmp13; /* -c13 */ + tmp11 += z4 - MULTIPLY(z2, FIX(0.424103948)); /* c3-c9-c13 */ + tmp12 += z4 - MULTIPLY(z3, FIX(2.373959773)); /* c3+c5-c13 */ + z4 = MULTIPLY(z3 - z2, FIX(1.405321284)); /* c1 */ + tmp14 += z4 + tmp13 - MULTIPLY(z3, FIX(1.6906431334)); /* c1+c9-c11 */ + tmp15 += z4 + MULTIPLY(z2, FIX(0.674957567)); /* c1+c11-c5 */ + + tmp13 = (z1 - z3) << PASS1_BITS; + + /* Final output stage */ + + wsptr[7*0] = (int) RIGHT_SHIFT(tmp20 + tmp10, CONST_BITS-PASS1_BITS); + wsptr[7*13] = (int) RIGHT_SHIFT(tmp20 - tmp10, CONST_BITS-PASS1_BITS); + wsptr[7*1] = (int) RIGHT_SHIFT(tmp21 + tmp11, CONST_BITS-PASS1_BITS); + wsptr[7*12] = (int) RIGHT_SHIFT(tmp21 - tmp11, CONST_BITS-PASS1_BITS); + wsptr[7*2] = (int) RIGHT_SHIFT(tmp22 + tmp12, CONST_BITS-PASS1_BITS); + wsptr[7*11] = (int) RIGHT_SHIFT(tmp22 - tmp12, CONST_BITS-PASS1_BITS); + wsptr[7*3] = (int) (tmp23 + tmp13); + wsptr[7*10] = (int) (tmp23 - tmp13); + wsptr[7*4] = (int) RIGHT_SHIFT(tmp24 + tmp14, CONST_BITS-PASS1_BITS); + wsptr[7*9] = (int) RIGHT_SHIFT(tmp24 - tmp14, CONST_BITS-PASS1_BITS); + wsptr[7*5] = (int) RIGHT_SHIFT(tmp25 + tmp15, CONST_BITS-PASS1_BITS); + wsptr[7*8] = (int) RIGHT_SHIFT(tmp25 - tmp15, CONST_BITS-PASS1_BITS); + wsptr[7*6] = (int) RIGHT_SHIFT(tmp26 + tmp16, CONST_BITS-PASS1_BITS); + wsptr[7*7] = (int) RIGHT_SHIFT(tmp26 - tmp16, CONST_BITS-PASS1_BITS); + } + + /* Pass 2: process 14 rows from work array, store into output array. + * 7-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/14). + */ + + wsptr = workspace; + for (ctr = 0; ctr < 14; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + tmp23 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + tmp23 <<= CONST_BITS; + + z1 = (INT32) wsptr[2]; + z2 = (INT32) wsptr[4]; + z3 = (INT32) wsptr[6]; + + tmp20 = MULTIPLY(z2 - z3, FIX(0.881747734)); /* c4 */ + tmp22 = MULTIPLY(z1 - z2, FIX(0.314692123)); /* c6 */ + tmp21 = tmp20 + tmp22 + tmp23 - MULTIPLY(z2, FIX(1.841218003)); /* c2+c4-c6 */ + tmp10 = z1 + z3; + z2 -= tmp10; + tmp10 = MULTIPLY(tmp10, FIX(1.274162392)) + tmp23; /* c2 */ + tmp20 += tmp10 - MULTIPLY(z3, FIX(0.077722536)); /* c2-c4-c6 */ + tmp22 += tmp10 - MULTIPLY(z1, FIX(2.470602249)); /* c2+c4+c6 */ + tmp23 += MULTIPLY(z2, FIX(1.414213562)); /* c0 */ + + /* Odd part */ + + z1 = (INT32) wsptr[1]; + z2 = (INT32) wsptr[3]; + z3 = (INT32) wsptr[5]; + + tmp11 = MULTIPLY(z1 + z2, FIX(0.935414347)); /* (c3+c1-c5)/2 */ + tmp12 = MULTIPLY(z1 - z2, FIX(0.170262339)); /* (c3+c5-c1)/2 */ + tmp10 = tmp11 - tmp12; + tmp11 += tmp12; + tmp12 = MULTIPLY(z2 + z3, - FIX(1.378756276)); /* -c1 */ + tmp11 += tmp12; + z2 = MULTIPLY(z1 + z3, FIX(0.613604268)); /* c5 */ + tmp10 += z2; + tmp12 += z2 + MULTIPLY(z3, FIX(1.870828693)); /* c3+c1-c5 */ + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += 7; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a 6x12 output block. + * + * 12-point IDCT in pass 1 (columns), 6-point in pass 2 (rows). + */ + +GLOBAL(void) +jpeg_idct_6x12 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15; + INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25; + INT32 z1, z2, z3, z4; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[6*12]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. + * 12-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/24). + */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 6; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + z3 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + z3 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + z3 += ONE << (CONST_BITS-PASS1_BITS-1); + + z4 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + z4 = MULTIPLY(z4, FIX(1.224744871)); /* c4 */ + + tmp10 = z3 + z4; + tmp11 = z3 - z4; + + z1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + z4 = MULTIPLY(z1, FIX(1.366025404)); /* c2 */ + z1 <<= CONST_BITS; + z2 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); + z2 <<= CONST_BITS; + + tmp12 = z1 - z2; + + tmp21 = z3 + tmp12; + tmp24 = z3 - tmp12; + + tmp12 = z4 + z2; + + tmp20 = tmp10 + tmp12; + tmp25 = tmp10 - tmp12; + + tmp12 = z4 - z1 - z2; + + tmp22 = tmp11 + tmp12; + tmp23 = tmp11 - tmp12; + + /* Odd part */ + + z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); + z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); + + tmp11 = MULTIPLY(z2, FIX(1.306562965)); /* c3 */ + tmp14 = MULTIPLY(z2, - FIX_0_541196100); /* -c9 */ + + tmp10 = z1 + z3; + tmp15 = MULTIPLY(tmp10 + z4, FIX(0.860918669)); /* c7 */ + tmp12 = tmp15 + MULTIPLY(tmp10, FIX(0.261052384)); /* c5-c7 */ + tmp10 = tmp12 + tmp11 + MULTIPLY(z1, FIX(0.280143716)); /* c1-c5 */ + tmp13 = MULTIPLY(z3 + z4, - FIX(1.045510580)); /* -(c7+c11) */ + tmp12 += tmp13 + tmp14 - MULTIPLY(z3, FIX(1.478575242)); /* c1+c5-c7-c11 */ + tmp13 += tmp15 - tmp11 + MULTIPLY(z4, FIX(1.586706681)); /* c1+c11 */ + tmp15 += tmp14 - MULTIPLY(z1, FIX(0.676326758)) - /* c7-c11 */ + MULTIPLY(z4, FIX(1.982889723)); /* c5+c7 */ + + z1 -= z4; + z2 -= z3; + z3 = MULTIPLY(z1 + z2, FIX_0_541196100); /* c9 */ + tmp11 = z3 + MULTIPLY(z1, FIX_0_765366865); /* c3-c9 */ + tmp14 = z3 - MULTIPLY(z2, FIX_1_847759065); /* c3+c9 */ + + /* Final output stage */ + + wsptr[6*0] = (int) RIGHT_SHIFT(tmp20 + tmp10, CONST_BITS-PASS1_BITS); + wsptr[6*11] = (int) RIGHT_SHIFT(tmp20 - tmp10, CONST_BITS-PASS1_BITS); + wsptr[6*1] = (int) RIGHT_SHIFT(tmp21 + tmp11, CONST_BITS-PASS1_BITS); + wsptr[6*10] = (int) RIGHT_SHIFT(tmp21 - tmp11, CONST_BITS-PASS1_BITS); + wsptr[6*2] = (int) RIGHT_SHIFT(tmp22 + tmp12, CONST_BITS-PASS1_BITS); + wsptr[6*9] = (int) RIGHT_SHIFT(tmp22 - tmp12, CONST_BITS-PASS1_BITS); + wsptr[6*3] = (int) RIGHT_SHIFT(tmp23 + tmp13, CONST_BITS-PASS1_BITS); + wsptr[6*8] = (int) RIGHT_SHIFT(tmp23 - tmp13, CONST_BITS-PASS1_BITS); + wsptr[6*4] = (int) RIGHT_SHIFT(tmp24 + tmp14, CONST_BITS-PASS1_BITS); + wsptr[6*7] = (int) RIGHT_SHIFT(tmp24 - tmp14, CONST_BITS-PASS1_BITS); + wsptr[6*5] = (int) RIGHT_SHIFT(tmp25 + tmp15, CONST_BITS-PASS1_BITS); + wsptr[6*6] = (int) RIGHT_SHIFT(tmp25 - tmp15, CONST_BITS-PASS1_BITS); + } + + /* Pass 2: process 12 rows from work array, store into output array. + * 6-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/12). + */ + + wsptr = workspace; + for (ctr = 0; ctr < 12; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + tmp10 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + tmp10 <<= CONST_BITS; + tmp12 = (INT32) wsptr[4]; + tmp20 = MULTIPLY(tmp12, FIX(0.707106781)); /* c4 */ + tmp11 = tmp10 + tmp20; + tmp21 = tmp10 - tmp20 - tmp20; + tmp20 = (INT32) wsptr[2]; + tmp10 = MULTIPLY(tmp20, FIX(1.224744871)); /* c2 */ + tmp20 = tmp11 + tmp10; + tmp22 = tmp11 - tmp10; + + /* Odd part */ + + z1 = (INT32) wsptr[1]; + z2 = (INT32) wsptr[3]; + z3 = (INT32) wsptr[5]; + tmp11 = MULTIPLY(z1 + z3, FIX(0.366025404)); /* c5 */ + tmp10 = tmp11 + ((z1 + z2) << CONST_BITS); + tmp12 = tmp11 + ((z3 - z2) << CONST_BITS); + tmp11 = (z1 - z2 - z3) << CONST_BITS; + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp10, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp11, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += 6; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a 5x10 output block. + * + * 10-point IDCT in pass 1 (columns), 5-point in pass 2 (rows). + */ + +GLOBAL(void) +jpeg_idct_5x10 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp10, tmp11, tmp12, tmp13, tmp14; + INT32 tmp20, tmp21, tmp22, tmp23, tmp24; + INT32 z1, z2, z3, z4, z5; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[5*10]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. + * 10-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/20). + */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 5; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + z3 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + z3 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + z3 += ONE << (CONST_BITS-PASS1_BITS-1); + z4 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + z1 = MULTIPLY(z4, FIX(1.144122806)); /* c4 */ + z2 = MULTIPLY(z4, FIX(0.437016024)); /* c8 */ + tmp10 = z3 + z1; + tmp11 = z3 - z2; + + tmp22 = RIGHT_SHIFT(z3 - ((z1 - z2) << 1), /* c0 = (c4-c8)*2 */ + CONST_BITS-PASS1_BITS); + + z2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + z3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); + + z1 = MULTIPLY(z2 + z3, FIX(0.831253876)); /* c6 */ + tmp12 = z1 + MULTIPLY(z2, FIX(0.513743148)); /* c2-c6 */ + tmp13 = z1 - MULTIPLY(z3, FIX(2.176250899)); /* c2+c6 */ + + tmp20 = tmp10 + tmp12; + tmp24 = tmp10 - tmp12; + tmp21 = tmp11 + tmp13; + tmp23 = tmp11 - tmp13; + + /* Odd part */ + + z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); + z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); + + tmp11 = z2 + z4; + tmp13 = z2 - z4; + + tmp12 = MULTIPLY(tmp13, FIX(0.309016994)); /* (c3-c7)/2 */ + z5 = z3 << CONST_BITS; + + z2 = MULTIPLY(tmp11, FIX(0.951056516)); /* (c3+c7)/2 */ + z4 = z5 + tmp12; + + tmp10 = MULTIPLY(z1, FIX(1.396802247)) + z2 + z4; /* c1 */ + tmp14 = MULTIPLY(z1, FIX(0.221231742)) - z2 + z4; /* c9 */ + + z2 = MULTIPLY(tmp11, FIX(0.587785252)); /* (c1-c9)/2 */ + z4 = z5 - tmp12 - (tmp13 << (CONST_BITS - 1)); + + tmp12 = (z1 - tmp13 - z3) << PASS1_BITS; + + tmp11 = MULTIPLY(z1, FIX(1.260073511)) - z2 - z4; /* c3 */ + tmp13 = MULTIPLY(z1, FIX(0.642039522)) - z2 + z4; /* c7 */ + + /* Final output stage */ + + wsptr[5*0] = (int) RIGHT_SHIFT(tmp20 + tmp10, CONST_BITS-PASS1_BITS); + wsptr[5*9] = (int) RIGHT_SHIFT(tmp20 - tmp10, CONST_BITS-PASS1_BITS); + wsptr[5*1] = (int) RIGHT_SHIFT(tmp21 + tmp11, CONST_BITS-PASS1_BITS); + wsptr[5*8] = (int) RIGHT_SHIFT(tmp21 - tmp11, CONST_BITS-PASS1_BITS); + wsptr[5*2] = (int) (tmp22 + tmp12); + wsptr[5*7] = (int) (tmp22 - tmp12); + wsptr[5*3] = (int) RIGHT_SHIFT(tmp23 + tmp13, CONST_BITS-PASS1_BITS); + wsptr[5*6] = (int) RIGHT_SHIFT(tmp23 - tmp13, CONST_BITS-PASS1_BITS); + wsptr[5*4] = (int) RIGHT_SHIFT(tmp24 + tmp14, CONST_BITS-PASS1_BITS); + wsptr[5*5] = (int) RIGHT_SHIFT(tmp24 - tmp14, CONST_BITS-PASS1_BITS); + } + + /* Pass 2: process 10 rows from work array, store into output array. + * 5-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/10). + */ + + wsptr = workspace; + for (ctr = 0; ctr < 10; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + tmp12 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + tmp12 <<= CONST_BITS; + tmp13 = (INT32) wsptr[2]; + tmp14 = (INT32) wsptr[4]; + z1 = MULTIPLY(tmp13 + tmp14, FIX(0.790569415)); /* (c2+c4)/2 */ + z2 = MULTIPLY(tmp13 - tmp14, FIX(0.353553391)); /* (c2-c4)/2 */ + z3 = tmp12 + z2; + tmp10 = z3 + z1; + tmp11 = z3 - z1; + tmp12 -= z2 << 2; + + /* Odd part */ + + z2 = (INT32) wsptr[1]; + z3 = (INT32) wsptr[3]; + + z1 = MULTIPLY(z2 + z3, FIX(0.831253876)); /* c3 */ + tmp13 = z1 + MULTIPLY(z2, FIX(0.513743148)); /* c1-c3 */ + tmp14 = z1 - MULTIPLY(z3, FIX(2.176250899)); /* c1+c3 */ + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp13, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp11 + tmp14, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp11 - tmp14, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += 5; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a 4x8 output block. + * + * 8-point IDCT in pass 1 (columns), 4-point in pass 2 (rows). + */ + +GLOBAL(void) +jpeg_idct_4x8 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp0, tmp1, tmp2, tmp3; + INT32 tmp10, tmp11, tmp12, tmp13; + INT32 z1, z2, z3; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[4*8]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. + * Note results are scaled up by sqrt(8) compared to a true IDCT; + * furthermore, we scale the results by 2**PASS1_BITS. + * 8-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/16). + */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 4; ctr > 0; ctr--) { + /* Due to quantization, we will usually find that many of the input + * coefficients are zero, especially the AC terms. We can exploit this + * by short-circuiting the IDCT calculation for any column in which all + * the AC terms are zero. In that case each output is equal to the + * DC coefficient (with scale factor as needed). + * With typical images and quantization tables, half or more of the + * column DCT calculations can be simplified this way. + */ + + if (inptr[DCTSIZE*1] == 0 && inptr[DCTSIZE*2] == 0 && + inptr[DCTSIZE*3] == 0 && inptr[DCTSIZE*4] == 0 && + inptr[DCTSIZE*5] == 0 && inptr[DCTSIZE*6] == 0 && + inptr[DCTSIZE*7] == 0) { + /* AC terms all zero */ + int dcval = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]) << PASS1_BITS; + + wsptr[4*0] = dcval; + wsptr[4*1] = dcval; + wsptr[4*2] = dcval; + wsptr[4*3] = dcval; + wsptr[4*4] = dcval; + wsptr[4*5] = dcval; + wsptr[4*6] = dcval; + wsptr[4*7] = dcval; + + inptr++; /* advance pointers to next column */ + quantptr++; + wsptr++; + continue; + } + + /* Even part: reverse the even part of the forward DCT. + * The rotator is c(-6). + */ + + z2 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + z3 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + z2 <<= CONST_BITS; + z3 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + z2 += ONE << (CONST_BITS-PASS1_BITS-1); + + tmp0 = z2 + z3; + tmp1 = z2 - z3; + + z2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + z3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); + + z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ + tmp2 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ + tmp3 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ + + tmp10 = tmp0 + tmp2; + tmp13 = tmp0 - tmp2; + tmp11 = tmp1 + tmp3; + tmp12 = tmp1 - tmp3; + + /* Odd part per figure 8; the matrix is unitary and hence its + * transpose is its inverse. i0..i3 are y7,y5,y3,y1 respectively. + */ + + tmp0 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); + tmp1 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); + tmp2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + tmp3 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + + z2 = tmp0 + tmp2; + z3 = tmp1 + tmp3; + + z1 = MULTIPLY(z2 + z3, FIX_1_175875602); /* c3 */ + z2 = MULTIPLY(z2, - FIX_1_961570560); /* -c3-c5 */ + z3 = MULTIPLY(z3, - FIX_0_390180644); /* -c3+c5 */ + z2 += z1; + z3 += z1; + + z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ + tmp0 = MULTIPLY(tmp0, FIX_0_298631336); /* -c1+c3+c5-c7 */ + tmp3 = MULTIPLY(tmp3, FIX_1_501321110); /* c1+c3-c5-c7 */ + tmp0 += z1 + z2; + tmp3 += z1 + z3; + + z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ + tmp1 = MULTIPLY(tmp1, FIX_2_053119869); /* c1+c3-c5+c7 */ + tmp2 = MULTIPLY(tmp2, FIX_3_072711026); /* c1+c3+c5-c7 */ + tmp1 += z1 + z3; + tmp2 += z1 + z2; + + /* Final output stage: inputs are tmp10..tmp13, tmp0..tmp3 */ + + wsptr[4*0] = (int) RIGHT_SHIFT(tmp10 + tmp3, CONST_BITS-PASS1_BITS); + wsptr[4*7] = (int) RIGHT_SHIFT(tmp10 - tmp3, CONST_BITS-PASS1_BITS); + wsptr[4*1] = (int) RIGHT_SHIFT(tmp11 + tmp2, CONST_BITS-PASS1_BITS); + wsptr[4*6] = (int) RIGHT_SHIFT(tmp11 - tmp2, CONST_BITS-PASS1_BITS); + wsptr[4*2] = (int) RIGHT_SHIFT(tmp12 + tmp1, CONST_BITS-PASS1_BITS); + wsptr[4*5] = (int) RIGHT_SHIFT(tmp12 - tmp1, CONST_BITS-PASS1_BITS); + wsptr[4*3] = (int) RIGHT_SHIFT(tmp13 + tmp0, CONST_BITS-PASS1_BITS); + wsptr[4*4] = (int) RIGHT_SHIFT(tmp13 - tmp0, CONST_BITS-PASS1_BITS); + + inptr++; /* advance pointers to next column */ + quantptr++; + wsptr++; + } + + /* Pass 2: process 8 rows from work array, store into output array. + * 4-point IDCT kernel, + * cK represents sqrt(2) * cos(K*pi/16) [refers to 8-point IDCT]. + */ + + wsptr = workspace; + for (ctr = 0; ctr < 8; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + tmp0 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + tmp2 = (INT32) wsptr[2]; + + tmp10 = (tmp0 + tmp2) << CONST_BITS; + tmp12 = (tmp0 - tmp2) << CONST_BITS; + + /* Odd part */ + /* Same rotation as in the even part of the 8x8 LL&M IDCT */ + + z2 = (INT32) wsptr[1]; + z3 = (INT32) wsptr[3]; + + z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ + tmp0 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ + tmp2 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp12 + tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12 - tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += 4; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a 3x6 output block. + * + * 6-point IDCT in pass 1 (columns), 3-point in pass 2 (rows). + */ + +GLOBAL(void) +jpeg_idct_3x6 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp0, tmp1, tmp2, tmp10, tmp11, tmp12; + INT32 z1, z2, z3; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + int * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + int workspace[3*6]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. + * 6-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/12). + */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 3; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + tmp0 <<= CONST_BITS; + /* Add fudge factor here for final descale. */ + tmp0 += ONE << (CONST_BITS-PASS1_BITS-1); + tmp2 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); + tmp10 = MULTIPLY(tmp2, FIX(0.707106781)); /* c4 */ + tmp1 = tmp0 + tmp10; + tmp11 = RIGHT_SHIFT(tmp0 - tmp10 - tmp10, CONST_BITS-PASS1_BITS); + tmp10 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + tmp0 = MULTIPLY(tmp10, FIX(1.224744871)); /* c2 */ + tmp10 = tmp1 + tmp0; + tmp12 = tmp1 - tmp0; + + /* Odd part */ + + z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); + tmp1 = MULTIPLY(z1 + z3, FIX(0.366025404)); /* c5 */ + tmp0 = tmp1 + ((z1 + z2) << CONST_BITS); + tmp2 = tmp1 + ((z3 - z2) << CONST_BITS); + tmp1 = (z1 - z2 - z3) << PASS1_BITS; + + /* Final output stage */ + + wsptr[3*0] = (int) RIGHT_SHIFT(tmp10 + tmp0, CONST_BITS-PASS1_BITS); + wsptr[3*5] = (int) RIGHT_SHIFT(tmp10 - tmp0, CONST_BITS-PASS1_BITS); + wsptr[3*1] = (int) (tmp11 + tmp1); + wsptr[3*4] = (int) (tmp11 - tmp1); + wsptr[3*2] = (int) RIGHT_SHIFT(tmp12 + tmp2, CONST_BITS-PASS1_BITS); + wsptr[3*3] = (int) RIGHT_SHIFT(tmp12 - tmp2, CONST_BITS-PASS1_BITS); + } + + /* Pass 2: process 6 rows from work array, store into output array. + * 3-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/6). + */ + + wsptr = workspace; + for (ctr = 0; ctr < 6; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + tmp0 = (INT32) wsptr[0] + + ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + + (ONE << (PASS1_BITS+2))); + tmp0 <<= CONST_BITS; + tmp2 = (INT32) wsptr[2]; + tmp12 = MULTIPLY(tmp2, FIX(0.707106781)); /* c2 */ + tmp10 = tmp0 + tmp12; + tmp2 = tmp0 - tmp12 - tmp12; + + /* Odd part */ + + tmp12 = (INT32) wsptr[1]; + tmp0 = MULTIPLY(tmp12, FIX(1.224744871)); /* c1 */ + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp0, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp2, + CONST_BITS+PASS1_BITS+3) + & RANGE_MASK]; + + wsptr += 3; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a 2x4 output block. + * + * 4-point IDCT in pass 1 (columns), 2-point in pass 2 (rows). + */ + +GLOBAL(void) +jpeg_idct_2x4 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + INT32 tmp0, tmp2, tmp10, tmp12; + INT32 z1, z2, z3; + JCOEFPTR inptr; + ISLOW_MULT_TYPE * quantptr; + INT32 * wsptr; + JSAMPROW outptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + int ctr; + INT32 workspace[2*4]; /* buffers data between passes */ + SHIFT_TEMPS + + /* Pass 1: process columns from input, store into work array. + * 4-point IDCT kernel, + * cK represents sqrt(2) * cos(K*pi/16) [refers to 8-point IDCT]. + */ + + inptr = coef_block; + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + wsptr = workspace; + for (ctr = 0; ctr < 2; ctr++, inptr++, quantptr++, wsptr++) { + /* Even part */ + + tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); + tmp2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); + + tmp10 = (tmp0 + tmp2) << CONST_BITS; + tmp12 = (tmp0 - tmp2) << CONST_BITS; + + /* Odd part */ + /* Same rotation as in the even part of the 8x8 LL&M IDCT */ + + z2 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); + z3 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); + + z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ + tmp0 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ + tmp2 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ + + /* Final output stage */ + + wsptr[2*0] = tmp10 + tmp0; + wsptr[2*3] = tmp10 - tmp0; + wsptr[2*1] = tmp12 + tmp2; + wsptr[2*2] = tmp12 - tmp2; + } + + /* Pass 2: process 4 rows from work array, store into output array. */ + + wsptr = workspace; + for (ctr = 0; ctr < 4; ctr++) { + outptr = output_buf[ctr] + output_col; + + /* Even part */ + + /* Add range center and fudge factor for final descale and range-limit. */ + tmp10 = wsptr[0] + + ((((INT32) RANGE_CENTER) << (CONST_BITS+3)) + + (ONE << (CONST_BITS+2))); + + /* Odd part */ + + tmp0 = wsptr[1]; + + /* Final output stage */ + + outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp0, CONST_BITS+3) + & RANGE_MASK]; + outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp0, CONST_BITS+3) + & RANGE_MASK]; + + wsptr += 2; /* advance pointer to next row */ + } +} + + +/* + * Perform dequantization and inverse DCT on one block of coefficients, + * producing a 1x2 output block. + * + * 2-point IDCT in pass 1 (columns), 1-point in pass 2 (rows). + */ + +GLOBAL(void) +jpeg_idct_1x2 (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col) +{ + DCTELEM tmp0, tmp1; + ISLOW_MULT_TYPE * quantptr; + JSAMPLE *range_limit = IDCT_range_limit(cinfo); + ISHIFT_TEMPS + + /* Process 1 column from input, store into output array. */ + + quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; + + /* Even part */ + + tmp0 = DEQUANTIZE(coef_block[DCTSIZE*0], quantptr[DCTSIZE*0]); + /* Add range center and fudge factor for final descale and range-limit. */ + tmp0 += (((DCTELEM) RANGE_CENTER) << 3) + (1 << 2); + + /* Odd part */ + + tmp1 = DEQUANTIZE(coef_block[DCTSIZE*1], quantptr[DCTSIZE*1]); + + /* Final output stage */ + + output_buf[0][output_col] = + range_limit[(int) IRIGHT_SHIFT(tmp0 + tmp1, 3) & RANGE_MASK]; + output_buf[1][output_col] = + range_limit[(int) IRIGHT_SHIFT(tmp0 - tmp1, 3) & RANGE_MASK]; +} + +#endif /* IDCT_SCALING_SUPPORTED */ +#endif /* DCT_ISLOW_SUPPORTED */ diff --git a/cmake/externals/libjpeg/jinclude.h b/cmake/externals/libjpeg/jinclude.h new file mode 100644 index 000000000..12ea8cd2f --- /dev/null +++ b/cmake/externals/libjpeg/jinclude.h @@ -0,0 +1,157 @@ +/* + * jinclude.h + * + * Copyright (C) 1991-1994, Thomas G. Lane. + * Modified 2017-2022 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file exists to provide a single place to fix any problems with + * including the wrong system include files. (Common problems are taken + * care of by the standard jconfig symbols, but on really weird systems + * you may have to edit this file.) + * + * NOTE: this file is NOT intended to be included by applications using + * the JPEG library. Most applications need only include jpeglib.h. + */ + + +/* Include auto-config file to find out which system include files we need. */ + +#include "jconfig.h" /* auto configuration options */ +#define JCONFIG_INCLUDED /* so that jpeglib.h doesn't do it again */ + +/* + * We need the NULL macro and size_t typedef. + * On an ANSI-conforming system it is sufficient to include . + * Otherwise, we get them from or ; we may have to + * pull in as well. + * Note that the core JPEG library does not require ; + * only the default error handler and data source/destination modules do. + * But we must pull it in because of the references to FILE in jpeglib.h. + * You can remove those references if you want to compile without . + */ + +#ifdef HAVE_STDDEF_H +#include +#endif + +#ifdef HAVE_STDLIB_H +#include +#endif + +#ifdef NEED_SYS_TYPES_H +#include +#endif + +#include + +/* + * We need memory copying and zeroing functions, plus strncpy(). + * ANSI and System V implementations declare these in . + * BSD doesn't have the mem() functions, but it does have bcopy()/bzero(). + * Some systems may declare memset and memcpy in . + * + * NOTE: we assume the size parameters to these functions are of type size_t. + * Change the casts in these macros if not! + */ + +#ifdef NEED_BSD_STRINGS + +#include +#define MEMZERO(target,size) bzero((void *)(target), (size_t)(size)) +#define MEMCOPY(dest,src,size) bcopy((const void *)(src), (void *)(dest), (size_t)(size)) + +#else /* not BSD, assume ANSI/SysV string lib */ + +#include +#define MEMZERO(target,size) memset((void *)(target), 0, (size_t)(size)) +#define MEMCOPY(dest,src,size) memcpy((void *)(dest), (const void *)(src), (size_t)(size)) + +#endif + +/* + * In ANSI C, and indeed any rational implementation, size_t is also the + * type returned by sizeof(). However, it seems there are some irrational + * implementations out there, in which sizeof() returns an int even though + * size_t is defined as long or unsigned long. To ensure consistent results + * we always use this SIZEOF() macro in place of using sizeof() directly. + */ + +#define SIZEOF(object) ((size_t) sizeof(object)) + +/* + * The modules that use fread() and fwrite() always invoke them through + * these macros. On some systems you may need to twiddle the argument casts. + * CAUTION: argument order is different from underlying functions! + * + * Furthermore, macros are provided for fflush() and ferror() in order + * to facilitate adaption by applications using an own FILE class. + * + * You can define your own custom file I/O functions in jconfig.h and + * #define JPEG_HAVE_FILE_IO_CUSTOM there to prevent redefinition here. + * + * You can #define JPEG_USE_FILE_IO_CUSTOM in jconfig.h to use custom file + * I/O functions implemented in Delphi VCL (Visual Component Library) + * in Vcl.Imaging.jpeg.pas for the TJPEGImage component utilizing + * the Delphi RTL (Run-Time Library) TMemoryStream component: + * + * procedure jpeg_stdio_src(var cinfo: jpeg_decompress_struct; + * input_file: TStream); external; + * + * procedure jpeg_stdio_dest(var cinfo: jpeg_compress_struct; + * output_file: TStream); external; + * + * function jfread(var buf; recsize, reccount: Integer; S: TStream): Integer; + * begin + * Result := S.Read(buf, recsize * reccount); + * end; + * + * function jfwrite(const buf; recsize, reccount: Integer; S: TStream): Integer; + * begin + * Result := S.Write(buf, recsize * reccount); + * end; + * + * function jfflush(S: TStream): Integer; + * begin + * Result := 0; + * end; + * + * function jferror(S: TStream): Integer; + * begin + * Result := 0; + * end; + * + * TMemoryStream of Delphi RTL has the distinctive feature to provide dynamic + * memory buffer management with a file/stream-based interface, particularly for + * the write (output) operation, which is easier to apply compared with direct + * implementations as given in jdatadst.c for memory destination. Those direct + * implementations of dynamic memory write tend to be more difficult to use, + * so providing an option like TMemoryStream may be a useful alternative. + * + * The CFile/CMemFile classes of the Microsoft Foundation Class (MFC) Library + * may be used in a similar fashion. + */ + +#ifndef JPEG_HAVE_FILE_IO_CUSTOM +#ifdef JPEG_USE_FILE_IO_CUSTOM +extern size_t jfread(void * __ptr, size_t __size, size_t __n, FILE * __stream); +extern size_t jfwrite(const void * __ptr, size_t __size, size_t __n, FILE * __stream); +extern int jfflush(FILE * __stream); +extern int jferror(FILE * __fp); + +#define JFREAD(file,buf,sizeofbuf) \ + ((size_t) jfread((void *) (buf), (size_t) 1, (size_t) (sizeofbuf), (file))) +#define JFWRITE(file,buf,sizeofbuf) \ + ((size_t) jfwrite((const void *) (buf), (size_t) 1, (size_t) (sizeofbuf), (file))) +#define JFFLUSH(file) jfflush(file) +#define JFERROR(file) jferror(file) +#else +#define JFREAD(file,buf,sizeofbuf) \ + ((size_t) fread((void *) (buf), (size_t) 1, (size_t) (sizeofbuf), (file))) +#define JFWRITE(file,buf,sizeofbuf) \ + ((size_t) fwrite((const void *) (buf), (size_t) 1, (size_t) (sizeofbuf), (file))) +#define JFFLUSH(file) fflush(file) +#define JFERROR(file) ferror(file) +#endif +#endif diff --git a/cmake/externals/libjpeg/jmemansi.c b/cmake/externals/libjpeg/jmemansi.c new file mode 100644 index 000000000..2d93e4962 --- /dev/null +++ b/cmake/externals/libjpeg/jmemansi.c @@ -0,0 +1,167 @@ +/* + * jmemansi.c + * + * Copyright (C) 1992-1996, Thomas G. Lane. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file provides a simple generic implementation of the system- + * dependent portion of the JPEG memory manager. This implementation + * assumes that you have the ANSI-standard library routine tmpfile(). + * Also, the problem of determining the amount of memory available + * is shoved onto the user. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" +#include "jmemsys.h" /* import the system-dependent declarations */ + +#ifndef HAVE_STDLIB_H /* should declare malloc(),free() */ +extern void * malloc JPP((size_t size)); +extern void free JPP((void *ptr)); +#endif + +#ifndef SEEK_SET /* pre-ANSI systems may not define this; */ +#define SEEK_SET 0 /* if not, assume 0 is correct */ +#endif + + +/* + * Memory allocation and freeing are controlled by the regular library + * routines malloc() and free(). + */ + +GLOBAL(void *) +jpeg_get_small (j_common_ptr cinfo, size_t sizeofobject) +{ + return (void *) malloc(sizeofobject); +} + +GLOBAL(void) +jpeg_free_small (j_common_ptr cinfo, void * object, size_t sizeofobject) +{ + free(object); +} + + +/* + * "Large" objects are treated the same as "small" ones. + * NB: although we include FAR keywords in the routine declarations, + * this file won't actually work in 80x86 small/medium model; at least, + * you probably won't be able to process useful-size images in only 64KB. + */ + +GLOBAL(void FAR *) +jpeg_get_large (j_common_ptr cinfo, size_t sizeofobject) +{ + return (void FAR *) malloc(sizeofobject); +} + +GLOBAL(void) +jpeg_free_large (j_common_ptr cinfo, void FAR * object, size_t sizeofobject) +{ + free(object); +} + + +/* + * This routine computes the total memory space available for allocation. + * It's impossible to do this in a portable way; our current solution is + * to make the user tell us (with a default value set at compile time). + * If you can actually get the available space, it's a good idea to subtract + * a slop factor of 5% or so. + */ + +#ifndef DEFAULT_MAX_MEM /* so can override from makefile */ +#define DEFAULT_MAX_MEM 1000000L /* default: one megabyte */ +#endif + +GLOBAL(long) +jpeg_mem_available (j_common_ptr cinfo, long min_bytes_needed, + long max_bytes_needed, long already_allocated) +{ + return cinfo->mem->max_memory_to_use - already_allocated; +} + + +/* + * Backing store (temporary file) management. + * Backing store objects are only used when the value returned by + * jpeg_mem_available is less than the total space needed. You can dispense + * with these routines if you have plenty of virtual memory; see jmemnobs.c. + */ + + +METHODDEF(void) +read_backing_store (j_common_ptr cinfo, backing_store_ptr info, + void FAR * buffer_address, + long file_offset, long byte_count) +{ + if (fseek(info->temp_file, file_offset, SEEK_SET)) + ERREXIT(cinfo, JERR_TFILE_SEEK); + if (JFREAD(info->temp_file, buffer_address, byte_count) + != (size_t) byte_count) + ERREXIT(cinfo, JERR_TFILE_READ); +} + + +METHODDEF(void) +write_backing_store (j_common_ptr cinfo, backing_store_ptr info, + void FAR * buffer_address, + long file_offset, long byte_count) +{ + if (fseek(info->temp_file, file_offset, SEEK_SET)) + ERREXIT(cinfo, JERR_TFILE_SEEK); + if (JFWRITE(info->temp_file, buffer_address, byte_count) + != (size_t) byte_count) + ERREXIT(cinfo, JERR_TFILE_WRITE); +} + + +METHODDEF(void) +close_backing_store (j_common_ptr cinfo, backing_store_ptr info) +{ + fclose(info->temp_file); + /* Since this implementation uses tmpfile() to create the file, + * no explicit file deletion is needed. + */ +} + + +/* + * Initial opening of a backing-store object. + * + * This version uses tmpfile(), which constructs a suitable file name + * behind the scenes. We don't have to use info->temp_name[] at all; + * indeed, we can't even find out the actual name of the temp file. + */ + +GLOBAL(void) +jpeg_open_backing_store (j_common_ptr cinfo, backing_store_ptr info, + long total_bytes_needed) +{ + if ((info->temp_file = tmpfile()) == NULL) + ERREXITS(cinfo, JERR_TFILE_CREATE, ""); + info->read_backing_store = read_backing_store; + info->write_backing_store = write_backing_store; + info->close_backing_store = close_backing_store; +} + + +/* + * These routines take care of any system-dependent initialization and + * cleanup required. + */ + +GLOBAL(long) +jpeg_mem_init (j_common_ptr cinfo) +{ + return DEFAULT_MAX_MEM; /* default for max_memory_to_use */ +} + +GLOBAL(void) +jpeg_mem_term (j_common_ptr cinfo) +{ + /* no work */ +} diff --git a/cmake/externals/libjpeg/jmemmgr.c b/cmake/externals/libjpeg/jmemmgr.c new file mode 100644 index 000000000..40377de47 --- /dev/null +++ b/cmake/externals/libjpeg/jmemmgr.c @@ -0,0 +1,1115 @@ +/* + * jmemmgr.c + * + * Copyright (C) 1991-1997, Thomas G. Lane. + * Modified 2011-2019 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains the JPEG system-independent memory management + * routines. This code is usable across a wide variety of machines; most + * of the system dependencies have been isolated in a separate file. + * The major functions provided here are: + * * pool-based allocation and freeing of memory; + * * policy decisions about how to divide available memory among the + * virtual arrays; + * * control logic for swapping virtual arrays between main memory and + * backing storage. + * The separate system-dependent file provides the actual backing-storage + * access code, and it contains the policy decision about how much total + * main memory to use. + * This file is system-dependent in the sense that some of its functions + * are unnecessary in some systems. For example, if there is enough virtual + * memory so that backing storage will never be used, much of the virtual + * array control logic could be removed. (Of course, if you have that much + * memory then you shouldn't care about a little bit of unused code...) + */ + +#define JPEG_INTERNALS +#define AM_MEMORY_MANAGER /* we define jvirt_Xarray_control structs */ +#include "jinclude.h" +#include "jpeglib.h" +#include "jmemsys.h" /* import the system-dependent declarations */ + +#ifndef NO_GETENV +#ifndef HAVE_STDLIB_H /* should declare getenv() */ +extern char * getenv JPP((const char * name)); +#endif +#endif + + +/* + * Some important notes: + * The allocation routines provided here must never return NULL. + * They should exit to error_exit if unsuccessful. + * + * It's not a good idea to try to merge the sarray and barray routines, + * even though they are textually almost the same, because samples are + * usually stored as bytes while coefficients are shorts or ints. Thus, + * in machines where byte pointers have a different representation from + * word pointers, the resulting machine code could not be the same. + */ + + +/* + * Many machines require storage alignment: longs must start on 4-byte + * boundaries, doubles on 8-byte boundaries, etc. On such machines, malloc() + * always returns pointers that are multiples of the worst-case alignment + * requirement, and we had better do so too. + * There isn't any really portable way to determine the worst-case alignment + * requirement. This module assumes that the alignment requirement is + * multiples of sizeof(ALIGN_TYPE). + * By default, we define ALIGN_TYPE as double. This is necessary on some + * workstations (where doubles really do need 8-byte alignment) and will work + * fine on nearly everything. If your machine has lesser alignment needs, + * you can save a few bytes by making ALIGN_TYPE smaller. + * The only place I know of where this will NOT work is certain Macintosh + * 680x0 compilers that define double as a 10-byte IEEE extended float. + * Doing 10-byte alignment is counterproductive because longwords won't be + * aligned well. Put "#define ALIGN_TYPE long" in jconfig.h if you have + * such a compiler. + */ + +#ifndef ALIGN_TYPE /* so can override from jconfig.h */ +#define ALIGN_TYPE double +#endif + + +/* + * We allocate objects from "pools", where each pool is gotten with a single + * request to jpeg_get_small() or jpeg_get_large(). There is no per-object + * overhead within a pool, except for alignment padding. Each pool has a + * header with a link to the next pool of the same class. + * Small and large pool headers are identical except that the latter's + * link pointer must be FAR on 80x86 machines. + * Notice that the "real" header fields are union'ed with a dummy ALIGN_TYPE + * field. This forces the compiler to make SIZEOF(small_pool_hdr) a multiple + * of the alignment requirement of ALIGN_TYPE. + */ + +typedef union small_pool_struct * small_pool_ptr; + +typedef union small_pool_struct { + struct { + small_pool_ptr next; /* next in list of pools */ + size_t bytes_used; /* how many bytes already used within pool */ + size_t bytes_left; /* bytes still available in this pool */ + } hdr; + ALIGN_TYPE dummy; /* included in union to ensure alignment */ +} small_pool_hdr; + +typedef union large_pool_struct FAR * large_pool_ptr; + +typedef union large_pool_struct { + struct { + large_pool_ptr next; /* next in list of pools */ + size_t bytes_used; /* how many bytes already used within pool */ + size_t bytes_left; /* bytes still available in this pool */ + } hdr; + ALIGN_TYPE dummy; /* included in union to ensure alignment */ +} large_pool_hdr; + + +/* + * Here is the full definition of a memory manager object. + */ + +typedef struct { + struct jpeg_memory_mgr pub; /* public fields */ + + /* Each pool identifier (lifetime class) names a linked list of pools. */ + small_pool_ptr small_list[JPOOL_NUMPOOLS]; + large_pool_ptr large_list[JPOOL_NUMPOOLS]; + + /* Since we only have one lifetime class of virtual arrays, only one + * linked list is necessary (for each datatype). Note that the virtual + * array control blocks being linked together are actually stored somewhere + * in the small-pool list. + */ + jvirt_sarray_ptr virt_sarray_list; + jvirt_barray_ptr virt_barray_list; + + /* This counts total space obtained from jpeg_get_small/large */ + size_t total_space_allocated; + + /* alloc_sarray and alloc_barray set this value for use by virtual + * array routines. + */ + JDIMENSION last_rowsperchunk; /* from most recent alloc_sarray/barray */ +} my_memory_mgr; + +typedef my_memory_mgr * my_mem_ptr; + + +/* + * The control blocks for virtual arrays. + * Note that these blocks are allocated in the "small" pool area. + * System-dependent info for the associated backing store (if any) is hidden + * inside the backing_store_info struct. + */ + +struct jvirt_sarray_control { + JSAMPARRAY mem_buffer; /* => the in-memory buffer */ + JDIMENSION rows_in_array; /* total virtual array height */ + JDIMENSION samplesperrow; /* width of array (and of memory buffer) */ + JDIMENSION maxaccess; /* max rows accessed by access_virt_sarray */ + JDIMENSION rows_in_mem; /* height of memory buffer */ + JDIMENSION rowsperchunk; /* allocation chunk size in mem_buffer */ + JDIMENSION cur_start_row; /* first logical row # in the buffer */ + JDIMENSION first_undef_row; /* row # of first uninitialized row */ + boolean pre_zero; /* pre-zero mode requested? */ + boolean dirty; /* do current buffer contents need written? */ + boolean b_s_open; /* is backing-store data valid? */ + jvirt_sarray_ptr next; /* link to next virtual sarray control block */ + backing_store_info b_s_info; /* System-dependent control info */ +}; + +struct jvirt_barray_control { + JBLOCKARRAY mem_buffer; /* => the in-memory buffer */ + JDIMENSION rows_in_array; /* total virtual array height */ + JDIMENSION blocksperrow; /* width of array (and of memory buffer) */ + JDIMENSION maxaccess; /* max rows accessed by access_virt_barray */ + JDIMENSION rows_in_mem; /* height of memory buffer */ + JDIMENSION rowsperchunk; /* allocation chunk size in mem_buffer */ + JDIMENSION cur_start_row; /* first logical row # in the buffer */ + JDIMENSION first_undef_row; /* row # of first uninitialized row */ + boolean pre_zero; /* pre-zero mode requested? */ + boolean dirty; /* do current buffer contents need written? */ + boolean b_s_open; /* is backing-store data valid? */ + jvirt_barray_ptr next; /* link to next virtual barray control block */ + backing_store_info b_s_info; /* System-dependent control info */ +}; + + +#ifdef MEM_STATS /* optional extra stuff for statistics */ + +LOCAL(void) +print_mem_stats (j_common_ptr cinfo, int pool_id) +{ + my_mem_ptr mem = (my_mem_ptr) cinfo->mem; + small_pool_ptr shdr_ptr; + large_pool_ptr lhdr_ptr; + + /* Since this is only a debugging stub, we can cheat a little by using + * fprintf directly rather than going through the trace message code. + * This is helpful because message parm array can't handle longs. + */ + fprintf(stderr, "Freeing pool %d, total space = %ld\n", + pool_id, (long) mem->total_space_allocated); + + for (lhdr_ptr = mem->large_list[pool_id]; lhdr_ptr != NULL; + lhdr_ptr = lhdr_ptr->hdr.next) { + fprintf(stderr, " Large chunk used %ld\n", + (long) lhdr_ptr->hdr.bytes_used); + } + + for (shdr_ptr = mem->small_list[pool_id]; shdr_ptr != NULL; + shdr_ptr = shdr_ptr->hdr.next) { + fprintf(stderr, " Small chunk used %ld free %ld\n", + (long) shdr_ptr->hdr.bytes_used, + (long) shdr_ptr->hdr.bytes_left); + } +} + +#endif /* MEM_STATS */ + + +LOCAL(noreturn_t) +out_of_memory (j_common_ptr cinfo, int which) +/* Report an out-of-memory error and stop execution */ +/* If we compiled MEM_STATS support, report alloc requests before dying */ +{ +#ifdef MEM_STATS + cinfo->err->trace_level = 2; /* force self_destruct to report stats */ +#endif + ERREXIT1(cinfo, JERR_OUT_OF_MEMORY, which); +} + + +/* + * Allocation of "small" objects. + * + * For these, we use pooled storage. When a new pool must be created, + * we try to get enough space for the current request plus a "slop" factor, + * where the slop will be the amount of leftover space in the new pool. + * The speed vs. space tradeoff is largely determined by the slop values. + * A different slop value is provided for each pool class (lifetime), + * and we also distinguish the first pool of a class from later ones. + * NOTE: the values given work fairly well on both 16- and 32-bit-int + * machines, but may be too small if longs are 64 bits or more. + */ + +static const size_t first_pool_slop[JPOOL_NUMPOOLS] = +{ + 1600, /* first PERMANENT pool */ + 16000 /* first IMAGE pool */ +}; + +static const size_t extra_pool_slop[JPOOL_NUMPOOLS] = +{ + 0, /* additional PERMANENT pools */ + 5000 /* additional IMAGE pools */ +}; + +#define MIN_SLOP 50 /* greater than 0 to avoid futile looping */ + + +METHODDEF(void *) +alloc_small (j_common_ptr cinfo, int pool_id, size_t sizeofobject) +/* Allocate a "small" object */ +{ + my_mem_ptr mem = (my_mem_ptr) cinfo->mem; + small_pool_ptr hdr_ptr, prev_hdr_ptr; + size_t odd_bytes, min_request, slop; + char * data_ptr; + + /* Check for unsatisfiable request (do now to ensure no overflow below) */ + if (sizeofobject > (size_t) MAX_ALLOC_CHUNK - SIZEOF(small_pool_hdr)) + out_of_memory(cinfo, 1); /* request exceeds malloc's ability */ + + /* Round up the requested size to a multiple of SIZEOF(ALIGN_TYPE) */ + odd_bytes = sizeofobject % SIZEOF(ALIGN_TYPE); + if (odd_bytes > 0) + sizeofobject += SIZEOF(ALIGN_TYPE) - odd_bytes; + + /* See if space is available in any existing pool */ + if (pool_id < 0 || pool_id >= JPOOL_NUMPOOLS) + ERREXIT1(cinfo, JERR_BAD_POOL_ID, pool_id); /* safety check */ + prev_hdr_ptr = NULL; + hdr_ptr = mem->small_list[pool_id]; + while (hdr_ptr != NULL) { + if (hdr_ptr->hdr.bytes_left >= sizeofobject) + break; /* found pool with enough space */ + prev_hdr_ptr = hdr_ptr; + hdr_ptr = hdr_ptr->hdr.next; + } + + /* Time to make a new pool? */ + if (hdr_ptr == NULL) { + /* min_request is what we need now, slop is what will be leftover */ + min_request = sizeofobject + SIZEOF(small_pool_hdr); + if (prev_hdr_ptr == NULL) /* first pool in class? */ + slop = first_pool_slop[pool_id]; + else + slop = extra_pool_slop[pool_id]; + /* Don't ask for more than MAX_ALLOC_CHUNK */ + if (slop > (size_t) MAX_ALLOC_CHUNK - min_request) + slop = (size_t) MAX_ALLOC_CHUNK - min_request; + /* Try to get space, if fail reduce slop and try again */ + for (;;) { + hdr_ptr = (small_pool_ptr) jpeg_get_small(cinfo, min_request + slop); + if (hdr_ptr != NULL) + break; + slop /= 2; + if (slop < MIN_SLOP) /* give up when it gets real small */ + out_of_memory(cinfo, 2); /* jpeg_get_small failed */ + } + mem->total_space_allocated += min_request + slop; + /* Success, initialize the new pool header and add to end of list */ + hdr_ptr->hdr.next = NULL; + hdr_ptr->hdr.bytes_used = 0; + hdr_ptr->hdr.bytes_left = sizeofobject + slop; + if (prev_hdr_ptr == NULL) /* first pool in class? */ + mem->small_list[pool_id] = hdr_ptr; + else + prev_hdr_ptr->hdr.next = hdr_ptr; + } + + /* OK, allocate the object from the current pool */ + data_ptr = (char *) (hdr_ptr + 1); /* point to first data byte in pool */ + data_ptr += hdr_ptr->hdr.bytes_used; /* point to place for object */ + hdr_ptr->hdr.bytes_used += sizeofobject; + hdr_ptr->hdr.bytes_left -= sizeofobject; + + return (void *) data_ptr; +} + + +/* + * Allocation of "large" objects. + * + * The external semantics of these are the same as "small" objects, + * except that FAR pointers are used on 80x86. However the pool + * management heuristics are quite different. We assume that each + * request is large enough that it may as well be passed directly to + * jpeg_get_large; the pool management just links everything together + * so that we can free it all on demand. + * Note: the major use of "large" objects is in JSAMPARRAY and JBLOCKARRAY + * structures. The routines that create these structures (see below) + * deliberately bunch rows together to ensure a large request size. + */ + +METHODDEF(void FAR *) +alloc_large (j_common_ptr cinfo, int pool_id, size_t sizeofobject) +/* Allocate a "large" object */ +{ + my_mem_ptr mem = (my_mem_ptr) cinfo->mem; + large_pool_ptr hdr_ptr; + size_t odd_bytes; + + /* Check for unsatisfiable request (do now to ensure no overflow below) */ + if (sizeofobject > (size_t) MAX_ALLOC_CHUNK - SIZEOF(large_pool_hdr)) + out_of_memory(cinfo, 3); /* request exceeds malloc's ability */ + + /* Round up the requested size to a multiple of SIZEOF(ALIGN_TYPE) */ + odd_bytes = sizeofobject % SIZEOF(ALIGN_TYPE); + if (odd_bytes > 0) + sizeofobject += SIZEOF(ALIGN_TYPE) - odd_bytes; + + /* Always make a new pool */ + if (pool_id < 0 || pool_id >= JPOOL_NUMPOOLS) + ERREXIT1(cinfo, JERR_BAD_POOL_ID, pool_id); /* safety check */ + + hdr_ptr = (large_pool_ptr) jpeg_get_large(cinfo, sizeofobject + + SIZEOF(large_pool_hdr)); + if (hdr_ptr == NULL) + out_of_memory(cinfo, 4); /* jpeg_get_large failed */ + mem->total_space_allocated += sizeofobject + SIZEOF(large_pool_hdr); + + /* Success, initialize the new pool header and add to list */ + hdr_ptr->hdr.next = mem->large_list[pool_id]; + /* We maintain space counts in each pool header for statistical purposes, + * even though they are not needed for allocation. + */ + hdr_ptr->hdr.bytes_used = sizeofobject; + hdr_ptr->hdr.bytes_left = 0; + mem->large_list[pool_id] = hdr_ptr; + + return (void FAR *) (hdr_ptr + 1); /* point to first data byte in pool */ +} + + +/* + * Creation of 2-D sample arrays. + * The pointers are in near heap, the samples themselves in FAR heap. + * + * To minimize allocation overhead and to allow I/O of large contiguous + * blocks, we allocate the sample rows in groups of as many rows as possible + * without exceeding MAX_ALLOC_CHUNK total bytes per allocation request. + * NB: the virtual array control routines, later in this file, know about + * this chunking of rows. The rowsperchunk value is left in the mem manager + * object so that it can be saved away if this sarray is the workspace for + * a virtual array. + */ + +METHODDEF(JSAMPARRAY) +alloc_sarray (j_common_ptr cinfo, int pool_id, + JDIMENSION samplesperrow, JDIMENSION numrows) +/* Allocate a 2-D sample array */ +{ + my_mem_ptr mem = (my_mem_ptr) cinfo->mem; + JSAMPARRAY result; + JSAMPROW workspace; + JDIMENSION rowsperchunk, currow, i; + long ltemp; + + /* Calculate max # of rows allowed in one allocation chunk */ + ltemp = (MAX_ALLOC_CHUNK - SIZEOF(large_pool_hdr)) / + ((long) samplesperrow * SIZEOF(JSAMPLE)); + if (ltemp <= 0) + ERREXIT(cinfo, JERR_WIDTH_OVERFLOW); + if (ltemp < (long) numrows) + rowsperchunk = (JDIMENSION) ltemp; + else + rowsperchunk = numrows; + mem->last_rowsperchunk = rowsperchunk; + + /* Get space for row pointers (small object) */ + result = (JSAMPARRAY) alloc_small(cinfo, pool_id, + (size_t) numrows * SIZEOF(JSAMPROW)); + + /* Get the rows themselves (large objects) */ + currow = 0; + while (currow < numrows) { + rowsperchunk = MIN(rowsperchunk, numrows - currow); + workspace = (JSAMPROW) alloc_large(cinfo, pool_id, + (size_t) rowsperchunk * (size_t) samplesperrow * SIZEOF(JSAMPLE)); + for (i = rowsperchunk; i > 0; i--) { + result[currow++] = workspace; + workspace += samplesperrow; + } + } + + return result; +} + + +/* + * Creation of 2-D coefficient-block arrays. + * This is essentially the same as the code for sample arrays, above. + */ + +METHODDEF(JBLOCKARRAY) +alloc_barray (j_common_ptr cinfo, int pool_id, + JDIMENSION blocksperrow, JDIMENSION numrows) +/* Allocate a 2-D coefficient-block array */ +{ + my_mem_ptr mem = (my_mem_ptr) cinfo->mem; + JBLOCKARRAY result; + JBLOCKROW workspace; + JDIMENSION rowsperchunk, currow, i; + long ltemp; + + /* Calculate max # of rows allowed in one allocation chunk */ + ltemp = (MAX_ALLOC_CHUNK - SIZEOF(large_pool_hdr)) / + ((long) blocksperrow * SIZEOF(JBLOCK)); + if (ltemp <= 0) + ERREXIT(cinfo, JERR_WIDTH_OVERFLOW); + if (ltemp < (long) numrows) + rowsperchunk = (JDIMENSION) ltemp; + else + rowsperchunk = numrows; + mem->last_rowsperchunk = rowsperchunk; + + /* Get space for row pointers (small object) */ + result = (JBLOCKARRAY) alloc_small(cinfo, pool_id, + (size_t) numrows * SIZEOF(JBLOCKROW)); + + /* Get the rows themselves (large objects) */ + currow = 0; + while (currow < numrows) { + rowsperchunk = MIN(rowsperchunk, numrows - currow); + workspace = (JBLOCKROW) alloc_large(cinfo, pool_id, + (size_t) rowsperchunk * (size_t) blocksperrow * SIZEOF(JBLOCK)); + for (i = rowsperchunk; i > 0; i--) { + result[currow++] = workspace; + workspace += blocksperrow; + } + } + + return result; +} + + +/* + * About virtual array management: + * + * The above "normal" array routines are only used to allocate strip buffers + * (as wide as the image, but just a few rows high). Full-image-sized buffers + * are handled as "virtual" arrays. The array is still accessed a strip at a + * time, but the memory manager must save the whole array for repeated + * accesses. The intended implementation is that there is a strip buffer in + * memory (as high as is possible given the desired memory limit), plus a + * backing file that holds the rest of the array. + * + * The request_virt_array routines are told the total size of the image and + * the maximum number of rows that will be accessed at once. The in-memory + * buffer must be at least as large as the maxaccess value. + * + * The request routines create control blocks but not the in-memory buffers. + * That is postponed until realize_virt_arrays is called. At that time the + * total amount of space needed is known (approximately, anyway), so free + * memory can be divided up fairly. + * + * The access_virt_array routines are responsible for making a specific strip + * area accessible (after reading or writing the backing file, if necessary). + * Note that the access routines are told whether the caller intends to modify + * the accessed strip; during a read-only pass this saves having to rewrite + * data to disk. The access routines are also responsible for pre-zeroing + * any newly accessed rows, if pre-zeroing was requested. + * + * In current usage, the access requests are usually for nonoverlapping + * strips; that is, successive access start_row numbers differ by exactly + * num_rows = maxaccess. This means we can get good performance with simple + * buffer dump/reload logic, by making the in-memory buffer be a multiple + * of the access height; then there will never be accesses across bufferload + * boundaries. The code will still work with overlapping access requests, + * but it doesn't handle bufferload overlaps very efficiently. + */ + + +METHODDEF(jvirt_sarray_ptr) +request_virt_sarray (j_common_ptr cinfo, int pool_id, boolean pre_zero, + JDIMENSION samplesperrow, JDIMENSION numrows, + JDIMENSION maxaccess) +/* Request a virtual 2-D sample array */ +{ + my_mem_ptr mem = (my_mem_ptr) cinfo->mem; + jvirt_sarray_ptr result; + + /* Only IMAGE-lifetime virtual arrays are currently supported */ + if (pool_id != JPOOL_IMAGE) + ERREXIT1(cinfo, JERR_BAD_POOL_ID, pool_id); /* safety check */ + + /* get control block */ + result = (jvirt_sarray_ptr) alloc_small(cinfo, pool_id, + SIZEOF(struct jvirt_sarray_control)); + + result->mem_buffer = NULL; /* marks array not yet realized */ + result->rows_in_array = numrows; + result->samplesperrow = samplesperrow; + result->maxaccess = maxaccess; + result->pre_zero = pre_zero; + result->b_s_open = FALSE; /* no associated backing-store object */ + result->next = mem->virt_sarray_list; /* add to list of virtual arrays */ + mem->virt_sarray_list = result; + + return result; +} + + +METHODDEF(jvirt_barray_ptr) +request_virt_barray (j_common_ptr cinfo, int pool_id, boolean pre_zero, + JDIMENSION blocksperrow, JDIMENSION numrows, + JDIMENSION maxaccess) +/* Request a virtual 2-D coefficient-block array */ +{ + my_mem_ptr mem = (my_mem_ptr) cinfo->mem; + jvirt_barray_ptr result; + + /* Only IMAGE-lifetime virtual arrays are currently supported */ + if (pool_id != JPOOL_IMAGE) + ERREXIT1(cinfo, JERR_BAD_POOL_ID, pool_id); /* safety check */ + + /* get control block */ + result = (jvirt_barray_ptr) alloc_small(cinfo, pool_id, + SIZEOF(struct jvirt_barray_control)); + + result->mem_buffer = NULL; /* marks array not yet realized */ + result->rows_in_array = numrows; + result->blocksperrow = blocksperrow; + result->maxaccess = maxaccess; + result->pre_zero = pre_zero; + result->b_s_open = FALSE; /* no associated backing-store object */ + result->next = mem->virt_barray_list; /* add to list of virtual arrays */ + mem->virt_barray_list = result; + + return result; +} + + +METHODDEF(void) +realize_virt_arrays (j_common_ptr cinfo) +/* Allocate the in-memory buffers for any unrealized virtual arrays */ +{ + my_mem_ptr mem = (my_mem_ptr) cinfo->mem; + long bytesperrow, space_per_minheight, maximum_space; + long avail_mem, minheights, max_minheights; + jvirt_sarray_ptr sptr; + jvirt_barray_ptr bptr; + + /* Compute the minimum space needed (maxaccess rows in each buffer) + * and the maximum space needed (full image height in each buffer). + * These may be of use to the system-dependent jpeg_mem_available routine. + */ + space_per_minheight = 0; + maximum_space = 0; + for (sptr = mem->virt_sarray_list; sptr != NULL; sptr = sptr->next) { + if (sptr->mem_buffer == NULL) { /* if not realized yet */ + bytesperrow = (long) sptr->samplesperrow * SIZEOF(JSAMPLE); + space_per_minheight += (long) sptr->maxaccess * bytesperrow; + maximum_space += (long) sptr->rows_in_array * bytesperrow; + } + } + for (bptr = mem->virt_barray_list; bptr != NULL; bptr = bptr->next) { + if (bptr->mem_buffer == NULL) { /* if not realized yet */ + bytesperrow = (long) bptr->blocksperrow * SIZEOF(JBLOCK); + space_per_minheight += (long) bptr->maxaccess * bytesperrow; + maximum_space += (long) bptr->rows_in_array * bytesperrow; + } + } + + if (space_per_minheight <= 0) + return; /* no unrealized arrays, no work */ + + /* Determine amount of memory to actually use; this is system-dependent. */ + avail_mem = jpeg_mem_available(cinfo, space_per_minheight, maximum_space, + (long) mem->total_space_allocated); + + /* If the maximum space needed is available, make all the buffers full + * height; otherwise parcel it out with the same number of minheights + * in each buffer. + */ + if (avail_mem >= maximum_space) + max_minheights = 1000000000L; + else { + max_minheights = avail_mem / space_per_minheight; + /* If there doesn't seem to be enough space, try to get the minimum + * anyway. This allows a "stub" implementation of jpeg_mem_available(). + */ + if (max_minheights <= 0) + max_minheights = 1; + } + + /* Allocate the in-memory buffers and initialize backing store as needed. */ + + for (sptr = mem->virt_sarray_list; sptr != NULL; sptr = sptr->next) { + if (sptr->mem_buffer == NULL) { /* if not realized yet */ + minheights = ((long) sptr->rows_in_array - 1L) / sptr->maxaccess + 1L; + if (minheights <= max_minheights) { + /* This buffer fits in memory */ + sptr->rows_in_mem = sptr->rows_in_array; + } else { + /* It doesn't fit in memory, create backing store. */ + sptr->rows_in_mem = (JDIMENSION) (max_minheights * sptr->maxaccess); + jpeg_open_backing_store(cinfo, & sptr->b_s_info, + (long) sptr->rows_in_array * + (long) sptr->samplesperrow * + (long) SIZEOF(JSAMPLE)); + sptr->b_s_open = TRUE; + } + sptr->mem_buffer = alloc_sarray(cinfo, JPOOL_IMAGE, + sptr->samplesperrow, sptr->rows_in_mem); + sptr->rowsperchunk = mem->last_rowsperchunk; + sptr->cur_start_row = 0; + sptr->first_undef_row = 0; + sptr->dirty = FALSE; + } + } + + for (bptr = mem->virt_barray_list; bptr != NULL; bptr = bptr->next) { + if (bptr->mem_buffer == NULL) { /* if not realized yet */ + minheights = ((long) bptr->rows_in_array - 1L) / bptr->maxaccess + 1L; + if (minheights <= max_minheights) { + /* This buffer fits in memory */ + bptr->rows_in_mem = bptr->rows_in_array; + } else { + /* It doesn't fit in memory, create backing store. */ + bptr->rows_in_mem = (JDIMENSION) (max_minheights * bptr->maxaccess); + jpeg_open_backing_store(cinfo, & bptr->b_s_info, + (long) bptr->rows_in_array * + (long) bptr->blocksperrow * + (long) SIZEOF(JBLOCK)); + bptr->b_s_open = TRUE; + } + bptr->mem_buffer = alloc_barray(cinfo, JPOOL_IMAGE, + bptr->blocksperrow, bptr->rows_in_mem); + bptr->rowsperchunk = mem->last_rowsperchunk; + bptr->cur_start_row = 0; + bptr->first_undef_row = 0; + bptr->dirty = FALSE; + } + } +} + + +LOCAL(void) +do_sarray_io (j_common_ptr cinfo, jvirt_sarray_ptr ptr, boolean writing) +/* Do backing store read or write of a virtual sample array */ +{ + long bytesperrow, file_offset, byte_count, rows, thisrow, i; + + bytesperrow = (long) ptr->samplesperrow * SIZEOF(JSAMPLE); + file_offset = (long) ptr->cur_start_row * bytesperrow; + /* Loop to read or write each allocation chunk in mem_buffer */ + for (i = 0; i < (long) ptr->rows_in_mem; i += ptr->rowsperchunk) { + /* One chunk, but check for short chunk at end of buffer */ + rows = MIN((long) ptr->rowsperchunk, (long) ptr->rows_in_mem - i); + /* Transfer no more than is currently defined */ + thisrow = (long) ptr->cur_start_row + i; + rows = MIN(rows, (long) ptr->first_undef_row - thisrow); + /* Transfer no more than fits in file */ + rows = MIN(rows, (long) ptr->rows_in_array - thisrow); + if (rows <= 0) /* this chunk might be past end of file! */ + break; + byte_count = rows * bytesperrow; + if (writing) + (*ptr->b_s_info.write_backing_store) (cinfo, & ptr->b_s_info, + (void FAR *) ptr->mem_buffer[i], + file_offset, byte_count); + else + (*ptr->b_s_info.read_backing_store) (cinfo, & ptr->b_s_info, + (void FAR *) ptr->mem_buffer[i], + file_offset, byte_count); + file_offset += byte_count; + } +} + + +LOCAL(void) +do_barray_io (j_common_ptr cinfo, jvirt_barray_ptr ptr, boolean writing) +/* Do backing store read or write of a virtual coefficient-block array */ +{ + long bytesperrow, file_offset, byte_count, rows, thisrow, i; + + bytesperrow = (long) ptr->blocksperrow * SIZEOF(JBLOCK); + file_offset = (long) ptr->cur_start_row * bytesperrow; + /* Loop to read or write each allocation chunk in mem_buffer */ + for (i = 0; i < (long) ptr->rows_in_mem; i += ptr->rowsperchunk) { + /* One chunk, but check for short chunk at end of buffer */ + rows = MIN((long) ptr->rowsperchunk, (long) ptr->rows_in_mem - i); + /* Transfer no more than is currently defined */ + thisrow = (long) ptr->cur_start_row + i; + rows = MIN(rows, (long) ptr->first_undef_row - thisrow); + /* Transfer no more than fits in file */ + rows = MIN(rows, (long) ptr->rows_in_array - thisrow); + if (rows <= 0) /* this chunk might be past end of file! */ + break; + byte_count = rows * bytesperrow; + if (writing) + (*ptr->b_s_info.write_backing_store) (cinfo, & ptr->b_s_info, + (void FAR *) ptr->mem_buffer[i], + file_offset, byte_count); + else + (*ptr->b_s_info.read_backing_store) (cinfo, & ptr->b_s_info, + (void FAR *) ptr->mem_buffer[i], + file_offset, byte_count); + file_offset += byte_count; + } +} + + +METHODDEF(JSAMPARRAY) +access_virt_sarray (j_common_ptr cinfo, jvirt_sarray_ptr ptr, + JDIMENSION start_row, JDIMENSION num_rows, + boolean writable) +/* Access the part of a virtual sample array starting at start_row */ +/* and extending for num_rows rows. writable is true if */ +/* caller intends to modify the accessed area. */ +{ + JDIMENSION end_row = start_row + num_rows; + JDIMENSION undef_row; + + /* debugging check */ + if (end_row > ptr->rows_in_array || num_rows > ptr->maxaccess || + ptr->mem_buffer == NULL) + ERREXIT(cinfo, JERR_BAD_VIRTUAL_ACCESS); + + /* Make the desired part of the virtual array accessible */ + if (start_row < ptr->cur_start_row || + end_row > ptr->cur_start_row + ptr->rows_in_mem) { + if (! ptr->b_s_open) + ERREXIT(cinfo, JERR_VIRTUAL_BUG); + /* Flush old buffer contents if necessary */ + if (ptr->dirty) { + do_sarray_io(cinfo, ptr, TRUE); + ptr->dirty = FALSE; + } + /* Decide what part of virtual array to access. + * Algorithm: if target address > current window, assume forward scan, + * load starting at target address. If target address < current window, + * assume backward scan, load so that target area is top of window. + * Note that when switching from forward write to forward read, will have + * start_row = 0, so the limiting case applies and we load from 0 anyway. + */ + if (start_row > ptr->cur_start_row) { + ptr->cur_start_row = start_row; + } else { + /* use long arithmetic here to avoid overflow & unsigned problems */ + long ltemp; + + ltemp = (long) end_row - (long) ptr->rows_in_mem; + if (ltemp < 0) + ltemp = 0; /* don't fall off front end of file */ + ptr->cur_start_row = (JDIMENSION) ltemp; + } + /* Read in the selected part of the array. + * During the initial write pass, we will do no actual read + * because the selected part is all undefined. + */ + do_sarray_io(cinfo, ptr, FALSE); + } + /* Ensure the accessed part of the array is defined; prezero if needed. + * To improve locality of access, we only prezero the part of the array + * that the caller is about to access, not the entire in-memory array. + */ + if (ptr->first_undef_row < end_row) { + if (ptr->first_undef_row < start_row) { + if (writable) /* writer skipped over a section of array */ + ERREXIT(cinfo, JERR_BAD_VIRTUAL_ACCESS); + undef_row = start_row; /* but reader is allowed to read ahead */ + } else { + undef_row = ptr->first_undef_row; + } + if (writable) + ptr->first_undef_row = end_row; + if (ptr->pre_zero) { + size_t bytesperrow = (size_t) ptr->samplesperrow * SIZEOF(JSAMPLE); + undef_row -= ptr->cur_start_row; /* make indexes relative to buffer */ + end_row -= ptr->cur_start_row; + while (undef_row < end_row) { + FMEMZERO((void FAR *) ptr->mem_buffer[undef_row], bytesperrow); + undef_row++; + } + } else { + if (! writable) /* reader looking at undefined data */ + ERREXIT(cinfo, JERR_BAD_VIRTUAL_ACCESS); + } + } + /* Flag the buffer dirty if caller will write in it */ + if (writable) + ptr->dirty = TRUE; + /* Return address of proper part of the buffer */ + return ptr->mem_buffer + (start_row - ptr->cur_start_row); +} + + +METHODDEF(JBLOCKARRAY) +access_virt_barray (j_common_ptr cinfo, jvirt_barray_ptr ptr, + JDIMENSION start_row, JDIMENSION num_rows, + boolean writable) +/* Access the part of a virtual block array starting at start_row */ +/* and extending for num_rows rows. writable is true if */ +/* caller intends to modify the accessed area. */ +{ + JDIMENSION end_row = start_row + num_rows; + JDIMENSION undef_row; + + /* debugging check */ + if (end_row > ptr->rows_in_array || num_rows > ptr->maxaccess || + ptr->mem_buffer == NULL) + ERREXIT(cinfo, JERR_BAD_VIRTUAL_ACCESS); + + /* Make the desired part of the virtual array accessible */ + if (start_row < ptr->cur_start_row || + end_row > ptr->cur_start_row + ptr->rows_in_mem) { + if (! ptr->b_s_open) + ERREXIT(cinfo, JERR_VIRTUAL_BUG); + /* Flush old buffer contents if necessary */ + if (ptr->dirty) { + do_barray_io(cinfo, ptr, TRUE); + ptr->dirty = FALSE; + } + /* Decide what part of virtual array to access. + * Algorithm: if target address > current window, assume forward scan, + * load starting at target address. If target address < current window, + * assume backward scan, load so that target area is top of window. + * Note that when switching from forward write to forward read, will have + * start_row = 0, so the limiting case applies and we load from 0 anyway. + */ + if (start_row > ptr->cur_start_row) { + ptr->cur_start_row = start_row; + } else { + /* use long arithmetic here to avoid overflow & unsigned problems */ + long ltemp; + + ltemp = (long) end_row - (long) ptr->rows_in_mem; + if (ltemp < 0) + ltemp = 0; /* don't fall off front end of file */ + ptr->cur_start_row = (JDIMENSION) ltemp; + } + /* Read in the selected part of the array. + * During the initial write pass, we will do no actual read + * because the selected part is all undefined. + */ + do_barray_io(cinfo, ptr, FALSE); + } + /* Ensure the accessed part of the array is defined; prezero if needed. + * To improve locality of access, we only prezero the part of the array + * that the caller is about to access, not the entire in-memory array. + */ + if (ptr->first_undef_row < end_row) { + if (ptr->first_undef_row < start_row) { + if (writable) /* writer skipped over a section of array */ + ERREXIT(cinfo, JERR_BAD_VIRTUAL_ACCESS); + undef_row = start_row; /* but reader is allowed to read ahead */ + } else { + undef_row = ptr->first_undef_row; + } + if (writable) + ptr->first_undef_row = end_row; + if (ptr->pre_zero) { + size_t bytesperrow = (size_t) ptr->blocksperrow * SIZEOF(JBLOCK); + undef_row -= ptr->cur_start_row; /* make indexes relative to buffer */ + end_row -= ptr->cur_start_row; + while (undef_row < end_row) { + FMEMZERO((void FAR *) ptr->mem_buffer[undef_row], bytesperrow); + undef_row++; + } + } else { + if (! writable) /* reader looking at undefined data */ + ERREXIT(cinfo, JERR_BAD_VIRTUAL_ACCESS); + } + } + /* Flag the buffer dirty if caller will write in it */ + if (writable) + ptr->dirty = TRUE; + /* Return address of proper part of the buffer */ + return ptr->mem_buffer + (start_row - ptr->cur_start_row); +} + + +/* + * Release all objects belonging to a specified pool. + */ + +METHODDEF(void) +free_pool (j_common_ptr cinfo, int pool_id) +{ + my_mem_ptr mem = (my_mem_ptr) cinfo->mem; + small_pool_ptr shdr_ptr; + large_pool_ptr lhdr_ptr; + size_t space_freed; + + if (pool_id < 0 || pool_id >= JPOOL_NUMPOOLS) + ERREXIT1(cinfo, JERR_BAD_POOL_ID, pool_id); /* safety check */ + +#ifdef MEM_STATS + if (cinfo->err->trace_level > 1) + print_mem_stats(cinfo, pool_id); /* print pool's memory usage statistics */ +#endif + + /* If freeing IMAGE pool, close any virtual arrays first */ + if (pool_id == JPOOL_IMAGE) { + jvirt_sarray_ptr sptr; + jvirt_barray_ptr bptr; + + for (sptr = mem->virt_sarray_list; sptr != NULL; sptr = sptr->next) { + if (sptr->b_s_open) { /* there may be no backing store */ + sptr->b_s_open = FALSE; /* prevent recursive close if error */ + (*sptr->b_s_info.close_backing_store) (cinfo, & sptr->b_s_info); + } + } + mem->virt_sarray_list = NULL; + for (bptr = mem->virt_barray_list; bptr != NULL; bptr = bptr->next) { + if (bptr->b_s_open) { /* there may be no backing store */ + bptr->b_s_open = FALSE; /* prevent recursive close if error */ + (*bptr->b_s_info.close_backing_store) (cinfo, & bptr->b_s_info); + } + } + mem->virt_barray_list = NULL; + } + + /* Release large objects */ + lhdr_ptr = mem->large_list[pool_id]; + mem->large_list[pool_id] = NULL; + + while (lhdr_ptr != NULL) { + large_pool_ptr next_lhdr_ptr = lhdr_ptr->hdr.next; + space_freed = lhdr_ptr->hdr.bytes_used + + lhdr_ptr->hdr.bytes_left + + SIZEOF(large_pool_hdr); + jpeg_free_large(cinfo, (void FAR *) lhdr_ptr, space_freed); + mem->total_space_allocated -= space_freed; + lhdr_ptr = next_lhdr_ptr; + } + + /* Release small objects */ + shdr_ptr = mem->small_list[pool_id]; + mem->small_list[pool_id] = NULL; + + while (shdr_ptr != NULL) { + small_pool_ptr next_shdr_ptr = shdr_ptr->hdr.next; + space_freed = shdr_ptr->hdr.bytes_used + + shdr_ptr->hdr.bytes_left + + SIZEOF(small_pool_hdr); + jpeg_free_small(cinfo, (void *) shdr_ptr, space_freed); + mem->total_space_allocated -= space_freed; + shdr_ptr = next_shdr_ptr; + } +} + + +/* + * Close up shop entirely. + * Note that this cannot be called unless cinfo->mem is non-NULL. + */ + +METHODDEF(void) +self_destruct (j_common_ptr cinfo) +{ + int pool; + + /* Close all backing store, release all memory. + * Releasing pools in reverse order might help avoid fragmentation + * with some (brain-damaged) malloc libraries. + */ + for (pool = JPOOL_NUMPOOLS-1; pool >= JPOOL_PERMANENT; pool--) { + free_pool(cinfo, pool); + } + + /* Release the memory manager control block too. */ + jpeg_free_small(cinfo, (void *) cinfo->mem, SIZEOF(my_memory_mgr)); + cinfo->mem = NULL; /* ensures I will be called only once */ + + jpeg_mem_term(cinfo); /* system-dependent cleanup */ +} + + +/* + * Memory manager initialization. + * When this is called, only the error manager pointer is valid in cinfo! + */ + +GLOBAL(void) +jinit_memory_mgr (j_common_ptr cinfo) +{ + my_mem_ptr mem; + long max_to_use; + int pool; + size_t test_mac; + + cinfo->mem = NULL; /* for safety if init fails */ + + /* Check for configuration errors. + * SIZEOF(ALIGN_TYPE) should be a power of 2; otherwise, it probably + * doesn't reflect any real hardware alignment requirement. + * The test is a little tricky: for X>0, X and X-1 have no one-bits + * in common if and only if X is a power of 2, ie has only one one-bit. + * Some compilers may give an "unreachable code" warning here; ignore it. + */ + if ((SIZEOF(ALIGN_TYPE) & (SIZEOF(ALIGN_TYPE)-1)) != 0) + ERREXIT(cinfo, JERR_BAD_ALIGN_TYPE); + /* MAX_ALLOC_CHUNK must be representable as type size_t, and must be + * a multiple of SIZEOF(ALIGN_TYPE). + * Again, an "unreachable code" warning may be ignored here. + * But a "constant too large" warning means you need to fix MAX_ALLOC_CHUNK. + */ + test_mac = (size_t) MAX_ALLOC_CHUNK; + if ((long) test_mac != MAX_ALLOC_CHUNK || + (MAX_ALLOC_CHUNK % SIZEOF(ALIGN_TYPE)) != 0) + ERREXIT(cinfo, JERR_BAD_ALLOC_CHUNK); + + max_to_use = jpeg_mem_init(cinfo); /* system-dependent initialization */ + + /* Attempt to allocate memory manager's control block */ + mem = (my_mem_ptr) jpeg_get_small(cinfo, SIZEOF(my_memory_mgr)); + + if (mem == NULL) { + jpeg_mem_term(cinfo); /* system-dependent cleanup */ + ERREXIT1(cinfo, JERR_OUT_OF_MEMORY, 0); + } + + /* OK, fill in the method pointers */ + mem->pub.alloc_small = alloc_small; + mem->pub.alloc_large = alloc_large; + mem->pub.alloc_sarray = alloc_sarray; + mem->pub.alloc_barray = alloc_barray; + mem->pub.request_virt_sarray = request_virt_sarray; + mem->pub.request_virt_barray = request_virt_barray; + mem->pub.realize_virt_arrays = realize_virt_arrays; + mem->pub.access_virt_sarray = access_virt_sarray; + mem->pub.access_virt_barray = access_virt_barray; + mem->pub.free_pool = free_pool; + mem->pub.self_destruct = self_destruct; + + /* Make MAX_ALLOC_CHUNK accessible to other modules */ + mem->pub.max_alloc_chunk = MAX_ALLOC_CHUNK; + + /* Initialize working state */ + mem->pub.max_memory_to_use = max_to_use; + + for (pool = JPOOL_NUMPOOLS-1; pool >= JPOOL_PERMANENT; pool--) { + mem->small_list[pool] = NULL; + mem->large_list[pool] = NULL; + } + mem->virt_sarray_list = NULL; + mem->virt_barray_list = NULL; + + mem->total_space_allocated = SIZEOF(my_memory_mgr); + + /* Declare ourselves open for business */ + cinfo->mem = &mem->pub; + + /* Check for an environment variable JPEGMEM; if found, override the + * default max_memory setting from jpeg_mem_init. Note that the + * surrounding application may again override this value. + * If your system doesn't support getenv(), define NO_GETENV to disable + * this feature. + */ +#ifndef NO_GETENV + { char * memenv; + + if ((memenv = getenv("JPEGMEM")) != NULL) { + char ch = 'x'; + + if (sscanf(memenv, "%ld%c", &max_to_use, &ch) > 0) { + if (ch == 'm' || ch == 'M') + max_to_use *= 1000L; + mem->pub.max_memory_to_use = max_to_use * 1000L; + } + } + } +#endif + +} diff --git a/cmake/externals/libjpeg/jmemnobs.c b/cmake/externals/libjpeg/jmemnobs.c new file mode 100644 index 000000000..a364cd822 --- /dev/null +++ b/cmake/externals/libjpeg/jmemnobs.c @@ -0,0 +1,113 @@ +/* + * jmemnobs.c + * + * Copyright (C) 1992-1996, Thomas G. Lane. + * Modified 2019 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file provides a really simple implementation of the system- + * dependent portion of the JPEG memory manager. This implementation + * assumes that no backing-store files are needed: all required space + * can be obtained from malloc(). + * This is very portable in the sense that it'll compile on almost anything, + * but you'd better have lots of main memory (or virtual memory) if you want + * to process big images. + * Note that the max_memory_to_use option is respected by this implementation. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" +#include "jmemsys.h" /* import the system-dependent declarations */ + +#ifndef HAVE_STDLIB_H /* should declare malloc(),free() */ +extern void * malloc JPP((size_t size)); +extern void free JPP((void *ptr)); +#endif + + +/* + * Memory allocation and freeing are controlled by the regular library + * routines malloc() and free(). + */ + +GLOBAL(void *) +jpeg_get_small (j_common_ptr cinfo, size_t sizeofobject) +{ + return (void *) malloc(sizeofobject); +} + +GLOBAL(void) +jpeg_free_small (j_common_ptr cinfo, void * object, size_t sizeofobject) +{ + free(object); +} + + +/* + * "Large" objects are treated the same as "small" ones. + * NB: although we include FAR keywords in the routine declarations, + * this file won't actually work in 80x86 small/medium model; at least, + * you probably won't be able to process useful-size images in only 64KB. + */ + +GLOBAL(void FAR *) +jpeg_get_large (j_common_ptr cinfo, size_t sizeofobject) +{ + return (void FAR *) malloc(sizeofobject); +} + +GLOBAL(void) +jpeg_free_large (j_common_ptr cinfo, void FAR * object, size_t sizeofobject) +{ + free(object); +} + + +/* + * This routine computes the total memory space available for allocation. + */ + +GLOBAL(long) +jpeg_mem_available (j_common_ptr cinfo, long min_bytes_needed, + long max_bytes_needed, long already_allocated) +{ + if (cinfo->mem->max_memory_to_use) + return cinfo->mem->max_memory_to_use - already_allocated; + + /* Here we say, "we got all you want bud!" */ + return max_bytes_needed; +} + + +/* + * Backing store (temporary file) management. + * Since jpeg_mem_available always promised the moon, + * this should never be called and we can just error out. + */ + +GLOBAL(void) +jpeg_open_backing_store (j_common_ptr cinfo, backing_store_ptr info, + long total_bytes_needed) +{ + ERREXIT(cinfo, JERR_NO_BACKING_STORE); +} + + +/* + * These routines take care of any system-dependent initialization and + * cleanup required. Here, there isn't any. + */ + +GLOBAL(long) +jpeg_mem_init (j_common_ptr cinfo) +{ + return 0; /* just set max_memory_to_use to 0 */ +} + +GLOBAL(void) +jpeg_mem_term (j_common_ptr cinfo) +{ + /* no work */ +} diff --git a/cmake/externals/libjpeg/jmemsys.h b/cmake/externals/libjpeg/jmemsys.h new file mode 100644 index 000000000..6c3c6d348 --- /dev/null +++ b/cmake/externals/libjpeg/jmemsys.h @@ -0,0 +1,198 @@ +/* + * jmemsys.h + * + * Copyright (C) 1992-1997, Thomas G. Lane. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This include file defines the interface between the system-independent + * and system-dependent portions of the JPEG memory manager. No other + * modules need include it. (The system-independent portion is jmemmgr.c; + * there are several different versions of the system-dependent portion.) + * + * This file works as-is for the system-dependent memory managers supplied + * in the IJG distribution. You may need to modify it if you write a + * custom memory manager. If system-dependent changes are needed in + * this file, the best method is to #ifdef them based on a configuration + * symbol supplied in jconfig.h, as we have done with USE_MSDOS_MEMMGR + * and USE_MAC_MEMMGR. + */ + + +/* Short forms of external names for systems with brain-damaged linkers. */ + +#ifdef NEED_SHORT_EXTERNAL_NAMES +#define jpeg_get_small jGetSmall +#define jpeg_free_small jFreeSmall +#define jpeg_get_large jGetLarge +#define jpeg_free_large jFreeLarge +#define jpeg_mem_available jMemAvail +#define jpeg_open_backing_store jOpenBackStore +#define jpeg_mem_init jMemInit +#define jpeg_mem_term jMemTerm +#endif /* NEED_SHORT_EXTERNAL_NAMES */ + + +/* + * These two functions are used to allocate and release small chunks of + * memory. (Typically the total amount requested through jpeg_get_small is + * no more than 20K or so; this will be requested in chunks of a few K each.) + * Behavior should be the same as for the standard library functions malloc + * and free; in particular, jpeg_get_small must return NULL on failure. + * On most systems, these ARE malloc and free. jpeg_free_small is passed the + * size of the object being freed, just in case it's needed. + * On an 80x86 machine using small-data memory model, these manage near heap. + */ + +EXTERN(void *) jpeg_get_small JPP((j_common_ptr cinfo, size_t sizeofobject)); +EXTERN(void) jpeg_free_small JPP((j_common_ptr cinfo, void * object, + size_t sizeofobject)); + +/* + * These two functions are used to allocate and release large chunks of + * memory (up to the total free space designated by jpeg_mem_available). + * The interface is the same as above, except that on an 80x86 machine, + * far pointers are used. On most other machines these are identical to + * the jpeg_get/free_small routines; but we keep them separate anyway, + * in case a different allocation strategy is desirable for large chunks. + */ + +EXTERN(void FAR *) jpeg_get_large JPP((j_common_ptr cinfo, + size_t sizeofobject)); +EXTERN(void) jpeg_free_large JPP((j_common_ptr cinfo, void FAR * object, + size_t sizeofobject)); + +/* + * The macro MAX_ALLOC_CHUNK designates the maximum number of bytes that may + * be requested in a single call to jpeg_get_large (and jpeg_get_small for that + * matter, but that case should never come into play). This macro is needed + * to model the 64Kb-segment-size limit of far addressing on 80x86 machines. + * On those machines, we expect that jconfig.h will provide a proper value. + * On machines with 32-bit flat address spaces, any large constant may be used. + * + * NB: jmemmgr.c expects that MAX_ALLOC_CHUNK will be representable as type + * size_t and will be a multiple of sizeof(align_type). + */ + +#ifndef MAX_ALLOC_CHUNK /* may be overridden in jconfig.h */ +#define MAX_ALLOC_CHUNK 1000000000L +#endif + +/* + * This routine computes the total space still available for allocation by + * jpeg_get_large. If more space than this is needed, backing store will be + * used. NOTE: any memory already allocated must not be counted. + * + * There is a minimum space requirement, corresponding to the minimum + * feasible buffer sizes; jmemmgr.c will request that much space even if + * jpeg_mem_available returns zero. The maximum space needed, enough to hold + * all working storage in memory, is also passed in case it is useful. + * Finally, the total space already allocated is passed. If no better + * method is available, cinfo->mem->max_memory_to_use - already_allocated + * is often a suitable calculation. + * + * It is OK for jpeg_mem_available to underestimate the space available + * (that'll just lead to more backing-store access than is really necessary). + * However, an overestimate will lead to failure. Hence it's wise to subtract + * a slop factor from the true available space. 5% should be enough. + * + * On machines with lots of virtual memory, any large constant may be returned. + * Conversely, zero may be returned to always use the minimum amount of memory. + */ + +EXTERN(long) jpeg_mem_available JPP((j_common_ptr cinfo, + long min_bytes_needed, + long max_bytes_needed, + long already_allocated)); + + +/* + * This structure holds whatever state is needed to access a single + * backing-store object. The read/write/close method pointers are called + * by jmemmgr.c to manipulate the backing-store object; all other fields + * are private to the system-dependent backing store routines. + */ + +#define TEMP_NAME_LENGTH 64 /* max length of a temporary file's name */ + + +#ifdef USE_MSDOS_MEMMGR /* DOS-specific junk */ + +typedef unsigned short XMSH; /* type of extended-memory handles */ +typedef unsigned short EMSH; /* type of expanded-memory handles */ + +typedef union { + short file_handle; /* DOS file handle if it's a temp file */ + XMSH xms_handle; /* handle if it's a chunk of XMS */ + EMSH ems_handle; /* handle if it's a chunk of EMS */ +} handle_union; + +#endif /* USE_MSDOS_MEMMGR */ + +#ifdef USE_MAC_MEMMGR /* Mac-specific junk */ +#include +#endif /* USE_MAC_MEMMGR */ + + +typedef struct backing_store_struct * backing_store_ptr; + +typedef struct backing_store_struct { + /* Methods for reading/writing/closing this backing-store object */ + JMETHOD(void, read_backing_store, (j_common_ptr cinfo, + backing_store_ptr info, + void FAR * buffer_address, + long file_offset, long byte_count)); + JMETHOD(void, write_backing_store, (j_common_ptr cinfo, + backing_store_ptr info, + void FAR * buffer_address, + long file_offset, long byte_count)); + JMETHOD(void, close_backing_store, (j_common_ptr cinfo, + backing_store_ptr info)); + + /* Private fields for system-dependent backing-store management */ +#ifdef USE_MSDOS_MEMMGR + /* For the MS-DOS manager (jmemdos.c), we need: */ + handle_union handle; /* reference to backing-store storage object */ + char temp_name[TEMP_NAME_LENGTH]; /* name if it's a file */ +#else +#ifdef USE_MAC_MEMMGR + /* For the Mac manager (jmemmac.c), we need: */ + short temp_file; /* file reference number to temp file */ + FSSpec tempSpec; /* the FSSpec for the temp file */ + char temp_name[TEMP_NAME_LENGTH]; /* name if it's a file */ +#else + /* For a typical implementation with temp files, we need: */ + FILE * temp_file; /* stdio reference to temp file */ + char temp_name[TEMP_NAME_LENGTH]; /* name of temp file */ +#endif +#endif +} backing_store_info; + + +/* + * Initial opening of a backing-store object. This must fill in the + * read/write/close pointers in the object. The read/write routines + * may take an error exit if the specified maximum file size is exceeded. + * (If jpeg_mem_available always returns a large value, this routine can + * just take an error exit.) + */ + +EXTERN(void) jpeg_open_backing_store JPP((j_common_ptr cinfo, + backing_store_ptr info, + long total_bytes_needed)); + + +/* + * These routines take care of any system-dependent initialization and + * cleanup required. jpeg_mem_init will be called before anything is + * allocated (and, therefore, nothing in cinfo is of use except the error + * manager pointer). It should return a suitable default value for + * max_memory_to_use; this may subsequently be overridden by the surrounding + * application. (Note that max_memory_to_use is only important if + * jpeg_mem_available chooses to consult it ... no one else will.) + * jpeg_mem_term may assume that all requested memory has been freed and that + * all opened backing-store objects have been closed. + */ + +EXTERN(long) jpeg_mem_init JPP((j_common_ptr cinfo)); +EXTERN(void) jpeg_mem_term JPP((j_common_ptr cinfo)); diff --git a/cmake/externals/libjpeg/jmorecfg.h b/cmake/externals/libjpeg/jmorecfg.h new file mode 100644 index 000000000..4638d6af2 --- /dev/null +++ b/cmake/externals/libjpeg/jmorecfg.h @@ -0,0 +1,457 @@ +/* + * jmorecfg.h + * + * Copyright (C) 1991-1997, Thomas G. Lane. + * Modified 1997-2022 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains additional configuration options that customize the + * JPEG software for special applications or support machine-dependent + * optimizations. Most users will not need to touch this file. + */ + + +/* + * Define BITS_IN_JSAMPLE as either + * 8 for 8-bit sample values (the usual setting) + * 9 for 9-bit sample values + * 10 for 10-bit sample values + * 11 for 11-bit sample values + * 12 for 12-bit sample values + * Only 8, 9, 10, 11, and 12 bits sample data precision are supported for + * full-feature DCT processing. Further depths up to 16-bit may be added + * later for the lossless modes of operation. + * Run-time selection and conversion of data precision will be added later + * and are currently not supported, sorry. + * Exception: The transcoding part (jpegtran) supports all settings in a + * single instance, since it operates on the level of DCT coefficients and + * not sample values. The DCT coefficients are of the same type (16 bits) + * in all cases (see below). + */ + +#define BITS_IN_JSAMPLE 8 /* use 8, 9, 10, 11, or 12 */ + + +/* + * Maximum number of components (color channels) allowed in JPEG image. + * To meet the letter of the JPEG spec, set this to 255. However, darn + * few applications need more than 4 channels (maybe 5 for CMYK + alpha + * mask). We recommend 10 as a reasonable compromise; use 4 if you are + * really short on memory. (Each allowed component costs a hundred or so + * bytes of storage, whether actually used in an image or not.) + */ + +#define MAX_COMPONENTS 10 /* maximum number of image components */ + + +/* + * Basic data types. + * You may need to change these if you have a machine with unusual data + * type sizes; for example, "char" not 8 bits, "short" not 16 bits, + * or "long" not 32 bits. We don't care whether "int" is 16 or 32 bits, + * but it had better be at least 16. + */ + +/* Representation of a single sample (pixel element value). + * We frequently allocate large arrays of these, so it's important to keep + * them small. But if you have memory to burn and access to char or short + * arrays is very slow on your hardware, you might want to change these. + */ + +#if BITS_IN_JSAMPLE == 8 +/* JSAMPLE should be the smallest type that will hold the values 0..255. + * You can use a signed char by having GETJSAMPLE mask it with 0xFF. + */ + +#ifdef HAVE_UNSIGNED_CHAR + +typedef unsigned char JSAMPLE; +#define GETJSAMPLE(value) ((int) (value)) + +#else /* not HAVE_UNSIGNED_CHAR */ + +typedef char JSAMPLE; +#ifdef CHAR_IS_UNSIGNED +#define GETJSAMPLE(value) ((int) (value)) +#else +#define GETJSAMPLE(value) ((int) (value) & 0xFF) +#endif /* CHAR_IS_UNSIGNED */ + +#endif /* HAVE_UNSIGNED_CHAR */ + +#define MAXJSAMPLE 255 +#define CENTERJSAMPLE 128 + +#endif /* BITS_IN_JSAMPLE == 8 */ + + +#if BITS_IN_JSAMPLE == 9 +/* JSAMPLE should be the smallest type that will hold the values 0..511. + * On nearly all machines "short" will do nicely. + */ + +typedef short JSAMPLE; +#define GETJSAMPLE(value) ((int) (value)) + +#define MAXJSAMPLE 511 +#define CENTERJSAMPLE 256 + +#endif /* BITS_IN_JSAMPLE == 9 */ + + +#if BITS_IN_JSAMPLE == 10 +/* JSAMPLE should be the smallest type that will hold the values 0..1023. + * On nearly all machines "short" will do nicely. + */ + +typedef short JSAMPLE; +#define GETJSAMPLE(value) ((int) (value)) + +#define MAXJSAMPLE 1023 +#define CENTERJSAMPLE 512 + +#endif /* BITS_IN_JSAMPLE == 10 */ + + +#if BITS_IN_JSAMPLE == 11 +/* JSAMPLE should be the smallest type that will hold the values 0..2047. + * On nearly all machines "short" will do nicely. + */ + +typedef short JSAMPLE; +#define GETJSAMPLE(value) ((int) (value)) + +#define MAXJSAMPLE 2047 +#define CENTERJSAMPLE 1024 + +#endif /* BITS_IN_JSAMPLE == 11 */ + + +#if BITS_IN_JSAMPLE == 12 +/* JSAMPLE should be the smallest type that will hold the values 0..4095. + * On nearly all machines "short" will do nicely. + */ + +typedef short JSAMPLE; +#define GETJSAMPLE(value) ((int) (value)) + +#define MAXJSAMPLE 4095 +#define CENTERJSAMPLE 2048 + +#endif /* BITS_IN_JSAMPLE == 12 */ + + +/* Representation of a DCT frequency coefficient. + * This should be a signed value of at least 16 bits; "short" is usually OK. + * Again, we allocate large arrays of these, but you can change to int + * if you have memory to burn and "short" is really slow. + */ + +typedef short JCOEF; + + +/* Compressed datastreams are represented as arrays of JOCTET. + * These must be EXACTLY 8 bits wide, at least once they are written to + * external storage. Note that when using the stdio data source/destination + * managers, this is also the data type passed to fread/fwrite. + */ + +#ifdef HAVE_UNSIGNED_CHAR + +typedef unsigned char JOCTET; +#define GETJOCTET(value) (value) + +#else /* not HAVE_UNSIGNED_CHAR */ + +typedef char JOCTET; +#ifdef CHAR_IS_UNSIGNED +#define GETJOCTET(value) (value) +#else +#define GETJOCTET(value) ((value) & 0xFF) +#endif /* CHAR_IS_UNSIGNED */ + +#endif /* HAVE_UNSIGNED_CHAR */ + + +/* These typedefs are used for various table entries and so forth. + * They must be at least as wide as specified; but making them too big + * won't cost a huge amount of memory, so we don't provide special + * extraction code like we did for JSAMPLE. (In other words, these + * typedefs live at a different point on the speed/space tradeoff curve.) + */ + +/* UINT8 must hold at least the values 0..255. */ + +#ifdef HAVE_UNSIGNED_CHAR +typedef unsigned char UINT8; +#else /* not HAVE_UNSIGNED_CHAR */ +#ifdef CHAR_IS_UNSIGNED +typedef char UINT8; +#else /* not CHAR_IS_UNSIGNED */ +typedef short UINT8; +#endif /* CHAR_IS_UNSIGNED */ +#endif /* HAVE_UNSIGNED_CHAR */ + +/* UINT16 must hold at least the values 0..65535. */ + +#ifdef HAVE_UNSIGNED_SHORT +typedef unsigned short UINT16; +#else /* not HAVE_UNSIGNED_SHORT */ +typedef unsigned int UINT16; +#endif /* HAVE_UNSIGNED_SHORT */ + +/* INT16 must hold at least the values -32768..32767. */ + +#ifndef XMD_H /* X11/xmd.h correctly defines INT16 */ +typedef short INT16; +#endif + +/* INT32 must hold at least signed 32-bit values. */ + +#ifndef XMD_H /* X11/xmd.h correctly defines INT32 */ +#ifndef _BASETSD_H_ /* Microsoft defines it in basetsd.h */ +#ifndef _BASETSD_H /* MinGW is slightly different */ +#ifndef QGLOBAL_H /* Qt defines it in qglobal.h */ +typedef long INT32; +#endif +#endif +#endif +#endif + +/* Datatype used for image dimensions. The JPEG standard only supports + * images up to 64K*64K due to 16-bit fields in SOF markers. Therefore + * "unsigned int" is sufficient on all machines. However, if you need to + * handle larger images and you don't mind deviating from the spec, you + * can change this datatype. + */ + +typedef unsigned int JDIMENSION; + +#define JPEG_MAX_DIMENSION 65500L /* a tad under 64K to prevent overflows */ + + +/* These macros are used in all function definitions and extern declarations. + * You could modify them if you need to change function linkage conventions; + * in particular, you'll need to do that to make the library a Windows DLL. + * Another application is to make all functions global for use with debuggers + * or code profilers that require it. + */ + +/* a function called through method pointers: */ +#define METHODDEF(type) static type +/* a function used only in its module: */ +#define LOCAL(type) static type +/* a function referenced thru EXTERNs: */ +#define GLOBAL(type) type +/* a reference to a GLOBAL function: */ +#define EXTERN(type) extern type + + +/* This macro is used to declare a "method", that is, a function pointer. + * We want to supply prototype parameters if the compiler can cope. + * Note that the arglist parameter must be parenthesized! + * Again, you can customize this if you need special linkage keywords. + */ + +#ifdef HAVE_PROTOTYPES +#define JMETHOD(type,methodname,arglist) type (*methodname) arglist +#else +#define JMETHOD(type,methodname,arglist) type (*methodname) () +#endif + + +/* The noreturn type identifier is used to declare functions + * which cannot return. + * Compilers can thus create more optimized code and perform + * better checks for warnings and errors. + * Static analyzer tools can make improved inferences about + * execution paths and are prevented from giving false alerts. + * + * Unfortunately, the proposed specifications of corresponding + * extensions in the Dec 2011 ISO C standard revision (C11), + * GCC, MSVC, etc. are not viable. + * Thus we introduce a user defined type to declare noreturn + * functions at least for clarity. A proper compiler would + * have a suitable noreturn type to match in place of void. + */ + +#ifndef HAVE_NORETURN_T +typedef void noreturn_t; +#endif + + +/* Here is the pseudo-keyword for declaring pointers that must be "far" + * on 80x86 machines. Most of the specialized coding for 80x86 is handled + * by just saying "FAR *" where such a pointer is needed. In a few places + * explicit coding is needed; see uses of the NEED_FAR_POINTERS symbol. + */ + +#ifndef FAR +#ifdef NEED_FAR_POINTERS +#define FAR far +#else +#define FAR +#endif +#endif + + +/* + * On a few systems, type boolean and/or its values FALSE, TRUE may appear + * in standard header files. Or you may have conflicts with application- + * specific header files that you want to include together with these files. + * Defining HAVE_BOOLEAN before including jpeglib.h should make it work. + */ + +#ifndef HAVE_BOOLEAN +#if defined FALSE || defined TRUE || defined QGLOBAL_H +/* Qt3 defines FALSE and TRUE as "const" variables in qglobal.h */ +typedef int boolean; +#ifndef FALSE /* in case these macros already exist */ +#define FALSE 0 /* values of boolean */ +#endif +#ifndef TRUE +#define TRUE 1 +#endif +#else +typedef enum { FALSE = 0, TRUE = 1 } boolean; +#endif +#endif + + +/* + * The remaining options affect code selection within the JPEG library, + * but they don't need to be visible to most applications using the library. + * To minimize application namespace pollution, the symbols won't be + * defined unless JPEG_INTERNALS or JPEG_INTERNAL_OPTIONS has been defined. + */ + +#ifdef JPEG_INTERNALS +#define JPEG_INTERNAL_OPTIONS +#endif + +#ifdef JPEG_INTERNAL_OPTIONS + + +/* + * These defines indicate whether to include various optional functions. + * Undefining some of these symbols will produce a smaller but less capable + * library. Note that you can leave certain source files out of the + * compilation/linking process if you've #undef'd the corresponding symbols. + * (You may HAVE to do that if your compiler doesn't like null source files.) + */ + +/* Capability options common to encoder and decoder: */ + +#define DCT_ISLOW_SUPPORTED /* slow but accurate integer algorithm */ +#define DCT_IFAST_SUPPORTED /* faster, less accurate integer method */ +#define DCT_FLOAT_SUPPORTED /* floating-point: accurate, fast on fast HW */ + +/* Encoder capability options: */ + +#define C_ARITH_CODING_SUPPORTED /* Arithmetic coding back end? */ +#define C_MULTISCAN_FILES_SUPPORTED /* Multiple-scan JPEG files? */ +#define C_PROGRESSIVE_SUPPORTED /* Progressive JPEG? (Requires MULTISCAN) */ +#define DCT_SCALING_SUPPORTED /* Input rescaling via DCT? (Requires DCT_ISLOW) */ +#define ENTROPY_OPT_SUPPORTED /* Optimization of entropy coding parms? */ +/* Note: if you selected more than 8-bit data precision, it is dangerous to + * turn off ENTROPY_OPT_SUPPORTED. The standard Huffman tables are only + * good for 8-bit precision, so arithmetic coding is recommended for higher + * precision. The Huffman encoder normally uses entropy optimization to + * compute usable tables for higher precision. Otherwise, you'll have to + * supply different default Huffman tables. + * The exact same statements apply for progressive JPEG: the default tables + * don't work for progressive mode. (This may get fixed, however.) + */ +#define INPUT_SMOOTHING_SUPPORTED /* Input image smoothing option? */ + +/* Decoder capability options: */ + +#define D_ARITH_CODING_SUPPORTED /* Arithmetic coding back end? */ +#define D_MULTISCAN_FILES_SUPPORTED /* Multiple-scan JPEG files? */ +#define D_PROGRESSIVE_SUPPORTED /* Progressive JPEG? (Requires MULTISCAN) */ +#define IDCT_SCALING_SUPPORTED /* Output rescaling via IDCT? (Requires DCT_ISLOW) */ +#define SAVE_MARKERS_SUPPORTED /* jpeg_save_markers() needed? */ +#define BLOCK_SMOOTHING_SUPPORTED /* Block smoothing? (Progressive only) */ +#undef UPSAMPLE_SCALING_SUPPORTED /* Output rescaling at upsample stage? */ +#define UPSAMPLE_MERGING_SUPPORTED /* Fast path for sloppy upsampling? */ +#define QUANT_1PASS_SUPPORTED /* 1-pass color quantization? */ +#define QUANT_2PASS_SUPPORTED /* 2-pass color quantization? */ + +/* more capability options later, no doubt */ + + +/* + * Ordering of RGB data in scanlines passed to or from the application. + * If your application wants to deal with data in the order B,G,R, just + * #define JPEG_USE_RGB_CUSTOM in jconfig.h, or define your own custom + * order in jconfig.h and #define JPEG_HAVE_RGB_CUSTOM. + * You can also deal with formats such as R,G,B,X (one extra byte per pixel) + * by changing RGB_PIXELSIZE. + * Note that changing the offsets will also change + * the order in which colormap data is organized. + * RESTRICTIONS: + * 1. The sample applications cjpeg,djpeg do NOT support modified RGB formats. + * 2. The color quantizer modules will not behave desirably if RGB_PIXELSIZE + * is not 3 (they don't understand about dummy color components!). + * So you can't use color quantization if you change that value. + */ + +#ifndef JPEG_HAVE_RGB_CUSTOM +#ifdef JPEG_USE_RGB_CUSTOM +#define RGB_RED 2 /* Offset of Red in an RGB scanline element */ +#define RGB_GREEN 1 /* Offset of Green */ +#define RGB_BLUE 0 /* Offset of Blue */ +#else +#define RGB_RED 0 /* Offset of Red in an RGB scanline element */ +#define RGB_GREEN 1 /* Offset of Green */ +#define RGB_BLUE 2 /* Offset of Blue */ +#endif +#define RGB_PIXELSIZE 3 /* JSAMPLEs per RGB scanline element */ +#endif + + +/* Definitions for speed-related optimizations. */ + + +/* If your compiler supports inline functions, define INLINE + * as the inline keyword; otherwise define it as empty. + */ + +#ifndef INLINE +#ifdef __GNUC__ /* for instance, GNU C knows about inline */ +#define INLINE __inline__ +#endif +#ifndef INLINE +#define INLINE /* default is to define it as empty */ +#endif +#endif + + +/* On some machines (notably 68000 series) "int" is 32 bits, but multiplying + * two 16-bit shorts is faster than multiplying two ints. Define MULTIPLIER + * as short on such a machine. MULTIPLIER must be at least 16 bits wide. + */ + +#ifndef MULTIPLIER +#define MULTIPLIER int /* type for fastest integer multiply */ +#endif + + +/* FAST_FLOAT should be either float or double, whichever is done faster + * by your compiler. (Note that this type is only used in the floating point + * DCT routines, so it only matters if you've defined DCT_FLOAT_SUPPORTED.) + * Typically, float is faster in ANSI C compilers, while double is faster in + * pre-ANSI compilers (because they insist on converting to double anyway). + * The code below therefore chooses float if we have ANSI-style prototypes. + */ + +#ifndef FAST_FLOAT +#ifdef HAVE_PROTOTYPES +#define FAST_FLOAT float +#else +#define FAST_FLOAT double +#endif +#endif + +#endif /* JPEG_INTERNAL_OPTIONS */ diff --git a/cmake/externals/libjpeg/jpegint.h b/cmake/externals/libjpeg/jpegint.h new file mode 100644 index 000000000..3528bff5b --- /dev/null +++ b/cmake/externals/libjpeg/jpegint.h @@ -0,0 +1,445 @@ +/* + * jpegint.h + * + * Copyright (C) 1991-1997, Thomas G. Lane. + * Modified 1997-2020 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file provides common declarations for the various JPEG modules. + * These declarations are considered internal to the JPEG library; most + * applications using the library shouldn't need to include this file. + */ + + +/* Declarations for both compression & decompression */ + +typedef enum { /* Operating modes for buffer controllers */ + JBUF_PASS_THRU, /* Plain stripwise operation */ + /* Remaining modes require a full-image buffer to have been created */ + JBUF_SAVE_SOURCE, /* Run source subobject only, save output */ + JBUF_CRANK_DEST, /* Run dest subobject only, using saved data */ + JBUF_SAVE_AND_PASS /* Run both subobjects, save output */ +} J_BUF_MODE; + +/* Values of global_state field (jdapi.c has some dependencies on ordering!) */ +#define CSTATE_START 100 /* after create_compress */ +#define CSTATE_SCANNING 101 /* start_compress done, write_scanlines OK */ +#define CSTATE_RAW_OK 102 /* start_compress done, write_raw_data OK */ +#define CSTATE_WRCOEFS 103 /* jpeg_write_coefficients done */ +#define DSTATE_START 200 /* after create_decompress */ +#define DSTATE_INHEADER 201 /* reading header markers, no SOS yet */ +#define DSTATE_READY 202 /* found SOS, ready for start_decompress */ +#define DSTATE_PRELOAD 203 /* reading multiscan file in start_decompress*/ +#define DSTATE_PRESCAN 204 /* performing dummy pass for 2-pass quant */ +#define DSTATE_SCANNING 205 /* start_decompress done, read_scanlines OK */ +#define DSTATE_RAW_OK 206 /* start_decompress done, read_raw_data OK */ +#define DSTATE_BUFIMAGE 207 /* expecting jpeg_start_output */ +#define DSTATE_BUFPOST 208 /* looking for SOS/EOI in jpeg_finish_output */ +#define DSTATE_RDCOEFS 209 /* reading file in jpeg_read_coefficients */ +#define DSTATE_STOPPING 210 /* looking for EOI in jpeg_finish_decompress */ + + +/* Declarations for compression modules */ + +/* Master control module */ +struct jpeg_comp_master { + JMETHOD(void, prepare_for_pass, (j_compress_ptr cinfo)); + JMETHOD(void, pass_startup, (j_compress_ptr cinfo)); + JMETHOD(void, finish_pass, (j_compress_ptr cinfo)); + + /* State variables made visible to other modules */ + boolean call_pass_startup; /* True if pass_startup must be called */ + boolean is_last_pass; /* True during last pass */ +}; + +/* Main buffer control (downsampled-data buffer) */ +struct jpeg_c_main_controller { + JMETHOD(void, start_pass, (j_compress_ptr cinfo, J_BUF_MODE pass_mode)); + JMETHOD(void, process_data, (j_compress_ptr cinfo, + JSAMPARRAY input_buf, JDIMENSION *in_row_ctr, + JDIMENSION in_rows_avail)); +}; + +/* Compression preprocessing (downsampling input buffer control) */ +struct jpeg_c_prep_controller { + JMETHOD(void, start_pass, (j_compress_ptr cinfo, J_BUF_MODE pass_mode)); + JMETHOD(void, pre_process_data, (j_compress_ptr cinfo, + JSAMPARRAY input_buf, + JDIMENSION *in_row_ctr, + JDIMENSION in_rows_avail, + JSAMPIMAGE output_buf, + JDIMENSION *out_row_group_ctr, + JDIMENSION out_row_groups_avail)); +}; + +/* Coefficient buffer control */ +struct jpeg_c_coef_controller { + JMETHOD(void, start_pass, (j_compress_ptr cinfo, J_BUF_MODE pass_mode)); + JMETHOD(boolean, compress_data, (j_compress_ptr cinfo, + JSAMPIMAGE input_buf)); +}; + +/* Colorspace conversion */ +struct jpeg_color_converter { + JMETHOD(void, start_pass, (j_compress_ptr cinfo)); + JMETHOD(void, color_convert, (j_compress_ptr cinfo, + JSAMPARRAY input_buf, JSAMPIMAGE output_buf, + JDIMENSION output_row, int num_rows)); +}; + +/* Downsampling */ +struct jpeg_downsampler { + JMETHOD(void, start_pass, (j_compress_ptr cinfo)); + JMETHOD(void, downsample, (j_compress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION in_row_index, + JSAMPIMAGE output_buf, + JDIMENSION out_row_group_index)); + + boolean need_context_rows; /* TRUE if need rows above & below */ +}; + +/* Forward DCT (also controls coefficient quantization) */ +typedef JMETHOD(void, forward_DCT_ptr, + (j_compress_ptr cinfo, jpeg_component_info * compptr, + JSAMPARRAY sample_data, JBLOCKROW coef_blocks, + JDIMENSION start_col, JDIMENSION num_blocks)); + +struct jpeg_forward_dct { + JMETHOD(void, start_pass, (j_compress_ptr cinfo)); + /* It is useful to allow each component to have a separate FDCT method. */ + forward_DCT_ptr forward_DCT[MAX_COMPONENTS]; +}; + +/* Entropy encoding */ +struct jpeg_entropy_encoder { + JMETHOD(void, start_pass, (j_compress_ptr cinfo, boolean gather_statistics)); + JMETHOD(boolean, encode_mcu, (j_compress_ptr cinfo, JBLOCKARRAY MCU_data)); + JMETHOD(void, finish_pass, (j_compress_ptr cinfo)); +}; + +/* Marker writing */ +struct jpeg_marker_writer { + JMETHOD(void, write_file_header, (j_compress_ptr cinfo)); + JMETHOD(void, write_frame_header, (j_compress_ptr cinfo)); + JMETHOD(void, write_scan_header, (j_compress_ptr cinfo)); + JMETHOD(void, write_file_trailer, (j_compress_ptr cinfo)); + JMETHOD(void, write_tables_only, (j_compress_ptr cinfo)); + /* These routines are exported to allow insertion of extra markers */ + /* Probably only COM and APPn markers should be written this way */ + JMETHOD(void, write_marker_header, (j_compress_ptr cinfo, int marker, + unsigned int datalen)); + JMETHOD(void, write_marker_byte, (j_compress_ptr cinfo, int val)); +}; + + +/* Declarations for decompression modules */ + +/* Master control module */ +struct jpeg_decomp_master { + JMETHOD(void, prepare_for_output_pass, (j_decompress_ptr cinfo)); + JMETHOD(void, finish_output_pass, (j_decompress_ptr cinfo)); + + /* State variables made visible to other modules */ + boolean is_dummy_pass; /* True during 1st pass for 2-pass quant */ +}; + +/* Input control module */ +struct jpeg_input_controller { + JMETHOD(int, consume_input, (j_decompress_ptr cinfo)); + JMETHOD(void, reset_input_controller, (j_decompress_ptr cinfo)); + JMETHOD(void, start_input_pass, (j_decompress_ptr cinfo)); + JMETHOD(void, finish_input_pass, (j_decompress_ptr cinfo)); + + /* State variables made visible to other modules */ + boolean has_multiple_scans; /* True if file has multiple scans */ + boolean eoi_reached; /* True when EOI has been consumed */ +}; + +/* Main buffer control (downsampled-data buffer) */ +struct jpeg_d_main_controller { + JMETHOD(void, start_pass, (j_decompress_ptr cinfo, J_BUF_MODE pass_mode)); + JMETHOD(void, process_data, (j_decompress_ptr cinfo, + JSAMPARRAY output_buf, JDIMENSION *out_row_ctr, + JDIMENSION out_rows_avail)); +}; + +/* Coefficient buffer control */ +struct jpeg_d_coef_controller { + JMETHOD(void, start_input_pass, (j_decompress_ptr cinfo)); + JMETHOD(int, consume_data, (j_decompress_ptr cinfo)); + JMETHOD(void, start_output_pass, (j_decompress_ptr cinfo)); + JMETHOD(int, decompress_data, (j_decompress_ptr cinfo, + JSAMPIMAGE output_buf)); + /* Pointer to array of coefficient virtual arrays, or NULL if none */ + jvirt_barray_ptr *coef_arrays; +}; + +/* Decompression postprocessing (color quantization buffer control) */ +struct jpeg_d_post_controller { + JMETHOD(void, start_pass, (j_decompress_ptr cinfo, J_BUF_MODE pass_mode)); + JMETHOD(void, post_process_data, (j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, + JDIMENSION *in_row_group_ctr, + JDIMENSION in_row_groups_avail, + JSAMPARRAY output_buf, + JDIMENSION *out_row_ctr, + JDIMENSION out_rows_avail)); +}; + +/* Marker reading & parsing */ +struct jpeg_marker_reader { + JMETHOD(void, reset_marker_reader, (j_decompress_ptr cinfo)); + /* Read markers until SOS or EOI. + * Returns same codes as are defined for jpeg_consume_input: + * JPEG_SUSPENDED, JPEG_REACHED_SOS, or JPEG_REACHED_EOI. + */ + JMETHOD(int, read_markers, (j_decompress_ptr cinfo)); + /* Read a restart marker --- exported for use by entropy decoder only */ + jpeg_marker_parser_method read_restart_marker; + + /* State of marker reader --- nominally internal, but applications + * supplying COM or APPn handlers might like to know the state. + */ + boolean saw_SOI; /* found SOI? */ + boolean saw_SOF; /* found SOF? */ + int next_restart_num; /* next restart number expected (0-7) */ + unsigned int discarded_bytes; /* # of bytes skipped looking for a marker */ +}; + +/* Entropy decoding */ +struct jpeg_entropy_decoder { + JMETHOD(void, start_pass, (j_decompress_ptr cinfo)); + JMETHOD(boolean, decode_mcu, (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data)); + JMETHOD(void, finish_pass, (j_decompress_ptr cinfo)); +}; + +/* Inverse DCT (also performs dequantization) */ +typedef JMETHOD(void, inverse_DCT_method_ptr, + (j_decompress_ptr cinfo, jpeg_component_info * compptr, + JCOEFPTR coef_block, + JSAMPARRAY output_buf, JDIMENSION output_col)); + +struct jpeg_inverse_dct { + JMETHOD(void, start_pass, (j_decompress_ptr cinfo)); + /* It is useful to allow each component to have a separate IDCT method. */ + inverse_DCT_method_ptr inverse_DCT[MAX_COMPONENTS]; +}; + +/* Upsampling (note that upsampler must also call color converter) */ +struct jpeg_upsampler { + JMETHOD(void, start_pass, (j_decompress_ptr cinfo)); + JMETHOD(void, upsample, (j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, + JDIMENSION *in_row_group_ctr, + JDIMENSION in_row_groups_avail, + JSAMPARRAY output_buf, + JDIMENSION *out_row_ctr, + JDIMENSION out_rows_avail)); + + boolean need_context_rows; /* TRUE if need rows above & below */ +}; + +/* Colorspace conversion */ +struct jpeg_color_deconverter { + JMETHOD(void, start_pass, (j_decompress_ptr cinfo)); + JMETHOD(void, color_convert, (j_decompress_ptr cinfo, + JSAMPIMAGE input_buf, JDIMENSION input_row, + JSAMPARRAY output_buf, int num_rows)); +}; + +/* Color quantization or color precision reduction */ +struct jpeg_color_quantizer { + JMETHOD(void, start_pass, (j_decompress_ptr cinfo, boolean is_pre_scan)); + JMETHOD(void, color_quantize, (j_decompress_ptr cinfo, + JSAMPARRAY input_buf, JSAMPARRAY output_buf, + int num_rows)); + JMETHOD(void, finish_pass, (j_decompress_ptr cinfo)); + JMETHOD(void, new_color_map, (j_decompress_ptr cinfo)); +}; + + +/* Definition of range extension bits for decompression processes. + * See the comments with prepare_range_limit_table (in jdmaster.c) + * for more info. + * The recommended default value for normal applications is 2. + * Applications with special requirements may use a different value. + * For example, Ghostscript wants to use 3 for proper handling of + * wacky images with oversize coefficient values. + */ + +#define RANGE_BITS 2 +#define RANGE_CENTER (CENTERJSAMPLE << RANGE_BITS) + + +/* Miscellaneous useful macros */ + +#undef MAX +#define MAX(a,b) ((a) > (b) ? (a) : (b)) +#undef MIN +#define MIN(a,b) ((a) < (b) ? (a) : (b)) + + +/* We assume that right shift corresponds to signed division by 2 with + * rounding towards minus infinity. This is correct for typical "arithmetic + * shift" instructions that shift in copies of the sign bit. But some + * C compilers implement >> with an unsigned shift. For these machines you + * must define RIGHT_SHIFT_IS_UNSIGNED. + * RIGHT_SHIFT provides a proper signed right shift of an INT32 quantity. + * It is only applied with constant shift counts. SHIFT_TEMPS must be + * included in the variables of any routine using RIGHT_SHIFT. + */ + +#ifdef RIGHT_SHIFT_IS_UNSIGNED +#define SHIFT_TEMPS INT32 shift_temp; +#define RIGHT_SHIFT(x,shft) \ + ((shift_temp = (x)) < 0 ? \ + (shift_temp >> (shft)) | ((~((INT32) 0)) << (32-(shft))) : \ + (shift_temp >> (shft))) +#else +#define SHIFT_TEMPS +#define RIGHT_SHIFT(x,shft) ((x) >> (shft)) +#endif + +/* Descale and correctly round an INT32 value that's scaled by N bits. + * We assume RIGHT_SHIFT rounds towards minus infinity, so adding + * the fudge factor is correct for either sign of X. + */ + +#define DESCALE(x,n) RIGHT_SHIFT((x) + ((INT32) 1 << ((n)-1)), n) + + +/* Short forms of external names for systems with brain-damaged linkers. */ + +#ifdef NEED_SHORT_EXTERNAL_NAMES +#define jinit_compress_master jICompress +#define jinit_c_master_control jICMaster +#define jinit_c_main_controller jICMainC +#define jinit_c_prep_controller jICPrepC +#define jinit_c_coef_controller jICCoefC +#define jinit_color_converter jICColor +#define jinit_downsampler jIDownsampler +#define jinit_forward_dct jIFDCT +#define jinit_huff_encoder jIHEncoder +#define jinit_arith_encoder jIAEncoder +#define jinit_marker_writer jIMWriter +#define jinit_master_decompress jIDMaster +#define jinit_d_main_controller jIDMainC +#define jinit_d_coef_controller jIDCoefC +#define jinit_d_post_controller jIDPostC +#define jinit_input_controller jIInCtlr +#define jinit_marker_reader jIMReader +#define jinit_huff_decoder jIHDecoder +#define jinit_arith_decoder jIADecoder +#define jinit_inverse_dct jIIDCT +#define jinit_upsampler jIUpsampler +#define jinit_color_deconverter jIDColor +#define jinit_1pass_quantizer jI1Quant +#define jinit_2pass_quantizer jI2Quant +#define jinit_merged_upsampler jIMUpsampler +#define jinit_memory_mgr jIMemMgr +#define jdiv_round_up jDivRound +#define jround_up jRound +#define jzero_far jZeroFar +#define jcopy_sample_rows jCopySamples +#define jcopy_block_row jCopyBlocks +#define jpeg_zigzag_order jZIGTable +#define jpeg_natural_order jZAGTable +#define jpeg_natural_order7 jZAG7Table +#define jpeg_natural_order6 jZAG6Table +#define jpeg_natural_order5 jZAG5Table +#define jpeg_natural_order4 jZAG4Table +#define jpeg_natural_order3 jZAG3Table +#define jpeg_natural_order2 jZAG2Table +#define jpeg_aritab jAriTab +#endif /* NEED_SHORT_EXTERNAL_NAMES */ + + +/* On normal machines we can apply MEMCOPY() and MEMZERO() to sample arrays + * and coefficient-block arrays. This won't work on 80x86 because the arrays + * are FAR and we're assuming a small-pointer memory model. However, some + * DOS compilers provide far-pointer versions of memcpy() and memset() even + * in the small-model libraries. These will be used if USE_FMEM is defined. + * Otherwise, the routines in jutils.c do it the hard way. + */ + +#ifndef NEED_FAR_POINTERS /* normal case, same as regular macro */ +#define FMEMZERO(target,size) MEMZERO(target,size) +#else /* 80x86 case */ +#ifdef USE_FMEM +#define FMEMZERO(target,size) _fmemset((void FAR *)(target), 0, (size_t)(size)) +#else +EXTERN(void) jzero_far JPP((void FAR * target, size_t bytestozero)); +#define FMEMZERO(target,size) jzero_far(target, size) +#endif +#endif + + +/* Compression module initialization routines */ +EXTERN(void) jinit_compress_master JPP((j_compress_ptr cinfo)); +EXTERN(void) jinit_c_master_control JPP((j_compress_ptr cinfo, + boolean transcode_only)); +EXTERN(void) jinit_c_main_controller JPP((j_compress_ptr cinfo, + boolean need_full_buffer)); +EXTERN(void) jinit_c_prep_controller JPP((j_compress_ptr cinfo, + boolean need_full_buffer)); +EXTERN(void) jinit_c_coef_controller JPP((j_compress_ptr cinfo, + boolean need_full_buffer)); +EXTERN(void) jinit_color_converter JPP((j_compress_ptr cinfo)); +EXTERN(void) jinit_downsampler JPP((j_compress_ptr cinfo)); +EXTERN(void) jinit_forward_dct JPP((j_compress_ptr cinfo)); +EXTERN(void) jinit_huff_encoder JPP((j_compress_ptr cinfo)); +EXTERN(void) jinit_arith_encoder JPP((j_compress_ptr cinfo)); +EXTERN(void) jinit_marker_writer JPP((j_compress_ptr cinfo)); +/* Decompression module initialization routines */ +EXTERN(void) jinit_master_decompress JPP((j_decompress_ptr cinfo)); +EXTERN(void) jinit_d_main_controller JPP((j_decompress_ptr cinfo, + boolean need_full_buffer)); +EXTERN(void) jinit_d_coef_controller JPP((j_decompress_ptr cinfo, + boolean need_full_buffer)); +EXTERN(void) jinit_d_post_controller JPP((j_decompress_ptr cinfo, + boolean need_full_buffer)); +EXTERN(void) jinit_input_controller JPP((j_decompress_ptr cinfo)); +EXTERN(void) jinit_marker_reader JPP((j_decompress_ptr cinfo)); +EXTERN(void) jinit_huff_decoder JPP((j_decompress_ptr cinfo)); +EXTERN(void) jinit_arith_decoder JPP((j_decompress_ptr cinfo)); +EXTERN(void) jinit_inverse_dct JPP((j_decompress_ptr cinfo)); +EXTERN(void) jinit_upsampler JPP((j_decompress_ptr cinfo)); +EXTERN(void) jinit_color_deconverter JPP((j_decompress_ptr cinfo)); +EXTERN(void) jinit_1pass_quantizer JPP((j_decompress_ptr cinfo)); +EXTERN(void) jinit_2pass_quantizer JPP((j_decompress_ptr cinfo)); +EXTERN(void) jinit_merged_upsampler JPP((j_decompress_ptr cinfo)); +/* Memory manager initialization */ +EXTERN(void) jinit_memory_mgr JPP((j_common_ptr cinfo)); + +/* Utility routines in jutils.c */ +EXTERN(long) jdiv_round_up JPP((long a, long b)); +EXTERN(long) jround_up JPP((long a, long b)); +EXTERN(void) jcopy_sample_rows JPP((JSAMPARRAY input_array, + JSAMPARRAY output_array, + int num_rows, JDIMENSION num_cols)); +EXTERN(void) jcopy_block_row JPP((JBLOCKROW input_row, JBLOCKROW output_row, + JDIMENSION num_blocks)); +/* Constant tables in jutils.c */ +#if 0 /* This table is not actually needed in v6a */ +extern const int jpeg_zigzag_order[]; /* natural coef order to zigzag order */ +#endif +extern const int jpeg_natural_order[]; /* zigzag coef order to natural order */ +extern const int jpeg_natural_order7[]; /* zz to natural order for 7x7 block */ +extern const int jpeg_natural_order6[]; /* zz to natural order for 6x6 block */ +extern const int jpeg_natural_order5[]; /* zz to natural order for 5x5 block */ +extern const int jpeg_natural_order4[]; /* zz to natural order for 4x4 block */ +extern const int jpeg_natural_order3[]; /* zz to natural order for 3x3 block */ +extern const int jpeg_natural_order2[]; /* zz to natural order for 2x2 block */ + +/* Arithmetic coding probability estimation tables in jaricom.c */ +extern const INT32 jpeg_aritab[]; + +/* Suppress undefined-structure complaints if necessary. */ + +#ifdef INCOMPLETE_TYPES_BROKEN +#ifndef AM_MEMORY_MANAGER /* only jmemmgr.c defines these */ +struct jvirt_sarray_control { long dummy; }; +struct jvirt_barray_control { long dummy; }; +#endif +#endif /* INCOMPLETE_TYPES_BROKEN */ diff --git a/cmake/externals/libjpeg/jpeglib.h b/cmake/externals/libjpeg/jpeglib.h new file mode 100644 index 000000000..e7e15ab2c --- /dev/null +++ b/cmake/externals/libjpeg/jpeglib.h @@ -0,0 +1,1183 @@ +/* + * jpeglib.h + * + * Copyright (C) 1991-1998, Thomas G. Lane. + * Modified 2002-2022 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file defines the application interface for the JPEG library. + * Most applications using the library need only include this file, + * and perhaps jerror.h if they want to know the exact error codes. + */ + +#ifndef JPEGLIB_H +#define JPEGLIB_H + +/* + * First we include the configuration files that record how this + * installation of the JPEG library is set up. jconfig.h can be + * generated automatically for many systems. jmorecfg.h contains + * manual configuration options that most people need not worry about. + */ + +#ifndef JCONFIG_INCLUDED /* in case jinclude.h already did */ +#include "jconfig.h" /* widely used configuration options */ +#endif +#include "jmorecfg.h" /* seldom changed options */ + + +#ifdef __cplusplus +#ifndef DONT_USE_EXTERN_C +extern "C" { +#endif +#endif + +/* Version IDs for the JPEG library. + * Might be useful for tests like "#if JPEG_LIB_VERSION >= 90". + */ + +#define JPEG_LIB_VERSION 90 /* Compatibility version 9.0 */ +#define JPEG_LIB_VERSION_MAJOR 9 +#define JPEG_LIB_VERSION_MINOR 6 + + +/* Various constants determining the sizes of things. + * All of these are specified by the JPEG standard, + * so don't change them if you want to be compatible. + */ + +#define DCTSIZE 8 /* The basic DCT block is 8x8 coefficients */ +#define DCTSIZE2 64 /* DCTSIZE squared; # of elements in a block */ +#define NUM_QUANT_TBLS 4 /* Quantization tables are numbered 0..3 */ +#define NUM_HUFF_TBLS 4 /* Huffman tables are numbered 0..3 */ +#define NUM_ARITH_TBLS 16 /* Arith-coding tables are numbered 0..15 */ +#define MAX_COMPS_IN_SCAN 4 /* JPEG limit on # of components in one scan */ +#define MAX_SAMP_FACTOR 4 /* JPEG limit on sampling factors */ +/* Unfortunately, some bozo at Adobe saw no reason to be bound by the standard; + * the PostScript DCT filter can emit files with many more than 10 blocks/MCU. + * If you happen to run across such a file, you can up D_MAX_BLOCKS_IN_MCU + * to handle it. We even let you do this from the jconfig.h file. However, + * we strongly discourage changing C_MAX_BLOCKS_IN_MCU; just because Adobe + * sometimes emits noncompliant files doesn't mean you should too. + */ +#define C_MAX_BLOCKS_IN_MCU 10 /* compressor's limit on blocks per MCU */ +#ifndef D_MAX_BLOCKS_IN_MCU +#define D_MAX_BLOCKS_IN_MCU 10 /* decompressor's limit on blocks per MCU */ +#endif + + +/* Data structures for images (arrays of samples and of DCT coefficients). + * On 80x86 machines, the image arrays are too big for near pointers, + * but the pointer arrays can fit in near memory. + */ + +typedef JSAMPLE FAR *JSAMPROW; /* ptr to one image row of pixel samples. */ +typedef JSAMPROW *JSAMPARRAY; /* ptr to some rows (a 2-D sample array) */ +typedef JSAMPARRAY *JSAMPIMAGE; /* a 3-D sample array: top index is color */ + +typedef JCOEF JBLOCK[DCTSIZE2]; /* one block of coefficients */ +typedef JBLOCK FAR *JBLOCKROW; /* pointer to one row of coefficient blocks */ +typedef JBLOCKROW *JBLOCKARRAY; /* a 2-D array of coefficient blocks */ +typedef JBLOCKARRAY *JBLOCKIMAGE; /* a 3-D array of coefficient blocks */ + +typedef JCOEF FAR *JCOEFPTR; /* useful in a couple of places */ + + +/* Types for JPEG compression parameters and working tables. */ + + +/* DCT coefficient quantization tables. */ + +typedef struct { + /* This array gives the coefficient quantizers in natural array order + * (not the zigzag order in which they are stored in a JPEG DQT marker). + * CAUTION: IJG versions prior to v6a kept this array in zigzag order. + */ + UINT16 quantval[DCTSIZE2]; /* quantization step for each coefficient */ + /* This field is used only during compression. It's initialized FALSE when + * the table is created, and set TRUE when it's been output to the file. + * You could suppress output of a table by setting this to TRUE. + * (See jpeg_suppress_tables for an example.) + */ + boolean sent_table; /* TRUE when table has been output */ +} JQUANT_TBL; + + +/* Huffman coding tables. */ + +typedef struct { + /* These two fields directly represent the contents of a JPEG DHT marker */ + UINT8 bits[17]; /* bits[k] = # of symbols with codes of */ + /* length k bits; bits[0] is unused */ + UINT8 huffval[256]; /* The symbols, in order of incr code length */ + /* This field is used only during compression. It's initialized FALSE when + * the table is created, and set TRUE when it's been output to the file. + * You could suppress output of a table by setting this to TRUE. + * (See jpeg_suppress_tables for an example.) + */ + boolean sent_table; /* TRUE when table has been output */ +} JHUFF_TBL; + + +/* Basic info about one component (color channel). */ + +typedef struct { + /* These values are fixed over the whole image. */ + /* For compression, they must be supplied by parameter setup; */ + /* for decompression, they are read from the SOF marker. */ + int component_id; /* identifier for this component (0..255) */ + int component_index; /* its index in SOF or cinfo->comp_info[] */ + int h_samp_factor; /* horizontal sampling factor (1..4) */ + int v_samp_factor; /* vertical sampling factor (1..4) */ + int quant_tbl_no; /* quantization table selector (0..3) */ + /* These values may vary between scans. */ + /* For compression, they must be supplied by parameter setup; */ + /* for decompression, they are read from the SOS marker. */ + /* The decompressor output side may not use these variables. */ + int dc_tbl_no; /* DC entropy table selector (0..3) */ + int ac_tbl_no; /* AC entropy table selector (0..3) */ + + /* Remaining fields should be treated as private by applications. */ + + /* These values are computed during compression or decompression startup: */ + /* Component's size in DCT blocks. + * Any dummy blocks added to complete an MCU are not counted; therefore + * these values do not depend on whether a scan is interleaved or not. + */ + JDIMENSION width_in_blocks; + JDIMENSION height_in_blocks; + /* Size of a DCT block in samples, + * reflecting any scaling we choose to apply during the DCT step. + * Values from 1 to 16 are supported. + * Note that different components may receive different DCT scalings. + */ + int DCT_h_scaled_size; + int DCT_v_scaled_size; + /* The downsampled dimensions are the component's actual, unpadded number + * of samples at the main buffer (preprocessing/compression interface); + * DCT scaling is included, so + * downsampled_width = + * ceil(image_width * Hi/Hmax * DCT_h_scaled_size/block_size) + * and similarly for height. + */ + JDIMENSION downsampled_width; /* actual width in samples */ + JDIMENSION downsampled_height; /* actual height in samples */ + /* For decompression, in cases where some of the components will be + * ignored (eg grayscale output from YCbCr image), we can skip most + * computations for the unused components. + * For compression, some of the components will need further quantization + * scale by factor of 2 after DCT (eg BG_YCC output from normal RGB input). + * The field is first set TRUE for decompression, FALSE for compression + * in initial_setup, and then adapted in color conversion setup. + */ + boolean component_needed; + + /* These values are computed before starting a scan of the component. */ + /* The decompressor output side may not use these variables. */ + int MCU_width; /* number of blocks per MCU, horizontally */ + int MCU_height; /* number of blocks per MCU, vertically */ + int MCU_blocks; /* MCU_width * MCU_height */ + int MCU_sample_width; /* MCU width in samples: MCU_width * DCT_h_scaled_size */ + int last_col_width; /* # of non-dummy blocks across in last MCU */ + int last_row_height; /* # of non-dummy blocks down in last MCU */ + + /* Saved quantization table for component; NULL if none yet saved. + * See jdinput.c comments about the need for this information. + * This field is currently used only for decompression. + */ + JQUANT_TBL * quant_table; + + /* Private per-component storage for DCT or IDCT subsystem. */ + void * dct_table; +} jpeg_component_info; + + +/* The script for encoding a multiple-scan file is an array of these: */ + +typedef struct { + int comps_in_scan; /* number of components encoded in this scan */ + int component_index[MAX_COMPS_IN_SCAN]; /* their SOF/comp_info[] indexes */ + int Ss, Se; /* progressive JPEG spectral selection parms */ + int Ah, Al; /* progressive JPEG successive approx. parms */ +} jpeg_scan_info; + +/* The decompressor can save APPn and COM markers in a list of these: */ + +typedef struct jpeg_marker_struct FAR * jpeg_saved_marker_ptr; + +struct jpeg_marker_struct { + jpeg_saved_marker_ptr next; /* next in list, or NULL */ + UINT8 marker; /* marker code: JPEG_COM, or JPEG_APP0+n */ + unsigned int original_length; /* # bytes of data in the file */ + unsigned int data_length; /* # bytes of data saved at data[] */ + JOCTET FAR * data; /* the data contained in the marker */ + /* the marker length word is not counted in data_length or original_length */ +}; + +/* Known color spaces. */ + +typedef enum { + JCS_UNKNOWN, /* error/unspecified */ + JCS_GRAYSCALE, /* monochrome */ + JCS_RGB, /* red/green/blue, standard RGB (sRGB) */ + JCS_YCbCr, /* Y/Cb/Cr (also known as YUV), standard YCC */ + JCS_CMYK, /* C/M/Y/K */ + JCS_YCCK, /* Y/Cb/Cr/K */ + JCS_BG_RGB, /* big gamut red/green/blue, bg-sRGB */ + JCS_BG_YCC /* big gamut Y/Cb/Cr, bg-sYCC */ +} J_COLOR_SPACE; + +/* Supported color transforms. */ + +typedef enum { + JCT_NONE = 0, + JCT_SUBTRACT_GREEN = 1 +} J_COLOR_TRANSFORM; + +/* DCT/IDCT algorithm options. */ + +typedef enum { + JDCT_ISLOW, /* slow but accurate integer algorithm */ + JDCT_IFAST, /* faster, less accurate integer method */ + JDCT_FLOAT /* floating-point: accurate, fast on fast HW */ +} J_DCT_METHOD; + +#ifndef JDCT_DEFAULT /* may be overridden in jconfig.h */ +#define JDCT_DEFAULT JDCT_ISLOW +#endif +#ifndef JDCT_FASTEST /* may be overridden in jconfig.h */ +#define JDCT_FASTEST JDCT_IFAST +#endif + +/* Dithering options for decompression. */ + +typedef enum { + JDITHER_NONE, /* no dithering */ + JDITHER_ORDERED, /* simple ordered dither */ + JDITHER_FS /* Floyd-Steinberg error diffusion dither */ +} J_DITHER_MODE; + + +/* Common fields between JPEG compression and decompression master structs. */ + +#define jpeg_common_fields \ + struct jpeg_error_mgr * err; /* Error handler module */\ + struct jpeg_memory_mgr * mem; /* Memory manager module */\ + struct jpeg_progress_mgr * progress; /* Progress monitor, or NULL if none */\ + void * client_data; /* Available for use by application */\ + boolean is_decompressor; /* So common code can tell which is which */\ + int global_state /* For checking call sequence validity */ + +/* Routines that are to be used by both halves of the library are declared + * to receive a pointer to this structure. There are no actual instances of + * jpeg_common_struct, only of jpeg_compress_struct and jpeg_decompress_struct. + */ +struct jpeg_common_struct { + jpeg_common_fields; /* Fields common to both master struct types */ + /* Additional fields follow in an actual jpeg_compress_struct or + * jpeg_decompress_struct. All three structs must agree on these + * initial fields! (This would be a lot cleaner in C++.) + */ +}; + +typedef struct jpeg_common_struct * j_common_ptr; +typedef struct jpeg_compress_struct * j_compress_ptr; +typedef struct jpeg_decompress_struct * j_decompress_ptr; + + +/* Master record for a compression instance */ + +struct jpeg_compress_struct { + jpeg_common_fields; /* Fields shared with jpeg_decompress_struct */ + + /* Destination for compressed data */ + struct jpeg_destination_mgr * dest; + + /* Description of source image --- these fields must be filled in by + * outer application before starting compression. in_color_space must + * be correct before you can even call jpeg_set_defaults(). + */ + + JDIMENSION image_width; /* input image width */ + JDIMENSION image_height; /* input image height */ + int input_components; /* # of color components in input image */ + J_COLOR_SPACE in_color_space; /* colorspace of input image */ + + double input_gamma; /* image gamma of input image */ + + /* Compression parameters --- these fields must be set before calling + * jpeg_start_compress(). We recommend calling jpeg_set_defaults() to + * initialize everything to reasonable defaults, then changing anything + * the application specifically wants to change. That way you won't get + * burnt when new parameters are added. Also note that there are several + * helper routines to simplify changing parameters. + */ + + unsigned int scale_num, scale_denom; /* fraction by which to scale image */ + + JDIMENSION jpeg_width; /* scaled JPEG image width */ + JDIMENSION jpeg_height; /* scaled JPEG image height */ + /* Dimensions of actual JPEG image that will be written to file, + * derived from input dimensions by scaling factors above. + * These fields are computed by jpeg_start_compress(). + * You can also use jpeg_calc_jpeg_dimensions() to determine these values + * in advance of calling jpeg_start_compress(). + */ + + int data_precision; /* bits of precision in image data */ + + int num_components; /* # of color components in JPEG image */ + J_COLOR_SPACE jpeg_color_space; /* colorspace of JPEG image */ + + jpeg_component_info * comp_info; + /* comp_info[i] describes component that appears i'th in SOF */ + + JQUANT_TBL * quant_tbl_ptrs[NUM_QUANT_TBLS]; + int q_scale_factor[NUM_QUANT_TBLS]; + /* ptrs to coefficient quantization tables, or NULL if not defined, + * and corresponding scale factors (percentage, initialized 100). + */ + + JHUFF_TBL * dc_huff_tbl_ptrs[NUM_HUFF_TBLS]; + JHUFF_TBL * ac_huff_tbl_ptrs[NUM_HUFF_TBLS]; + /* ptrs to Huffman coding tables, or NULL if not defined */ + + UINT8 arith_dc_L[NUM_ARITH_TBLS]; /* L values for DC arith-coding tables */ + UINT8 arith_dc_U[NUM_ARITH_TBLS]; /* U values for DC arith-coding tables */ + UINT8 arith_ac_K[NUM_ARITH_TBLS]; /* Kx values for AC arith-coding tables */ + + int num_scans; /* # of entries in scan_info array */ + const jpeg_scan_info * scan_info; /* script for multi-scan file, or NULL */ + /* The default value of scan_info is NULL, which causes a single-scan + * sequential JPEG file to be emitted. To create a multi-scan file, + * set num_scans and scan_info to point to an array of scan definitions. + */ + + boolean raw_data_in; /* TRUE=caller supplies downsampled data */ + boolean arith_code; /* TRUE=arithmetic coding, FALSE=Huffman */ + boolean optimize_coding; /* TRUE=optimize entropy encoding parms */ + boolean CCIR601_sampling; /* TRUE=first samples are cosited */ + boolean do_fancy_downsampling; /* TRUE=apply fancy downsampling */ + int smoothing_factor; /* 1..100, or 0 for no input smoothing */ + J_DCT_METHOD dct_method; /* DCT algorithm selector */ + + /* The restart interval can be specified in absolute MCUs by setting + * restart_interval, or in MCU rows by setting restart_in_rows + * (in which case the correct restart_interval will be figured + * for each scan). + */ + unsigned int restart_interval; /* MCUs per restart, or 0 for no restart */ + int restart_in_rows; /* if > 0, MCU rows per restart interval */ + + /* Parameters controlling emission of special markers. */ + + boolean write_JFIF_header; /* should a JFIF marker be written? */ + UINT8 JFIF_major_version; /* What to write for the JFIF version number */ + UINT8 JFIF_minor_version; + /* These three values are not used by the JPEG code, merely copied */ + /* into the JFIF APP0 marker. density_unit can be 0 for unknown, */ + /* 1 for dots/inch, or 2 for dots/cm. Note that the pixel aspect */ + /* ratio is defined by X_density/Y_density even when density_unit=0. */ + UINT8 density_unit; /* JFIF code for pixel size units */ + UINT16 X_density; /* Horizontal pixel density */ + UINT16 Y_density; /* Vertical pixel density */ + boolean write_Adobe_marker; /* should an Adobe marker be written? */ + + J_COLOR_TRANSFORM color_transform; + /* Color transform identifier, writes LSE marker if nonzero */ + + /* State variable: index of next scanline to be written to + * jpeg_write_scanlines(). Application may use this to control its + * processing loop, e.g., "while (next_scanline < image_height)". + */ + + JDIMENSION next_scanline; /* 0 .. image_height-1 */ + + /* Remaining fields are known throughout compressor, but generally + * should not be touched by a surrounding application. + */ + + /* + * These fields are computed during compression startup + */ + boolean progressive_mode; /* TRUE if scan script uses progressive mode */ + int max_h_samp_factor; /* largest h_samp_factor */ + int max_v_samp_factor; /* largest v_samp_factor */ + + int min_DCT_h_scaled_size; /* smallest DCT_h_scaled_size of any component */ + int min_DCT_v_scaled_size; /* smallest DCT_v_scaled_size of any component */ + + JDIMENSION total_iMCU_rows; /* # of iMCU rows to be input to coef ctlr */ + /* The coefficient controller receives data in units of MCU rows as defined + * for fully interleaved scans (whether the JPEG file is interleaved or not). + * There are v_samp_factor * DCT_v_scaled_size sample rows of each component + * in an "iMCU" (interleaved MCU) row. + */ + + /* + * These fields are valid during any one scan. + * They describe the components and MCUs actually appearing in the scan. + */ + int comps_in_scan; /* # of JPEG components in this scan */ + jpeg_component_info * cur_comp_info[MAX_COMPS_IN_SCAN]; + /* *cur_comp_info[i] describes component that appears i'th in SOS */ + + JDIMENSION MCUs_per_row; /* # of MCUs across the image */ + JDIMENSION MCU_rows_in_scan; /* # of MCU rows in the image */ + + int blocks_in_MCU; /* # of DCT blocks per MCU */ + int MCU_membership[C_MAX_BLOCKS_IN_MCU]; + /* MCU_membership[i] is index in cur_comp_info of component owning */ + /* i'th block in an MCU */ + + int Ss, Se, Ah, Al; /* progressive JPEG parameters for scan */ + + int block_size; /* the basic DCT block size: 1..16 */ + const int * natural_order; /* natural-order position array */ + int lim_Se; /* min( Se, DCTSIZE2-1 ) */ + + /* + * Links to compression subobjects (methods and private variables of modules) + */ + struct jpeg_comp_master * master; + struct jpeg_c_main_controller * main; + struct jpeg_c_prep_controller * prep; + struct jpeg_c_coef_controller * coef; + struct jpeg_marker_writer * marker; + struct jpeg_color_converter * cconvert; + struct jpeg_downsampler * downsample; + struct jpeg_forward_dct * fdct; + struct jpeg_entropy_encoder * entropy; + jpeg_scan_info * script_space; /* workspace for jpeg_simple_progression */ + int script_space_size; +}; + + +/* Master record for a decompression instance */ + +struct jpeg_decompress_struct { + jpeg_common_fields; /* Fields shared with jpeg_compress_struct */ + + /* Source of compressed data */ + struct jpeg_source_mgr * src; + + /* Basic description of image --- filled in by jpeg_read_header(). */ + /* Application may inspect these values to decide how to process image. */ + + JDIMENSION image_width; /* nominal image width (from SOF marker) */ + JDIMENSION image_height; /* nominal image height */ + int num_components; /* # of color components in JPEG image */ + J_COLOR_SPACE jpeg_color_space; /* colorspace of JPEG image */ + + /* Decompression processing parameters --- these fields must be set before + * calling jpeg_start_decompress(). Note that jpeg_read_header() initializes + * them to default values. + */ + + J_COLOR_SPACE out_color_space; /* colorspace for output */ + + unsigned int scale_num, scale_denom; /* fraction by which to scale image */ + + double output_gamma; /* image gamma wanted in output */ + + boolean buffered_image; /* TRUE=multiple output passes */ + boolean raw_data_out; /* TRUE=downsampled data wanted */ + + J_DCT_METHOD dct_method; /* IDCT algorithm selector */ + boolean do_fancy_upsampling; /* TRUE=apply fancy upsampling */ + boolean do_block_smoothing; /* TRUE=apply interblock smoothing */ + + boolean quantize_colors; /* TRUE=colormapped output wanted */ + /* the following are ignored if not quantize_colors: */ + J_DITHER_MODE dither_mode; /* type of color dithering to use */ + boolean two_pass_quantize; /* TRUE=use two-pass color quantization */ + int desired_number_of_colors; /* max # colors to use in created colormap */ + /* these are significant only in buffered-image mode: */ + boolean enable_1pass_quant; /* enable future use of 1-pass quantizer */ + boolean enable_external_quant;/* enable future use of external colormap */ + boolean enable_2pass_quant; /* enable future use of 2-pass quantizer */ + + /* Description of actual output image that will be returned to application. + * These fields are computed by jpeg_start_decompress(). + * You can also use jpeg_calc_output_dimensions() to determine these values + * in advance of calling jpeg_start_decompress(). + */ + + JDIMENSION output_width; /* scaled image width */ + JDIMENSION output_height; /* scaled image height */ + int out_color_components; /* # of color components in out_color_space */ + int output_components; /* # of color components returned */ + /* output_components is 1 (a colormap index) when quantizing colors; + * otherwise it equals out_color_components. + */ + int rec_outbuf_height; /* min recommended height of scanline buffer */ + /* If the buffer passed to jpeg_read_scanlines() is less than this many rows + * high, space and time will be wasted due to unnecessary data copying. + * Usually rec_outbuf_height will be 1 or 2, at most 4. + */ + + /* When quantizing colors, the output colormap is described by these fields. + * The application can supply a colormap by setting colormap non-NULL before + * calling jpeg_start_decompress; otherwise a colormap is created during + * jpeg_start_decompress or jpeg_start_output. + * The map has out_color_components rows and actual_number_of_colors columns. + */ + int actual_number_of_colors; /* number of entries in use */ + JSAMPARRAY colormap; /* The color map as a 2-D pixel array */ + + /* State variables: these variables indicate the progress of decompression. + * The application may examine these but must not modify them. + */ + + /* Row index of next scanline to be read from jpeg_read_scanlines(). + * Application may use this to control its processing loop, e.g., + * "while (output_scanline < output_height)". + */ + JDIMENSION output_scanline; /* 0 .. output_height-1 */ + + /* Current input scan number and number of iMCU rows completed in scan. + * These indicate the progress of the decompressor input side. + */ + int input_scan_number; /* Number of SOS markers seen so far */ + JDIMENSION input_iMCU_row; /* Number of iMCU rows completed */ + + /* The "output scan number" is the notional scan being displayed by the + * output side. The decompressor will not allow output scan/row number + * to get ahead of input scan/row, but it can fall arbitrarily far behind. + */ + int output_scan_number; /* Nominal scan number being displayed */ + JDIMENSION output_iMCU_row; /* Number of iMCU rows read */ + + /* Current progression status. coef_bits[c][i] indicates the precision + * with which component c's DCT coefficient i (in zigzag order) is known. + * It is -1 when no data has yet been received, otherwise it is the point + * transform (shift) value for the most recent scan of the coefficient + * (thus, 0 at completion of the progression). + * This pointer is NULL when reading a non-progressive file. + */ + int (*coef_bits)[DCTSIZE2]; /* -1 or current Al value for each coef */ + + /* Internal JPEG parameters --- the application usually need not look at + * these fields. Note that the decompressor output side may not use + * any parameters that can change between scans. + */ + + /* Quantization and Huffman tables are carried forward across input + * datastreams when processing abbreviated JPEG datastreams. + */ + + JQUANT_TBL * quant_tbl_ptrs[NUM_QUANT_TBLS]; + /* ptrs to coefficient quantization tables, or NULL if not defined */ + + JHUFF_TBL * dc_huff_tbl_ptrs[NUM_HUFF_TBLS]; + JHUFF_TBL * ac_huff_tbl_ptrs[NUM_HUFF_TBLS]; + /* ptrs to Huffman coding tables, or NULL if not defined */ + + /* These parameters are never carried across datastreams, since they + * are given in SOF/SOS markers or defined to be reset by SOI. + */ + + int data_precision; /* bits of precision in image data */ + + jpeg_component_info * comp_info; + /* comp_info[i] describes component that appears i'th in SOF */ + + boolean is_baseline; /* TRUE if Baseline SOF0 encountered */ + boolean progressive_mode; /* TRUE if SOFn specifies progressive mode */ + boolean arith_code; /* TRUE=arithmetic coding, FALSE=Huffman */ + + UINT8 arith_dc_L[NUM_ARITH_TBLS]; /* L values for DC arith-coding tables */ + UINT8 arith_dc_U[NUM_ARITH_TBLS]; /* U values for DC arith-coding tables */ + UINT8 arith_ac_K[NUM_ARITH_TBLS]; /* Kx values for AC arith-coding tables */ + + unsigned int restart_interval; /* MCUs per restart interval, or 0 for no restart */ + + /* These fields record data obtained from optional markers recognized by + * the JPEG library. + */ + boolean saw_JFIF_marker; /* TRUE iff a JFIF APP0 marker was found */ + /* Data copied from JFIF marker; only valid if saw_JFIF_marker is TRUE: */ + UINT8 JFIF_major_version; /* JFIF version number */ + UINT8 JFIF_minor_version; + UINT8 density_unit; /* JFIF code for pixel size units */ + UINT16 X_density; /* Horizontal pixel density */ + UINT16 Y_density; /* Vertical pixel density */ + boolean saw_Adobe_marker; /* TRUE iff an Adobe APP14 marker was found */ + UINT8 Adobe_transform; /* Color transform code from Adobe marker */ + + J_COLOR_TRANSFORM color_transform; + /* Color transform identifier derived from LSE marker, otherwise zero */ + + boolean CCIR601_sampling; /* TRUE=first samples are cosited */ + + /* Aside from the specific data retained from APPn markers known to the + * library, the uninterpreted contents of any or all APPn and COM markers + * can be saved in a list for examination by the application. + */ + jpeg_saved_marker_ptr marker_list; /* Head of list of saved markers */ + + /* Remaining fields are known throughout decompressor, but generally + * should not be touched by a surrounding application. + */ + + /* + * These fields are computed during decompression startup + */ + int max_h_samp_factor; /* largest h_samp_factor */ + int max_v_samp_factor; /* largest v_samp_factor */ + + int min_DCT_h_scaled_size; /* smallest DCT_h_scaled_size of any component */ + int min_DCT_v_scaled_size; /* smallest DCT_v_scaled_size of any component */ + + JDIMENSION total_iMCU_rows; /* # of iMCU rows in image */ + /* The coefficient controller's input and output progress is measured in + * units of "iMCU" (interleaved MCU) rows. These are the same as MCU rows + * in fully interleaved JPEG scans, but are used whether the scan is + * interleaved or not. We define an iMCU row as v_samp_factor DCT block + * rows of each component. Therefore, the IDCT output contains + * v_samp_factor * DCT_v_scaled_size sample rows of a component per iMCU row. + */ + + JSAMPLE * sample_range_limit; /* table for fast range-limiting */ + + /* + * These fields are valid during any one scan. + * They describe the components and MCUs actually appearing in the scan. + * Note that the decompressor output side must not use these fields. + */ + int comps_in_scan; /* # of JPEG components in this scan */ + jpeg_component_info * cur_comp_info[MAX_COMPS_IN_SCAN]; + /* *cur_comp_info[i] describes component that appears i'th in SOS */ + + JDIMENSION MCUs_per_row; /* # of MCUs across the image */ + JDIMENSION MCU_rows_in_scan; /* # of MCU rows in the image */ + + int blocks_in_MCU; /* # of DCT blocks per MCU */ + int MCU_membership[D_MAX_BLOCKS_IN_MCU]; + /* MCU_membership[i] is index in cur_comp_info of component owning */ + /* i'th block in an MCU */ + + int Ss, Se, Ah, Al; /* progressive JPEG parameters for scan */ + + /* These fields are derived from Se of first SOS marker. + */ + int block_size; /* the basic DCT block size: 1..16 */ + const int * natural_order; /* natural-order position array for entropy decode */ + int lim_Se; /* min( Se, DCTSIZE2-1 ) for entropy decode */ + + /* This field is shared between entropy decoder and marker parser. + * It is either zero or the code of a JPEG marker that has been + * read from the data source, but has not yet been processed. + */ + int unread_marker; + + /* + * Links to decompression subobjects (methods, private variables of modules) + */ + struct jpeg_decomp_master * master; + struct jpeg_d_main_controller * main; + struct jpeg_d_coef_controller * coef; + struct jpeg_d_post_controller * post; + struct jpeg_input_controller * inputctl; + struct jpeg_marker_reader * marker; + struct jpeg_entropy_decoder * entropy; + struct jpeg_inverse_dct * idct; + struct jpeg_upsampler * upsample; + struct jpeg_color_deconverter * cconvert; + struct jpeg_color_quantizer * cquantize; +}; + + +/* "Object" declarations for JPEG modules that may be supplied or called + * directly by the surrounding application. + * As with all objects in the JPEG library, these structs only define the + * publicly visible methods and state variables of a module. Additional + * private fields may exist after the public ones. + */ + + +/* Error handler object */ + +struct jpeg_error_mgr { + /* Error exit handler: does not return to caller */ + JMETHOD(noreturn_t, error_exit, (j_common_ptr cinfo)); + /* Conditionally emit a trace or warning message */ + JMETHOD(void, emit_message, (j_common_ptr cinfo, int msg_level)); + /* Routine that actually outputs a trace or error message */ + JMETHOD(void, output_message, (j_common_ptr cinfo)); + /* Format a message string for the most recent JPEG error or message */ + JMETHOD(void, format_message, (j_common_ptr cinfo, char * buffer)); +#define JMSG_LENGTH_MAX 200 /* recommended size of format_message buffer */ + /* Reset error state variables at start of a new image */ + JMETHOD(void, reset_error_mgr, (j_common_ptr cinfo)); + + /* The message ID code and any parameters are saved here. + * A message can have one string parameter or up to 8 int parameters. + */ + int msg_code; +#define JMSG_STR_PARM_MAX 80 + union { + int i[8]; + char s[JMSG_STR_PARM_MAX]; + } msg_parm; + + /* Standard state variables for error facility */ + + int trace_level; /* max msg_level that will be displayed */ + + /* For recoverable corrupt-data errors, we emit a warning message, + * but keep going unless emit_message chooses to abort. emit_message + * should count warnings in num_warnings. The surrounding application + * can check for bad data by seeing if num_warnings is nonzero at the + * end of processing. + */ + long num_warnings; /* number of corrupt-data warnings */ + + /* These fields point to the table(s) of error message strings. + * An application can change the table pointer to switch to a different + * message list (typically, to change the language in which errors are + * reported). Some applications may wish to add additional error codes + * that will be handled by the JPEG library error mechanism; the second + * table pointer is used for this purpose. + * + * First table includes all errors generated by JPEG library itself. + * Error code 0 is reserved for a "no such error string" message. + */ + const char * const * jpeg_message_table; /* Library errors */ + int last_jpeg_message; /* Table contains strings 0..last_jpeg_message */ + /* Second table can be added by application (see cjpeg/djpeg for example). + * It contains strings numbered first_addon_message..last_addon_message. + */ + const char * const * addon_message_table; /* Non-library errors */ + int first_addon_message; /* code for first string in addon table */ + int last_addon_message; /* code for last string in addon table */ +}; + + +/* Progress monitor object */ + +struct jpeg_progress_mgr { + JMETHOD(void, progress_monitor, (j_common_ptr cinfo)); + + long pass_counter; /* work units completed in this pass */ + long pass_limit; /* total number of work units in this pass */ + int completed_passes; /* passes completed so far */ + int total_passes; /* total number of passes expected */ +}; + + +/* Data destination object for compression */ + +struct jpeg_destination_mgr { + JOCTET * next_output_byte; /* => next byte to write in buffer */ + size_t free_in_buffer; /* # of byte spaces remaining in buffer */ + + JMETHOD(void, init_destination, (j_compress_ptr cinfo)); + JMETHOD(boolean, empty_output_buffer, (j_compress_ptr cinfo)); + JMETHOD(void, term_destination, (j_compress_ptr cinfo)); +}; + + +/* Data source object for decompression */ + +struct jpeg_source_mgr { + const JOCTET * next_input_byte; /* => next byte to read from buffer */ + size_t bytes_in_buffer; /* # of bytes remaining in buffer */ + + JMETHOD(void, init_source, (j_decompress_ptr cinfo)); + JMETHOD(boolean, fill_input_buffer, (j_decompress_ptr cinfo)); + JMETHOD(void, skip_input_data, (j_decompress_ptr cinfo, long num_bytes)); + JMETHOD(boolean, resync_to_restart, (j_decompress_ptr cinfo, int desired)); + JMETHOD(void, term_source, (j_decompress_ptr cinfo)); +}; + + +/* Memory manager object. + * Allocates "small" objects (a few K total), "large" objects (tens of K), + * and "really big" objects (virtual arrays with backing store if needed). + * The memory manager does not allow individual objects to be freed; rather, + * each created object is assigned to a pool, and whole pools can be freed + * at once. This is faster and more convenient than remembering exactly what + * to free, especially where malloc()/free() are not too speedy. + * NB: alloc routines never return NULL. They exit to error_exit if not + * successful. + */ + +#define JPOOL_PERMANENT 0 /* lasts until master record is destroyed */ +#define JPOOL_IMAGE 1 /* lasts until done with image/datastream */ +#define JPOOL_NUMPOOLS 2 + +typedef struct jvirt_sarray_control * jvirt_sarray_ptr; +typedef struct jvirt_barray_control * jvirt_barray_ptr; + + +struct jpeg_memory_mgr { + /* Method pointers */ + JMETHOD(void *, alloc_small, (j_common_ptr cinfo, int pool_id, + size_t sizeofobject)); + JMETHOD(void FAR *, alloc_large, (j_common_ptr cinfo, int pool_id, + size_t sizeofobject)); + JMETHOD(JSAMPARRAY, alloc_sarray, (j_common_ptr cinfo, int pool_id, + JDIMENSION samplesperrow, + JDIMENSION numrows)); + JMETHOD(JBLOCKARRAY, alloc_barray, (j_common_ptr cinfo, int pool_id, + JDIMENSION blocksperrow, + JDIMENSION numrows)); + JMETHOD(jvirt_sarray_ptr, request_virt_sarray, (j_common_ptr cinfo, + int pool_id, + boolean pre_zero, + JDIMENSION samplesperrow, + JDIMENSION numrows, + JDIMENSION maxaccess)); + JMETHOD(jvirt_barray_ptr, request_virt_barray, (j_common_ptr cinfo, + int pool_id, + boolean pre_zero, + JDIMENSION blocksperrow, + JDIMENSION numrows, + JDIMENSION maxaccess)); + JMETHOD(void, realize_virt_arrays, (j_common_ptr cinfo)); + JMETHOD(JSAMPARRAY, access_virt_sarray, (j_common_ptr cinfo, + jvirt_sarray_ptr ptr, + JDIMENSION start_row, + JDIMENSION num_rows, + boolean writable)); + JMETHOD(JBLOCKARRAY, access_virt_barray, (j_common_ptr cinfo, + jvirt_barray_ptr ptr, + JDIMENSION start_row, + JDIMENSION num_rows, + boolean writable)); + JMETHOD(void, free_pool, (j_common_ptr cinfo, int pool_id)); + JMETHOD(void, self_destruct, (j_common_ptr cinfo)); + + /* Limit on memory allocation for this JPEG object. (Note that this is + * merely advisory, not a guaranteed maximum; it only affects the space + * used for virtual-array buffers.) May be changed by outer application + * after creating the JPEG object. + */ + long max_memory_to_use; + + /* Maximum allocation request accepted by alloc_large. */ + long max_alloc_chunk; +}; + + +/* Routine signature for application-supplied marker processing methods. + * Need not pass marker code since it is stored in cinfo->unread_marker. + */ +typedef JMETHOD(boolean, jpeg_marker_parser_method, (j_decompress_ptr cinfo)); + + +/* Declarations for routines called by application. + * The JPP macro hides prototype parameters from compilers that can't cope. + * Note JPP requires double parentheses. + */ + +#ifdef HAVE_PROTOTYPES +#define JPP(arglist) arglist +#else +#define JPP(arglist) () +#endif + + +/* Short forms of external names for systems with brain-damaged linkers. + * We shorten external names to be unique in the first six letters, which + * is good enough for all known systems. + * (If your compiler itself needs names to be unique in less than 15 + * characters, you are out of luck. Get a better compiler.) + */ + +#ifdef NEED_SHORT_EXTERNAL_NAMES +#define jpeg_std_error jStdError +#define jpeg_CreateCompress jCreaCompress +#define jpeg_CreateDecompress jCreaDecompress +#define jpeg_destroy_compress jDestCompress +#define jpeg_destroy_decompress jDestDecompress +#define jpeg_stdio_dest jStdDest +#define jpeg_stdio_src jStdSrc +#define jpeg_mem_dest jMemDest +#define jpeg_mem_src jMemSrc +#define jpeg_set_defaults jSetDefaults +#define jpeg_set_colorspace jSetColorspace +#define jpeg_default_colorspace jDefColorspace +#define jpeg_set_quality jSetQuality +#define jpeg_set_linear_quality jSetLQuality +#define jpeg_default_qtables jDefQTables +#define jpeg_add_quant_table jAddQuantTable +#define jpeg_quality_scaling jQualityScaling +#define jpeg_simple_progression jSimProgress +#define jpeg_suppress_tables jSuppressTables +#define jpeg_alloc_quant_table jAlcQTable +#define jpeg_alloc_huff_table jAlcHTable +#define jpeg_std_huff_table jStdHTable +#define jpeg_start_compress jStrtCompress +#define jpeg_write_scanlines jWrtScanlines +#define jpeg_finish_compress jFinCompress +#define jpeg_calc_jpeg_dimensions jCjpegDimensions +#define jpeg_write_raw_data jWrtRawData +#define jpeg_write_marker jWrtMarker +#define jpeg_write_m_header jWrtMHeader +#define jpeg_write_m_byte jWrtMByte +#define jpeg_write_tables jWrtTables +#define jpeg_read_header jReadHeader +#define jpeg_start_decompress jStrtDecompress +#define jpeg_read_scanlines jReadScanlines +#define jpeg_finish_decompress jFinDecompress +#define jpeg_read_raw_data jReadRawData +#define jpeg_has_multiple_scans jHasMultScn +#define jpeg_start_output jStrtOutput +#define jpeg_finish_output jFinOutput +#define jpeg_input_complete jInComplete +#define jpeg_new_colormap jNewCMap +#define jpeg_consume_input jConsumeInput +#define jpeg_core_output_dimensions jCoreDimensions +#define jpeg_calc_output_dimensions jCalcDimensions +#define jpeg_save_markers jSaveMarkers +#define jpeg_set_marker_processor jSetMarker +#define jpeg_read_coefficients jReadCoefs +#define jpeg_write_coefficients jWrtCoefs +#define jpeg_copy_critical_parameters jCopyCrit +#define jpeg_abort_compress jAbrtCompress +#define jpeg_abort_decompress jAbrtDecompress +#define jpeg_abort jAbort +#define jpeg_destroy jDestroy +#define jpeg_resync_to_restart jResyncRestart +#endif /* NEED_SHORT_EXTERNAL_NAMES */ + + +/* Default error-management setup */ +EXTERN(struct jpeg_error_mgr *) jpeg_std_error + JPP((struct jpeg_error_mgr * err)); + +/* Initialization of JPEG compression objects. + * jpeg_create_compress() and jpeg_create_decompress() are the exported + * names that applications should call. These expand to calls on + * jpeg_CreateCompress and jpeg_CreateDecompress with additional information + * passed for version mismatch checking. + * NB: you must set up the error-manager BEFORE calling jpeg_create_xxx. + */ +#define jpeg_create_compress(cinfo) \ + jpeg_CreateCompress((cinfo), JPEG_LIB_VERSION, \ + (size_t) sizeof(struct jpeg_compress_struct)) +#define jpeg_create_decompress(cinfo) \ + jpeg_CreateDecompress((cinfo), JPEG_LIB_VERSION, \ + (size_t) sizeof(struct jpeg_decompress_struct)) +EXTERN(void) jpeg_CreateCompress JPP((j_compress_ptr cinfo, + int version, size_t structsize)); +EXTERN(void) jpeg_CreateDecompress JPP((j_decompress_ptr cinfo, + int version, size_t structsize)); +/* Destruction of JPEG compression objects */ +EXTERN(void) jpeg_destroy_compress JPP((j_compress_ptr cinfo)); +EXTERN(void) jpeg_destroy_decompress JPP((j_decompress_ptr cinfo)); + +/* Standard data source and destination managers: stdio streams. */ +/* Caller is responsible for opening the file before and closing after. */ +EXTERN(void) jpeg_stdio_dest JPP((j_compress_ptr cinfo, FILE * outfile)); +EXTERN(void) jpeg_stdio_src JPP((j_decompress_ptr cinfo, FILE * infile)); + +/* Data source and destination managers: memory buffers. */ +EXTERN(void) jpeg_mem_dest JPP((j_compress_ptr cinfo, + unsigned char ** outbuffer, + size_t * outsize)); +EXTERN(void) jpeg_mem_src JPP((j_decompress_ptr cinfo, + const unsigned char * inbuffer, + size_t insize)); + +/* Default parameter setup for compression */ +EXTERN(void) jpeg_set_defaults JPP((j_compress_ptr cinfo)); +/* Compression parameter setup aids */ +EXTERN(void) jpeg_set_colorspace JPP((j_compress_ptr cinfo, + J_COLOR_SPACE colorspace)); +EXTERN(void) jpeg_default_colorspace JPP((j_compress_ptr cinfo)); +EXTERN(void) jpeg_set_quality JPP((j_compress_ptr cinfo, int quality, + boolean force_baseline)); +EXTERN(void) jpeg_set_linear_quality JPP((j_compress_ptr cinfo, + int scale_factor, + boolean force_baseline)); +EXTERN(void) jpeg_default_qtables JPP((j_compress_ptr cinfo, + boolean force_baseline)); +EXTERN(void) jpeg_add_quant_table JPP((j_compress_ptr cinfo, int which_tbl, + const unsigned int *basic_table, + int scale_factor, + boolean force_baseline)); +EXTERN(int) jpeg_quality_scaling JPP((int quality)); +EXTERN(void) jpeg_simple_progression JPP((j_compress_ptr cinfo)); +EXTERN(void) jpeg_suppress_tables JPP((j_compress_ptr cinfo, + boolean suppress)); +EXTERN(JQUANT_TBL *) jpeg_alloc_quant_table JPP((j_common_ptr cinfo)); +EXTERN(JHUFF_TBL *) jpeg_alloc_huff_table JPP((j_common_ptr cinfo)); +EXTERN(JHUFF_TBL *) jpeg_std_huff_table JPP((j_common_ptr cinfo, + boolean isDC, int tblno)); + +/* Main entry points for compression */ +EXTERN(void) jpeg_start_compress JPP((j_compress_ptr cinfo, + boolean write_all_tables)); +EXTERN(JDIMENSION) jpeg_write_scanlines JPP((j_compress_ptr cinfo, + JSAMPARRAY scanlines, + JDIMENSION num_lines)); +EXTERN(void) jpeg_finish_compress JPP((j_compress_ptr cinfo)); + +/* Precalculate JPEG dimensions for current compression parameters. */ +EXTERN(void) jpeg_calc_jpeg_dimensions JPP((j_compress_ptr cinfo)); + +/* Replaces jpeg_write_scanlines when writing raw downsampled data. */ +EXTERN(JDIMENSION) jpeg_write_raw_data JPP((j_compress_ptr cinfo, + JSAMPIMAGE data, + JDIMENSION num_lines)); + +/* Write a special marker. See libjpeg.txt concerning safe usage. */ +EXTERN(void) jpeg_write_marker + JPP((j_compress_ptr cinfo, int marker, + const JOCTET * dataptr, unsigned int datalen)); +/* Same, but piecemeal. */ +EXTERN(void) jpeg_write_m_header + JPP((j_compress_ptr cinfo, int marker, unsigned int datalen)); +EXTERN(void) jpeg_write_m_byte + JPP((j_compress_ptr cinfo, int val)); + +/* Alternate compression function: just write an abbreviated table file */ +EXTERN(void) jpeg_write_tables JPP((j_compress_ptr cinfo)); + +/* Decompression startup: read start of JPEG datastream to see what's there */ +EXTERN(int) jpeg_read_header JPP((j_decompress_ptr cinfo, + boolean require_image)); +/* Return value is one of: */ +#define JPEG_SUSPENDED 0 /* Suspended due to lack of input data */ +#define JPEG_HEADER_OK 1 /* Found valid image datastream */ +#define JPEG_HEADER_TABLES_ONLY 2 /* Found valid table-specs-only datastream */ +/* If you pass require_image = TRUE (normal case), you need not check for + * a TABLES_ONLY return code; an abbreviated file will cause an error exit. + * JPEG_SUSPENDED is only possible if you use a data source module that can + * give a suspension return (the stdio source module doesn't). + */ + +/* Main entry points for decompression */ +EXTERN(boolean) jpeg_start_decompress JPP((j_decompress_ptr cinfo)); +EXTERN(JDIMENSION) jpeg_read_scanlines JPP((j_decompress_ptr cinfo, + JSAMPARRAY scanlines, + JDIMENSION max_lines)); +EXTERN(boolean) jpeg_finish_decompress JPP((j_decompress_ptr cinfo)); + +/* Replaces jpeg_read_scanlines when reading raw downsampled data. */ +EXTERN(JDIMENSION) jpeg_read_raw_data JPP((j_decompress_ptr cinfo, + JSAMPIMAGE data, + JDIMENSION max_lines)); + +/* Additional entry points for buffered-image mode. */ +EXTERN(boolean) jpeg_has_multiple_scans JPP((j_decompress_ptr cinfo)); +EXTERN(boolean) jpeg_start_output JPP((j_decompress_ptr cinfo, + int scan_number)); +EXTERN(boolean) jpeg_finish_output JPP((j_decompress_ptr cinfo)); +EXTERN(boolean) jpeg_input_complete JPP((j_decompress_ptr cinfo)); +EXTERN(void) jpeg_new_colormap JPP((j_decompress_ptr cinfo)); +EXTERN(int) jpeg_consume_input JPP((j_decompress_ptr cinfo)); +/* Return value is one of: */ +/* #define JPEG_SUSPENDED 0 Suspended due to lack of input data */ +#define JPEG_REACHED_SOS 1 /* Reached start of new scan */ +#define JPEG_REACHED_EOI 2 /* Reached end of image */ +#define JPEG_ROW_COMPLETED 3 /* Completed one iMCU row */ +#define JPEG_SCAN_COMPLETED 4 /* Completed last iMCU row of a scan */ + +/* Precalculate output dimensions for current decompression parameters. */ +EXTERN(void) jpeg_core_output_dimensions JPP((j_decompress_ptr cinfo)); +EXTERN(void) jpeg_calc_output_dimensions JPP((j_decompress_ptr cinfo)); + +/* Control saving of COM and APPn markers into marker_list. */ +EXTERN(void) jpeg_save_markers + JPP((j_decompress_ptr cinfo, int marker_code, + unsigned int length_limit)); + +/* Install a special processing method for COM or APPn markers. */ +EXTERN(void) jpeg_set_marker_processor + JPP((j_decompress_ptr cinfo, int marker_code, + jpeg_marker_parser_method routine)); + +/* Read or write raw DCT coefficients --- useful for lossless transcoding. */ +EXTERN(jvirt_barray_ptr *) jpeg_read_coefficients JPP((j_decompress_ptr cinfo)); +EXTERN(void) jpeg_write_coefficients JPP((j_compress_ptr cinfo, + jvirt_barray_ptr * coef_arrays)); +EXTERN(void) jpeg_copy_critical_parameters JPP((j_decompress_ptr srcinfo, + j_compress_ptr dstinfo)); + +/* If you choose to abort compression or decompression before completing + * jpeg_finish_(de)compress, then you need to clean up to release memory, + * temporary files, etc. You can just call jpeg_destroy_(de)compress + * if you're done with the JPEG object, but if you want to clean it up and + * reuse it, call this: + */ +EXTERN(void) jpeg_abort_compress JPP((j_compress_ptr cinfo)); +EXTERN(void) jpeg_abort_decompress JPP((j_decompress_ptr cinfo)); + +/* Generic versions of jpeg_abort and jpeg_destroy that work on either + * flavor of JPEG object. These may be more convenient in some places. + */ +EXTERN(void) jpeg_abort JPP((j_common_ptr cinfo)); +EXTERN(void) jpeg_destroy JPP((j_common_ptr cinfo)); + +/* Default restart-marker-resync procedure for use by data source modules */ +EXTERN(boolean) jpeg_resync_to_restart JPP((j_decompress_ptr cinfo, + int desired)); + + +/* These marker codes are exported since applications and data source modules + * are likely to want to use them. + */ + +#define JPEG_RST0 0xD0 /* RST0 marker code */ +#define JPEG_EOI 0xD9 /* EOI marker code */ +#define JPEG_APP0 0xE0 /* APP0 marker code */ +#define JPEG_COM 0xFE /* COM marker code */ + + +/* If we have a brain-damaged compiler that emits warnings (or worse, errors) + * for structure definitions that are never filled in, keep it quiet by + * supplying dummy definitions for the various substructures. + */ + +#ifdef INCOMPLETE_TYPES_BROKEN +#ifndef JPEG_INTERNALS /* will be defined in jpegint.h */ +struct jvirt_sarray_control { long dummy; }; +struct jvirt_barray_control { long dummy; }; +struct jpeg_comp_master { long dummy; }; +struct jpeg_c_main_controller { long dummy; }; +struct jpeg_c_prep_controller { long dummy; }; +struct jpeg_c_coef_controller { long dummy; }; +struct jpeg_marker_writer { long dummy; }; +struct jpeg_color_converter { long dummy; }; +struct jpeg_downsampler { long dummy; }; +struct jpeg_forward_dct { long dummy; }; +struct jpeg_entropy_encoder { long dummy; }; +struct jpeg_decomp_master { long dummy; }; +struct jpeg_d_main_controller { long dummy; }; +struct jpeg_d_coef_controller { long dummy; }; +struct jpeg_d_post_controller { long dummy; }; +struct jpeg_input_controller { long dummy; }; +struct jpeg_marker_reader { long dummy; }; +struct jpeg_entropy_decoder { long dummy; }; +struct jpeg_inverse_dct { long dummy; }; +struct jpeg_upsampler { long dummy; }; +struct jpeg_color_deconverter { long dummy; }; +struct jpeg_color_quantizer { long dummy; }; +#endif /* JPEG_INTERNALS */ +#endif /* INCOMPLETE_TYPES_BROKEN */ + + +/* + * The JPEG library modules define JPEG_INTERNALS before including this file. + * The internal structure declarations are read only when that is true. + * Applications using the library should not include jpegint.h, but may wish + * to include jerror.h. + */ + +#ifdef JPEG_INTERNALS +#include "jpegint.h" /* fetch private declarations */ +#include "jerror.h" /* fetch error codes too */ +#endif + +#ifdef __cplusplus +#ifndef DONT_USE_EXTERN_C +} +#endif +#endif + +#endif /* JPEGLIB_H */ diff --git a/cmake/externals/libjpeg/jquant1.c b/cmake/externals/libjpeg/jquant1.c new file mode 100644 index 000000000..60b1843e7 --- /dev/null +++ b/cmake/externals/libjpeg/jquant1.c @@ -0,0 +1,851 @@ +/* + * jquant1.c + * + * Copyright (C) 1991-1996, Thomas G. Lane. + * Modified 2011-2020 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains 1-pass color quantization (color mapping) routines. + * These routines provide mapping to a fixed color map using equally spaced + * color values. Optional Floyd-Steinberg or ordered dithering is available. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + +#ifdef QUANT_1PASS_SUPPORTED + + +/* + * The main purpose of 1-pass quantization is to provide a fast, if not very + * high quality, colormapped output capability. A 2-pass quantizer usually + * gives better visual quality; however, for quantized grayscale output this + * quantizer is perfectly adequate. Dithering is highly recommended with this + * quantizer, though you can turn it off if you really want to. + * + * In 1-pass quantization the colormap must be chosen in advance of seeing the + * image. We use a map consisting of all combinations of Ncolors[i] color + * values for the i'th component. The Ncolors[] values are chosen so that + * their product, the total number of colors, is no more than that requested. + * (In most cases, the product will be somewhat less.) + * + * Since the colormap is orthogonal, the representative value for each color + * component can be determined without considering the other components; + * then these indexes can be combined into a colormap index by a standard + * N-dimensional-array-subscript calculation. Most of the arithmetic involved + * can be precalculated and stored in the lookup table colorindex[]. + * colorindex[i][j] maps pixel value j in component i to the nearest + * representative value (grid plane) for that component; this index is + * multiplied by the array stride for component i, so that the + * index of the colormap entry closest to a given pixel value is just + * sum( colorindex[component-number][pixel-component-value] ) + * Aside from being fast, this scheme allows for variable spacing between + * representative values with no additional lookup cost. + * + * If gamma correction has been applied in color conversion, it might be wise + * to adjust the color grid spacing so that the representative colors are + * equidistant in linear space. At this writing, gamma correction is not + * implemented by jdcolor, so nothing is done here. + */ + + +/* Declarations for ordered dithering. + * + * We use a standard 16x16 ordered dither array. The basic concept of ordered + * dithering is described in many references, for instance Dale Schumacher's + * chapter II.2 of Graphics Gems II (James Arvo, ed. Academic Press, 1991). + * In place of Schumacher's comparisons against a "threshold" value, we add a + * "dither" value to the input pixel and then round the result to the nearest + * output value. The dither value is equivalent to (0.5 - threshold) times + * the distance between output values. For ordered dithering, we assume that + * the output colors are equally spaced; if not, results will probably be + * worse, since the dither may be too much or too little at a given point. + * + * The normal calculation would be to form pixel value + dither, range-limit + * this to 0..MAXJSAMPLE, and then index into the colorindex table as usual. + * We can skip the separate range-limiting step by extending the colorindex + * table in both directions. + */ + +#define ODITHER_SIZE 16 /* dimension of dither matrix */ +/* NB: if ODITHER_SIZE is not a power of 2, ODITHER_MASK uses will break */ +#define ODITHER_CELLS (ODITHER_SIZE*ODITHER_SIZE) /* # cells in matrix */ +#define ODITHER_MASK (ODITHER_SIZE-1) /* mask for wrapping around counters */ + +typedef int ODITHER_MATRIX[ODITHER_SIZE][ODITHER_SIZE]; +typedef int (*ODITHER_MATRIX_PTR)[ODITHER_SIZE]; + +static const UINT8 base_dither_matrix[ODITHER_SIZE][ODITHER_SIZE] = { + /* Bayer's order-4 dither array. Generated by the code given in + * Stephen Hawley's article "Ordered Dithering" in Graphics Gems I. + * The values in this array must range from 0 to ODITHER_CELLS-1. + */ + { 0,192, 48,240, 12,204, 60,252, 3,195, 51,243, 15,207, 63,255 }, + { 128, 64,176,112,140, 76,188,124,131, 67,179,115,143, 79,191,127 }, + { 32,224, 16,208, 44,236, 28,220, 35,227, 19,211, 47,239, 31,223 }, + { 160, 96,144, 80,172,108,156, 92,163, 99,147, 83,175,111,159, 95 }, + { 8,200, 56,248, 4,196, 52,244, 11,203, 59,251, 7,199, 55,247 }, + { 136, 72,184,120,132, 68,180,116,139, 75,187,123,135, 71,183,119 }, + { 40,232, 24,216, 36,228, 20,212, 43,235, 27,219, 39,231, 23,215 }, + { 168,104,152, 88,164,100,148, 84,171,107,155, 91,167,103,151, 87 }, + { 2,194, 50,242, 14,206, 62,254, 1,193, 49,241, 13,205, 61,253 }, + { 130, 66,178,114,142, 78,190,126,129, 65,177,113,141, 77,189,125 }, + { 34,226, 18,210, 46,238, 30,222, 33,225, 17,209, 45,237, 29,221 }, + { 162, 98,146, 82,174,110,158, 94,161, 97,145, 81,173,109,157, 93 }, + { 10,202, 58,250, 6,198, 54,246, 9,201, 57,249, 5,197, 53,245 }, + { 138, 74,186,122,134, 70,182,118,137, 73,185,121,133, 69,181,117 }, + { 42,234, 26,218, 38,230, 22,214, 41,233, 25,217, 37,229, 21,213 }, + { 170,106,154, 90,166,102,150, 86,169,105,153, 89,165,101,149, 85 } +}; + + +/* Declarations for Floyd-Steinberg dithering. + * + * Errors are accumulated into the array fserrors[], at a resolution of + * 1/16th of a pixel count. The error at a given pixel is propagated + * to its not-yet-processed neighbors using the standard F-S fractions, + * ... (here) 7/16 + * 3/16 5/16 1/16 + * We work left-to-right on even rows, right-to-left on odd rows. + * + * We can get away with a single array (holding one row's worth of errors) + * by using it to store the current row's errors at pixel columns not yet + * processed, but the next row's errors at columns already processed. We + * need only a few extra variables to hold the errors immediately around the + * current column. (If we are lucky, those variables are in registers, but + * even if not, they're probably cheaper to access than array elements are.) + * + * The fserrors[] array is indexed [component#][position]. + * We provide (#columns + 2) entries per component; the extra entry at each + * end saves us from special-casing the first and last pixels. + * + * Note: on a wide image, we might not have enough room in a PC's near data + * segment to hold the error array; so it is allocated with alloc_large. + */ + +#if BITS_IN_JSAMPLE == 8 +typedef INT16 FSERROR; /* 16 bits should be enough */ +typedef int LOCFSERROR; /* use 'int' for calculation temps */ +#else +typedef INT32 FSERROR; /* may need more than 16 bits */ +typedef INT32 LOCFSERROR; /* be sure calculation temps are big enough */ +#endif + +typedef FSERROR FAR *FSERRPTR; /* pointer to error array (in FAR storage!) */ + + +/* Private subobject */ + +#define MAX_Q_COMPS 4 /* max components I can handle */ + +typedef struct { + struct jpeg_color_quantizer pub; /* public fields */ + + /* Initially allocated colormap is saved here */ + JSAMPARRAY sv_colormap; /* The color map as a 2-D pixel array */ + int sv_actual; /* number of entries in use */ + + JSAMPARRAY colorindex; /* Precomputed mapping for speed */ + /* colorindex[i][j] = index of color closest to pixel value j in component i, + * premultiplied as described above. Since colormap indexes must fit into + * JSAMPLEs, the entries of this array will too. + */ + boolean is_padded; /* is the colorindex padded for odither? */ + + int Ncolors[MAX_Q_COMPS]; /* # of values alloced to each component */ + + /* Variables for ordered dithering */ + int row_index; /* cur row's vertical index in dither matrix */ + ODITHER_MATRIX_PTR odither[MAX_Q_COMPS]; /* one dither array per component */ + + /* Variables for Floyd-Steinberg dithering */ + FSERRPTR fserrors[MAX_Q_COMPS]; /* accumulated errors */ + boolean on_odd_row; /* flag to remember which row we are on */ +} my_cquantizer; + +typedef my_cquantizer * my_cquantize_ptr; + + +/* + * Policy-making subroutines for create_colormap and create_colorindex. + * These routines determine the colormap to be used. The rest of the module + * only assumes that the colormap is orthogonal. + * + * * select_ncolors decides how to divvy up the available colors + * among the components. + * * output_value defines the set of representative values for a component. + * * largest_input_value defines the mapping from input values to + * representative values for a component. + * Note that the latter two routines may impose different policies for + * different components, though this is not currently done. + */ + + +LOCAL(int) +select_ncolors (j_decompress_ptr cinfo, int Ncolors[]) +/* Determine allocation of desired colors to components, */ +/* and fill in Ncolors[] array to indicate choice. */ +/* Return value is total number of colors (product of Ncolors[] values). */ +{ + int nc = cinfo->out_color_components; /* number of color components */ + int max_colors = cinfo->desired_number_of_colors; + int total_colors, iroot, i, j; + boolean changed; + long temp; + static const int RGB_order[3] = { RGB_GREEN, RGB_RED, RGB_BLUE }; + + /* We can allocate at least the nc'th root of max_colors per component. */ + /* Compute floor(nc'th root of max_colors). */ + iroot = 1; + do { + iroot++; + temp = iroot; /* set temp = iroot ** nc */ + for (i = 1; i < nc; i++) + temp *= iroot; + } while (temp <= (long) max_colors); /* repeat till iroot exceeds root */ + iroot--; /* now iroot = floor(root) */ + + /* Must have at least 2 color values per component */ + if (iroot < 2) + ERREXIT1(cinfo, JERR_QUANT_FEW_COLORS, (int) temp); + + /* Initialize to iroot color values for each component */ + total_colors = 1; + for (i = 0; i < nc; i++) { + Ncolors[i] = iroot; + total_colors *= iroot; + } + /* We may be able to increment the count for one or more components without + * exceeding max_colors, though we know not all can be incremented. + * Sometimes, the first component can be incremented more than once! + * (Example: for 16 colors, we start at 2*2*2, go to 3*2*2, then 4*2*2.) + * In RGB colorspace, try to increment G first, then R, then B. + */ + do { + changed = FALSE; + for (i = 0; i < nc; i++) { + j = (cinfo->out_color_space == JCS_RGB ? RGB_order[i] : i); + /* calculate new total_colors if Ncolors[j] is incremented */ + temp = total_colors / Ncolors[j]; + temp *= Ncolors[j]+1; /* done in long arith to avoid oflo */ + if (temp > (long) max_colors) + break; /* won't fit, done with this pass */ + Ncolors[j]++; /* OK, apply the increment */ + total_colors = (int) temp; + changed = TRUE; + } + } while (changed); + + return total_colors; +} + + +LOCAL(int) +output_value (j_decompress_ptr cinfo, int ci, int j, int maxj) +/* Return j'th output value, where j will range from 0 to maxj */ +/* The output values must fall in 0..MAXJSAMPLE in increasing order */ +{ + /* We always provide values 0 and MAXJSAMPLE for each component; + * any additional values are equally spaced between these limits. + * (Forcing the upper and lower values to the limits ensures that + * dithering can't produce a color outside the selected gamut.) + */ + return (int) (((INT32) j * MAXJSAMPLE + maxj/2) / maxj); +} + + +LOCAL(int) +largest_input_value (j_decompress_ptr cinfo, int ci, int j, int maxj) +/* Return largest input value that should map to j'th output value */ +/* Must have largest(j=0) >= 0, and largest(j=maxj) >= MAXJSAMPLE */ +{ + /* Breakpoints are halfway between values returned by output_value */ + return (int) (((INT32) (2*j + 1) * MAXJSAMPLE + maxj) / (2*maxj)); +} + + +/* + * Create the colormap. + */ + +LOCAL(void) +create_colormap (j_decompress_ptr cinfo) +{ + my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; + JSAMPARRAY colormap; /* Created colormap */ + int total_colors; /* Number of distinct output colors */ + int i,j,k, nci, blksize, blkdist, ptr, val; + + /* Select number of colors for each component */ + total_colors = select_ncolors(cinfo, cquantize->Ncolors); + + /* Report selected color counts */ + if (cinfo->out_color_components == 3) + TRACEMS4(cinfo, 1, JTRC_QUANT_3_NCOLORS, + total_colors, cquantize->Ncolors[0], + cquantize->Ncolors[1], cquantize->Ncolors[2]); + else + TRACEMS1(cinfo, 1, JTRC_QUANT_NCOLORS, total_colors); + + /* Allocate and fill in the colormap. */ + /* The colors are ordered in the map in standard row-major order, */ + /* i.e. rightmost (highest-indexed) color changes most rapidly. */ + + colormap = (*cinfo->mem->alloc_sarray) ((j_common_ptr) cinfo, JPOOL_IMAGE, + (JDIMENSION) total_colors, (JDIMENSION) cinfo->out_color_components); + + /* blksize is number of adjacent repeated entries for a component */ + /* blkdist is distance between groups of identical entries for a component */ + blkdist = total_colors; + + for (i = 0; i < cinfo->out_color_components; i++) { + /* fill in colormap entries for i'th color component */ + nci = cquantize->Ncolors[i]; /* # of distinct values for this color */ + blksize = blkdist / nci; + for (j = 0; j < nci; j++) { + /* Compute j'th output value (out of nci) for component */ + val = output_value(cinfo, i, j, nci-1); + /* Fill in all colormap entries that have this value of this component */ + for (ptr = j * blksize; ptr < total_colors; ptr += blkdist) { + /* fill in blksize entries beginning at ptr */ + for (k = 0; k < blksize; k++) + colormap[i][ptr+k] = (JSAMPLE) val; + } + } + blkdist = blksize; /* blksize of this color is blkdist of next */ + } + + /* Save the colormap in private storage, + * where it will survive color quantization mode changes. + */ + cquantize->sv_colormap = colormap; + cquantize->sv_actual = total_colors; +} + + +/* + * Create the color index table. + */ + +LOCAL(void) +create_colorindex (j_decompress_ptr cinfo) +{ + my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; + JSAMPROW indexptr; + int i,j,k, nci, blksize, val, pad; + + /* For ordered dither, we pad the color index tables by MAXJSAMPLE in + * each direction (input index values can be -MAXJSAMPLE .. 2*MAXJSAMPLE). + * This is not necessary in the other dithering modes. However, we + * flag whether it was done in case user changes dithering mode. + */ + if (cinfo->dither_mode == JDITHER_ORDERED) { + pad = MAXJSAMPLE*2; + cquantize->is_padded = TRUE; + } else { + pad = 0; + cquantize->is_padded = FALSE; + } + + cquantize->colorindex = (*cinfo->mem->alloc_sarray) + ((j_common_ptr) cinfo, JPOOL_IMAGE, + (JDIMENSION) (MAXJSAMPLE+1 + pad), + (JDIMENSION) cinfo->out_color_components); + + /* blksize is number of adjacent repeated entries for a component */ + blksize = cquantize->sv_actual; + + for (i = 0; i < cinfo->out_color_components; i++) { + /* fill in colorindex entries for i'th color component */ + nci = cquantize->Ncolors[i]; /* # of distinct values for this color */ + blksize = blksize / nci; + + /* adjust colorindex pointers to provide padding at negative indexes. */ + if (pad) + cquantize->colorindex[i] += MAXJSAMPLE; + + /* in loop, val = index of current output value, */ + /* and k = largest j that maps to current val */ + indexptr = cquantize->colorindex[i]; + val = 0; + k = largest_input_value(cinfo, i, 0, nci-1); + for (j = 0; j <= MAXJSAMPLE; j++) { + while (j > k) /* advance val if past boundary */ + k = largest_input_value(cinfo, i, ++val, nci-1); + /* premultiply so that no multiplication needed in main processing */ + indexptr[j] = (JSAMPLE) (val * blksize); + } + /* Pad at both ends if necessary */ + if (pad) + for (j = 1; j <= MAXJSAMPLE; j++) { + indexptr[-j] = indexptr[0]; + indexptr[MAXJSAMPLE+j] = indexptr[MAXJSAMPLE]; + } + } +} + + +/* + * Create an ordered-dither array for a component having ncolors + * distinct output values. + */ + +LOCAL(ODITHER_MATRIX_PTR) +make_odither_array (j_decompress_ptr cinfo, int ncolors) +{ + ODITHER_MATRIX_PTR odither; + int j,k; + INT32 num,den; + + odither = (ODITHER_MATRIX_PTR) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(ODITHER_MATRIX)); + /* The inter-value distance for this color is MAXJSAMPLE/(ncolors-1). + * Hence the dither value for the matrix cell with fill order f + * (f=0..N-1) should be (N-1-2*f)/(2*N) * MAXJSAMPLE/(ncolors-1). + * On 16-bit-int machine, be careful to avoid overflow. + */ + den = 2 * ODITHER_CELLS * ((INT32) (ncolors - 1)); + for (j = 0; j < ODITHER_SIZE; j++) { + for (k = 0; k < ODITHER_SIZE; k++) { + num = ((INT32) (ODITHER_CELLS-1 - 2*((int)base_dither_matrix[j][k]))) + * MAXJSAMPLE; + /* Ensure round towards zero despite C's lack of consistency + * about rounding negative values in integer division... + */ + odither[j][k] = (int) (num<0 ? -((-num)/den) : num/den); + } + } + return odither; +} + + +/* + * Create the ordered-dither tables. + * Components having the same number of representative colors may + * share a dither table. + */ + +LOCAL(void) +create_odither_tables (j_decompress_ptr cinfo) +{ + my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; + ODITHER_MATRIX_PTR odither; + int i, j, nci; + + for (i = 0; i < cinfo->out_color_components; i++) { + nci = cquantize->Ncolors[i]; /* # of distinct values for this color */ + odither = NULL; /* search for matching prior component */ + for (j = 0; j < i; j++) { + if (nci == cquantize->Ncolors[j]) { + odither = cquantize->odither[j]; + break; + } + } + if (odither == NULL) /* need a new table? */ + odither = make_odither_array(cinfo, nci); + cquantize->odither[i] = odither; + } +} + + +/* + * Map some rows of pixels to the output colormapped representation. + */ + +METHODDEF(void) +color_quantize (j_decompress_ptr cinfo, JSAMPARRAY input_buf, + JSAMPARRAY output_buf, int num_rows) +/* General case, no dithering */ +{ + my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; + JSAMPARRAY colorindex = cquantize->colorindex; + register int pixcode, ci; + register JSAMPROW ptrin, ptrout; + int row; + JDIMENSION col; + JDIMENSION width = cinfo->output_width; + register int nc = cinfo->out_color_components; + + for (row = 0; row < num_rows; row++) { + ptrin = input_buf[row]; + ptrout = output_buf[row]; + for (col = width; col > 0; col--) { + pixcode = 0; + for (ci = 0; ci < nc; ci++) { + pixcode += GETJSAMPLE(colorindex[ci][GETJSAMPLE(*ptrin++)]); + } + *ptrout++ = (JSAMPLE) pixcode; + } + } +} + + +METHODDEF(void) +color_quantize3 (j_decompress_ptr cinfo, JSAMPARRAY input_buf, + JSAMPARRAY output_buf, int num_rows) +/* Fast path for out_color_components==3, no dithering */ +{ + my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; + register int pixcode; + register JSAMPROW ptrin, ptrout; + JSAMPROW colorindex0 = cquantize->colorindex[0]; + JSAMPROW colorindex1 = cquantize->colorindex[1]; + JSAMPROW colorindex2 = cquantize->colorindex[2]; + int row; + JDIMENSION col; + JDIMENSION width = cinfo->output_width; + + for (row = 0; row < num_rows; row++) { + ptrin = input_buf[row]; + ptrout = output_buf[row]; + for (col = width; col > 0; col--) { + pixcode = GETJSAMPLE(colorindex0[GETJSAMPLE(*ptrin++)]); + pixcode += GETJSAMPLE(colorindex1[GETJSAMPLE(*ptrin++)]); + pixcode += GETJSAMPLE(colorindex2[GETJSAMPLE(*ptrin++)]); + *ptrout++ = (JSAMPLE) pixcode; + } + } +} + + +METHODDEF(void) +quantize_ord_dither (j_decompress_ptr cinfo, JSAMPARRAY input_buf, + JSAMPARRAY output_buf, int num_rows) +/* General case, with ordered dithering */ +{ + my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; + register JSAMPROW input_ptr; + register JSAMPROW output_ptr; + JSAMPROW colorindex_ci; + int * dither; /* points to active row of dither matrix */ + int row_index, col_index; /* current indexes into dither matrix */ + int nc = cinfo->out_color_components; + int ci; + int row; + JDIMENSION col; + JDIMENSION width = cinfo->output_width; + + for (row = 0; row < num_rows; row++) { + /* Initialize output values to 0 so can process components separately */ + FMEMZERO((void FAR *) output_buf[row], (size_t) width * SIZEOF(JSAMPLE)); + row_index = cquantize->row_index; + for (ci = 0; ci < nc; ci++) { + input_ptr = input_buf[row] + ci; + output_ptr = output_buf[row]; + colorindex_ci = cquantize->colorindex[ci]; + dither = cquantize->odither[ci][row_index]; + col_index = 0; + + for (col = width; col > 0; col--) { + /* Form pixel value + dither, range-limit to 0..MAXJSAMPLE, + * select output value, accumulate into output code for this pixel. + * Range-limiting need not be done explicitly, as we have extended + * the colorindex table to produce the right answers for out-of-range + * inputs. The maximum dither is +- MAXJSAMPLE; this sets the + * required amount of padding. + */ + *output_ptr += colorindex_ci[GETJSAMPLE(*input_ptr)+dither[col_index]]; + input_ptr += nc; + output_ptr++; + col_index = (col_index + 1) & ODITHER_MASK; + } + } + /* Advance row index for next row */ + row_index = (row_index + 1) & ODITHER_MASK; + cquantize->row_index = row_index; + } +} + + +METHODDEF(void) +quantize3_ord_dither (j_decompress_ptr cinfo, JSAMPARRAY input_buf, + JSAMPARRAY output_buf, int num_rows) +/* Fast path for out_color_components==3, with ordered dithering */ +{ + my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; + register int pixcode; + register JSAMPROW input_ptr; + register JSAMPROW output_ptr; + JSAMPROW colorindex0 = cquantize->colorindex[0]; + JSAMPROW colorindex1 = cquantize->colorindex[1]; + JSAMPROW colorindex2 = cquantize->colorindex[2]; + int * dither0; /* points to active row of dither matrix */ + int * dither1; + int * dither2; + int row_index, col_index; /* current indexes into dither matrix */ + int row; + JDIMENSION col; + JDIMENSION width = cinfo->output_width; + + for (row = 0; row < num_rows; row++) { + row_index = cquantize->row_index; + input_ptr = input_buf[row]; + output_ptr = output_buf[row]; + dither0 = cquantize->odither[0][row_index]; + dither1 = cquantize->odither[1][row_index]; + dither2 = cquantize->odither[2][row_index]; + col_index = 0; + + for (col = width; col > 0; col--) { + pixcode = GETJSAMPLE(colorindex0[GETJSAMPLE(*input_ptr++) + + dither0[col_index]]); + pixcode += GETJSAMPLE(colorindex1[GETJSAMPLE(*input_ptr++) + + dither1[col_index]]); + pixcode += GETJSAMPLE(colorindex2[GETJSAMPLE(*input_ptr++) + + dither2[col_index]]); + *output_ptr++ = (JSAMPLE) pixcode; + col_index = (col_index + 1) & ODITHER_MASK; + } + row_index = (row_index + 1) & ODITHER_MASK; + cquantize->row_index = row_index; + } +} + + +METHODDEF(void) +quantize_fs_dither (j_decompress_ptr cinfo, JSAMPARRAY input_buf, + JSAMPARRAY output_buf, int num_rows) +/* General case, with Floyd-Steinberg dithering */ +{ + my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; + register LOCFSERROR cur; /* current error or pixel value */ + LOCFSERROR belowerr; /* error for pixel below cur */ + LOCFSERROR bpreverr; /* error for below/prev col */ + LOCFSERROR bnexterr; /* error for below/next col */ + LOCFSERROR delta; + register FSERRPTR errorptr; /* => fserrors[] at column before current */ + register JSAMPROW input_ptr; + register JSAMPROW output_ptr; + JSAMPROW colorindex_ci; + JSAMPROW colormap_ci; + int pixcode; + int nc = cinfo->out_color_components; + int dir; /* 1 for left-to-right, -1 for right-to-left */ + int dirnc; /* dir * nc */ + int ci; + int row; + JDIMENSION col; + JDIMENSION width = cinfo->output_width; + JSAMPLE *range_limit = cinfo->sample_range_limit; + SHIFT_TEMPS + + for (row = 0; row < num_rows; row++) { + /* Initialize output values to 0 so can process components separately */ + FMEMZERO((void FAR *) output_buf[row], (size_t) width * SIZEOF(JSAMPLE)); + for (ci = 0; ci < nc; ci++) { + input_ptr = input_buf[row] + ci; + output_ptr = output_buf[row]; + if (cquantize->on_odd_row) { + /* work right to left in this row */ + input_ptr += (width-1) * nc; /* so point to rightmost pixel */ + output_ptr += width-1; + dir = -1; + dirnc = -nc; + errorptr = cquantize->fserrors[ci] + (width+1); /* => entry after last column */ + } else { + /* work left to right in this row */ + dir = 1; + dirnc = nc; + errorptr = cquantize->fserrors[ci]; /* => entry before first column */ + } + colorindex_ci = cquantize->colorindex[ci]; + colormap_ci = cquantize->sv_colormap[ci]; + /* Preset error values: no error propagated to first pixel from left */ + cur = 0; + /* and no error propagated to row below yet */ + belowerr = bpreverr = 0; + + for (col = width; col > 0; col--) { + /* cur holds the error propagated from the previous pixel on the + * current line. Add the error propagated from the previous line + * to form the complete error correction term for this pixel, and + * round the error term (which is expressed * 16) to an integer. + * RIGHT_SHIFT rounds towards minus infinity, so adding 8 is correct + * for either sign of the error value. + * Note: errorptr points to *previous* column's array entry. + */ + cur = RIGHT_SHIFT(cur + errorptr[dir] + 8, 4); + /* Form pixel value + error, and range-limit to 0..MAXJSAMPLE. + * The maximum error is +- MAXJSAMPLE; this sets the required size + * of the range_limit array. + */ + cur += GETJSAMPLE(*input_ptr); + cur = GETJSAMPLE(range_limit[cur]); + /* Select output value, accumulate into output code for this pixel */ + pixcode = GETJSAMPLE(colorindex_ci[cur]); + *output_ptr += (JSAMPLE) pixcode; + /* Compute actual representation error at this pixel */ + /* Note: we can do this even though we don't have the final */ + /* pixel code, because the colormap is orthogonal. */ + cur -= GETJSAMPLE(colormap_ci[pixcode]); + /* Compute error fractions to be propagated to adjacent pixels. + * Add these into the running sums, and simultaneously shift the + * next-line error sums left by 1 column. + */ + bnexterr = cur; + delta = cur * 2; + cur += delta; /* form error * 3 */ + errorptr[0] = (FSERROR) (bpreverr + cur); + cur += delta; /* form error * 5 */ + bpreverr = belowerr + cur; + belowerr = bnexterr; + cur += delta; /* form error * 7 */ + /* At this point cur contains the 7/16 error value to be propagated + * to the next pixel on the current line, and all the errors for the + * next line have been shifted over. We are therefore ready to move on. + */ + input_ptr += dirnc; /* advance input ptr to next column */ + output_ptr += dir; /* advance output ptr to next column */ + errorptr += dir; /* advance errorptr to current column */ + } + /* Post-loop cleanup: we must unload the final error value into the + * final fserrors[] entry. Note we need not unload belowerr because + * it is for the dummy column before or after the actual array. + */ + errorptr[0] = (FSERROR) bpreverr; /* unload prev err into array */ + } + cquantize->on_odd_row = (cquantize->on_odd_row ? FALSE : TRUE); + } +} + + +/* + * Allocate workspace for Floyd-Steinberg errors. + */ + +LOCAL(void) +alloc_fs_workspace (j_decompress_ptr cinfo) +{ + my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; + size_t arraysize; + int i; + + arraysize = ((size_t) cinfo->output_width + (size_t) 2) * SIZEOF(FSERROR); + for (i = 0; i < cinfo->out_color_components; i++) { + cquantize->fserrors[i] = (FSERRPTR) (*cinfo->mem->alloc_large) + ((j_common_ptr) cinfo, JPOOL_IMAGE, arraysize); + } +} + + +/* + * Initialize for one-pass color quantization. + */ + +METHODDEF(void) +start_pass_1_quant (j_decompress_ptr cinfo, boolean is_pre_scan) +{ + my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; + size_t arraysize; + int i; + + /* Install my colormap. */ + cinfo->colormap = cquantize->sv_colormap; + cinfo->actual_number_of_colors = cquantize->sv_actual; + + /* Initialize for desired dithering mode. */ + switch (cinfo->dither_mode) { + case JDITHER_NONE: + if (cinfo->out_color_components == 3) + cquantize->pub.color_quantize = color_quantize3; + else + cquantize->pub.color_quantize = color_quantize; + break; + case JDITHER_ORDERED: + if (cinfo->out_color_components == 3) + cquantize->pub.color_quantize = quantize3_ord_dither; + else + cquantize->pub.color_quantize = quantize_ord_dither; + cquantize->row_index = 0; /* initialize state for ordered dither */ + /* If user changed to ordered dither from another mode, + * we must recreate the color index table with padding. + * This will cost extra space, but probably isn't very likely. + */ + if (! cquantize->is_padded) + create_colorindex(cinfo); + /* Create ordered-dither tables if we didn't already. */ + if (cquantize->odither[0] == NULL) + create_odither_tables(cinfo); + break; + case JDITHER_FS: + cquantize->pub.color_quantize = quantize_fs_dither; + cquantize->on_odd_row = FALSE; /* initialize state for F-S dither */ + /* Allocate Floyd-Steinberg workspace if didn't already. */ + if (cquantize->fserrors[0] == NULL) + alloc_fs_workspace(cinfo); + /* Initialize the propagated errors to zero. */ + arraysize = ((size_t) cinfo->output_width + (size_t) 2) * SIZEOF(FSERROR); + for (i = 0; i < cinfo->out_color_components; i++) + FMEMZERO((void FAR *) cquantize->fserrors[i], arraysize); + break; + default: + ERREXIT(cinfo, JERR_NOT_COMPILED); + } +} + + +/* + * Finish up at the end of the pass. + */ + +METHODDEF(void) +finish_pass_1_quant (j_decompress_ptr cinfo) +{ + /* no work in 1-pass case */ +} + + +/* + * Switch to a new external colormap between output passes. + * Shouldn't get to this module! + */ + +METHODDEF(void) +new_color_map_1_quant (j_decompress_ptr cinfo) +{ + ERREXIT(cinfo, JERR_MODE_CHANGE); +} + + +/* + * Module initialization routine for 1-pass color quantization. + */ + +GLOBAL(void) +jinit_1pass_quantizer (j_decompress_ptr cinfo) +{ + my_cquantize_ptr cquantize; + + cquantize = (my_cquantize_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_cquantizer)); + cinfo->cquantize = &cquantize->pub; + cquantize->pub.start_pass = start_pass_1_quant; + cquantize->pub.finish_pass = finish_pass_1_quant; + cquantize->pub.new_color_map = new_color_map_1_quant; + cquantize->fserrors[0] = NULL; /* Flag FS workspace not allocated */ + cquantize->odither[0] = NULL; /* Also flag odither arrays not allocated */ + + /* Make sure my internal arrays won't overflow */ + if (cinfo->out_color_components > MAX_Q_COMPS) + ERREXIT1(cinfo, JERR_QUANT_COMPONENTS, MAX_Q_COMPS); + /* Make sure colormap indexes can be represented by JSAMPLEs */ + if (cinfo->desired_number_of_colors > (MAXJSAMPLE+1)) + ERREXIT1(cinfo, JERR_QUANT_MANY_COLORS, MAXJSAMPLE+1); + + /* Create the colormap and color index table. */ + create_colormap(cinfo); + create_colorindex(cinfo); + + /* Allocate Floyd-Steinberg workspace now if requested. + * We do this now since it is FAR storage and may affect the memory + * manager's space calculations. If the user changes to FS dither + * mode in a later pass, we will allocate the space then, and will + * possibly overrun the max_memory_to_use setting. + */ + if (cinfo->dither_mode == JDITHER_FS) + alloc_fs_workspace(cinfo); +} + +#endif /* QUANT_1PASS_SUPPORTED */ diff --git a/cmake/externals/libjpeg/jquant2.c b/cmake/externals/libjpeg/jquant2.c new file mode 100644 index 000000000..662b9bcef --- /dev/null +++ b/cmake/externals/libjpeg/jquant2.c @@ -0,0 +1,1311 @@ +/* + * jquant2.c + * + * Copyright (C) 1991-1996, Thomas G. Lane. + * Modified 2011-2020 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains 2-pass color quantization (color mapping) routines. + * These routines provide selection of a custom color map for an image, + * followed by mapping of the image to that color map, with optional + * Floyd-Steinberg dithering. + * It is also possible to use just the second pass to map to an arbitrary + * externally-given color map. + * + * Note: ordered dithering is not supported, since there isn't any fast + * way to compute intercolor distances; it's unclear that ordered dither's + * fundamental assumptions even hold with an irregularly spaced color map. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + +#ifdef QUANT_2PASS_SUPPORTED + + +/* + * This module implements the well-known Heckbert paradigm for color + * quantization. Most of the ideas used here can be traced back to + * Heckbert's seminal paper + * Heckbert, Paul. "Color Image Quantization for Frame Buffer Display", + * Proc. SIGGRAPH '82, Computer Graphics v.16 #3 (July 1982), pp 297-304. + * + * In the first pass over the image, we accumulate a histogram showing the + * usage count of each possible color. To keep the histogram to a reasonable + * size, we reduce the precision of the input; typical practice is to retain + * 5 or 6 bits per color, so that 8 or 4 different input values are counted + * in the same histogram cell. + * + * Next, the color-selection step begins with a box representing the whole + * color space, and repeatedly splits the "largest" remaining box until we + * have as many boxes as desired colors. Then the mean color in each + * remaining box becomes one of the possible output colors. + * + * The second pass over the image maps each input pixel to the closest output + * color (optionally after applying a Floyd-Steinberg dithering correction). + * This mapping is logically trivial, but making it go fast enough requires + * considerable care. + * + * Heckbert-style quantizers vary a good deal in their policies for choosing + * the "largest" box and deciding where to cut it. The particular policies + * used here have proved out well in experimental comparisons, but better ones + * may yet be found. + * + * In earlier versions of the IJG code, this module quantized in YCbCr color + * space, processing the raw upsampled data without a color conversion step. + * This allowed the color conversion math to be done only once per colormap + * entry, not once per pixel. However, that optimization precluded other + * useful optimizations (such as merging color conversion with upsampling) + * and it also interfered with desired capabilities such as quantizing to an + * externally-supplied colormap. We have therefore abandoned that approach. + * The present code works in the post-conversion color space, typically RGB. + * + * To improve the visual quality of the results, we actually work in scaled + * RGB space, giving G distances more weight than R, and R in turn more than + * B. To do everything in integer math, we must use integer scale factors. + * The 2/3/1 scale factors used here correspond loosely to the relative + * weights of the colors in the NTSC grayscale equation. + * If you want to use this code to quantize a non-RGB color space, you'll + * probably need to change these scale factors. + */ + +#define R_SCALE 2 /* scale R distances by this much */ +#define G_SCALE 3 /* scale G distances by this much */ +#define B_SCALE 1 /* and B by this much */ + +/* Relabel R/G/B as components 0/1/2, respecting the RGB ordering defined + * in jmorecfg.h. As the code stands, it will do the right thing for R,G,B + * and B,G,R orders. If you define some other weird order in jmorecfg.h, + * you'll get compile errors until you extend this logic. In that case + * you'll probably want to tweak the histogram sizes too. + */ + +#if RGB_RED == 0 +#define C0_SCALE R_SCALE +#endif +#if RGB_BLUE == 0 +#define C0_SCALE B_SCALE +#endif +#if RGB_GREEN == 1 +#define C1_SCALE G_SCALE +#endif +#if RGB_RED == 2 +#define C2_SCALE R_SCALE +#endif +#if RGB_BLUE == 2 +#define C2_SCALE B_SCALE +#endif + + +/* + * First we have the histogram data structure and routines for creating it. + * + * The number of bits of precision can be adjusted by changing these symbols. + * We recommend keeping 6 bits for G and 5 each for R and B. + * If you have plenty of memory and cycles, 6 bits all around gives marginally + * better results; if you are short of memory, 5 bits all around will save + * some space but degrade the results. + * To maintain a fully accurate histogram, we'd need to allocate a "long" + * (preferably unsigned long) for each cell. In practice this is overkill; + * we can get by with 16 bits per cell. Few of the cell counts will overflow, + * and clamping those that do overflow to the maximum value will give close- + * enough results. This reduces the recommended histogram size from 256Kb + * to 128Kb, which is a useful savings on PC-class machines. + * (In the second pass the histogram space is re-used for pixel mapping data; + * in that capacity, each cell must be able to store zero to the number of + * desired colors. 16 bits/cell is plenty for that too.) + * Since the JPEG code is intended to run in small memory model on 80x86 + * machines, we can't just allocate the histogram in one chunk. Instead + * of a true 3-D array, we use a row of pointers to 2-D arrays. Each + * pointer corresponds to a C0 value (typically 2^5 = 32 pointers) and + * each 2-D array has 2^6*2^5 = 2048 or 2^6*2^6 = 4096 entries. Note that + * on 80x86 machines, the pointer row is in near memory but the actual + * arrays are in far memory (same arrangement as we use for image arrays). + */ + +#define MAXNUMCOLORS (MAXJSAMPLE+1) /* maximum size of colormap */ + +/* These will do the right thing for either R,G,B or B,G,R color order, + * but you may not like the results for other color orders. + */ +#define HIST_C0_BITS 5 /* bits of precision in R/B histogram */ +#define HIST_C1_BITS 6 /* bits of precision in G histogram */ +#define HIST_C2_BITS 5 /* bits of precision in B/R histogram */ + +/* Number of elements along histogram axes. */ +#define HIST_C0_ELEMS (1<cquantize; + register JSAMPROW ptr; + register histptr histp; + register hist3d histogram = cquantize->histogram; + int row; + JDIMENSION col; + JDIMENSION width = cinfo->output_width; + + for (row = 0; row < num_rows; row++) { + ptr = input_buf[row]; + for (col = width; col > 0; col--) { + /* get pixel value and index into the histogram */ + histp = & histogram[GETJSAMPLE(ptr[0]) >> C0_SHIFT] + [GETJSAMPLE(ptr[1]) >> C1_SHIFT] + [GETJSAMPLE(ptr[2]) >> C2_SHIFT]; + /* increment, check for overflow and undo increment if so. */ + if (++(*histp) <= 0) + (*histp)--; + ptr += 3; + } + } +} + + +/* + * Next we have the really interesting routines: selection of a colormap + * given the completed histogram. + * These routines work with a list of "boxes", each representing a rectangular + * subset of the input color space (to histogram precision). + */ + +typedef struct { + /* The bounds of the box (inclusive); expressed as histogram indexes */ + int c0min, c0max; + int c1min, c1max; + int c2min, c2max; + /* The volume (actually 2-norm) of the box */ + INT32 volume; + /* The number of nonzero histogram cells within this box */ + long colorcount; +} box; + +typedef box * boxptr; + + +LOCAL(boxptr) +find_biggest_color_pop (boxptr boxlist, int numboxes) +/* Find the splittable box with the largest color population */ +/* Returns NULL if no splittable boxes remain */ +{ + register boxptr boxp; + register int i; + register long maxc = 0; + boxptr which = NULL; + + for (i = 0, boxp = boxlist; i < numboxes; i++, boxp++) { + if (boxp->colorcount > maxc && boxp->volume > 0) { + which = boxp; + maxc = boxp->colorcount; + } + } + return which; +} + + +LOCAL(boxptr) +find_biggest_volume (boxptr boxlist, int numboxes) +/* Find the splittable box with the largest (scaled) volume */ +/* Returns NULL if no splittable boxes remain */ +{ + register boxptr boxp; + register int i; + register INT32 maxv = 0; + boxptr which = NULL; + + for (i = 0, boxp = boxlist; i < numboxes; i++, boxp++) { + if (boxp->volume > maxv) { + which = boxp; + maxv = boxp->volume; + } + } + return which; +} + + +LOCAL(void) +update_box (j_decompress_ptr cinfo, boxptr boxp) +/* Shrink the min/max bounds of a box to enclose only nonzero elements, */ +/* and recompute its volume and population */ +{ + my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; + hist3d histogram = cquantize->histogram; + histptr histp; + int c0,c1,c2; + int c0min,c0max,c1min,c1max,c2min,c2max; + INT32 dist0,dist1,dist2; + long ccount; + + c0min = boxp->c0min; c0max = boxp->c0max; + c1min = boxp->c1min; c1max = boxp->c1max; + c2min = boxp->c2min; c2max = boxp->c2max; + + if (c0max > c0min) + for (c0 = c0min; c0 <= c0max; c0++) + for (c1 = c1min; c1 <= c1max; c1++) { + histp = & histogram[c0][c1][c2min]; + for (c2 = c2min; c2 <= c2max; c2++) + if (*histp++ != 0) { + boxp->c0min = c0min = c0; + goto have_c0min; + } + } + have_c0min: + if (c0max > c0min) + for (c0 = c0max; c0 >= c0min; c0--) + for (c1 = c1min; c1 <= c1max; c1++) { + histp = & histogram[c0][c1][c2min]; + for (c2 = c2min; c2 <= c2max; c2++) + if (*histp++ != 0) { + boxp->c0max = c0max = c0; + goto have_c0max; + } + } + have_c0max: + if (c1max > c1min) + for (c1 = c1min; c1 <= c1max; c1++) + for (c0 = c0min; c0 <= c0max; c0++) { + histp = & histogram[c0][c1][c2min]; + for (c2 = c2min; c2 <= c2max; c2++) + if (*histp++ != 0) { + boxp->c1min = c1min = c1; + goto have_c1min; + } + } + have_c1min: + if (c1max > c1min) + for (c1 = c1max; c1 >= c1min; c1--) + for (c0 = c0min; c0 <= c0max; c0++) { + histp = & histogram[c0][c1][c2min]; + for (c2 = c2min; c2 <= c2max; c2++) + if (*histp++ != 0) { + boxp->c1max = c1max = c1; + goto have_c1max; + } + } + have_c1max: + if (c2max > c2min) + for (c2 = c2min; c2 <= c2max; c2++) + for (c0 = c0min; c0 <= c0max; c0++) { + histp = & histogram[c0][c1min][c2]; + for (c1 = c1min; c1 <= c1max; c1++, histp += HIST_C2_ELEMS) + if (*histp != 0) { + boxp->c2min = c2min = c2; + goto have_c2min; + } + } + have_c2min: + if (c2max > c2min) + for (c2 = c2max; c2 >= c2min; c2--) + for (c0 = c0min; c0 <= c0max; c0++) { + histp = & histogram[c0][c1min][c2]; + for (c1 = c1min; c1 <= c1max; c1++, histp += HIST_C2_ELEMS) + if (*histp != 0) { + boxp->c2max = c2max = c2; + goto have_c2max; + } + } + have_c2max: + + /* Update box volume. + * We use 2-norm rather than real volume here; this biases the method + * against making long narrow boxes, and it has the side benefit that + * a box is splittable iff norm > 0. + * Since the differences are expressed in histogram-cell units, + * we have to shift back to JSAMPLE units to get consistent distances; + * after which, we scale according to the selected distance scale factors. + */ + dist0 = ((c0max - c0min) << C0_SHIFT) * C0_SCALE; + dist1 = ((c1max - c1min) << C1_SHIFT) * C1_SCALE; + dist2 = ((c2max - c2min) << C2_SHIFT) * C2_SCALE; + boxp->volume = dist0*dist0 + dist1*dist1 + dist2*dist2; + + /* Now scan remaining volume of box and compute population */ + ccount = 0; + for (c0 = c0min; c0 <= c0max; c0++) + for (c1 = c1min; c1 <= c1max; c1++) { + histp = & histogram[c0][c1][c2min]; + for (c2 = c2min; c2 <= c2max; c2++, histp++) + if (*histp != 0) { + ccount++; + } + } + boxp->colorcount = ccount; +} + + +LOCAL(int) +median_cut (j_decompress_ptr cinfo, boxptr boxlist, int numboxes, + int desired_colors) +/* Repeatedly select and split the largest box until we have enough boxes */ +{ + int n,lb; + int c0,c1,c2,cmax; + register boxptr b1,b2; + + while (numboxes < desired_colors) { + /* Select box to split. + * Current algorithm: by population for first half, then by volume. + */ + if (numboxes*2 <= desired_colors) { + b1 = find_biggest_color_pop(boxlist, numboxes); + } else { + b1 = find_biggest_volume(boxlist, numboxes); + } + if (b1 == NULL) /* no splittable boxes left! */ + break; + b2 = &boxlist[numboxes]; /* where new box will go */ + /* Copy the color bounds to the new box. */ + b2->c0max = b1->c0max; b2->c1max = b1->c1max; b2->c2max = b1->c2max; + b2->c0min = b1->c0min; b2->c1min = b1->c1min; b2->c2min = b1->c2min; + /* Choose which axis to split the box on. + * Current algorithm: longest scaled axis. + * See notes in update_box about scaling distances. + */ + c0 = ((b1->c0max - b1->c0min) << C0_SHIFT) * C0_SCALE; + c1 = ((b1->c1max - b1->c1min) << C1_SHIFT) * C1_SCALE; + c2 = ((b1->c2max - b1->c2min) << C2_SHIFT) * C2_SCALE; + /* We want to break any ties in favor of green, then red, blue last. + * This code does the right thing for R,G,B or B,G,R color orders only. + */ +#if RGB_RED == 0 + cmax = c1; n = 1; + if (c0 > cmax) { cmax = c0; n = 0; } + if (c2 > cmax) { n = 2; } +#else + cmax = c1; n = 1; + if (c2 > cmax) { cmax = c2; n = 2; } + if (c0 > cmax) { n = 0; } +#endif + /* Choose split point along selected axis, and update box bounds. + * Current algorithm: split at halfway point. + * (Since the box has been shrunk to minimum volume, + * any split will produce two nonempty subboxes.) + * Note that lb value is max for lower box, so must be < old max. + */ + switch (n) { + case 0: + lb = (b1->c0max + b1->c0min) / 2; + b1->c0max = lb; + b2->c0min = lb+1; + break; + case 1: + lb = (b1->c1max + b1->c1min) / 2; + b1->c1max = lb; + b2->c1min = lb+1; + break; + case 2: + lb = (b1->c2max + b1->c2min) / 2; + b1->c2max = lb; + b2->c2min = lb+1; + break; + } + /* Update stats for boxes */ + update_box(cinfo, b1); + update_box(cinfo, b2); + numboxes++; + } + return numboxes; +} + + +LOCAL(void) +compute_color (j_decompress_ptr cinfo, boxptr boxp, int icolor) +/* Compute representative color for a box, put it in colormap[icolor] */ +{ + /* Current algorithm: mean weighted by pixels (not colors) */ + /* Note it is important to get the rounding correct! */ + my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; + hist3d histogram = cquantize->histogram; + histptr histp; + int c0,c1,c2; + int c0min,c0max,c1min,c1max,c2min,c2max; + long count; + long total = 0; + long c0total = 0; + long c1total = 0; + long c2total = 0; + + c0min = boxp->c0min; c0max = boxp->c0max; + c1min = boxp->c1min; c1max = boxp->c1max; + c2min = boxp->c2min; c2max = boxp->c2max; + + for (c0 = c0min; c0 <= c0max; c0++) + for (c1 = c1min; c1 <= c1max; c1++) { + histp = & histogram[c0][c1][c2min]; + for (c2 = c2min; c2 <= c2max; c2++) { + if ((count = *histp++) != 0) { + total += count; + c0total += ((c0 << C0_SHIFT) + ((1<>1)) * count; + c1total += ((c1 << C1_SHIFT) + ((1<>1)) * count; + c2total += ((c2 << C2_SHIFT) + ((1<>1)) * count; + } + } + } + + cinfo->colormap[0][icolor] = (JSAMPLE) ((c0total + (total>>1)) / total); + cinfo->colormap[1][icolor] = (JSAMPLE) ((c1total + (total>>1)) / total); + cinfo->colormap[2][icolor] = (JSAMPLE) ((c2total + (total>>1)) / total); +} + + +LOCAL(void) +select_colors (j_decompress_ptr cinfo, int desired_colors) +/* Master routine for color selection */ +{ + boxptr boxlist; + int numboxes; + int i; + + /* Allocate workspace for box list */ + boxlist = (boxptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, desired_colors * SIZEOF(box)); + /* Initialize one box containing whole space */ + numboxes = 1; + boxlist[0].c0min = 0; + boxlist[0].c0max = MAXJSAMPLE >> C0_SHIFT; + boxlist[0].c1min = 0; + boxlist[0].c1max = MAXJSAMPLE >> C1_SHIFT; + boxlist[0].c2min = 0; + boxlist[0].c2max = MAXJSAMPLE >> C2_SHIFT; + /* Shrink it to actually-used volume and set its statistics */ + update_box(cinfo, & boxlist[0]); + /* Perform median-cut to produce final box list */ + numboxes = median_cut(cinfo, boxlist, numboxes, desired_colors); + /* Compute the representative color for each box, fill colormap */ + for (i = 0; i < numboxes; i++) + compute_color(cinfo, & boxlist[i], i); + cinfo->actual_number_of_colors = numboxes; + TRACEMS1(cinfo, 1, JTRC_QUANT_SELECTED, numboxes); +} + + +/* + * These routines are concerned with the time-critical task of mapping input + * colors to the nearest color in the selected colormap. + * + * We re-use the histogram space as an "inverse color map", essentially a + * cache for the results of nearest-color searches. All colors within a + * histogram cell will be mapped to the same colormap entry, namely the one + * closest to the cell's center. This may not be quite the closest entry to + * the actual input color, but it's almost as good. A zero in the cache + * indicates we haven't found the nearest color for that cell yet; the array + * is cleared to zeroes before starting the mapping pass. When we find the + * nearest color for a cell, its colormap index plus one is recorded in the + * cache for future use. The pass2 scanning routines call fill_inverse_cmap + * when they need to use an unfilled entry in the cache. + * + * Our method of efficiently finding nearest colors is based on the "locally + * sorted search" idea described by Heckbert and on the incremental distance + * calculation described by Spencer W. Thomas in chapter III.1 of Graphics + * Gems II (James Arvo, ed. Academic Press, 1991). Thomas points out that + * the distances from a given colormap entry to each cell of the histogram can + * be computed quickly using an incremental method: the differences between + * distances to adjacent cells themselves differ by a constant. This allows a + * fairly fast implementation of the "brute force" approach of computing the + * distance from every colormap entry to every histogram cell. Unfortunately, + * it needs a work array to hold the best-distance-so-far for each histogram + * cell (because the inner loop has to be over cells, not colormap entries). + * The work array elements have to be INT32s, so the work array would need + * 256Kb at our recommended precision. This is not feasible in DOS machines. + * + * To get around these problems, we apply Thomas' method to compute the + * nearest colors for only the cells within a small subbox of the histogram. + * The work array need be only as big as the subbox, so the memory usage + * problem is solved. Furthermore, we need not fill subboxes that are never + * referenced in pass2; many images use only part of the color gamut, so a + * fair amount of work is saved. An additional advantage of this + * approach is that we can apply Heckbert's locality criterion to quickly + * eliminate colormap entries that are far away from the subbox; typically + * three-fourths of the colormap entries are rejected by Heckbert's criterion, + * and we need not compute their distances to individual cells in the subbox. + * The speed of this approach is heavily influenced by the subbox size: too + * small means too much overhead, too big loses because Heckbert's criterion + * can't eliminate as many colormap entries. Empirically the best subbox + * size seems to be about 1/512th of the histogram (1/8th in each direction). + * + * Thomas' article also describes a refined method which is asymptotically + * faster than the brute-force method, but it is also far more complex and + * cannot efficiently be applied to small subboxes. It is therefore not + * useful for programs intended to be portable to DOS machines. On machines + * with plenty of memory, filling the whole histogram in one shot with Thomas' + * refined method might be faster than the present code --- but then again, + * it might not be any faster, and it's certainly more complicated. + */ + + +/* log2(histogram cells in update box) for each axis; this can be adjusted */ +#define BOX_C0_LOG (HIST_C0_BITS-3) +#define BOX_C1_LOG (HIST_C1_BITS-3) +#define BOX_C2_LOG (HIST_C2_BITS-3) + +#define BOX_C0_ELEMS (1<actual_number_of_colors; + int maxc0, maxc1, maxc2; + int centerc0, centerc1, centerc2; + int i, x, ncolors; + INT32 minmaxdist, min_dist, max_dist, tdist; + INT32 mindist[MAXNUMCOLORS]; /* min distance to colormap entry i */ + + /* Compute true coordinates of update box's upper corner and center. + * Actually we compute the coordinates of the center of the upper-corner + * histogram cell, which are the upper bounds of the volume we care about. + * Note that since ">>" rounds down, the "center" values may be closer to + * min than to max; hence comparisons to them must be "<=", not "<". + */ + maxc0 = minc0 + ((1 << BOX_C0_SHIFT) - (1 << C0_SHIFT)); + centerc0 = (minc0 + maxc0) >> 1; + maxc1 = minc1 + ((1 << BOX_C1_SHIFT) - (1 << C1_SHIFT)); + centerc1 = (minc1 + maxc1) >> 1; + maxc2 = minc2 + ((1 << BOX_C2_SHIFT) - (1 << C2_SHIFT)); + centerc2 = (minc2 + maxc2) >> 1; + + /* For each color in colormap, find: + * 1. its minimum squared-distance to any point in the update box + * (zero if color is within update box); + * 2. its maximum squared-distance to any point in the update box. + * Both of these can be found by considering only the corners of the box. + * We save the minimum distance for each color in mindist[]; + * only the smallest maximum distance is of interest. + */ + minmaxdist = 0x7FFFFFFFL; + + for (i = 0; i < numcolors; i++) { + /* We compute the squared-c0-distance term, then add in the other two. */ + x = GETJSAMPLE(cinfo->colormap[0][i]); + if (x < minc0) { + tdist = (x - minc0) * C0_SCALE; + min_dist = tdist*tdist; + tdist = (x - maxc0) * C0_SCALE; + max_dist = tdist*tdist; + } else if (x > maxc0) { + tdist = (x - maxc0) * C0_SCALE; + min_dist = tdist*tdist; + tdist = (x - minc0) * C0_SCALE; + max_dist = tdist*tdist; + } else { + /* within cell range so no contribution to min_dist */ + min_dist = 0; + if (x <= centerc0) { + tdist = (x - maxc0) * C0_SCALE; + max_dist = tdist*tdist; + } else { + tdist = (x - minc0) * C0_SCALE; + max_dist = tdist*tdist; + } + } + + x = GETJSAMPLE(cinfo->colormap[1][i]); + if (x < minc1) { + tdist = (x - minc1) * C1_SCALE; + min_dist += tdist*tdist; + tdist = (x - maxc1) * C1_SCALE; + max_dist += tdist*tdist; + } else if (x > maxc1) { + tdist = (x - maxc1) * C1_SCALE; + min_dist += tdist*tdist; + tdist = (x - minc1) * C1_SCALE; + max_dist += tdist*tdist; + } else { + /* within cell range so no contribution to min_dist */ + if (x <= centerc1) { + tdist = (x - maxc1) * C1_SCALE; + max_dist += tdist*tdist; + } else { + tdist = (x - minc1) * C1_SCALE; + max_dist += tdist*tdist; + } + } + + x = GETJSAMPLE(cinfo->colormap[2][i]); + if (x < minc2) { + tdist = (x - minc2) * C2_SCALE; + min_dist += tdist*tdist; + tdist = (x - maxc2) * C2_SCALE; + max_dist += tdist*tdist; + } else if (x > maxc2) { + tdist = (x - maxc2) * C2_SCALE; + min_dist += tdist*tdist; + tdist = (x - minc2) * C2_SCALE; + max_dist += tdist*tdist; + } else { + /* within cell range so no contribution to min_dist */ + if (x <= centerc2) { + tdist = (x - maxc2) * C2_SCALE; + max_dist += tdist*tdist; + } else { + tdist = (x - minc2) * C2_SCALE; + max_dist += tdist*tdist; + } + } + + mindist[i] = min_dist; /* save away the results */ + if (max_dist < minmaxdist) + minmaxdist = max_dist; + } + + /* Now we know that no cell in the update box is more than minmaxdist + * away from some colormap entry. Therefore, only colors that are + * within minmaxdist of some part of the box need be considered. + */ + ncolors = 0; + for (i = 0; i < numcolors; i++) { + if (mindist[i] <= minmaxdist) + colorlist[ncolors++] = (JSAMPLE) i; + } + return ncolors; +} + + +LOCAL(void) +find_best_colors (j_decompress_ptr cinfo, int minc0, int minc1, int minc2, + int numcolors, JSAMPLE colorlist[], JSAMPLE bestcolor[]) +/* Find the closest colormap entry for each cell in the update box, + * given the list of candidate colors prepared by find_nearby_colors. + * Return the indexes of the closest entries in the bestcolor[] array. + * This routine uses Thomas' incremental distance calculation method to + * find the distance from a colormap entry to successive cells in the box. + */ +{ + int ic0, ic1, ic2; + int i, icolor; + register INT32 * bptr; /* pointer into bestdist[] array */ + JSAMPLE * cptr; /* pointer into bestcolor[] array */ + INT32 dist0, dist1; /* initial distance values */ + register INT32 dist2; /* current distance in inner loop */ + INT32 xx0, xx1; /* distance increments */ + register INT32 xx2; + INT32 inc0, inc1, inc2; /* initial values for increments */ + /* This array holds the distance to the nearest-so-far color for each cell */ + INT32 bestdist[BOX_C0_ELEMS * BOX_C1_ELEMS * BOX_C2_ELEMS]; + + /* Initialize best-distance for each cell of the update box */ + bptr = bestdist; + for (i = BOX_C0_ELEMS*BOX_C1_ELEMS*BOX_C2_ELEMS-1; i >= 0; i--) + *bptr++ = 0x7FFFFFFFL; + + /* For each color selected by find_nearby_colors, + * compute its distance to the center of each cell in the box. + * If that's less than best-so-far, update best distance and color number. + */ + + /* Nominal steps between cell centers ("x" in Thomas article) */ +#define STEP_C0 ((1 << C0_SHIFT) * C0_SCALE) +#define STEP_C1 ((1 << C1_SHIFT) * C1_SCALE) +#define STEP_C2 ((1 << C2_SHIFT) * C2_SCALE) + + for (i = 0; i < numcolors; i++) { + icolor = GETJSAMPLE(colorlist[i]); + /* Compute (square of) distance from minc0/c1/c2 to this color */ + inc0 = (minc0 - GETJSAMPLE(cinfo->colormap[0][icolor])) * C0_SCALE; + dist0 = inc0*inc0; + inc1 = (minc1 - GETJSAMPLE(cinfo->colormap[1][icolor])) * C1_SCALE; + dist0 += inc1*inc1; + inc2 = (minc2 - GETJSAMPLE(cinfo->colormap[2][icolor])) * C2_SCALE; + dist0 += inc2*inc2; + /* Form the initial difference increments */ + inc0 = inc0 * (2 * STEP_C0) + STEP_C0 * STEP_C0; + inc1 = inc1 * (2 * STEP_C1) + STEP_C1 * STEP_C1; + inc2 = inc2 * (2 * STEP_C2) + STEP_C2 * STEP_C2; + /* Now loop over all cells in box, updating distance per Thomas method */ + bptr = bestdist; + cptr = bestcolor; + xx0 = inc0; + for (ic0 = BOX_C0_ELEMS-1; ic0 >= 0; ic0--) { + dist1 = dist0; + xx1 = inc1; + for (ic1 = BOX_C1_ELEMS-1; ic1 >= 0; ic1--) { + dist2 = dist1; + xx2 = inc2; + for (ic2 = BOX_C2_ELEMS-1; ic2 >= 0; ic2--) { + if (dist2 < *bptr) { + *bptr = dist2; + *cptr = (JSAMPLE) icolor; + } + dist2 += xx2; + xx2 += 2 * STEP_C2 * STEP_C2; + bptr++; + cptr++; + } + dist1 += xx1; + xx1 += 2 * STEP_C1 * STEP_C1; + } + dist0 += xx0; + xx0 += 2 * STEP_C0 * STEP_C0; + } + } +} + + +LOCAL(void) +fill_inverse_cmap (j_decompress_ptr cinfo, int c0, int c1, int c2) +/* Fill the inverse-colormap entries in the update box that contains */ +/* histogram cell c0/c1/c2. (Only that one cell MUST be filled, but */ +/* we can fill as many others as we wish.) */ +{ + my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; + hist3d histogram = cquantize->histogram; + int minc0, minc1, minc2; /* lower left corner of update box */ + int ic0, ic1, ic2; + register JSAMPLE * cptr; /* pointer into bestcolor[] array */ + register histptr cachep; /* pointer into main cache array */ + /* This array lists the candidate colormap indexes. */ + JSAMPLE colorlist[MAXNUMCOLORS]; + int numcolors; /* number of candidate colors */ + /* This array holds the actually closest colormap index for each cell. */ + JSAMPLE bestcolor[BOX_C0_ELEMS * BOX_C1_ELEMS * BOX_C2_ELEMS]; + + /* Convert cell coordinates to update box ID */ + c0 >>= BOX_C0_LOG; + c1 >>= BOX_C1_LOG; + c2 >>= BOX_C2_LOG; + + /* Compute true coordinates of update box's origin corner. + * Actually we compute the coordinates of the center of the corner + * histogram cell, which are the lower bounds of the volume we care about. + */ + minc0 = (c0 << BOX_C0_SHIFT) + ((1 << C0_SHIFT) >> 1); + minc1 = (c1 << BOX_C1_SHIFT) + ((1 << C1_SHIFT) >> 1); + minc2 = (c2 << BOX_C2_SHIFT) + ((1 << C2_SHIFT) >> 1); + + /* Determine which colormap entries are close enough to be candidates + * for the nearest entry to some cell in the update box. + */ + numcolors = find_nearby_colors(cinfo, minc0, minc1, minc2, colorlist); + + /* Determine the actually nearest colors. */ + find_best_colors(cinfo, minc0, minc1, minc2, numcolors, colorlist, + bestcolor); + + /* Save the best color numbers (plus 1) in the main cache array */ + c0 <<= BOX_C0_LOG; /* convert ID back to base cell indexes */ + c1 <<= BOX_C1_LOG; + c2 <<= BOX_C2_LOG; + cptr = bestcolor; + for (ic0 = 0; ic0 < BOX_C0_ELEMS; ic0++) { + for (ic1 = 0; ic1 < BOX_C1_ELEMS; ic1++) { + cachep = & histogram[c0+ic0][c1+ic1][c2]; + for (ic2 = 0; ic2 < BOX_C2_ELEMS; ic2++) { + *cachep++ = (histcell) (GETJSAMPLE(*cptr++) + 1); + } + } + } +} + + +/* + * Map some rows of pixels to the output colormapped representation. + */ + +METHODDEF(void) +pass2_no_dither (j_decompress_ptr cinfo, + JSAMPARRAY input_buf, JSAMPARRAY output_buf, int num_rows) +/* This version performs no dithering */ +{ + my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; + hist3d histogram = cquantize->histogram; + register JSAMPROW inptr, outptr; + register histptr cachep; + register int c0, c1, c2; + int row; + JDIMENSION col; + JDIMENSION width = cinfo->output_width; + + for (row = 0; row < num_rows; row++) { + inptr = input_buf[row]; + outptr = output_buf[row]; + for (col = width; col > 0; col--) { + /* get pixel value and index into the cache */ + c0 = GETJSAMPLE(*inptr++) >> C0_SHIFT; + c1 = GETJSAMPLE(*inptr++) >> C1_SHIFT; + c2 = GETJSAMPLE(*inptr++) >> C2_SHIFT; + cachep = & histogram[c0][c1][c2]; + /* If we have not seen this color before, find nearest colormap entry */ + /* and update the cache */ + if (*cachep == 0) + fill_inverse_cmap(cinfo, c0,c1,c2); + /* Now emit the colormap index for this cell */ + *outptr++ = (JSAMPLE) (*cachep - 1); + } + } +} + + +METHODDEF(void) +pass2_fs_dither (j_decompress_ptr cinfo, + JSAMPARRAY input_buf, JSAMPARRAY output_buf, int num_rows) +/* This version performs Floyd-Steinberg dithering */ +{ + my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; + hist3d histogram = cquantize->histogram; + register LOCFSERROR cur0, cur1, cur2; /* current error or pixel value */ + LOCFSERROR belowerr0, belowerr1, belowerr2; /* error for pixel below cur */ + LOCFSERROR bpreverr0, bpreverr1, bpreverr2; /* error for below/prev col */ + register FSERRPTR errorptr; /* => fserrors[] at column before current */ + JSAMPROW inptr; /* => current input pixel */ + JSAMPROW outptr; /* => current output pixel */ + histptr cachep; + int dir; /* +1 or -1 depending on direction */ + int dir3; /* 3*dir, for advancing inptr & errorptr */ + int row; + JDIMENSION col; + JDIMENSION width = cinfo->output_width; + JSAMPLE *range_limit = cinfo->sample_range_limit; + int *error_limit = cquantize->error_limiter; + JSAMPROW colormap0 = cinfo->colormap[0]; + JSAMPROW colormap1 = cinfo->colormap[1]; + JSAMPROW colormap2 = cinfo->colormap[2]; + SHIFT_TEMPS + + for (row = 0; row < num_rows; row++) { + inptr = input_buf[row]; + outptr = output_buf[row]; + if (cquantize->on_odd_row) { + /* work right to left in this row */ + inptr += (width-1) * 3; /* so point to rightmost pixel */ + outptr += width-1; + dir = -1; + dir3 = -3; + errorptr = cquantize->fserrors + (width+1)*3; /* => entry after last column */ + cquantize->on_odd_row = FALSE; /* flip for next time */ + } else { + /* work left to right in this row */ + dir = 1; + dir3 = 3; + errorptr = cquantize->fserrors; /* => entry before first real column */ + cquantize->on_odd_row = TRUE; /* flip for next time */ + } + /* Preset error values: no error propagated to first pixel from left */ + cur0 = cur1 = cur2 = 0; + /* and no error propagated to row below yet */ + belowerr0 = belowerr1 = belowerr2 = 0; + bpreverr0 = bpreverr1 = bpreverr2 = 0; + + for (col = width; col > 0; col--) { + /* curN holds the error propagated from the previous pixel on the + * current line. Add the error propagated from the previous line + * to form the complete error correction term for this pixel, and + * round the error term (which is expressed * 16) to an integer. + * RIGHT_SHIFT rounds towards minus infinity, so adding 8 is correct + * for either sign of the error value. + * Note: errorptr points to *previous* column's array entry. + */ + cur0 = RIGHT_SHIFT(cur0 + errorptr[dir3+0] + 8, 4); + cur1 = RIGHT_SHIFT(cur1 + errorptr[dir3+1] + 8, 4); + cur2 = RIGHT_SHIFT(cur2 + errorptr[dir3+2] + 8, 4); + /* Limit the error using transfer function set by init_error_limit. + * See comments with init_error_limit for rationale. + */ + cur0 = error_limit[cur0]; + cur1 = error_limit[cur1]; + cur2 = error_limit[cur2]; + /* Form pixel value + error, and range-limit to 0..MAXJSAMPLE. + * The maximum error is +- MAXJSAMPLE (or less with error limiting); + * this sets the required size of the range_limit array. + */ + cur0 += GETJSAMPLE(inptr[0]); + cur1 += GETJSAMPLE(inptr[1]); + cur2 += GETJSAMPLE(inptr[2]); + cur0 = GETJSAMPLE(range_limit[cur0]); + cur1 = GETJSAMPLE(range_limit[cur1]); + cur2 = GETJSAMPLE(range_limit[cur2]); + /* Index into the cache with adjusted pixel value */ + cachep = & histogram[cur0>>C0_SHIFT][cur1>>C1_SHIFT][cur2>>C2_SHIFT]; + /* If we have not seen this color before, find nearest colormap */ + /* entry and update the cache */ + if (*cachep == 0) + fill_inverse_cmap(cinfo, cur0>>C0_SHIFT,cur1>>C1_SHIFT,cur2>>C2_SHIFT); + /* Now emit the colormap index for this cell */ + { register int pixcode = *cachep - 1; + *outptr = (JSAMPLE) pixcode; + /* Compute representation error for this pixel */ + cur0 -= GETJSAMPLE(colormap0[pixcode]); + cur1 -= GETJSAMPLE(colormap1[pixcode]); + cur2 -= GETJSAMPLE(colormap2[pixcode]); + } + /* Compute error fractions to be propagated to adjacent pixels. + * Add these into the running sums, and simultaneously shift the + * next-line error sums left by 1 column. + */ + { register LOCFSERROR bnexterr, delta; + + bnexterr = cur0; /* Process component 0 */ + delta = cur0 * 2; + cur0 += delta; /* form error * 3 */ + errorptr[0] = (FSERROR) (bpreverr0 + cur0); + cur0 += delta; /* form error * 5 */ + bpreverr0 = belowerr0 + cur0; + belowerr0 = bnexterr; + cur0 += delta; /* form error * 7 */ + bnexterr = cur1; /* Process component 1 */ + delta = cur1 * 2; + cur1 += delta; /* form error * 3 */ + errorptr[1] = (FSERROR) (bpreverr1 + cur1); + cur1 += delta; /* form error * 5 */ + bpreverr1 = belowerr1 + cur1; + belowerr1 = bnexterr; + cur1 += delta; /* form error * 7 */ + bnexterr = cur2; /* Process component 2 */ + delta = cur2 * 2; + cur2 += delta; /* form error * 3 */ + errorptr[2] = (FSERROR) (bpreverr2 + cur2); + cur2 += delta; /* form error * 5 */ + bpreverr2 = belowerr2 + cur2; + belowerr2 = bnexterr; + cur2 += delta; /* form error * 7 */ + } + /* At this point curN contains the 7/16 error value to be propagated + * to the next pixel on the current line, and all the errors for the + * next line have been shifted over. We are therefore ready to move on. + */ + inptr += dir3; /* Advance pixel pointers to next column */ + outptr += dir; + errorptr += dir3; /* advance errorptr to current column */ + } + /* Post-loop cleanup: we must unload the final error values into the + * final fserrors[] entry. Note we need not unload belowerrN because + * it is for the dummy column before or after the actual array. + */ + errorptr[0] = (FSERROR) bpreverr0; /* unload prev errs into array */ + errorptr[1] = (FSERROR) bpreverr1; + errorptr[2] = (FSERROR) bpreverr2; + } +} + + +/* + * Initialize the error-limiting transfer function (lookup table). + * The raw F-S error computation can potentially compute error values of up to + * +- MAXJSAMPLE. But we want the maximum correction applied to a pixel to be + * much less, otherwise obviously wrong pixels will be created. (Typical + * effects include weird fringes at color-area boundaries, isolated bright + * pixels in a dark area, etc.) The standard advice for avoiding this problem + * is to ensure that the "corners" of the color cube are allocated as output + * colors; then repeated errors in the same direction cannot cause cascading + * error buildup. However, that only prevents the error from getting + * completely out of hand; Aaron Giles reports that error limiting improves + * the results even with corner colors allocated. + * A simple clamping of the error values to about +- MAXJSAMPLE/8 works pretty + * well, but the smoother transfer function used below is even better. Thanks + * to Aaron Giles for this idea. + */ + +LOCAL(void) +init_error_limit (j_decompress_ptr cinfo) +/* Allocate and fill in the error_limiter table */ +{ + my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; + int * table; + int in, out; + + table = (int *) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE*2+1) * SIZEOF(int)); + table += MAXJSAMPLE; /* so can index -MAXJSAMPLE .. +MAXJSAMPLE */ + cquantize->error_limiter = table; + +#define STEPSIZE ((MAXJSAMPLE+1)/16) + /* Map errors 1:1 up to +- MAXJSAMPLE/16 */ + out = 0; + for (in = 0; in < STEPSIZE; in++, out++) { + table[in] = out; table[-in] = -out; + } + /* Map errors 1:2 up to +- 3*MAXJSAMPLE/16 */ + for (; in < STEPSIZE*3; in++, out += (in&1) ? 0 : 1) { + table[in] = out; table[-in] = -out; + } + /* Clamp the rest to final out value (which is (MAXJSAMPLE+1)/8) */ + for (; in <= MAXJSAMPLE; in++) { + table[in] = out; table[-in] = -out; + } +#undef STEPSIZE +} + + +/* + * Finish up at the end of each pass. + */ + +METHODDEF(void) +finish_pass1 (j_decompress_ptr cinfo) +{ + my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; + + /* Select the representative colors and fill in cinfo->colormap */ + cinfo->colormap = cquantize->sv_colormap; + select_colors(cinfo, cquantize->desired); + /* Force next pass to zero the color index table */ + cquantize->needs_zeroed = TRUE; +} + + +METHODDEF(void) +finish_pass2 (j_decompress_ptr cinfo) +{ + /* no work */ +} + + +/* + * Initialize for each processing pass. + */ + +METHODDEF(void) +start_pass_2_quant (j_decompress_ptr cinfo, boolean is_pre_scan) +{ + my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; + hist3d histogram = cquantize->histogram; + int i; + + /* Only F-S dithering or no dithering is supported. */ + /* If user asks for ordered dither, give him F-S. */ + if (cinfo->dither_mode != JDITHER_NONE) + cinfo->dither_mode = JDITHER_FS; + + if (is_pre_scan) { + /* Set up method pointers */ + cquantize->pub.color_quantize = prescan_quantize; + cquantize->pub.finish_pass = finish_pass1; + cquantize->needs_zeroed = TRUE; /* Always zero histogram */ + } else { + /* Set up method pointers */ + if (cinfo->dither_mode == JDITHER_FS) + cquantize->pub.color_quantize = pass2_fs_dither; + else + cquantize->pub.color_quantize = pass2_no_dither; + cquantize->pub.finish_pass = finish_pass2; + + /* Make sure color count is acceptable */ + i = cinfo->actual_number_of_colors; + if (i < 1) + ERREXIT1(cinfo, JERR_QUANT_FEW_COLORS, 1); + if (i > MAXNUMCOLORS) + ERREXIT1(cinfo, JERR_QUANT_MANY_COLORS, MAXNUMCOLORS); + + if (cinfo->dither_mode == JDITHER_FS) { + size_t arraysize = ((size_t) cinfo->output_width + (size_t) 2) + * (3 * SIZEOF(FSERROR)); + /* Allocate Floyd-Steinberg workspace if we didn't already. */ + if (cquantize->fserrors == NULL) + cquantize->fserrors = (FSERRPTR) (*cinfo->mem->alloc_large) + ((j_common_ptr) cinfo, JPOOL_IMAGE, arraysize); + /* Initialize the propagated errors to zero. */ + FMEMZERO((void FAR *) cquantize->fserrors, arraysize); + /* Make the error-limit table if we didn't already. */ + if (cquantize->error_limiter == NULL) + init_error_limit(cinfo); + cquantize->on_odd_row = FALSE; + } + + } + /* Zero the histogram or inverse color map, if necessary */ + if (cquantize->needs_zeroed) { + for (i = 0; i < HIST_C0_ELEMS; i++) { + FMEMZERO((void FAR *) histogram[i], + HIST_C1_ELEMS*HIST_C2_ELEMS * SIZEOF(histcell)); + } + cquantize->needs_zeroed = FALSE; + } +} + + +/* + * Switch to a new external colormap between output passes. + */ + +METHODDEF(void) +new_color_map_2_quant (j_decompress_ptr cinfo) +{ + my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; + + /* Reset the inverse color map */ + cquantize->needs_zeroed = TRUE; +} + + +/* + * Module initialization routine for 2-pass color quantization. + */ + +GLOBAL(void) +jinit_2pass_quantizer (j_decompress_ptr cinfo) +{ + my_cquantize_ptr cquantize; + int i; + + cquantize = (my_cquantize_ptr) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_cquantizer)); + cinfo->cquantize = &cquantize->pub; + cquantize->pub.start_pass = start_pass_2_quant; + cquantize->pub.new_color_map = new_color_map_2_quant; + cquantize->fserrors = NULL; /* flag optional arrays not allocated */ + cquantize->error_limiter = NULL; + + /* Make sure jdmaster didn't give me a case I can't handle */ + if (cinfo->out_color_components != 3) + ERREXIT(cinfo, JERR_NOTIMPL); + + /* Allocate the histogram/inverse colormap storage */ + cquantize->histogram = (hist3d) (*cinfo->mem->alloc_small) + ((j_common_ptr) cinfo, JPOOL_IMAGE, HIST_C0_ELEMS * SIZEOF(hist2d)); + for (i = 0; i < HIST_C0_ELEMS; i++) { + cquantize->histogram[i] = (hist2d) (*cinfo->mem->alloc_large) + ((j_common_ptr) cinfo, JPOOL_IMAGE, + HIST_C1_ELEMS*HIST_C2_ELEMS * SIZEOF(histcell)); + } + cquantize->needs_zeroed = TRUE; /* histogram is garbage now */ + + /* Allocate storage for the completed colormap, if required. + * We do this now since it is FAR storage and may affect + * the memory manager's space calculations. + */ + if (cinfo->enable_2pass_quant) { + /* Make sure color count is acceptable */ + int desired = cinfo->desired_number_of_colors; + /* Lower bound on # of colors ... somewhat arbitrary as long as > 0 */ + if (desired < 8) + ERREXIT1(cinfo, JERR_QUANT_FEW_COLORS, 8); + /* Make sure colormap indexes can be represented by JSAMPLEs */ + if (desired > MAXNUMCOLORS) + ERREXIT1(cinfo, JERR_QUANT_MANY_COLORS, MAXNUMCOLORS); + cquantize->sv_colormap = (*cinfo->mem->alloc_sarray) + ((j_common_ptr) cinfo, JPOOL_IMAGE, + (JDIMENSION) desired, (JDIMENSION) 3); + cquantize->desired = desired; + } else + cquantize->sv_colormap = NULL; + + /* Only F-S dithering or no dithering is supported. */ + /* If user asks for ordered dither, give him F-S. */ + if (cinfo->dither_mode != JDITHER_NONE) + cinfo->dither_mode = JDITHER_FS; + + /* Allocate Floyd-Steinberg workspace if necessary. + * This isn't really needed until pass 2, but again it is FAR storage. + * Although we will cope with a later change in dither_mode, + * we do not promise to honor max_memory_to_use if dither_mode changes. + */ + if (cinfo->dither_mode == JDITHER_FS) { + cquantize->fserrors = (FSERRPTR) (*cinfo->mem->alloc_large) + ((j_common_ptr) cinfo, JPOOL_IMAGE, + ((size_t) cinfo->output_width + (size_t) 2) * (3 * SIZEOF(FSERROR))); + /* Might as well create the error-limiting table too. */ + init_error_limit(cinfo); + } +} + +#endif /* QUANT_2PASS_SUPPORTED */ diff --git a/cmake/externals/libjpeg/jutils.c b/cmake/externals/libjpeg/jutils.c new file mode 100644 index 000000000..31e16dfb5 --- /dev/null +++ b/cmake/externals/libjpeg/jutils.c @@ -0,0 +1,224 @@ +/* + * jutils.c + * + * Copyright (C) 1991-1996, Thomas G. Lane. + * Modified 2009-2020 by Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains tables and miscellaneous utility routines needed + * for both compression and decompression. + * Note we prefix all global names with "j" to minimize conflicts with + * a surrounding application. + */ + +#define JPEG_INTERNALS +#include "jinclude.h" +#include "jpeglib.h" + + +/* + * jpeg_zigzag_order[i] is the zigzag-order position of the i'th element + * of a DCT block read in natural order (left to right, top to bottom). + */ + +#if 0 /* This table is not actually needed in v6a */ + +const int jpeg_zigzag_order[DCTSIZE2] = { + 0, 1, 5, 6, 14, 15, 27, 28, + 2, 4, 7, 13, 16, 26, 29, 42, + 3, 8, 12, 17, 25, 30, 41, 43, + 9, 11, 18, 24, 31, 40, 44, 53, + 10, 19, 23, 32, 39, 45, 52, 54, + 20, 22, 33, 38, 46, 51, 55, 60, + 21, 34, 37, 47, 50, 56, 59, 61, + 35, 36, 48, 49, 57, 58, 62, 63 +}; + +#endif + +/* + * jpeg_natural_order[i] is the natural-order position of the i'th element + * of zigzag order. + * + * When reading corrupted data, the Huffman decoders could attempt + * to reference an entry beyond the end of this array (if the decoded + * zero run length reaches past the end of the block). To prevent + * wild stores without adding an inner-loop test, we put some extra + * "63"s after the real entries. This will cause the extra coefficient + * to be stored in location 63 of the block, not somewhere random. + * The worst case would be a run-length of 15, which means we need 16 + * fake entries. + */ + +const int jpeg_natural_order[DCTSIZE2+16] = { + 0, 1, 8, 16, 9, 2, 3, 10, + 17, 24, 32, 25, 18, 11, 4, 5, + 12, 19, 26, 33, 40, 48, 41, 34, + 27, 20, 13, 6, 7, 14, 21, 28, + 35, 42, 49, 56, 57, 50, 43, 36, + 29, 22, 15, 23, 30, 37, 44, 51, + 58, 59, 52, 45, 38, 31, 39, 46, + 53, 60, 61, 54, 47, 55, 62, 63, + 63, 63, 63, 63, 63, 63, 63, 63, /* extra entries for safety in decoder */ + 63, 63, 63, 63, 63, 63, 63, 63 +}; + +const int jpeg_natural_order7[7*7+16] = { + 0, 1, 8, 16, 9, 2, 3, 10, + 17, 24, 32, 25, 18, 11, 4, 5, + 12, 19, 26, 33, 40, 48, 41, 34, + 27, 20, 13, 6, 14, 21, 28, 35, + 42, 49, 50, 43, 36, 29, 22, 30, + 37, 44, 51, 52, 45, 38, 46, 53, + 54, + 63, 63, 63, 63, 63, 63, 63, 63, /* extra entries for safety in decoder */ + 63, 63, 63, 63, 63, 63, 63, 63 +}; + +const int jpeg_natural_order6[6*6+16] = { + 0, 1, 8, 16, 9, 2, 3, 10, + 17, 24, 32, 25, 18, 11, 4, 5, + 12, 19, 26, 33, 40, 41, 34, 27, + 20, 13, 21, 28, 35, 42, 43, 36, + 29, 37, 44, 45, + 63, 63, 63, 63, 63, 63, 63, 63, /* extra entries for safety in decoder */ + 63, 63, 63, 63, 63, 63, 63, 63 +}; + +const int jpeg_natural_order5[5*5+16] = { + 0, 1, 8, 16, 9, 2, 3, 10, + 17, 24, 32, 25, 18, 11, 4, 12, + 19, 26, 33, 34, 27, 20, 28, 35, + 36, + 63, 63, 63, 63, 63, 63, 63, 63, /* extra entries for safety in decoder */ + 63, 63, 63, 63, 63, 63, 63, 63 +}; + +const int jpeg_natural_order4[4*4+16] = { + 0, 1, 8, 16, 9, 2, 3, 10, + 17, 24, 25, 18, 11, 19, 26, 27, + 63, 63, 63, 63, 63, 63, 63, 63, /* extra entries for safety in decoder */ + 63, 63, 63, 63, 63, 63, 63, 63 +}; + +const int jpeg_natural_order3[3*3+16] = { + 0, 1, 8, 16, 9, 2, 10, 17, + 18, + 63, 63, 63, 63, 63, 63, 63, 63, /* extra entries for safety in decoder */ + 63, 63, 63, 63, 63, 63, 63, 63 +}; + +const int jpeg_natural_order2[2*2+16] = { + 0, 1, 8, 9, + 63, 63, 63, 63, 63, 63, 63, 63, /* extra entries for safety in decoder */ + 63, 63, 63, 63, 63, 63, 63, 63 +}; + + +/* + * Arithmetic utilities + */ + +GLOBAL(long) +jdiv_round_up (long a, long b) +/* Compute a/b rounded up to next integer, ie, ceil(a/b) */ +/* Assumes a >= 0, b > 0 */ +{ + return (a + b - 1L) / b; +} + + +GLOBAL(long) +jround_up (long a, long b) +/* Compute a rounded up to next multiple of b, ie, ceil(a/b)*b */ +/* Assumes a >= 0, b > 0 */ +{ + a += b - 1L; + return a - (a % b); +} + + +/* On normal machines we can apply MEMCOPY() and MEMZERO() to sample arrays + * and coefficient-block arrays. This won't work on 80x86 because the arrays + * are FAR and we're assuming a small-pointer memory model. However, some + * DOS compilers provide far-pointer versions of memcpy() and memset() even + * in the small-model libraries. These will be used if USE_FMEM is defined. + * Otherwise, the routines below do it the hard way. (The performance cost + * is not all that great, because these routines aren't very heavily used.) + */ + +#ifndef NEED_FAR_POINTERS /* normal case, same as regular macro */ +#define FMEMCOPY(dest,src,size) MEMCOPY(dest,src,size) +#else /* 80x86 case, define if we can */ +#ifdef USE_FMEM +#define FMEMCOPY(dest,src,size) _fmemcpy((void FAR *)(dest), (const void FAR *)(src), (size_t)(size)) +#else +/* This function is for use by the FMEMZERO macro defined in jpegint.h. + * Do not call this function directly, use the FMEMZERO macro instead. + */ +GLOBAL(void) +jzero_far (void FAR * target, size_t bytestozero) +/* Zero out a chunk of FAR memory. */ +/* This might be sample-array data, block-array data, or alloc_large data. */ +{ + register char FAR * ptr = (char FAR *) target; + register size_t count; + + for (count = bytestozero; count > 0; count--) { + *ptr++ = 0; + } +} +#endif +#endif + + +GLOBAL(void) +jcopy_sample_rows (JSAMPARRAY input_array, + JSAMPARRAY output_array, + int num_rows, JDIMENSION num_cols) +/* Copy some rows of samples from one place to another. + * num_rows rows are copied from *input_array++ to *output_array++; + * these areas may overlap for duplication. + * The source and destination arrays must be at least as wide as num_cols. + */ +{ + register JSAMPROW inptr, outptr; +#ifdef FMEMCOPY + register size_t count = (size_t) num_cols * SIZEOF(JSAMPLE); +#else + register JDIMENSION count; +#endif + register int row; + + for (row = num_rows; row > 0; row--) { + inptr = *input_array++; + outptr = *output_array++; +#ifdef FMEMCOPY + FMEMCOPY(outptr, inptr, count); +#else + for (count = num_cols; count > 0; count--) + *outptr++ = *inptr++; /* needn't bother with GETJSAMPLE() here */ +#endif + } +} + + +GLOBAL(void) +jcopy_block_row (JBLOCKROW input_row, JBLOCKROW output_row, + JDIMENSION num_blocks) +/* Copy a row of coefficient blocks from one place to another. */ +{ +#ifdef FMEMCOPY + FMEMCOPY(output_row, input_row, (size_t) num_blocks * (DCTSIZE2 * SIZEOF(JCOEF))); +#else + register JCOEFPTR inptr, outptr; + register long count; + + inptr = (JCOEFPTR) input_row; + outptr = (JCOEFPTR) output_row; + for (count = (long) num_blocks * DCTSIZE2; count > 0; count--) { + *outptr++ = *inptr++; + } +#endif +} diff --git a/cmake/externals/libjpeg/jversion.h b/cmake/externals/libjpeg/jversion.h new file mode 100644 index 000000000..df53ef5e5 --- /dev/null +++ b/cmake/externals/libjpeg/jversion.h @@ -0,0 +1,14 @@ +/* + * jversion.h + * + * Copyright (C) 1991-2024, Thomas G. Lane, Guido Vollbeding. + * This file is part of the Independent JPEG Group's software. + * For conditions of distribution and use, see the accompanying README file. + * + * This file contains software version identification. + */ + + +#define JVERSION "9f 14-Jan-2024" + +#define JCOPYRIGHT "Copyright (C) 2024, Thomas G. Lane, Guido Vollbeding" diff --git a/cmake/externals/libspng/CMakeLists.txt b/cmake/externals/libspng/CMakeLists.txt new file mode 100644 index 000000000..c277888e5 --- /dev/null +++ b/cmake/externals/libspng/CMakeLists.txt @@ -0,0 +1,39 @@ +# ---------------------------------------------------------------------------- +# CMake file for libspng. See root CMakeLists.txt +# +# ---------------------------------------------------------------------------- + +project(SPNG) + +set(CURR_INCLUDE_DIR "${CMAKE_CURRENT_LIST_DIR}") +set_property(GLOBAL PROPERTY SPNG_INCLUDE_DIR ${CURR_INCLUDE_DIR}) + +file(GLOB_RECURSE spng_headers RELATIVE "${CMAKE_CURRENT_LIST_DIR}" "*.h") +file(GLOB_RECURSE spng_sources RELATIVE "${CMAKE_CURRENT_LIST_DIR}" "*.c") + +message(STATUS "libspng will be used as PNG codec") + +# ---------------------------------------------------------------------------------- +# Define the library target: +# ---------------------------------------------------------------------------------- + +if(MSVC) + add_definitions(-D_CRT_SECURE_NO_DEPRECATE) +endif(MSVC) + +add_library(SPNG STATIC ${OPENCV_3RDPARTY_EXCLUDE_FROM_ALL} ${spng_headers} ${spng_sources}) +target_link_libraries(SPNG ${ZLIB_LIBRARIES}) + +set_target_properties(SPNG + PROPERTIES OUTPUT_NAME SPNG + DEBUG_POSTFIX "${OPENCV_DEBUG_POSTFIX}" + COMPILE_PDB_NAME SPNG + COMPILE_PDB_NAME_DEBUG "SPNG${OPENCV_DEBUG_POSTFIX}" + ARCHIVE_OUTPUT_DIRECTORY "${3P_LIBRARY_OUTPUT_PATH}" + ) + +target_compile_definitions(SPNG PUBLIC SPNG_STATIC) + +if(ENABLE_SOLUTION_FOLDERS) + set_target_properties(SPNG PROPERTIES FOLDER "3rdparty") +endif() \ No newline at end of file diff --git a/cmake/externals/libspng/LICENSE b/cmake/externals/libspng/LICENSE new file mode 100644 index 000000000..a29d5a2fa --- /dev/null +++ b/cmake/externals/libspng/LICENSE @@ -0,0 +1,25 @@ +BSD 2-Clause License + +Copyright (c) 2018-2023, Randy +All rights reserved. + +Redistribution and use in source and binary forms, with or without +modification, are permitted provided that the following conditions are met: + +* Redistributions of source code must retain the above copyright notice, this + list of conditions and the following disclaimer. + +* Redistributions in binary form must reproduce the above copyright notice, + this list of conditions and the following disclaimer in the documentation + and/or other materials provided with the distribution. + +THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" +AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE +DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE +FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL +DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR +SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER +CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, +OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. \ No newline at end of file diff --git a/cmake/externals/libspng/spng.c b/cmake/externals/libspng/spng.c new file mode 100644 index 000000000..b22b7110a --- /dev/null +++ b/cmake/externals/libspng/spng.c @@ -0,0 +1,6980 @@ +/* SPDX-License-Identifier: (BSD-2-Clause AND libpng-2.0) */ + +#define SPNG__BUILD + +#include "spng.h" + +#include +#include +#include +#include + +#define ZLIB_CONST + +#ifdef __FRAMAC__ + #define SPNG_DISABLE_OPT + #include "tests/framac_stubs.h" +#else + #ifdef SPNG_USE_MINIZ + #include + #else + #include + #endif +#endif + +#ifdef SPNG_MULTITHREADING + #include +#endif + +/* Not build options, edit at your own risk! */ +#define SPNG_READ_SIZE (8192) +#define SPNG_WRITE_SIZE SPNG_READ_SIZE +#define SPNG_MAX_CHUNK_COUNT (1000) + +#define SPNG_TARGET_CLONES(x) + +#ifndef SPNG_DISABLE_OPT + + #if defined(__i386__) || defined(__x86_64__) || defined(_M_IX86) || defined(_M_X64) + #define SPNG_X86 + + #if defined(__x86_64__) || defined(_M_X64) + #define SPNG_X86_64 + #endif + + #elif defined(__aarch64__) || defined(_M_ARM64) /* || defined(__ARM_NEON) */ + #define SPNG_ARM /* NOTE: only arm64 builds are tested! */ + #else + #pragma message "disabling SIMD optimizations for unknown target" + #define SPNG_DISABLE_OPT + #endif + + #if defined(SPNG_X86_64) && defined(SPNG_ENABLE_TARGET_CLONES) + #undef SPNG_TARGET_CLONES + #define SPNG_TARGET_CLONES(x) __attribute__((target_clones(x))) + #else + #define SPNG_TARGET_CLONES(x) + #endif + + #ifndef SPNG_DISABLE_OPT + static void defilter_sub3(size_t rowbytes, unsigned char *row); + static void defilter_sub4(size_t rowbytes, unsigned char *row); + static void defilter_avg3(size_t rowbytes, unsigned char *row, const unsigned char *prev); + static void defilter_avg4(size_t rowbytes, unsigned char *row, const unsigned char *prev); + static void defilter_paeth3(size_t rowbytes, unsigned char *row, const unsigned char *prev); + static void defilter_paeth4(size_t rowbytes, unsigned char *row, const unsigned char *prev); + + #if defined(SPNG_ARM) + static uint32_t expand_palette_rgba8_neon(unsigned char *row, const unsigned char *scanline, const unsigned char *plte, uint32_t width); + static uint32_t expand_palette_rgb8_neon(unsigned char *row, const unsigned char *scanline, const unsigned char *plte, uint32_t width); + #endif + #endif +#endif + +#if defined(_MSC_VER) + #pragma warning(push) + #pragma warning(disable: 4244) +#endif + +#if (defined(__BYTE_ORDER__) && __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__) || defined(__BIG_ENDIAN__) + #define SPNG_BIG_ENDIAN +#else + #define SPNG_LITTLE_ENDIAN +#endif + +enum spng_state +{ + SPNG_STATE_INVALID = 0, + SPNG_STATE_INIT = 1, /* No PNG buffer/stream is set */ + SPNG_STATE_INPUT, /* Decoder input PNG was set */ + SPNG_STATE_OUTPUT = SPNG_STATE_INPUT, /* Encoder output was set */ + SPNG_STATE_IHDR, /* IHDR was read/written */ + SPNG_STATE_FIRST_IDAT, /* Encoded up to / reached first IDAT */ + SPNG_STATE_DECODE_INIT, /* Decoder is ready for progressive reads */ + SPNG_STATE_ENCODE_INIT = SPNG_STATE_DECODE_INIT, + SPNG_STATE_EOI, /* Reached the last scanline/row */ + SPNG_STATE_LAST_IDAT, /* Reached last IDAT, set at end of decode_image() */ + SPNG_STATE_AFTER_IDAT, /* */ + SPNG_STATE_IEND, /* Reached IEND */ +}; + +enum spng__internal +{ + SPNG__IO_SIGNAL = 1 << 9, + SPNG__CTX_FLAGS_ALL = (SPNG_CTX_IGNORE_ADLER32 | SPNG_CTX_ENCODER) +}; + +#define SPNG_STR(x) _SPNG_STR(x) +#define _SPNG_STR(x) #x + +#define SPNG_VERSION_STRING SPNG_STR(SPNG_VERSION_MAJOR) "." \ + SPNG_STR(SPNG_VERSION_MINOR) "." \ + SPNG_STR(SPNG_VERSION_PATCH) + +#define SPNG_GET_CHUNK_BOILERPLATE(chunk) \ + if(ctx == NULL) return 1; \ + int ret = read_chunks(ctx, 0); \ + if(ret) return ret; \ + if(!ctx->stored.chunk) return SPNG_ECHUNKAVAIL; \ + if(chunk == NULL) return 1 + +#define SPNG_SET_CHUNK_BOILERPLATE(chunk) \ + if(ctx == NULL || chunk == NULL) return 1; \ + if(ctx->data == NULL && !ctx->encode_only) return SPNG_ENOSRC; \ + int ret = read_chunks(ctx, 0); \ + if(ret) return ret + +/* Determine if the spng_option can be overriden/optimized */ +#define spng__optimize(option) (ctx->optimize_option & (1 << option)) + +struct spng_subimage +{ + uint32_t width; + uint32_t height; + size_t out_width; /* byte width based on output format */ + size_t scanline_width; +}; + +struct spng_text2 +{ + int type; + char *keyword; + char *text; + + size_t text_length; + + uint8_t compression_flag; /* iTXt only */ + char *language_tag; /* iTXt only */ + char *translated_keyword; /* iTXt only */ + + size_t cache_usage; + char user_keyword_storage[80]; +}; + +struct decode_flags +{ + unsigned apply_trns: 1; + unsigned apply_gamma: 1; + unsigned use_sbit: 1; + unsigned indexed: 1; + unsigned do_scaling: 1; + unsigned interlaced: 1; + unsigned same_layout: 1; + unsigned zerocopy: 1; + unsigned unpack: 1; +}; + +struct encode_flags +{ + unsigned interlace: 1; + unsigned same_layout: 1; + unsigned to_bigendian: 1; + unsigned progressive: 1; + unsigned finalize: 1; + + enum spng_filter_choice filter_choice; +}; + +struct spng_chunk_bitfield +{ + unsigned ihdr: 1; + unsigned plte: 1; + unsigned chrm: 1; + unsigned iccp: 1; + unsigned gama: 1; + unsigned sbit: 1; + unsigned srgb: 1; + unsigned text: 1; + unsigned bkgd: 1; + unsigned hist: 1; + unsigned trns: 1; + unsigned phys: 1; + unsigned splt: 1; + unsigned time: 1; + unsigned offs: 1; + unsigned exif: 1; + unsigned unknown: 1; +}; + +/* Packed sample iterator */ +struct spng__iter +{ + const uint8_t mask; + unsigned shift_amount; + const unsigned initial_shift, bit_depth; + const unsigned char *samples; +}; + +union spng__decode_plte +{ + struct spng_plte_entry rgba[256]; + unsigned char rgb[256 * 3]; + unsigned char raw[256 * 4]; + uint32_t align_this; +}; + +struct spng__zlib_options +{ + int compression_level; + int window_bits; + int mem_level; + int strategy; + int data_type; +}; + +typedef void spng__undo(spng_ctx *ctx); + +struct spng_ctx +{ + size_t data_size; + size_t bytes_read; + size_t stream_buf_size; + unsigned char *stream_buf; + const unsigned char *data; + + /* User-defined pointers for streaming */ + spng_read_fn *read_fn; + spng_write_fn *write_fn; + void *stream_user_ptr; + + /* Used for buffer reads */ + const unsigned char *png_base; + size_t bytes_left; + size_t last_read_size; + + /* Used for encoding */ + int user_owns_out_png; + unsigned char *out_png; + unsigned char *write_ptr; + size_t out_png_size; + size_t bytes_encoded; + + /* These are updated by read/write_header()/read_chunk_bytes() */ + struct spng_chunk current_chunk; + uint32_t cur_chunk_bytes_left; + uint32_t cur_actual_crc; + + struct spng_alloc alloc; + + enum spng_ctx_flags flags; + enum spng_format fmt; + + enum spng_state state; + + unsigned streaming: 1; + unsigned internal_buffer: 1; /* encoding to internal buffer */ + + unsigned inflate: 1; + unsigned deflate: 1; + unsigned encode_only: 1; + unsigned strict: 1; + unsigned discard: 1; + unsigned skip_crc: 1; + unsigned keep_unknown: 1; + unsigned prev_was_idat: 1; + + struct spng__zlib_options image_options; + struct spng__zlib_options text_options; + + spng__undo *undo; + + /* input file contains this chunk */ + struct spng_chunk_bitfield file; + + /* chunk was stored with spng_set_*() */ + struct spng_chunk_bitfield user; + + /* chunk was stored by reading or with spng_set_*() */ + struct spng_chunk_bitfield stored; + + /* used to reset the above in case of an error */ + struct spng_chunk_bitfield prev_stored; + + struct spng_chunk first_idat, last_idat; + + uint32_t max_width, max_height; + + size_t max_chunk_size; + size_t chunk_cache_limit; + size_t chunk_cache_usage; + uint32_t chunk_count_limit; + uint32_t chunk_count_total; + + int crc_action_critical; + int crc_action_ancillary; + + uint32_t optimize_option; + + struct spng_ihdr ihdr; + + struct spng_plte plte; + + struct spng_chrm_int chrm_int; + struct spng_iccp iccp; + + uint32_t gama; + + struct spng_sbit sbit; + + uint8_t srgb_rendering_intent; + + uint32_t n_text; + struct spng_text2 *text_list; + + struct spng_bkgd bkgd; + struct spng_hist hist; + struct spng_trns trns; + struct spng_phys phys; + + uint32_t n_splt; + struct spng_splt *splt_list; + + struct spng_time time; + struct spng_offs offs; + struct spng_exif exif; + + uint32_t n_chunks; + struct spng_unknown_chunk *chunk_list; + + struct spng_subimage subimage[7]; + + z_stream zstream; + unsigned char *scanline_buf, *prev_scanline_buf, *row_buf, *filtered_scanline_buf; + unsigned char *scanline, *prev_scanline, *row, *filtered_scanline; + + /* based on fmt */ + size_t image_size; /* may be zero */ + size_t image_width; + + unsigned bytes_per_pixel; /* derived from ihdr */ + unsigned pixel_size; /* derived from spng_format+ihdr */ + int widest_pass; + int last_pass; /* last non-empty pass */ + + uint16_t *gamma_lut; /* points to either _lut8 or _lut16 */ + uint16_t *gamma_lut16; + uint16_t gamma_lut8[256]; + unsigned char trns_px[8]; + union spng__decode_plte decode_plte; + struct spng_sbit decode_sb; + struct decode_flags decode_flags; + struct spng_row_info row_info; + + struct encode_flags encode_flags; +}; + +static const uint32_t spng_u32max = INT32_MAX; + +static const uint32_t adam7_x_start[7] = { 0, 4, 0, 2, 0, 1, 0 }; +static const uint32_t adam7_y_start[7] = { 0, 0, 4, 0, 2, 0, 1 }; +static const uint32_t adam7_x_delta[7] = { 8, 8, 4, 4, 2, 2, 1 }; +static const uint32_t adam7_y_delta[7] = { 8, 8, 8, 4, 4, 2, 2 }; + +static const uint8_t spng_signature[8] = { 137, 80, 78, 71, 13, 10, 26, 10 }; + +static const uint8_t type_ihdr[4] = { 73, 72, 68, 82 }; +static const uint8_t type_plte[4] = { 80, 76, 84, 69 }; +static const uint8_t type_idat[4] = { 73, 68, 65, 84 }; +static const uint8_t type_iend[4] = { 73, 69, 78, 68 }; + +static const uint8_t type_trns[4] = { 116, 82, 78, 83 }; +static const uint8_t type_chrm[4] = { 99, 72, 82, 77 }; +static const uint8_t type_gama[4] = { 103, 65, 77, 65 }; +static const uint8_t type_iccp[4] = { 105, 67, 67, 80 }; +static const uint8_t type_sbit[4] = { 115, 66, 73, 84 }; +static const uint8_t type_srgb[4] = { 115, 82, 71, 66 }; +static const uint8_t type_text[4] = { 116, 69, 88, 116 }; +static const uint8_t type_ztxt[4] = { 122, 84, 88, 116 }; +static const uint8_t type_itxt[4] = { 105, 84, 88, 116 }; +static const uint8_t type_bkgd[4] = { 98, 75, 71, 68 }; +static const uint8_t type_hist[4] = { 104, 73, 83, 84 }; +static const uint8_t type_phys[4] = { 112, 72, 89, 115 }; +static const uint8_t type_splt[4] = { 115, 80, 76, 84 }; +static const uint8_t type_time[4] = { 116, 73, 77, 69 }; + +static const uint8_t type_offs[4] = { 111, 70, 70, 115 }; +static const uint8_t type_exif[4] = { 101, 88, 73, 102 }; + +static inline void *spng__malloc(spng_ctx *ctx, size_t size) +{ + return ctx->alloc.malloc_fn(size); +} + +static inline void *spng__calloc(spng_ctx *ctx, size_t nmemb, size_t size) +{ + return ctx->alloc.calloc_fn(nmemb, size); +} + +static inline void *spng__realloc(spng_ctx *ctx, void *ptr, size_t size) +{ + return ctx->alloc.realloc_fn(ptr, size); +} + +static inline void spng__free(spng_ctx *ctx, void *ptr) +{ + ctx->alloc.free_fn(ptr); +} + +#if defined(SPNG_USE_MINIZ) +static void *spng__zalloc(void *opaque, size_t items, size_t size) +#else +static void *spng__zalloc(void *opaque, uInt items, uInt size) +#endif +{ + spng_ctx *ctx = opaque; + + if(size > SIZE_MAX / items) return NULL; + + size_t len = (size_t)items * size; + + return spng__malloc(ctx, len); +} + +static void spng__zfree(void *opqaue, void *ptr) +{ + spng_ctx *ctx = opqaue; + spng__free(ctx, ptr); +} + +static inline uint16_t read_u16(const void *src) +{ + const unsigned char *data = src; + + return (data[0] & 0xFFU) << 8 | (data[1] & 0xFFU); +} + +static inline uint32_t read_u32(const void *src) +{ + const unsigned char *data = src; + + return (data[0] & 0xFFUL) << 24 | (data[1] & 0xFFUL) << 16 | + (data[2] & 0xFFUL) << 8 | (data[3] & 0xFFUL); +} + +static inline int32_t read_s32(const void *src) +{ + int32_t ret = (int32_t)read_u32(src); + + return ret; +} + +static inline void write_u16(void *dest, uint16_t x) +{ + unsigned char *data = dest; + + data[0] = x >> 8; + data[1] = x & 0xFF; +} + +static inline void write_u32(void *dest, uint32_t x) +{ + unsigned char *data = dest; + + data[0] = (x >> 24); + data[1] = (x >> 16) & 0xFF; + data[2] = (x >> 8) & 0xFF; + data[3] = x & 0xFF; +} + +static inline void write_s32(void *dest, int32_t x) +{ + uint32_t n = x; + write_u32(dest, n); +} + +/* Returns an iterator for 1,2,4,8-bit samples */ +static struct spng__iter spng__iter_init(unsigned bit_depth, const unsigned char *samples) +{ + struct spng__iter iter = + { + .mask = (uint32_t)(1 << bit_depth) - 1, + .shift_amount = 8 - bit_depth, + .initial_shift = 8 - bit_depth, + .bit_depth = bit_depth, + .samples = samples + }; + + return iter; +} + +/* Returns the current sample unpacked, iterates to the next one */ +static inline uint8_t get_sample(struct spng__iter *iter) +{ + uint8_t x = (iter->samples[0] >> iter->shift_amount) & iter->mask; + + iter->shift_amount -= iter->bit_depth; + + if(iter->shift_amount > 7) + { + iter->shift_amount = iter->initial_shift; + iter->samples++; + } + + return x; +} + +static void u16_row_to_host(void *row, size_t size) +{ + uint16_t *px = row; + size_t i, n = size / 2; + + for(i=0; i < n; i++) + { + px[i] = read_u16(&px[i]); + } +} + +static void u16_row_to_bigendian(void *row, size_t size) +{ + uint16_t *px = (uint16_t*)row; + size_t i, n = size / 2; + + for(i=0; i < n; i++) + { + write_u16(&px[i], px[i]); + } +} + +static void rgb8_row_to_rgba8(const unsigned char *row, unsigned char *out, uint32_t n) +{ + uint32_t i; + for(i=0; i < n; i++) + { + memcpy(out + i * 4, row + i * 3, 3); + out[i*4+3] = 255; + } +} + +static unsigned num_channels(const struct spng_ihdr *ihdr) +{ + switch(ihdr->color_type) + { + case SPNG_COLOR_TYPE_TRUECOLOR: return 3; + case SPNG_COLOR_TYPE_GRAYSCALE_ALPHA: return 2; + case SPNG_COLOR_TYPE_TRUECOLOR_ALPHA: return 4; + case SPNG_COLOR_TYPE_GRAYSCALE: + case SPNG_COLOR_TYPE_INDEXED: + return 1; + default: return 0; + } +} + +/* Calculate scanline width in bits, round up to the nearest byte */ +static int calculate_scanline_width(const struct spng_ihdr *ihdr, uint32_t width, size_t *scanline_width) +{ + if(ihdr == NULL || !width) return SPNG_EINTERNAL; + + size_t res = num_channels(ihdr) * ihdr->bit_depth; + + if(res > SIZE_MAX / width) return SPNG_EOVERFLOW; + res = res * width; + + res += 15; /* Filter byte + 7 for rounding */ + + if(res < 15) return SPNG_EOVERFLOW; + + res /= 8; + + if(res > UINT32_MAX) return SPNG_EOVERFLOW; + + *scanline_width = res; + + return 0; +} + +static int calculate_subimages(struct spng_ctx *ctx) +{ + if(ctx == NULL) return SPNG_EINTERNAL; + + struct spng_ihdr *ihdr = &ctx->ihdr; + struct spng_subimage *sub = ctx->subimage; + + if(ihdr->interlace_method == 1) + { + sub[0].width = (ihdr->width + 7) >> 3; + sub[0].height = (ihdr->height + 7) >> 3; + sub[1].width = (ihdr->width + 3) >> 3; + sub[1].height = (ihdr->height + 7) >> 3; + sub[2].width = (ihdr->width + 3) >> 2; + sub[2].height = (ihdr->height + 3) >> 3; + sub[3].width = (ihdr->width + 1) >> 2; + sub[3].height = (ihdr->height + 3) >> 2; + sub[4].width = (ihdr->width + 1) >> 1; + sub[4].height = (ihdr->height + 1) >> 2; + sub[5].width = ihdr->width >> 1; + sub[5].height = (ihdr->height + 1) >> 1; + sub[6].width = ihdr->width; + sub[6].height = ihdr->height >> 1; + } + else + { + sub[0].width = ihdr->width; + sub[0].height = ihdr->height; + } + + int i; + for(i=0; i < 7; i++) + { + if(sub[i].width == 0 || sub[i].height == 0) continue; + + int ret = calculate_scanline_width(ihdr, sub[i].width, &sub[i].scanline_width); + if(ret) return ret; + + if(sub[ctx->widest_pass].scanline_width < sub[i].scanline_width) ctx->widest_pass = i; + + ctx->last_pass = i; + } + + return 0; +} + +static int check_decode_fmt(const struct spng_ihdr *ihdr, const int fmt) +{ + switch(fmt) + { + case SPNG_FMT_RGBA8: + case SPNG_FMT_RGBA16: + case SPNG_FMT_RGB8: + case SPNG_FMT_PNG: + case SPNG_FMT_RAW: + return 0; + case SPNG_FMT_G8: + case SPNG_FMT_GA8: + if(ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE && ihdr->bit_depth <= 8) return 0; + else return SPNG_EFMT; + case SPNG_FMT_GA16: + if(ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE && ihdr->bit_depth == 16) return 0; + else return SPNG_EFMT; + default: return SPNG_EFMT; + } +} + +static int calculate_image_width(const struct spng_ihdr *ihdr, int fmt, size_t *len) +{ + if(ihdr == NULL || len == NULL) return SPNG_EINTERNAL; + + size_t res = ihdr->width; + unsigned bytes_per_pixel; + + switch(fmt) + { + case SPNG_FMT_RGBA8: + case SPNG_FMT_GA16: + bytes_per_pixel = 4; + break; + case SPNG_FMT_RGBA16: + bytes_per_pixel = 8; + break; + case SPNG_FMT_RGB8: + bytes_per_pixel = 3; + break; + case SPNG_FMT_PNG: + case SPNG_FMT_RAW: + { + int ret = calculate_scanline_width(ihdr, ihdr->width, &res); + if(ret) return ret; + + res -= 1; /* exclude filter byte */ + bytes_per_pixel = 1; + break; + } + case SPNG_FMT_G8: + bytes_per_pixel = 1; + break; + case SPNG_FMT_GA8: + bytes_per_pixel = 2; + break; + default: return SPNG_EINTERNAL; + } + + if(res > SIZE_MAX / bytes_per_pixel) return SPNG_EOVERFLOW; + res = res * bytes_per_pixel; + + *len = res; + + return 0; +} + +static int calculate_image_size(const struct spng_ihdr *ihdr, int fmt, size_t *len) +{ + if(ihdr == NULL || len == NULL) return SPNG_EINTERNAL; + + size_t res = 0; + + int ret = calculate_image_width(ihdr, fmt, &res); + if(ret) return ret; + + if(res > SIZE_MAX / ihdr->height) return SPNG_EOVERFLOW; + res = res * ihdr->height; + + *len = res; + + return 0; +} + +static int increase_cache_usage(spng_ctx *ctx, size_t bytes, int new_chunk) +{ + if(ctx == NULL || !bytes) return SPNG_EINTERNAL; + + if(new_chunk) + { + ctx->chunk_count_total++; + if(ctx->chunk_count_total < 1) return SPNG_EOVERFLOW; + + if(ctx->chunk_count_total > ctx->chunk_count_limit) return SPNG_ECHUNK_LIMITS; + } + + size_t new_usage = ctx->chunk_cache_usage + bytes; + + if(new_usage < ctx->chunk_cache_usage) return SPNG_EOVERFLOW; + + if(new_usage > ctx->chunk_cache_limit) return SPNG_ECHUNK_LIMITS; + + ctx->chunk_cache_usage = new_usage; + + return 0; +} + +static int decrease_cache_usage(spng_ctx *ctx, size_t usage) +{ + if(ctx == NULL || !usage) return SPNG_EINTERNAL; + if(usage > ctx->chunk_cache_usage) return SPNG_EINTERNAL; + + ctx->chunk_cache_usage -= usage; + + return 0; +} + +static int is_critical_chunk(struct spng_chunk *chunk) +{ + if(chunk == NULL) return 0; + if((chunk->type[0] & (1 << 5)) == 0) return 1; + + return 0; +} + +static int decode_err(spng_ctx *ctx, int err) +{ + ctx->state = SPNG_STATE_INVALID; + + return err; +} + +static int encode_err(spng_ctx *ctx, int err) +{ + ctx->state = SPNG_STATE_INVALID; + + return err; +} + +static inline int read_data(spng_ctx *ctx, size_t bytes) +{ + if(ctx == NULL) return SPNG_EINTERNAL; + if(!bytes) return 0; + + if(ctx->streaming && (bytes > SPNG_READ_SIZE)) return SPNG_EINTERNAL; + + int ret = ctx->read_fn(ctx, ctx->stream_user_ptr, ctx->stream_buf, bytes); + + if(ret) + { + if(ret > 0 || ret < SPNG_IO_ERROR) ret = SPNG_IO_ERROR; + + return ret; + } + + ctx->bytes_read += bytes; + if(ctx->bytes_read < bytes) return SPNG_EOVERFLOW; + + return 0; +} + +/* Ensure there is enough space for encoding starting at ctx->write_ptr */ +static int require_bytes(spng_ctx *ctx, size_t bytes) +{ + if(ctx == NULL) return SPNG_EINTERNAL; + + if(ctx->streaming) + { + if(bytes > ctx->stream_buf_size) + { + size_t new_size = ctx->stream_buf_size; + + /* Start at default IDAT size + header + crc */ + if(new_size < (SPNG_WRITE_SIZE + 12)) new_size = SPNG_WRITE_SIZE + 12; + + if(new_size < bytes) new_size = bytes; + + void *temp = spng__realloc(ctx, ctx->stream_buf, new_size); + + if(temp == NULL) return encode_err(ctx, SPNG_EMEM); + + ctx->stream_buf = temp; + ctx->stream_buf_size = bytes; + ctx->write_ptr = ctx->stream_buf; + } + + return 0; + } + + if(!ctx->internal_buffer) return SPNG_ENODST; + + size_t required = ctx->bytes_encoded + bytes; + if(required < bytes) return SPNG_EOVERFLOW; + + if(required > ctx->out_png_size) + { + size_t new_size = ctx->out_png_size; + + /* Start with a size that doesn't require a realloc() 100% of the time */ + if(new_size < (SPNG_WRITE_SIZE * 2)) new_size = SPNG_WRITE_SIZE * 2; + + /* Prefer the next power of two over the requested size */ + while(new_size < required) + { + if(new_size / SIZE_MAX > 2) return encode_err(ctx, SPNG_EOVERFLOW); + + new_size *= 2; + } + + void *temp = spng__realloc(ctx, ctx->out_png, new_size); + + if(temp == NULL) return encode_err(ctx, SPNG_EMEM); + + ctx->out_png = temp; + ctx->out_png_size = new_size; + ctx->write_ptr = ctx->out_png + ctx->bytes_encoded; + } + + return 0; +} + +static int write_data(spng_ctx *ctx, const void *data, size_t bytes) +{ + if(ctx == NULL) return SPNG_EINTERNAL; + if(!bytes) return 0; + + if(ctx->streaming) + { + if(bytes > SPNG_WRITE_SIZE) return SPNG_EINTERNAL; + + int ret = ctx->write_fn(ctx, ctx->stream_user_ptr, (void*)data, bytes); + + if(ret) + { + if(ret > 0 || ret < SPNG_IO_ERROR) ret = SPNG_IO_ERROR; + + return encode_err(ctx, ret); + } + } + else + { + int ret = require_bytes(ctx, bytes); + if(ret) return encode_err(ctx, ret); + + memcpy(ctx->write_ptr, data, bytes); + + ctx->write_ptr += bytes; + } + + ctx->bytes_encoded += bytes; + if(ctx->bytes_encoded < bytes) return SPNG_EOVERFLOW; + + return 0; +} + +static int write_header(spng_ctx *ctx, const uint8_t chunk_type[4], size_t chunk_length, unsigned char **data) +{ + if(ctx == NULL || chunk_type == NULL) return SPNG_EINTERNAL; + if(chunk_length > spng_u32max) return SPNG_EINTERNAL; + + size_t total = chunk_length + 12; + + int ret = require_bytes(ctx, total); + if(ret) return ret; + + uint32_t crc = crc32(0, NULL, 0); + ctx->current_chunk.crc = crc32(crc, chunk_type, 4); + + memcpy(&ctx->current_chunk.type, chunk_type, 4); + ctx->current_chunk.length = (uint32_t)chunk_length; + + if(!data) return SPNG_EINTERNAL; + + if(ctx->streaming) *data = ctx->stream_buf + 8; + else *data = ctx->write_ptr + 8; + + return 0; +} + +static int trim_chunk(spng_ctx *ctx, uint32_t length) +{ + if(length > spng_u32max) return SPNG_EINTERNAL; + if(length > ctx->current_chunk.length) return SPNG_EINTERNAL; + + ctx->current_chunk.length = length; + + return 0; +} + +static int finish_chunk(spng_ctx *ctx) +{ + if(ctx == NULL) return SPNG_EINTERNAL; + + struct spng_chunk *chunk = &ctx->current_chunk; + + unsigned char *header; + unsigned char *chunk_data; + + if(ctx->streaming) + { + chunk_data = ctx->stream_buf + 8; + header = ctx->stream_buf; + } + else + { + chunk_data = ctx->write_ptr + 8; + header = ctx->write_ptr; + } + + write_u32(header, chunk->length); + memcpy(header + 4, chunk->type, 4); + + chunk->crc = crc32(chunk->crc, chunk_data, chunk->length); + + write_u32(chunk_data + chunk->length, chunk->crc); + + if(ctx->streaming) + { + const unsigned char *ptr = ctx->stream_buf; + uint32_t bytes_left = chunk->length + 12; + uint32_t len = 0; + + while(bytes_left) + { + ptr += len; + len = SPNG_WRITE_SIZE; + + if(len > bytes_left) len = bytes_left; + + int ret = write_data(ctx, ptr, len); + if(ret) return ret; + + bytes_left -= len; + } + } + else + { + ctx->bytes_encoded += chunk->length; + if(ctx->bytes_encoded < chunk->length) return SPNG_EOVERFLOW; + + ctx->bytes_encoded += 12; + if(ctx->bytes_encoded < 12) return SPNG_EOVERFLOW; + + ctx->write_ptr += chunk->length + 12; + } + + return 0; +} + +static int write_chunk(spng_ctx *ctx, const uint8_t type[4], const void *data, size_t length) +{ + if(ctx == NULL || type == NULL) return SPNG_EINTERNAL; + if(length && data == NULL) return SPNG_EINTERNAL; + + unsigned char *write_ptr; + + int ret = write_header(ctx, type, length, &write_ptr); + if(ret) return ret; + + if(length) memcpy(write_ptr, data, length); + + return finish_chunk(ctx); +} + +static int write_iend(spng_ctx *ctx) +{ + unsigned char iend_chunk[12] = { 0, 0, 0, 0, 73, 69, 78, 68, 174, 66, 96, 130 }; + return write_data(ctx, iend_chunk, 12); +} + +static int write_unknown_chunks(spng_ctx *ctx, enum spng_location location) +{ + if(!ctx->stored.unknown) return 0; + + const struct spng_unknown_chunk *chunk = ctx->chunk_list; + + uint32_t i; + for(i=0; i < ctx->n_chunks; i++, chunk++) + { + if(chunk->location != location) continue; + + int ret = write_chunk(ctx, chunk->type, chunk->data, chunk->length); + if(ret) return ret; + } + + return 0; +} + +/* Read and check the current chunk's crc, + returns -SPNG_CRC_DISCARD if the chunk should be discarded */ +static inline int read_and_check_crc(spng_ctx *ctx) +{ + if(ctx == NULL) return SPNG_EINTERNAL; + + int ret; + ret = read_data(ctx, 4); + if(ret) return ret; + + ctx->current_chunk.crc = read_u32(ctx->data); + + if(ctx->skip_crc) return 0; + + if(ctx->cur_actual_crc != ctx->current_chunk.crc) + { + if(is_critical_chunk(&ctx->current_chunk)) + { + if(ctx->crc_action_critical == SPNG_CRC_USE) return 0; + } + else + { + if(ctx->crc_action_ancillary == SPNG_CRC_USE) return 0; + if(ctx->crc_action_ancillary == SPNG_CRC_DISCARD) return -SPNG_CRC_DISCARD; + } + + return SPNG_ECHUNK_CRC; + } + + return 0; +} + +/* Read and validate the current chunk's crc and the next chunk header */ +static inline int read_header(spng_ctx *ctx) +{ + if(ctx == NULL) return SPNG_EINTERNAL; + + int ret; + struct spng_chunk chunk = { 0 }; + + ret = read_and_check_crc(ctx); + if(ret) + { + if(ret == -SPNG_CRC_DISCARD) + { + ctx->discard = 1; + } + else return ret; + } + + ret = read_data(ctx, 8); + if(ret) return ret; + + chunk.offset = ctx->bytes_read - 8; + + chunk.length = read_u32(ctx->data); + + memcpy(&chunk.type, ctx->data + 4, 4); + + if(chunk.length > spng_u32max) return SPNG_ECHUNK_STDLEN; + + ctx->cur_chunk_bytes_left = chunk.length; + + if(is_critical_chunk(&chunk) && ctx->crc_action_critical == SPNG_CRC_USE) ctx->skip_crc = 1; + else if(ctx->crc_action_ancillary == SPNG_CRC_USE) ctx->skip_crc = 1; + else ctx->skip_crc = 0; + + if(!ctx->skip_crc) + { + ctx->cur_actual_crc = crc32(0, NULL, 0); + ctx->cur_actual_crc = crc32(ctx->cur_actual_crc, chunk.type, 4); + } + + ctx->current_chunk = chunk; + + return 0; +} + +/* Read chunk bytes and update crc */ +static int read_chunk_bytes(spng_ctx *ctx, uint32_t bytes) +{ + if(ctx == NULL) return SPNG_EINTERNAL; + if(!ctx->cur_chunk_bytes_left || !bytes) return SPNG_EINTERNAL; + if(bytes > ctx->cur_chunk_bytes_left) return SPNG_EINTERNAL; /* XXX: more specific error? */ + + int ret; + + ret = read_data(ctx, bytes); + if(ret) return ret; + + if(!ctx->skip_crc) ctx->cur_actual_crc = crc32(ctx->cur_actual_crc, ctx->data, bytes); + + ctx->cur_chunk_bytes_left -= bytes; + + return ret; +} + +/* read_chunk_bytes() + read_data() with custom output buffer */ +static int read_chunk_bytes2(spng_ctx *ctx, void *out, uint32_t bytes) +{ + if(ctx == NULL) return SPNG_EINTERNAL; + if(!ctx->cur_chunk_bytes_left || !bytes) return SPNG_EINTERNAL; + if(bytes > ctx->cur_chunk_bytes_left) return SPNG_EINTERNAL; /* XXX: more specific error? */ + + int ret; + uint32_t len = bytes; + + if(ctx->streaming && len > SPNG_READ_SIZE) len = SPNG_READ_SIZE; + + while(bytes) + { + if(len > bytes) len = bytes; + + ret = ctx->read_fn(ctx, ctx->stream_user_ptr, out, len); + if(ret) return ret; + + if(!ctx->streaming) memcpy(out, ctx->data, len); + + ctx->bytes_read += len; + if(ctx->bytes_read < len) return SPNG_EOVERFLOW; + + if(!ctx->skip_crc) ctx->cur_actual_crc = crc32(ctx->cur_actual_crc, out, len); + + ctx->cur_chunk_bytes_left -= len; + + out = (char*)out + len; + bytes -= len; + len = SPNG_READ_SIZE; + } + + return 0; +} + +static int discard_chunk_bytes(spng_ctx *ctx, uint32_t bytes) +{ + if(ctx == NULL) return SPNG_EINTERNAL; + if(!bytes) return 0; + + int ret; + + if(ctx->streaming) /* Do small, consecutive reads */ + { + while(bytes) + { + uint32_t len = SPNG_READ_SIZE; + + if(len > bytes) len = bytes; + + ret = read_chunk_bytes(ctx, len); + if(ret) return ret; + + bytes -= len; + } + } + else + { + ret = read_chunk_bytes(ctx, bytes); + if(ret) return ret; + } + + return 0; +} + +static int spng__inflate_init(spng_ctx *ctx, int window_bits) +{ + if(ctx->zstream.state) inflateEnd(&ctx->zstream); + + ctx->inflate = 1; + + ctx->zstream.zalloc = spng__zalloc; + ctx->zstream.zfree = spng__zfree; + ctx->zstream.opaque = ctx; + + if(inflateInit2(&ctx->zstream, window_bits) != Z_OK) return SPNG_EZLIB_INIT; + +#if ZLIB_VERNUM >= 0x1290 && !defined(SPNG_USE_MINIZ) + + int validate = 1; + + if(ctx->flags & SPNG_CTX_IGNORE_ADLER32) validate = 0; + + if(is_critical_chunk(&ctx->current_chunk)) + { + if(ctx->crc_action_critical == SPNG_CRC_USE) validate = 0; + } + else /* ancillary */ + { + if(ctx->crc_action_ancillary == SPNG_CRC_USE) validate = 0; + } + + if(inflateValidate(&ctx->zstream, validate)) return SPNG_EZLIB_INIT; + +#else /* This requires zlib >= 1.2.11 */ + #pragma message ("inflateValidate() not available, SPNG_CTX_IGNORE_ADLER32 will be ignored") +#endif + + return 0; +} + +static int spng__deflate_init(spng_ctx *ctx, struct spng__zlib_options *options) +{ + if(ctx->zstream.state) deflateEnd(&ctx->zstream); + + ctx->deflate = 1; + + z_stream *zstream = &ctx->zstream; + zstream->zalloc = spng__zalloc; + zstream->zfree = spng__zfree; + zstream->opaque = ctx; + zstream->data_type = options->data_type; + + int ret = deflateInit2(zstream, options->compression_level, Z_DEFLATED, options->window_bits, options->mem_level, options->strategy); + + if(ret != Z_OK) return SPNG_EZLIB_INIT; + + return 0; +} + +/* Inflate a zlib stream starting with start_buf if non-NULL, + continuing from the datastream till an end marker, + allocating and writing the inflated stream to *out, + leaving "extra" bytes at the end, final buffer length is *len. + + Takes into account the chunk size and cache limits. +*/ +static int spng__inflate_stream(spng_ctx *ctx, char **out, size_t *len, size_t extra, const void *start_buf, size_t start_len) +{ + int ret = spng__inflate_init(ctx, 15); + if(ret) return ret; + + size_t max = ctx->chunk_cache_limit - ctx->chunk_cache_usage; + + if(ctx->max_chunk_size < max) max = ctx->max_chunk_size; + + if(extra > max) return SPNG_ECHUNK_LIMITS; + max -= extra; + + uint32_t read_size; + size_t size = 8 * 1024; + void *t, *buf = spng__malloc(ctx, size); + + if(buf == NULL) return SPNG_EMEM; + + z_stream *stream = &ctx->zstream; + + if(start_buf != NULL && start_len) + { + stream->avail_in = (uInt)start_len; + stream->next_in = start_buf; + } + else + { + stream->avail_in = 0; + stream->next_in = NULL; + } + + stream->avail_out = (uInt)size; + stream->next_out = buf; + + while(ret != Z_STREAM_END) + { + ret = inflate(stream, Z_NO_FLUSH); + + if(ret == Z_STREAM_END) break; + + if(ret != Z_OK && ret != Z_BUF_ERROR) + { + ret = SPNG_EZLIB; + goto err; + } + + if(!stream->avail_out) /* Resize buffer */ + { + /* overflow or reached chunk/cache limit */ + if( (2 > SIZE_MAX / size) || (size > max / 2) ) + { + ret = SPNG_ECHUNK_LIMITS; + goto err; + } + + size *= 2; + + t = spng__realloc(ctx, buf, size); + if(t == NULL) goto mem; + + buf = t; + + stream->avail_out = (uInt)size / 2; + stream->next_out = (unsigned char*)buf + size / 2; + } + else if(!stream->avail_in) /* Read more chunk bytes */ + { + read_size = ctx->cur_chunk_bytes_left; + if(ctx->streaming && read_size > SPNG_READ_SIZE) read_size = SPNG_READ_SIZE; + + ret = read_chunk_bytes(ctx, read_size); + + if(ret) + { + if(!read_size) ret = SPNG_EZLIB; + + goto err; + } + + stream->avail_in = read_size; + stream->next_in = ctx->data; + } + } + + size = stream->total_out; + + if(!size) + { + ret = SPNG_EZLIB; + goto err; + } + + size += extra; + if(size < extra) goto mem; + + t = spng__realloc(ctx, buf, size); + if(t == NULL) goto mem; + + buf = t; + + (void)increase_cache_usage(ctx, size, 0); + + *out = buf; + *len = size; + + return 0; + +mem: + ret = SPNG_EMEM; +err: + spng__free(ctx, buf); + return ret; +} + +/* Read at least one byte from the IDAT stream */ +static int read_idat_bytes(spng_ctx *ctx, uint32_t *bytes_read) +{ + if(ctx == NULL || bytes_read == NULL) return SPNG_EINTERNAL; + if(memcmp(ctx->current_chunk.type, type_idat, 4)) return SPNG_EIDAT_TOO_SHORT; + + int ret; + uint32_t len; + + while(!ctx->cur_chunk_bytes_left) + { + ret = read_header(ctx); + if(ret) return ret; + + if(memcmp(ctx->current_chunk.type, type_idat, 4)) return SPNG_EIDAT_TOO_SHORT; + } + + if(ctx->streaming) + {/* TODO: estimate bytes to read for progressive reads */ + len = SPNG_READ_SIZE; + if(len > ctx->cur_chunk_bytes_left) len = ctx->cur_chunk_bytes_left; + } + else len = ctx->current_chunk.length; + + ret = read_chunk_bytes(ctx, len); + + *bytes_read = len; + + return ret; +} + +static int read_scanline_bytes(spng_ctx *ctx, unsigned char *dest, size_t len) +{ + if(ctx == NULL || dest == NULL) return SPNG_EINTERNAL; + + int ret = Z_OK; + uint32_t bytes_read; + + z_stream *zstream = &ctx->zstream; + + zstream->avail_out = (uInt)len; + zstream->next_out = dest; + + while(zstream->avail_out != 0) + { + ret = inflate(zstream, Z_NO_FLUSH); + + if(ret == Z_OK) continue; + + if(ret == Z_STREAM_END) /* Reached an end-marker */ + { + if(zstream->avail_out != 0) return SPNG_EIDAT_TOO_SHORT; + } + else if(ret == Z_BUF_ERROR) /* Read more IDAT bytes */ + { + ret = read_idat_bytes(ctx, &bytes_read); + if(ret) return ret; + + zstream->avail_in = bytes_read; + zstream->next_in = ctx->data; + } + else return SPNG_EIDAT_STREAM; + } + + return 0; +} + +static uint8_t paeth(uint8_t a, uint8_t b, uint8_t c) +{ + int16_t p = a + b - c; + int16_t pa = abs(p - a); + int16_t pb = abs(p - b); + int16_t pc = abs(p - c); + + if(pa <= pb && pa <= pc) return a; + else if(pb <= pc) return b; + + return c; +} + +SPNG_TARGET_CLONES("default,avx2") +static void defilter_up(size_t bytes, unsigned char *row, const unsigned char *prev) +{ + size_t i; + for(i=0; i < bytes; i++) + { + row[i] += prev[i]; + } +} + +/* Defilter *scanline in-place. + *prev_scanline and *scanline should point to the first pixel, + scanline_width is the width of the scanline including the filter byte. +*/ +static int defilter_scanline(const unsigned char *prev_scanline, unsigned char *scanline, + size_t scanline_width, unsigned bytes_per_pixel, unsigned filter) +{ + if(prev_scanline == NULL || scanline == NULL || !scanline_width) return SPNG_EINTERNAL; + + size_t i; + scanline_width--; + + if(filter == 0) return 0; + +#ifndef SPNG_DISABLE_OPT + if(filter == SPNG_FILTER_UP) goto no_opt; + + if(bytes_per_pixel == 4) + { + if(filter == SPNG_FILTER_SUB) + defilter_sub4(scanline_width, scanline); + else if(filter == SPNG_FILTER_AVERAGE) + defilter_avg4(scanline_width, scanline, prev_scanline); + else if(filter == SPNG_FILTER_PAETH) + defilter_paeth4(scanline_width, scanline, prev_scanline); + else return SPNG_EFILTER; + + return 0; + } + else if(bytes_per_pixel == 3) + { + if(filter == SPNG_FILTER_SUB) + defilter_sub3(scanline_width, scanline); + else if(filter == SPNG_FILTER_AVERAGE) + defilter_avg3(scanline_width, scanline, prev_scanline); + else if(filter == SPNG_FILTER_PAETH) + defilter_paeth3(scanline_width, scanline, prev_scanline); + else return SPNG_EFILTER; + + return 0; + } +no_opt: +#endif + + if(filter == SPNG_FILTER_UP) + { + defilter_up(scanline_width, scanline, prev_scanline); + return 0; + } + + for(i=0; i < scanline_width; i++) + { + uint8_t x, a, b, c; + + if(i >= bytes_per_pixel) + { + a = scanline[i - bytes_per_pixel]; + b = prev_scanline[i]; + c = prev_scanline[i - bytes_per_pixel]; + } + else /* First pixel in row */ + { + a = 0; + b = prev_scanline[i]; + c = 0; + } + + x = scanline[i]; + + switch(filter) + { + case SPNG_FILTER_SUB: + { + x = x + a; + break; + } + case SPNG_FILTER_AVERAGE: + { + uint16_t avg = (a + b) / 2; + x = x + avg; + break; + } + case SPNG_FILTER_PAETH: + { + x = x + paeth(a,b,c); + break; + } + } + + scanline[i] = x; + } + + return 0; +} + +static int filter_scanline(unsigned char *filtered, const unsigned char *prev_scanline, const unsigned char *scanline, + size_t scanline_width, unsigned bytes_per_pixel, const unsigned filter) +{ + if(prev_scanline == NULL || scanline == NULL || scanline_width <= 1) return SPNG_EINTERNAL; + + if(filter > 4) return SPNG_EFILTER; + if(filter == 0) return 0; + + scanline_width--; + + uint32_t i; + for(i=0; i < scanline_width; i++) + { + uint8_t x, a, b, c; + + if(i >= bytes_per_pixel) + { + a = scanline[i - bytes_per_pixel]; + b = prev_scanline[i]; + c = prev_scanline[i - bytes_per_pixel]; + } + else /* first pixel in row */ + { + a = 0; + b = prev_scanline[i]; + c = 0; + } + + x = scanline[i]; + + switch(filter) + { + case SPNG_FILTER_SUB: + { + x = x - a; + break; + } + case SPNG_FILTER_UP: + { + x = x - b; + break; + } + case SPNG_FILTER_AVERAGE: + { + uint16_t avg = (a + b) / 2; + x = x - avg; + break; + } + case SPNG_FILTER_PAETH: + { + x = x - paeth(a,b,c); + break; + } + } + + filtered[i] = x; + } + + return 0; +} + +static int32_t filter_sum(const unsigned char *prev_scanline, const unsigned char *scanline, + size_t size, unsigned bytes_per_pixel, const unsigned filter) +{ + /* prevent potential over/underflow, bails out at a width of ~8M pixels for RGBA8 */ + if(size > (INT32_MAX / 128)) return INT32_MAX; + + uint32_t i; + int32_t sum = 0; + uint8_t x, a, b, c; + + for(i=0; i < size; i++) + { + if(i >= bytes_per_pixel) + { + a = scanline[i - bytes_per_pixel]; + b = prev_scanline[i]; + c = prev_scanline[i - bytes_per_pixel]; + } + else /* first pixel in row */ + { + a = 0; + b = prev_scanline[i]; + c = 0; + } + + x = scanline[i]; + + switch(filter) + { + case SPNG_FILTER_NONE: + { + break; + } + case SPNG_FILTER_SUB: + { + x = x - a; + break; + } + case SPNG_FILTER_UP: + { + x = x - b; + break; + } + case SPNG_FILTER_AVERAGE: + { + uint16_t avg = (a + b) / 2; + x = x - avg; + break; + } + case SPNG_FILTER_PAETH: + { + x = x - paeth(a,b,c); + break; + } + } + + sum += 128 - abs((int)x - 128); + } + + return sum; +} + +static unsigned get_best_filter(const unsigned char *prev_scanline, const unsigned char *scanline, + size_t scanline_width, unsigned bytes_per_pixel, const int choices) +{ + if(!choices) return SPNG_FILTER_NONE; + + scanline_width--; + + int i; + unsigned int best_filter = 0; + enum spng_filter_choice flag; + int32_t sum, best_score = INT32_MAX; + int32_t filter_scores[5] = { INT32_MAX, INT32_MAX, INT32_MAX, INT32_MAX, INT32_MAX }; + + if( !(choices & (choices - 1)) ) + {/* only one choice/bit is set */ + for(i=0; i < 5; i++) + { + if(choices == 1 << (i + 3)) return i; + } + } + + for(i=0; i < 5; i++) + { + flag = 1 << (i + 3); + + if(choices & flag) sum = filter_sum(prev_scanline, scanline, scanline_width, bytes_per_pixel, i); + else continue; + + filter_scores[i] = abs(sum); + + if(filter_scores[i] < best_score) + { + best_score = filter_scores[i]; + best_filter = i; + } + } + + return best_filter; +} + +/* Scale "sbits" significant bits in "sample" from "bit_depth" to "target" + + "bit_depth" must be a valid PNG depth + "sbits" must be less than or equal to "bit_depth" + "target" must be between 1 and 16 +*/ +static uint16_t sample_to_target(uint16_t sample, unsigned bit_depth, unsigned sbits, unsigned target) +{ + if(bit_depth == sbits) + { + if(target == sbits) return sample; /* No scaling */ + }/* bit_depth > sbits */ + else sample = sample >> (bit_depth - sbits); /* Shift significant bits to bottom */ + + /* Downscale */ + if(target < sbits) return sample >> (sbits - target); + + /* Upscale using left bit replication */ + int8_t shift_amount = target - sbits; + uint16_t sample_bits = sample; + sample = 0; + + while(shift_amount >= 0) + { + sample = sample | (sample_bits << shift_amount); + shift_amount -= sbits; + } + + int8_t partial = shift_amount + (int8_t)sbits; + + if(partial != 0) sample = sample | (sample_bits >> abs(shift_amount)); + + return sample; +} + +static inline void gamma_correct_row(unsigned char *row, uint32_t pixels, int fmt, const uint16_t *gamma_lut) +{ + uint32_t i; + + if(fmt == SPNG_FMT_RGBA8) + { + unsigned char *px; + for(i=0; i < pixels; i++) + { + px = row + i * 4; + + px[0] = gamma_lut[px[0]]; + px[1] = gamma_lut[px[1]]; + px[2] = gamma_lut[px[2]]; + } + } + else if(fmt == SPNG_FMT_RGBA16) + { + for(i=0; i < pixels; i++) + { + uint16_t px[4]; + memcpy(px, row + i * 8, 8); + + px[0] = gamma_lut[px[0]]; + px[1] = gamma_lut[px[1]]; + px[2] = gamma_lut[px[2]]; + + memcpy(row + i * 8, px, 8); + } + } + else if(fmt == SPNG_FMT_RGB8) + { + unsigned char *px; + for(i=0; i < pixels; i++) + { + px = row + i * 3; + + px[0] = gamma_lut[px[0]]; + px[1] = gamma_lut[px[1]]; + px[2] = gamma_lut[px[2]]; + } + } +} + +/* Apply transparency to output row */ +static inline void trns_row(unsigned char *row, + const unsigned char *scanline, + const unsigned char *trns, + unsigned scanline_stride, + struct spng_ihdr *ihdr, + uint32_t pixels, + int fmt) +{ + uint32_t i; + unsigned row_stride; + unsigned depth = ihdr->bit_depth; + + if(fmt == SPNG_FMT_RGBA8) + { + if(ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE) return; /* already applied in the decoding loop */ + + row_stride = 4; + for(i=0; i < pixels; i++, scanline+=scanline_stride, row+=row_stride) + { + if(!memcmp(scanline, trns, scanline_stride)) row[3] = 0; + } + } + else if(fmt == SPNG_FMT_RGBA16) + { + if(ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE) return; /* already applied in the decoding loop */ + + row_stride = 8; + for(i=0; i < pixels; i++, scanline+=scanline_stride, row+=row_stride) + { + if(!memcmp(scanline, trns, scanline_stride)) memset(row + 6, 0, 2); + } + } + else if(fmt == SPNG_FMT_GA8) + { + row_stride = 2; + + if(depth == 16) + { + for(i=0; i < pixels; i++, scanline+=scanline_stride, row+=row_stride) + { + if(!memcmp(scanline, trns, scanline_stride)) memset(row + 1, 0, 1); + } + } + else /* depth <= 8 */ + { + struct spng__iter iter = spng__iter_init(depth, scanline); + + for(i=0; i < pixels; i++, row+=row_stride) + { + if(trns[0] == get_sample(&iter)) row[1] = 0; + } + } + } + else if(fmt == SPNG_FMT_GA16) + { + row_stride = 4; + + if(depth == 16) + { + for(i=0; i< pixels; i++, scanline+=scanline_stride, row+=row_stride) + { + if(!memcmp(scanline, trns, 2)) memset(row + 2, 0, 2); + } + } + else + { + struct spng__iter iter = spng__iter_init(depth, scanline); + + for(i=0; i< pixels; i++, row+=row_stride) + { + if(trns[0] == get_sample(&iter)) memset(row + 2, 0, 2); + } + } + } + else return; +} + +static inline void scale_row(unsigned char *row, uint32_t pixels, int fmt, unsigned depth, const struct spng_sbit *sbit) +{ + uint32_t i; + + if(fmt == SPNG_FMT_RGBA8) + { + unsigned char px[4]; + for(i=0; i < pixels; i++) + { + memcpy(px, row + i * 4, 4); + + px[0] = sample_to_target(px[0], depth, sbit->red_bits, 8); + px[1] = sample_to_target(px[1], depth, sbit->green_bits, 8); + px[2] = sample_to_target(px[2], depth, sbit->blue_bits, 8); + px[3] = sample_to_target(px[3], depth, sbit->alpha_bits, 8); + + memcpy(row + i * 4, px, 4); + } + } + else if(fmt == SPNG_FMT_RGBA16) + { + uint16_t px[4]; + for(i=0; i < pixels; i++) + { + memcpy(px, row + i * 8, 8); + + px[0] = sample_to_target(px[0], depth, sbit->red_bits, 16); + px[1] = sample_to_target(px[1], depth, sbit->green_bits, 16); + px[2] = sample_to_target(px[2], depth, sbit->blue_bits, 16); + px[3] = sample_to_target(px[3], depth, sbit->alpha_bits, 16); + + memcpy(row + i * 8, px, 8); + } + } + else if(fmt == SPNG_FMT_RGB8) + { + unsigned char px[4]; + for(i=0; i < pixels; i++) + { + memcpy(px, row + i * 3, 3); + + px[0] = sample_to_target(px[0], depth, sbit->red_bits, 8); + px[1] = sample_to_target(px[1], depth, sbit->green_bits, 8); + px[2] = sample_to_target(px[2], depth, sbit->blue_bits, 8); + + memcpy(row + i * 3, px, 3); + } + } + else if(fmt == SPNG_FMT_G8) + { + for(i=0; i < pixels; i++) + { + row[i] = sample_to_target(row[i], depth, sbit->grayscale_bits, 8); + } + } + else if(fmt == SPNG_FMT_GA8) + { + for(i=0; i < pixels; i++) + { + row[i*2] = sample_to_target(row[i*2], depth, sbit->grayscale_bits, 8); + } + } +} + +/* Expand to *row using 8-bit palette indices from *scanline */ +static void expand_row(unsigned char *row, + const unsigned char *scanline, + const union spng__decode_plte *decode_plte, + uint32_t width, + int fmt) +{ + uint32_t i = 0; + unsigned char *px; + unsigned char entry; + const struct spng_plte_entry *plte = decode_plte->rgba; + +#if defined(SPNG_ARM) + if(fmt == SPNG_FMT_RGBA8) i = expand_palette_rgba8_neon(row, scanline, decode_plte->raw, width); + else if(fmt == SPNG_FMT_RGB8) + { + i = expand_palette_rgb8_neon(row, scanline, decode_plte->raw, width); + + for(; i < width; i++) + {/* In this case the LUT is 3 bytes packed */ + px = row + i * 3; + entry = scanline[i]; + px[0] = decode_plte->raw[entry * 3 + 0]; + px[1] = decode_plte->raw[entry * 3 + 1]; + px[2] = decode_plte->raw[entry * 3 + 2]; + } + return; + } +#endif + + if(fmt == SPNG_FMT_RGBA8) + { + for(; i < width; i++) + { + px = row + i * 4; + entry = scanline[i]; + px[0] = plte[entry].red; + px[1] = plte[entry].green; + px[2] = plte[entry].blue; + px[3] = plte[entry].alpha; + } + } + else if(fmt == SPNG_FMT_RGB8) + { + for(; i < width; i++) + { + px = row + i * 3; + entry = scanline[i]; + px[0] = plte[entry].red; + px[1] = plte[entry].green; + px[2] = plte[entry].blue; + } + } +} + +/* Unpack 1/2/4/8-bit samples to G8/GA8/GA16 or G16 -> GA16 */ +static void unpack_scanline(unsigned char *out, const unsigned char *scanline, uint32_t width, unsigned bit_depth, int fmt) +{ + struct spng__iter iter = spng__iter_init(bit_depth, scanline); + uint32_t i; + uint16_t sample, alpha = 65535; + + + if(fmt == SPNG_FMT_GA8) goto ga8; + else if(fmt == SPNG_FMT_GA16) goto ga16; + + /* 1/2/4-bit -> 8-bit */ + for(i=0; i < width; i++) out[i] = get_sample(&iter); + + return; + +ga8: + /* 1/2/4/8-bit -> GA8 */ + for(i=0; i < width; i++) + { + out[i*2] = get_sample(&iter); + out[i*2 + 1] = 255; + } + + return; + +ga16: + + /* 16 -> GA16 */ + if(bit_depth == 16) + { + for(i=0; i < width; i++) + { + memcpy(out + i * 4, scanline + i * 2, 2); + memcpy(out + i * 4 + 2, &alpha, 2); + } + return; + } + + /* 1/2/4/8-bit -> GA16 */ + for(i=0; i < width; i++) + { + sample = get_sample(&iter); + memcpy(out + i * 4, &sample, 2); + memcpy(out + i * 4 + 2, &alpha, 2); + } +} + +static int check_ihdr(const struct spng_ihdr *ihdr, uint32_t max_width, uint32_t max_height) +{ + if(ihdr->width > spng_u32max || !ihdr->width) return SPNG_EWIDTH; + if(ihdr->height > spng_u32max || !ihdr->height) return SPNG_EHEIGHT; + + if(ihdr->width > max_width) return SPNG_EUSER_WIDTH; + if(ihdr->height > max_height) return SPNG_EUSER_HEIGHT; + + switch(ihdr->color_type) + { + case SPNG_COLOR_TYPE_GRAYSCALE: + { + if( !(ihdr->bit_depth == 1 || ihdr->bit_depth == 2 || + ihdr->bit_depth == 4 || ihdr->bit_depth == 8 || + ihdr->bit_depth == 16) ) + return SPNG_EBIT_DEPTH; + + break; + } + case SPNG_COLOR_TYPE_TRUECOLOR: + case SPNG_COLOR_TYPE_GRAYSCALE_ALPHA: + case SPNG_COLOR_TYPE_TRUECOLOR_ALPHA: + { + if( !(ihdr->bit_depth == 8 || ihdr->bit_depth == 16) ) + return SPNG_EBIT_DEPTH; + + break; + } + case SPNG_COLOR_TYPE_INDEXED: + { + if( !(ihdr->bit_depth == 1 || ihdr->bit_depth == 2 || + ihdr->bit_depth == 4 || ihdr->bit_depth == 8) ) + return SPNG_EBIT_DEPTH; + + break; + } + default: return SPNG_ECOLOR_TYPE; + } + + if(ihdr->compression_method) return SPNG_ECOMPRESSION_METHOD; + if(ihdr->filter_method) return SPNG_EFILTER_METHOD; + + if(ihdr->interlace_method > 1) return SPNG_EINTERLACE_METHOD; + + return 0; +} + +static int check_plte(const struct spng_plte *plte, const struct spng_ihdr *ihdr) +{ + if(plte == NULL || ihdr == NULL) return 1; + + if(plte->n_entries == 0) return 1; + if(plte->n_entries > 256) return 1; + + if(ihdr->color_type == SPNG_COLOR_TYPE_INDEXED) + { + if(plte->n_entries > (1U << ihdr->bit_depth)) return 1; + } + + return 0; +} + +static int check_sbit(const struct spng_sbit *sbit, const struct spng_ihdr *ihdr) +{ + if(sbit == NULL || ihdr == NULL) return 1; + + if(ihdr->color_type == 0) + { + if(sbit->grayscale_bits == 0) return SPNG_ESBIT; + if(sbit->grayscale_bits > ihdr->bit_depth) return SPNG_ESBIT; + } + else if(ihdr->color_type == 2 || ihdr->color_type == 3) + { + if(sbit->red_bits == 0) return SPNG_ESBIT; + if(sbit->green_bits == 0) return SPNG_ESBIT; + if(sbit->blue_bits == 0) return SPNG_ESBIT; + + uint8_t bit_depth; + if(ihdr->color_type == 3) bit_depth = 8; + else bit_depth = ihdr->bit_depth; + + if(sbit->red_bits > bit_depth) return SPNG_ESBIT; + if(sbit->green_bits > bit_depth) return SPNG_ESBIT; + if(sbit->blue_bits > bit_depth) return SPNG_ESBIT; + } + else if(ihdr->color_type == 4) + { + if(sbit->grayscale_bits == 0) return SPNG_ESBIT; + if(sbit->alpha_bits == 0) return SPNG_ESBIT; + + if(sbit->grayscale_bits > ihdr->bit_depth) return SPNG_ESBIT; + if(sbit->alpha_bits > ihdr->bit_depth) return SPNG_ESBIT; + } + else if(ihdr->color_type == 6) + { + if(sbit->red_bits == 0) return SPNG_ESBIT; + if(sbit->green_bits == 0) return SPNG_ESBIT; + if(sbit->blue_bits == 0) return SPNG_ESBIT; + if(sbit->alpha_bits == 0) return SPNG_ESBIT; + + if(sbit->red_bits > ihdr->bit_depth) return SPNG_ESBIT; + if(sbit->green_bits > ihdr->bit_depth) return SPNG_ESBIT; + if(sbit->blue_bits > ihdr->bit_depth) return SPNG_ESBIT; + if(sbit->alpha_bits > ihdr->bit_depth) return SPNG_ESBIT; + } + + return 0; +} + +static int check_chrm_int(const struct spng_chrm_int *chrm_int) +{ + if(chrm_int == NULL) return 1; + + if(chrm_int->white_point_x > spng_u32max || + chrm_int->white_point_y > spng_u32max || + chrm_int->red_x > spng_u32max || + chrm_int->red_y > spng_u32max || + chrm_int->green_x > spng_u32max || + chrm_int->green_y > spng_u32max || + chrm_int->blue_x > spng_u32max || + chrm_int->blue_y > spng_u32max) return SPNG_ECHRM; + + return 0; +} + +static int check_phys(const struct spng_phys *phys) +{ + if(phys == NULL) return 1; + + if(phys->unit_specifier > 1) return SPNG_EPHYS; + + if(phys->ppu_x > spng_u32max) return SPNG_EPHYS; + if(phys->ppu_y > spng_u32max) return SPNG_EPHYS; + + return 0; +} + +static int check_time(const struct spng_time *time) +{ + if(time == NULL) return 1; + + if(time->month == 0 || time->month > 12) return 1; + if(time->day == 0 || time->day > 31) return 1; + if(time->hour > 23) return 1; + if(time->minute > 59) return 1; + if(time->second > 60) return 1; + + return 0; +} + +static int check_offs(const struct spng_offs *offs) +{ + if(offs == NULL) return 1; + + if(offs->unit_specifier > 1) return 1; + + return 0; +} + +static int check_exif(const struct spng_exif *exif) +{ + if(exif == NULL) return 1; + if(exif->data == NULL) return 1; + + if(exif->length < 4) return SPNG_ECHUNK_SIZE; + if(exif->length > spng_u32max) return SPNG_ECHUNK_STDLEN; + + const uint8_t exif_le[4] = { 73, 73, 42, 0 }; + const uint8_t exif_be[4] = { 77, 77, 0, 42 }; + + if(memcmp(exif->data, exif_le, 4) && memcmp(exif->data, exif_be, 4)) return 1; + + return 0; +} + +/* Validate PNG keyword */ +static int check_png_keyword(const char *str) +{ + if(str == NULL) return 1; + size_t len = strlen(str); + const char *end = str + len; + + if(!len) return 1; + if(len > 79) return 1; + if(str[0] == ' ') return 1; /* Leading space */ + if(end[-1] == ' ') return 1; /* Trailing space */ + if(strstr(str, " ") != NULL) return 1; /* Consecutive spaces */ + + uint8_t c; + while(str != end) + { + memcpy(&c, str, 1); + + if( (c >= 32 && c <= 126) || (c >= 161) ) str++; + else return 1; /* Invalid character */ + } + + return 0; +} + +/* Validate PNG text *str up to 'len' bytes */ +static int check_png_text(const char *str, size_t len) +{/* XXX: are consecutive newlines permitted? */ + if(str == NULL || len == 0) return 1; + + uint8_t c; + size_t i = 0; + while(i < len) + { + memcpy(&c, str + i, 1); + + if( (c >= 32 && c <= 126) || (c >= 161) || c == 10) i++; + else return 1; /* Invalid character */ + } + + return 0; +} + +/* Returns non-zero for standard chunks which are stored without allocating memory */ +static int is_small_chunk(uint8_t type[4]) +{ + if(!memcmp(type, type_plte, 4)) return 1; + else if(!memcmp(type, type_chrm, 4)) return 1; + else if(!memcmp(type, type_gama, 4)) return 1; + else if(!memcmp(type, type_sbit, 4)) return 1; + else if(!memcmp(type, type_srgb, 4)) return 1; + else if(!memcmp(type, type_bkgd, 4)) return 1; + else if(!memcmp(type, type_trns, 4)) return 1; + else if(!memcmp(type, type_hist, 4)) return 1; + else if(!memcmp(type, type_phys, 4)) return 1; + else if(!memcmp(type, type_time, 4)) return 1; + else if(!memcmp(type, type_offs, 4)) return 1; + else return 0; +} + +static int read_ihdr(spng_ctx *ctx) +{ + int ret; + struct spng_chunk *chunk = &ctx->current_chunk; + const unsigned char *data; + + chunk->offset = 8; + chunk->length = 13; + size_t sizeof_sig_ihdr = 29; + + ret = read_data(ctx, sizeof_sig_ihdr); + if(ret) return ret; + + data = ctx->data; + + if(memcmp(data, spng_signature, sizeof(spng_signature))) return SPNG_ESIGNATURE; + + chunk->length = read_u32(data + 8); + memcpy(&chunk->type, data + 12, 4); + + if(chunk->length != 13) return SPNG_EIHDR_SIZE; + if(memcmp(chunk->type, type_ihdr, 4)) return SPNG_ENOIHDR; + + ctx->cur_actual_crc = crc32(0, NULL, 0); + ctx->cur_actual_crc = crc32(ctx->cur_actual_crc, data + 12, 17); + + ctx->ihdr.width = read_u32(data + 16); + ctx->ihdr.height = read_u32(data + 20); + ctx->ihdr.bit_depth = data[24]; + ctx->ihdr.color_type = data[25]; + ctx->ihdr.compression_method = data[26]; + ctx->ihdr.filter_method = data[27]; + ctx->ihdr.interlace_method = data[28]; + + ret = check_ihdr(&ctx->ihdr, ctx->max_width, ctx->max_height); + if(ret) return ret; + + ctx->file.ihdr = 1; + ctx->stored.ihdr = 1; + + if(ctx->ihdr.bit_depth < 8) ctx->bytes_per_pixel = 1; + else ctx->bytes_per_pixel = num_channels(&ctx->ihdr) * (ctx->ihdr.bit_depth / 8); + + ret = calculate_subimages(ctx); + if(ret) return ret; + + return 0; +} + +static void splt_undo(spng_ctx *ctx) +{ + struct spng_splt *splt = &ctx->splt_list[ctx->n_splt - 1]; + + spng__free(ctx, splt->entries); + + decrease_cache_usage(ctx, sizeof(struct spng_splt)); + decrease_cache_usage(ctx, splt->n_entries * sizeof(struct spng_splt_entry)); + + splt->entries = NULL; + + ctx->n_splt--; +} + +static void text_undo(spng_ctx *ctx) +{ + struct spng_text2 *text = &ctx->text_list[ctx->n_text - 1]; + + spng__free(ctx, text->keyword); + if(text->compression_flag) spng__free(ctx, text->text); + + decrease_cache_usage(ctx, text->cache_usage); + decrease_cache_usage(ctx, sizeof(struct spng_text2)); + + text->keyword = NULL; + text->text = NULL; + + ctx->n_text--; +} + +static void chunk_undo(spng_ctx *ctx) +{ + struct spng_unknown_chunk *chunk = &ctx->chunk_list[ctx->n_chunks - 1]; + + spng__free(ctx, chunk->data); + + decrease_cache_usage(ctx, chunk->length); + decrease_cache_usage(ctx, sizeof(struct spng_unknown_chunk)); + + chunk->data = NULL; + + ctx->n_chunks--; +} + +static int read_non_idat_chunks(spng_ctx *ctx) +{ + int ret; + struct spng_chunk chunk; + const unsigned char *data; + + ctx->discard = 0; + ctx->undo = NULL; + ctx->prev_stored = ctx->stored; + + while( !(ret = read_header(ctx))) + { + if(ctx->discard) + { + if(ctx->undo) ctx->undo(ctx); + + ctx->stored = ctx->prev_stored; + } + + ctx->discard = 0; + ctx->undo = NULL; + + ctx->prev_stored = ctx->stored; + chunk = ctx->current_chunk; + + if(!memcmp(chunk.type, type_idat, 4)) + { + if(ctx->state < SPNG_STATE_FIRST_IDAT) + { + if(ctx->ihdr.color_type == 3 && !ctx->stored.plte) return SPNG_ENOPLTE; + + ctx->first_idat = chunk; + return 0; + } + + if(ctx->prev_was_idat) + { + /* Ignore extra IDAT's */ + ret = discard_chunk_bytes(ctx, chunk.length); + if(ret) return ret; + + continue; + } + else return SPNG_ECHUNK_POS; /* IDAT chunk not at the end of the IDAT sequence */ + } + + ctx->prev_was_idat = 0; + + if(is_small_chunk(chunk.type)) + { + /* None of the known chunks can be zero length */ + if(!chunk.length) return SPNG_ECHUNK_SIZE; + + /* The largest of these chunks is PLTE with 256 entries */ + ret = read_chunk_bytes(ctx, chunk.length > 768 ? 768 : chunk.length); + if(ret) return ret; + } + + data = ctx->data; + + if(is_critical_chunk(&chunk)) + { + if(!memcmp(chunk.type, type_plte, 4)) + { + if(ctx->file.trns || ctx->file.hist || ctx->file.bkgd) return SPNG_ECHUNK_POS; + if(chunk.length % 3 != 0) return SPNG_ECHUNK_SIZE; + + ctx->plte.n_entries = chunk.length / 3; + + if(check_plte(&ctx->plte, &ctx->ihdr)) return SPNG_ECHUNK_SIZE; /* XXX: EPLTE? */ + + size_t i; + for(i=0; i < ctx->plte.n_entries; i++) + { + ctx->plte.entries[i].red = data[i * 3]; + ctx->plte.entries[i].green = data[i * 3 + 1]; + ctx->plte.entries[i].blue = data[i * 3 + 2]; + } + + ctx->file.plte = 1; + ctx->stored.plte = 1; + } + else if(!memcmp(chunk.type, type_iend, 4)) + { + if(ctx->state == SPNG_STATE_AFTER_IDAT) + { + if(chunk.length) return SPNG_ECHUNK_SIZE; + + ret = read_and_check_crc(ctx); + if(ret == -SPNG_CRC_DISCARD) ret = 0; + + return ret; + } + else return SPNG_ECHUNK_POS; + } + else if(!memcmp(chunk.type, type_ihdr, 4)) return SPNG_ECHUNK_POS; + else return SPNG_ECHUNK_UNKNOWN_CRITICAL; + } + else if(!memcmp(chunk.type, type_chrm, 4)) /* Ancillary chunks */ + { + if(ctx->file.plte) return SPNG_ECHUNK_POS; + if(ctx->state == SPNG_STATE_AFTER_IDAT) return SPNG_ECHUNK_POS; + if(ctx->file.chrm) return SPNG_EDUP_CHRM; + + if(chunk.length != 32) return SPNG_ECHUNK_SIZE; + + ctx->chrm_int.white_point_x = read_u32(data); + ctx->chrm_int.white_point_y = read_u32(data + 4); + ctx->chrm_int.red_x = read_u32(data + 8); + ctx->chrm_int.red_y = read_u32(data + 12); + ctx->chrm_int.green_x = read_u32(data + 16); + ctx->chrm_int.green_y = read_u32(data + 20); + ctx->chrm_int.blue_x = read_u32(data + 24); + ctx->chrm_int.blue_y = read_u32(data + 28); + + if(check_chrm_int(&ctx->chrm_int)) return SPNG_ECHRM; + + ctx->file.chrm = 1; + ctx->stored.chrm = 1; + } + else if(!memcmp(chunk.type, type_gama, 4)) + { + if(ctx->file.plte) return SPNG_ECHUNK_POS; + if(ctx->state == SPNG_STATE_AFTER_IDAT) return SPNG_ECHUNK_POS; + if(ctx->file.gama) return SPNG_EDUP_GAMA; + + if(chunk.length != 4) return SPNG_ECHUNK_SIZE; + + ctx->gama = read_u32(data); + + if(!ctx->gama) return SPNG_EGAMA; + if(ctx->gama > spng_u32max) return SPNG_EGAMA; + + ctx->file.gama = 1; + ctx->stored.gama = 1; + } + else if(!memcmp(chunk.type, type_sbit, 4)) + { + if(ctx->file.plte) return SPNG_ECHUNK_POS; + if(ctx->state == SPNG_STATE_AFTER_IDAT) return SPNG_ECHUNK_POS; + if(ctx->file.sbit) return SPNG_EDUP_SBIT; + + if(ctx->ihdr.color_type == 0) + { + if(chunk.length != 1) return SPNG_ECHUNK_SIZE; + + ctx->sbit.grayscale_bits = data[0]; + } + else if(ctx->ihdr.color_type == 2 || ctx->ihdr.color_type == 3) + { + if(chunk.length != 3) return SPNG_ECHUNK_SIZE; + + ctx->sbit.red_bits = data[0]; + ctx->sbit.green_bits = data[1]; + ctx->sbit.blue_bits = data[2]; + } + else if(ctx->ihdr.color_type == 4) + { + if(chunk.length != 2) return SPNG_ECHUNK_SIZE; + + ctx->sbit.grayscale_bits = data[0]; + ctx->sbit.alpha_bits = data[1]; + } + else if(ctx->ihdr.color_type == 6) + { + if(chunk.length != 4) return SPNG_ECHUNK_SIZE; + + ctx->sbit.red_bits = data[0]; + ctx->sbit.green_bits = data[1]; + ctx->sbit.blue_bits = data[2]; + ctx->sbit.alpha_bits = data[3]; + } + + if(check_sbit(&ctx->sbit, &ctx->ihdr)) return SPNG_ESBIT; + + ctx->file.sbit = 1; + ctx->stored.sbit = 1; + } + else if(!memcmp(chunk.type, type_srgb, 4)) + { + if(ctx->file.plte) return SPNG_ECHUNK_POS; + if(ctx->state == SPNG_STATE_AFTER_IDAT) return SPNG_ECHUNK_POS; + if(ctx->file.srgb) return SPNG_EDUP_SRGB; + + if(chunk.length != 1) return SPNG_ECHUNK_SIZE; + + ctx->srgb_rendering_intent = data[0]; + + if(ctx->srgb_rendering_intent > 3) return SPNG_ESRGB; + + ctx->file.srgb = 1; + ctx->stored.srgb = 1; + } + else if(!memcmp(chunk.type, type_bkgd, 4)) + { + if(ctx->state == SPNG_STATE_AFTER_IDAT) return SPNG_ECHUNK_POS; + if(ctx->file.bkgd) return SPNG_EDUP_BKGD; + + if(ctx->ihdr.color_type == 0 || ctx->ihdr.color_type == 4) + { + if(chunk.length != 2) return SPNG_ECHUNK_SIZE; + + ctx->bkgd.gray = read_u16(data); + } + else if(ctx->ihdr.color_type == 2 || ctx->ihdr.color_type == 6) + { + if(chunk.length != 6) return SPNG_ECHUNK_SIZE; + + ctx->bkgd.red = read_u16(data); + ctx->bkgd.green = read_u16(data + 2); + ctx->bkgd.blue = read_u16(data + 4); + } + else if(ctx->ihdr.color_type == 3) + { + if(chunk.length != 1) return SPNG_ECHUNK_SIZE; + if(!ctx->file.plte) return SPNG_EBKGD_NO_PLTE; + + ctx->bkgd.plte_index = data[0]; + if(ctx->bkgd.plte_index >= ctx->plte.n_entries) return SPNG_EBKGD_PLTE_IDX; + } + + ctx->file.bkgd = 1; + ctx->stored.bkgd = 1; + } + else if(!memcmp(chunk.type, type_trns, 4)) + { + if(ctx->state == SPNG_STATE_AFTER_IDAT) return SPNG_ECHUNK_POS; + if(ctx->file.trns) return SPNG_EDUP_TRNS; + if(!chunk.length) return SPNG_ECHUNK_SIZE; + + if(ctx->ihdr.color_type == 0) + { + if(chunk.length != 2) return SPNG_ECHUNK_SIZE; + + ctx->trns.gray = read_u16(data); + } + else if(ctx->ihdr.color_type == 2) + { + if(chunk.length != 6) return SPNG_ECHUNK_SIZE; + + ctx->trns.red = read_u16(data); + ctx->trns.green = read_u16(data + 2); + ctx->trns.blue = read_u16(data + 4); + } + else if(ctx->ihdr.color_type == 3) + { + if(chunk.length > ctx->plte.n_entries) return SPNG_ECHUNK_SIZE; + if(!ctx->file.plte) return SPNG_ETRNS_NO_PLTE; + + memcpy(ctx->trns.type3_alpha, data, chunk.length); + ctx->trns.n_type3_entries = chunk.length; + } + + if(ctx->ihdr.color_type == 4 || ctx->ihdr.color_type == 6) return SPNG_ETRNS_COLOR_TYPE; + + ctx->file.trns = 1; + ctx->stored.trns = 1; + } + else if(!memcmp(chunk.type, type_hist, 4)) + { + if(!ctx->file.plte) return SPNG_EHIST_NO_PLTE; + if(ctx->state == SPNG_STATE_AFTER_IDAT) return SPNG_ECHUNK_POS; + if(ctx->file.hist) return SPNG_EDUP_HIST; + + if( (chunk.length / 2) != (ctx->plte.n_entries) ) return SPNG_ECHUNK_SIZE; + + size_t k; + for(k=0; k < (chunk.length / 2); k++) + { + ctx->hist.frequency[k] = read_u16(data + k*2); + } + + ctx->file.hist = 1; + ctx->stored.hist = 1; + } + else if(!memcmp(chunk.type, type_phys, 4)) + { + if(ctx->state == SPNG_STATE_AFTER_IDAT) return SPNG_ECHUNK_POS; + if(ctx->file.phys) return SPNG_EDUP_PHYS; + + if(chunk.length != 9) return SPNG_ECHUNK_SIZE; + + ctx->phys.ppu_x = read_u32(data); + ctx->phys.ppu_y = read_u32(data + 4); + ctx->phys.unit_specifier = data[8]; + + if(check_phys(&ctx->phys)) return SPNG_EPHYS; + + ctx->file.phys = 1; + ctx->stored.phys = 1; + } + else if(!memcmp(chunk.type, type_time, 4)) + { + if(ctx->file.time) return SPNG_EDUP_TIME; + + if(chunk.length != 7) return SPNG_ECHUNK_SIZE; + + struct spng_time time; + + time.year = read_u16(data); + time.month = data[2]; + time.day = data[3]; + time.hour = data[4]; + time.minute = data[5]; + time.second = data[6]; + + if(check_time(&time)) return SPNG_ETIME; + + ctx->file.time = 1; + + if(!ctx->user.time) ctx->time = time; + + ctx->stored.time = 1; + } + else if(!memcmp(chunk.type, type_offs, 4)) + { + if(ctx->state == SPNG_STATE_AFTER_IDAT) return SPNG_ECHUNK_POS; + if(ctx->file.offs) return SPNG_EDUP_OFFS; + + if(chunk.length != 9) return SPNG_ECHUNK_SIZE; + + ctx->offs.x = read_s32(data); + ctx->offs.y = read_s32(data + 4); + ctx->offs.unit_specifier = data[8]; + + if(check_offs(&ctx->offs)) return SPNG_EOFFS; + + ctx->file.offs = 1; + ctx->stored.offs = 1; + } + else /* Arbitrary-length chunk */ + { + + if(!memcmp(chunk.type, type_exif, 4)) + { + if(ctx->file.exif) return SPNG_EDUP_EXIF; + if(!chunk.length) return SPNG_EEXIF; + + ctx->file.exif = 1; + + if(ctx->user.exif) goto discard; + + if(increase_cache_usage(ctx, chunk.length, 1)) return SPNG_ECHUNK_LIMITS; + + struct spng_exif exif; + + exif.length = chunk.length; + + exif.data = spng__malloc(ctx, chunk.length); + if(exif.data == NULL) return SPNG_EMEM; + + ret = read_chunk_bytes2(ctx, exif.data, chunk.length); + if(ret) + { + spng__free(ctx, exif.data); + return ret; + } + + if(check_exif(&exif)) + { + spng__free(ctx, exif.data); + return SPNG_EEXIF; + } + + ctx->exif = exif; + + ctx->stored.exif = 1; + } + else if(!memcmp(chunk.type, type_iccp, 4)) + {/* TODO: add test file with color profile */ + if(ctx->file.plte) return SPNG_ECHUNK_POS; + if(ctx->state == SPNG_STATE_AFTER_IDAT) return SPNG_ECHUNK_POS; + if(ctx->file.iccp) return SPNG_EDUP_ICCP; + if(!chunk.length) return SPNG_ECHUNK_SIZE; + + ctx->file.iccp = 1; + + uint32_t peek_bytes = 81 > chunk.length ? chunk.length : 81; + + ret = read_chunk_bytes(ctx, peek_bytes); + if(ret) return ret; + + unsigned char *keyword_nul = memchr(ctx->data, '\0', peek_bytes); + if(keyword_nul == NULL) return SPNG_EICCP_NAME; + + uint32_t keyword_len = keyword_nul - ctx->data; + + if(keyword_len > 79) return SPNG_EICCP_NAME; + + memcpy(ctx->iccp.profile_name, ctx->data, keyword_len); + + if(check_png_keyword(ctx->iccp.profile_name)) return SPNG_EICCP_NAME; + + if(chunk.length < (keyword_len + 2)) return SPNG_ECHUNK_SIZE; + + if(ctx->data[keyword_len + 1] != 0) return SPNG_EICCP_COMPRESSION_METHOD; + + ret = spng__inflate_stream(ctx, &ctx->iccp.profile, &ctx->iccp.profile_len, 0, ctx->data + keyword_len + 2, peek_bytes - (keyword_len + 2)); + + if(ret) return ret; + + ctx->stored.iccp = 1; + } + else if(!memcmp(chunk.type, type_text, 4) || + !memcmp(chunk.type, type_ztxt, 4) || + !memcmp(chunk.type, type_itxt, 4)) + { + if(!chunk.length) return SPNG_ECHUNK_SIZE; + + ctx->file.text = 1; + + if(ctx->user.text) goto discard; + + if(increase_cache_usage(ctx, sizeof(struct spng_text2), 1)) return SPNG_ECHUNK_LIMITS; + + ctx->n_text++; + if(ctx->n_text < 1) return SPNG_EOVERFLOW; + if(sizeof(struct spng_text2) > SIZE_MAX / ctx->n_text) return SPNG_EOVERFLOW; + + void *buf = spng__realloc(ctx, ctx->text_list, ctx->n_text * sizeof(struct spng_text2)); + if(buf == NULL) return SPNG_EMEM; + ctx->text_list = buf; + + struct spng_text2 *text = &ctx->text_list[ctx->n_text - 1]; + memset(text, 0, sizeof(struct spng_text2)); + + ctx->undo = text_undo; + + uint32_t text_offset = 0, language_tag_offset = 0, translated_keyword_offset = 0; + uint32_t peek_bytes = 256; /* enough for 3 80-byte keywords and some text bytes */ + uint32_t keyword_len; + + if(peek_bytes > chunk.length) peek_bytes = chunk.length; + + ret = read_chunk_bytes(ctx, peek_bytes); + if(ret) return ret; + + data = ctx->data; + + const unsigned char *zlib_stream = NULL; + const unsigned char *peek_end = data + peek_bytes; + const unsigned char *keyword_nul = memchr(data, 0, chunk.length > 80 ? 80 : chunk.length); + + if(keyword_nul == NULL) return SPNG_ETEXT_KEYWORD; + + keyword_len = keyword_nul - data; + + if(!memcmp(chunk.type, type_text, 4)) + { + text->type = SPNG_TEXT; + + text->text_length = chunk.length - keyword_len - 1; + + text_offset = keyword_len; + + /* increment past nul if there is a text field */ + if(text->text_length) text_offset++; + } + else if(!memcmp(chunk.type, type_ztxt, 4)) + { + text->type = SPNG_ZTXT; + + if((peek_bytes - keyword_len) <= 2) return SPNG_EZTXT; + + if(keyword_nul[1]) return SPNG_EZTXT_COMPRESSION_METHOD; + + text->compression_flag = 1; + + text_offset = keyword_len + 2; + } + else if(!memcmp(chunk.type, type_itxt, 4)) + { + text->type = SPNG_ITXT; + + /* at least two 1-byte fields, two >=0 length strings, and one byte of (compressed) text */ + if((peek_bytes - keyword_len) < 5) return SPNG_EITXT; + + text->compression_flag = keyword_nul[1]; + + if(text->compression_flag > 1) return SPNG_EITXT_COMPRESSION_FLAG; + + if(keyword_nul[2]) return SPNG_EITXT_COMPRESSION_METHOD; + + language_tag_offset = keyword_len + 3; + + const unsigned char *term; + term = memchr(data + language_tag_offset, 0, peek_bytes - language_tag_offset); + if(term == NULL) return SPNG_EITXT_LANG_TAG; + + if((peek_end - term) < 2) return SPNG_EITXT; + + translated_keyword_offset = term - data + 1; + + zlib_stream = memchr(data + translated_keyword_offset, 0, peek_bytes - translated_keyword_offset); + if(zlib_stream == NULL) return SPNG_EITXT; + if(zlib_stream == peek_end) return SPNG_EITXT; + + text_offset = zlib_stream - data + 1; + text->text_length = chunk.length - text_offset; + } + else return SPNG_EINTERNAL; + + + if(text->compression_flag) + { + /* cache usage = peek_bytes + decompressed text size + nul */ + if(increase_cache_usage(ctx, peek_bytes, 0)) return SPNG_ECHUNK_LIMITS; + + text->keyword = spng__calloc(ctx, 1, peek_bytes); + if(text->keyword == NULL) return SPNG_EMEM; + + memcpy(text->keyword, data, peek_bytes); + + zlib_stream = ctx->data + text_offset; + + ret = spng__inflate_stream(ctx, &text->text, &text->text_length, 1, zlib_stream, peek_bytes - text_offset); + + if(ret) return ret; + + text->text[text->text_length - 1] = '\0'; + text->cache_usage = text->text_length + peek_bytes; + } + else + { + if(increase_cache_usage(ctx, chunk.length + 1, 0)) return SPNG_ECHUNK_LIMITS; + + text->keyword = spng__malloc(ctx, chunk.length + 1); + if(text->keyword == NULL) return SPNG_EMEM; + + memcpy(text->keyword, data, peek_bytes); + + if(chunk.length > peek_bytes) + { + ret = read_chunk_bytes2(ctx, text->keyword + peek_bytes, chunk.length - peek_bytes); + if(ret) return ret; + } + + text->text = text->keyword + text_offset; + + text->text_length = chunk.length - text_offset; + + text->text[text->text_length] = '\0'; + text->cache_usage = chunk.length + 1; + } + + if(check_png_keyword(text->keyword)) return SPNG_ETEXT_KEYWORD; + + text->text_length = strlen(text->text); + + if(text->type != SPNG_ITXT) + { + language_tag_offset = keyword_len; + translated_keyword_offset = keyword_len; + + if(ctx->strict && check_png_text(text->text, text->text_length)) + { + if(text->type == SPNG_ZTXT) return SPNG_EZTXT; + else return SPNG_ETEXT; + } + } + + text->language_tag = text->keyword + language_tag_offset; + text->translated_keyword = text->keyword + translated_keyword_offset; + + ctx->stored.text = 1; + } + else if(!memcmp(chunk.type, type_splt, 4)) + { + if(ctx->state == SPNG_STATE_AFTER_IDAT) return SPNG_ECHUNK_POS; + if(ctx->user.splt) goto discard; /* XXX: could check profile names for uniqueness */ + if(!chunk.length) return SPNG_ECHUNK_SIZE; + + ctx->file.splt = 1; + + /* chunk.length + sizeof(struct spng_splt) + splt->n_entries * sizeof(struct spng_splt_entry) */ + if(increase_cache_usage(ctx, chunk.length + sizeof(struct spng_splt), 1)) return SPNG_ECHUNK_LIMITS; + + ctx->n_splt++; + if(ctx->n_splt < 1) return SPNG_EOVERFLOW; + if(sizeof(struct spng_splt) > SIZE_MAX / ctx->n_splt) return SPNG_EOVERFLOW; + + void *buf = spng__realloc(ctx, ctx->splt_list, ctx->n_splt * sizeof(struct spng_splt)); + if(buf == NULL) return SPNG_EMEM; + ctx->splt_list = buf; + + struct spng_splt *splt = &ctx->splt_list[ctx->n_splt - 1]; + + memset(splt, 0, sizeof(struct spng_splt)); + + ctx->undo = splt_undo; + + void *t = spng__malloc(ctx, chunk.length); + if(t == NULL) return SPNG_EMEM; + + splt->entries = t; /* simplifies error handling */ + data = t; + + ret = read_chunk_bytes2(ctx, t, chunk.length); + if(ret) return ret; + + uint32_t keyword_len = chunk.length < 80 ? chunk.length : 80; + + const unsigned char *keyword_nul = memchr(data, 0, keyword_len); + if(keyword_nul == NULL) return SPNG_ESPLT_NAME; + + keyword_len = keyword_nul - data; + + memcpy(splt->name, data, keyword_len); + + if(check_png_keyword(splt->name)) return SPNG_ESPLT_NAME; + + uint32_t j; + for(j=0; j < (ctx->n_splt - 1); j++) + { + if(!strcmp(ctx->splt_list[j].name, splt->name)) return SPNG_ESPLT_DUP_NAME; + } + + if( (chunk.length - keyword_len) <= 2) return SPNG_ECHUNK_SIZE; + + splt->sample_depth = data[keyword_len + 1]; + + uint32_t entries_len = chunk.length - keyword_len - 2; + if(!entries_len) return SPNG_ECHUNK_SIZE; + + if(splt->sample_depth == 16) + { + if(entries_len % 10 != 0) return SPNG_ECHUNK_SIZE; + splt->n_entries = entries_len / 10; + } + else if(splt->sample_depth == 8) + { + if(entries_len % 6 != 0) return SPNG_ECHUNK_SIZE; + splt->n_entries = entries_len / 6; + } + else return SPNG_ESPLT_DEPTH; + + if(!splt->n_entries) return SPNG_ECHUNK_SIZE; + + size_t list_size = splt->n_entries; + + if(list_size > SIZE_MAX / sizeof(struct spng_splt_entry)) return SPNG_EOVERFLOW; + + list_size *= sizeof(struct spng_splt_entry); + + if(increase_cache_usage(ctx, list_size, 0)) return SPNG_ECHUNK_LIMITS; + + splt->entries = spng__malloc(ctx, list_size); + if(splt->entries == NULL) + { + spng__free(ctx, t); + return SPNG_EMEM; + } + + data = (unsigned char*)t + keyword_len + 2; + + uint32_t k; + if(splt->sample_depth == 16) + { + for(k=0; k < splt->n_entries; k++) + { + splt->entries[k].red = read_u16(data + k * 10); + splt->entries[k].green = read_u16(data + k * 10 + 2); + splt->entries[k].blue = read_u16(data + k * 10 + 4); + splt->entries[k].alpha = read_u16(data + k * 10 + 6); + splt->entries[k].frequency = read_u16(data + k * 10 + 8); + } + } + else if(splt->sample_depth == 8) + { + for(k=0; k < splt->n_entries; k++) + { + splt->entries[k].red = data[k * 6]; + splt->entries[k].green = data[k * 6 + 1]; + splt->entries[k].blue = data[k * 6 + 2]; + splt->entries[k].alpha = data[k * 6 + 3]; + splt->entries[k].frequency = read_u16(data + k * 6 + 4); + } + } + + spng__free(ctx, t); + decrease_cache_usage(ctx, chunk.length); + + ctx->stored.splt = 1; + } + else /* Unknown chunk */ + { + ctx->file.unknown = 1; + + if(!ctx->keep_unknown) goto discard; + if(ctx->user.unknown) goto discard; + + if(increase_cache_usage(ctx, chunk.length + sizeof(struct spng_unknown_chunk), 1)) return SPNG_ECHUNK_LIMITS; + + ctx->n_chunks++; + if(ctx->n_chunks < 1) return SPNG_EOVERFLOW; + if(sizeof(struct spng_unknown_chunk) > SIZE_MAX / ctx->n_chunks) return SPNG_EOVERFLOW; + + void *buf = spng__realloc(ctx, ctx->chunk_list, ctx->n_chunks * sizeof(struct spng_unknown_chunk)); + if(buf == NULL) return SPNG_EMEM; + ctx->chunk_list = buf; + + struct spng_unknown_chunk *chunkp = &ctx->chunk_list[ctx->n_chunks - 1]; + + memset(chunkp, 0, sizeof(struct spng_unknown_chunk)); + + ctx->undo = chunk_undo; + + memcpy(chunkp->type, chunk.type, 4); + + if(ctx->state < SPNG_STATE_FIRST_IDAT) + { + if(ctx->file.plte) chunkp->location = SPNG_AFTER_PLTE; + else chunkp->location = SPNG_AFTER_IHDR; + } + else if(ctx->state >= SPNG_STATE_AFTER_IDAT) chunkp->location = SPNG_AFTER_IDAT; + + if(chunk.length > 0) + { + void *t = spng__malloc(ctx, chunk.length); + if(t == NULL) return SPNG_EMEM; + + ret = read_chunk_bytes2(ctx, t, chunk.length); + if(ret) + { + spng__free(ctx, t); + return ret; + } + + chunkp->length = chunk.length; + chunkp->data = t; + } + + ctx->stored.unknown = 1; + } + +discard: + ret = discard_chunk_bytes(ctx, ctx->cur_chunk_bytes_left); + if(ret) return ret; + } + + } + + return ret; +} + +/* Read chunks before or after the IDAT chunks depending on state */ +static int read_chunks(spng_ctx *ctx, int only_ihdr) +{ + if(ctx == NULL) return SPNG_EINTERNAL; + if(!ctx->state) return SPNG_EBADSTATE; + if(ctx->data == NULL) + { + if(ctx->encode_only) return 0; + else return SPNG_EINTERNAL; + } + + int ret = 0; + + if(ctx->state == SPNG_STATE_INPUT) + { + ret = read_ihdr(ctx); + + if(ret) return decode_err(ctx, ret); + + ctx->state = SPNG_STATE_IHDR; + } + + if(only_ihdr) return 0; + + if(ctx->state == SPNG_STATE_EOI) + { + ctx->state = SPNG_STATE_AFTER_IDAT; + ctx->prev_was_idat = 1; + } + + while(ctx->state < SPNG_STATE_FIRST_IDAT || ctx->state == SPNG_STATE_AFTER_IDAT) + { + ret = read_non_idat_chunks(ctx); + + if(!ret) + { + if(ctx->state < SPNG_STATE_FIRST_IDAT) ctx->state = SPNG_STATE_FIRST_IDAT; + else if(ctx->state == SPNG_STATE_AFTER_IDAT) ctx->state = SPNG_STATE_IEND; + } + else + { + switch(ret) + { + case SPNG_ECHUNK_POS: + case SPNG_ECHUNK_SIZE: /* size != expected size, SPNG_ECHUNK_STDLEN = invalid size */ + case SPNG_EDUP_PLTE: + case SPNG_EDUP_CHRM: + case SPNG_EDUP_GAMA: + case SPNG_EDUP_ICCP: + case SPNG_EDUP_SBIT: + case SPNG_EDUP_SRGB: + case SPNG_EDUP_BKGD: + case SPNG_EDUP_HIST: + case SPNG_EDUP_TRNS: + case SPNG_EDUP_PHYS: + case SPNG_EDUP_TIME: + case SPNG_EDUP_OFFS: + case SPNG_EDUP_EXIF: + case SPNG_ECHRM: + case SPNG_ETRNS_COLOR_TYPE: + case SPNG_ETRNS_NO_PLTE: + case SPNG_EGAMA: + case SPNG_EICCP_NAME: + case SPNG_EICCP_COMPRESSION_METHOD: + case SPNG_ESBIT: + case SPNG_ESRGB: + case SPNG_ETEXT: + case SPNG_ETEXT_KEYWORD: + case SPNG_EZTXT: + case SPNG_EZTXT_COMPRESSION_METHOD: + case SPNG_EITXT: + case SPNG_EITXT_COMPRESSION_FLAG: + case SPNG_EITXT_COMPRESSION_METHOD: + case SPNG_EITXT_LANG_TAG: + case SPNG_EITXT_TRANSLATED_KEY: + case SPNG_EBKGD_NO_PLTE: + case SPNG_EBKGD_PLTE_IDX: + case SPNG_EHIST_NO_PLTE: + case SPNG_EPHYS: + case SPNG_ESPLT_NAME: + case SPNG_ESPLT_DUP_NAME: + case SPNG_ESPLT_DEPTH: + case SPNG_ETIME: + case SPNG_EOFFS: + case SPNG_EEXIF: + case SPNG_EZLIB: + { + if(!ctx->strict && !is_critical_chunk(&ctx->current_chunk)) + { + ret = discard_chunk_bytes(ctx, ctx->cur_chunk_bytes_left); + if(ret) return decode_err(ctx, ret); + + if(ctx->undo) ctx->undo(ctx); + + ctx->stored = ctx->prev_stored; + + ctx->discard = 0; + ctx->undo = NULL; + + continue; + } + else return decode_err(ctx, ret); + + break; + } + default: return decode_err(ctx, ret); + } + } + } + + return ret; +} + +static int read_scanline(spng_ctx *ctx) +{ + int ret, pass = ctx->row_info.pass; + struct spng_row_info *ri = &ctx->row_info; + const struct spng_subimage *sub = ctx->subimage; + size_t scanline_width = sub[pass].scanline_width; + uint32_t scanline_idx = ri->scanline_idx; + + uint8_t next_filter = 0; + + if(scanline_idx == (sub[pass].height - 1) && ri->pass == ctx->last_pass) + { + ret = read_scanline_bytes(ctx, ctx->scanline, scanline_width - 1); + } + else + { + ret = read_scanline_bytes(ctx, ctx->scanline, scanline_width); + if(ret) return ret; + + next_filter = ctx->scanline[scanline_width - 1]; + if(next_filter > 4) ret = SPNG_EFILTER; + } + + if(ret) return ret; + + if(!scanline_idx && ri->filter > 1) + { + /* prev_scanline is all zeros for the first scanline */ + memset(ctx->prev_scanline, 0, scanline_width); + } + + if(ctx->ihdr.bit_depth == 16 && ctx->fmt != SPNG_FMT_RAW) u16_row_to_host(ctx->scanline, scanline_width - 1); + + ret = defilter_scanline(ctx->prev_scanline, ctx->scanline, scanline_width, ctx->bytes_per_pixel, ri->filter); + if(ret) return ret; + + ri->filter = next_filter; + + return 0; +} + +static int update_row_info(spng_ctx *ctx) +{ + int interlacing = ctx->ihdr.interlace_method; + struct spng_row_info *ri = &ctx->row_info; + const struct spng_subimage *sub = ctx->subimage; + + if(ri->scanline_idx == (sub[ri->pass].height - 1)) /* Last scanline */ + { + if(ri->pass == ctx->last_pass) + { + ctx->state = SPNG_STATE_EOI; + + return SPNG_EOI; + } + + ri->scanline_idx = 0; + ri->pass++; + + /* Skip empty passes */ + while( (!sub[ri->pass].width || !sub[ri->pass].height) && (ri->pass < ctx->last_pass)) ri->pass++; + } + else + { + ri->row_num++; + ri->scanline_idx++; + } + + if(interlacing) ri->row_num = adam7_y_start[ri->pass] + ri->scanline_idx * adam7_y_delta[ri->pass]; + + return 0; +} + +int spng_decode_scanline(spng_ctx *ctx, void *out, size_t len) +{ + if(ctx == NULL || out == NULL) return 1; + + if(ctx->state >= SPNG_STATE_EOI) return SPNG_EOI; + + struct decode_flags f = ctx->decode_flags; + + struct spng_row_info *ri = &ctx->row_info; + const struct spng_subimage *sub = ctx->subimage; + + const struct spng_ihdr *ihdr = &ctx->ihdr; + const uint16_t *gamma_lut = ctx->gamma_lut; + unsigned char *trns_px = ctx->trns_px; + const struct spng_sbit *sb = &ctx->decode_sb; + const struct spng_plte_entry *plte = ctx->decode_plte.rgba; + struct spng__iter iter = spng__iter_init(ihdr->bit_depth, ctx->scanline); + + const unsigned char *scanline; + + const int pass = ri->pass; + const int fmt = ctx->fmt; + const size_t scanline_width = sub[pass].scanline_width; + const uint32_t width = sub[pass].width; + uint32_t k; + uint8_t r_8, g_8, b_8, a_8, gray_8; + uint16_t r_16, g_16, b_16, a_16, gray_16; + r_8=0; g_8=0; b_8=0; a_8=0; gray_8=0; + r_16=0; g_16=0; b_16=0; a_16=0; gray_16=0; + size_t pixel_size = 4; /* SPNG_FMT_RGBA8 */ + size_t pixel_offset = 0; + unsigned char *pixel; + unsigned processing_depth = ihdr->bit_depth; + + if(f.indexed) processing_depth = 8; + + if(fmt == SPNG_FMT_RGBA16) pixel_size = 8; + else if(fmt == SPNG_FMT_RGB8) pixel_size = 3; + + if(len < sub[pass].out_width) return SPNG_EBUFSIZ; + + int ret = read_scanline(ctx); + + if(ret) return decode_err(ctx, ret); + + scanline = ctx->scanline; + + for(k=0; k < width; k++) + { + pixel = (unsigned char*)out + pixel_offset; + pixel_offset += pixel_size; + + if(f.same_layout) + { + if(f.zerocopy) break; + + memcpy(out, scanline, scanline_width - 1); + break; + } + + if(f.unpack) + { + unpack_scanline(out, scanline, width, ihdr->bit_depth, fmt); + break; + } + + if(ihdr->color_type == SPNG_COLOR_TYPE_TRUECOLOR) + { + if(ihdr->bit_depth == 16) + { + memcpy(&r_16, scanline + (k * 6), 2); + memcpy(&g_16, scanline + (k * 6) + 2, 2); + memcpy(&b_16, scanline + (k * 6) + 4, 2); + + a_16 = 65535; + } + else /* == 8 */ + { + if(fmt == SPNG_FMT_RGBA8) + { + rgb8_row_to_rgba8(scanline, out, width); + break; + } + + r_8 = scanline[k * 3]; + g_8 = scanline[k * 3 + 1]; + b_8 = scanline[k * 3 + 2]; + + a_8 = 255; + } + } + else if(ihdr->color_type == SPNG_COLOR_TYPE_INDEXED) + { + uint8_t entry = 0; + + if(ihdr->bit_depth == 8) + { + if(fmt & (SPNG_FMT_RGBA8 | SPNG_FMT_RGB8)) + { + expand_row(out, scanline, &ctx->decode_plte, width, fmt); + break; + } + + entry = scanline[k]; + } + else /* < 8 */ + { + entry = get_sample(&iter); + } + + if(fmt & (SPNG_FMT_RGBA8 | SPNG_FMT_RGB8)) + { + pixel[0] = plte[entry].red; + pixel[1] = plte[entry].green; + pixel[2] = plte[entry].blue; + if(fmt == SPNG_FMT_RGBA8) pixel[3] = plte[entry].alpha; + + continue; + } + else /* RGBA16 */ + { + r_16 = plte[entry].red; + g_16 = plte[entry].green; + b_16 = plte[entry].blue; + a_16 = plte[entry].alpha; + + r_16 = (r_16 << 8) | r_16; + g_16 = (g_16 << 8) | g_16; + b_16 = (b_16 << 8) | b_16; + a_16 = (a_16 << 8) | a_16; + + memcpy(pixel, &r_16, 2); + memcpy(pixel + 2, &g_16, 2); + memcpy(pixel + 4, &b_16, 2); + memcpy(pixel + 6, &a_16, 2); + + continue; + } + } + else if(ihdr->color_type == SPNG_COLOR_TYPE_TRUECOLOR_ALPHA) + { + if(ihdr->bit_depth == 16) + { + memcpy(&r_16, scanline + (k * 8), 2); + memcpy(&g_16, scanline + (k * 8) + 2, 2); + memcpy(&b_16, scanline + (k * 8) + 4, 2); + memcpy(&a_16, scanline + (k * 8) + 6, 2); + } + else /* == 8 */ + { + r_8 = scanline[k * 4]; + g_8 = scanline[k * 4 + 1]; + b_8 = scanline[k * 4 + 2]; + a_8 = scanline[k * 4 + 3]; + } + } + else if(ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE) + { + if(ihdr->bit_depth == 16) + { + memcpy(&gray_16, scanline + k * 2, 2); + + if(f.apply_trns && ctx->trns.gray == gray_16) a_16 = 0; + else a_16 = 65535; + + r_16 = gray_16; + g_16 = gray_16; + b_16 = gray_16; + } + else /* <= 8 */ + { + gray_8 = get_sample(&iter); + + if(f.apply_trns && ctx->trns.gray == gray_8) a_8 = 0; + else a_8 = 255; + + r_8 = gray_8; g_8 = gray_8; b_8 = gray_8; + } + } + else if(ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE_ALPHA) + { + if(ihdr->bit_depth == 16) + { + memcpy(&gray_16, scanline + (k * 4), 2); + memcpy(&a_16, scanline + (k * 4) + 2, 2); + + r_16 = gray_16; + g_16 = gray_16; + b_16 = gray_16; + } + else /* == 8 */ + { + gray_8 = scanline[k * 2]; + a_8 = scanline[k * 2 + 1]; + + r_8 = gray_8; + g_8 = gray_8; + b_8 = gray_8; + } + } + + + if(fmt & (SPNG_FMT_RGBA8 | SPNG_FMT_RGB8)) + { + if(ihdr->bit_depth == 16) + { + r_8 = r_16 >> 8; + g_8 = g_16 >> 8; + b_8 = b_16 >> 8; + a_8 = a_16 >> 8; + } + + pixel[0] = r_8; + pixel[1] = g_8; + pixel[2] = b_8; + + if(fmt == SPNG_FMT_RGBA8) pixel[3] = a_8; + } + else if(fmt == SPNG_FMT_RGBA16) + { + if(ihdr->bit_depth != 16) + { + r_16 = r_8; + g_16 = g_8; + b_16 = b_8; + a_16 = a_8; + } + + memcpy(pixel, &r_16, 2); + memcpy(pixel + 2, &g_16, 2); + memcpy(pixel + 4, &b_16, 2); + memcpy(pixel + 6, &a_16, 2); + } + }/* for(k=0; k < width; k++) */ + + if(f.apply_trns) trns_row(out, scanline, trns_px, ctx->bytes_per_pixel, &ctx->ihdr, width, fmt); + + if(f.do_scaling) scale_row(out, width, fmt, processing_depth, sb); + + if(f.apply_gamma) gamma_correct_row(out, width, fmt, gamma_lut); + + /* The previous scanline is always defiltered */ + void *t = ctx->prev_scanline; + ctx->prev_scanline = ctx->scanline; + ctx->scanline = t; + + ret = update_row_info(ctx); + + if(ret == SPNG_EOI) + { + if(ctx->cur_chunk_bytes_left) /* zlib stream ended before an IDAT chunk boundary */ + {/* Discard the rest of the chunk */ + int error = discard_chunk_bytes(ctx, ctx->cur_chunk_bytes_left); + if(error) return decode_err(ctx, error); + } + + ctx->last_idat = ctx->current_chunk; + } + + return ret; +} + +int spng_decode_row(spng_ctx *ctx, void *out, size_t len) +{ + if(ctx == NULL || out == NULL) return 1; + if(ctx->state >= SPNG_STATE_EOI) return SPNG_EOI; + if(len < ctx->image_width) return SPNG_EBUFSIZ; + + const struct spng_ihdr *ihdr = &ctx->ihdr; + int ret, pass = ctx->row_info.pass; + unsigned char *outptr = out; + + if(!ihdr->interlace_method || pass == 6) return spng_decode_scanline(ctx, out, len); + + ret = spng_decode_scanline(ctx, ctx->row, ctx->image_width); + if(ret && ret != SPNG_EOI) return ret; + + uint32_t k; + unsigned pixel_size = 4; /* RGBA8 */ + if(ctx->fmt == SPNG_FMT_RGBA16) pixel_size = 8; + else if(ctx->fmt == SPNG_FMT_RGB8) pixel_size = 3; + else if(ctx->fmt == SPNG_FMT_G8) pixel_size = 1; + else if(ctx->fmt == SPNG_FMT_GA8) pixel_size = 2; + else if(ctx->fmt & (SPNG_FMT_PNG | SPNG_FMT_RAW)) + { + if(ihdr->bit_depth < 8) + { + struct spng__iter iter = spng__iter_init(ihdr->bit_depth, ctx->row); + const uint8_t samples_per_byte = 8 / ihdr->bit_depth; + uint8_t sample; + + for(k=0; k < ctx->subimage[pass].width; k++) + { + sample = get_sample(&iter); + + size_t ioffset = adam7_x_start[pass] + k * adam7_x_delta[pass]; + + sample = sample << (iter.initial_shift - ioffset * ihdr->bit_depth % 8); + + ioffset /= samples_per_byte; + + outptr[ioffset] |= sample; + } + + return 0; + } + else pixel_size = ctx->bytes_per_pixel; + } + + for(k=0; k < ctx->subimage[pass].width; k++) + { + size_t ioffset = (adam7_x_start[pass] + (size_t) k * adam7_x_delta[pass]) * pixel_size; + + memcpy(outptr + ioffset, ctx->row + k * pixel_size, pixel_size); + } + + return 0; +} + +int spng_decode_chunks(spng_ctx *ctx) +{ + if(ctx == NULL) return 1; + if(ctx->encode_only) return SPNG_ECTXTYPE; + if(ctx->state < SPNG_STATE_INPUT) return SPNG_ENOSRC; + if(ctx->state == SPNG_STATE_IEND) return 0; + + return read_chunks(ctx, 0); +} + +int spng_decode_image(spng_ctx *ctx, void *out, size_t len, int fmt, int flags) +{ + if(ctx == NULL) return 1; + if(ctx->encode_only) return SPNG_ECTXTYPE; + if(ctx->state >= SPNG_STATE_EOI) return SPNG_EOI; + + const struct spng_ihdr *ihdr = &ctx->ihdr; + + int ret = read_chunks(ctx, 0); + if(ret) return decode_err(ctx, ret); + + ret = check_decode_fmt(ihdr, fmt); + if(ret) return ret; + + ret = calculate_image_width(ihdr, fmt, &ctx->image_width); + if(ret) return decode_err(ctx, ret); + + if(ctx->image_width > SIZE_MAX / ihdr->height) ctx->image_size = 0; /* overflow */ + else ctx->image_size = ctx->image_width * ihdr->height; + + if( !(flags & SPNG_DECODE_PROGRESSIVE) ) + { + if(out == NULL) return 1; + if(!ctx->image_size) return SPNG_EOVERFLOW; + if(len < ctx->image_size) return SPNG_EBUFSIZ; + } + + uint32_t bytes_read = 0; + + ret = read_idat_bytes(ctx, &bytes_read); + if(ret) return decode_err(ctx, ret); + + if(bytes_read > 1) + { + int valid = read_u16(ctx->data) % 31 ? 0 : 1; + + unsigned flg = ctx->data[1]; + unsigned flevel = flg >> 6; + int compression_level = Z_DEFAULT_COMPRESSION; + + if(flevel == 0) compression_level = 0; /* fastest */ + else if(flevel == 1) compression_level = 1; /* fast */ + else if(flevel == 2) compression_level = 6; /* default */ + else if(flevel == 3) compression_level = 9; /* slowest, max compression */ + + if(valid) ctx->image_options.compression_level = compression_level; + } + + ret = spng__inflate_init(ctx, ctx->image_options.window_bits); + if(ret) return decode_err(ctx, ret); + + ctx->zstream.avail_in = bytes_read; + ctx->zstream.next_in = ctx->data; + + size_t scanline_buf_size = ctx->subimage[ctx->widest_pass].scanline_width; + + scanline_buf_size += 32; + + if(scanline_buf_size < 32) return SPNG_EOVERFLOW; + + ctx->scanline_buf = spng__malloc(ctx, scanline_buf_size); + ctx->prev_scanline_buf = spng__malloc(ctx, scanline_buf_size); + + ctx->scanline = ctx->scanline_buf; + ctx->prev_scanline = ctx->prev_scanline_buf; + + struct decode_flags f = {0}; + + ctx->fmt = fmt; + + if(ihdr->color_type == SPNG_COLOR_TYPE_INDEXED) f.indexed = 1; + + unsigned processing_depth = ihdr->bit_depth; + + if(f.indexed) processing_depth = 8; + + if(ihdr->interlace_method) + { + f.interlaced = 1; + ctx->row_buf = spng__malloc(ctx, ctx->image_width); + ctx->row = ctx->row_buf; + + if(ctx->row == NULL) return decode_err(ctx, SPNG_EMEM); + } + + if(ctx->scanline == NULL || ctx->prev_scanline == NULL) + { + return decode_err(ctx, SPNG_EMEM); + } + + f.do_scaling = 1; + if(f.indexed) f.do_scaling = 0; + + unsigned depth_target = 8; /* FMT_RGBA8, G8 */ + if(fmt == SPNG_FMT_RGBA16) depth_target = 16; + + if(flags & SPNG_DECODE_TRNS && ctx->stored.trns) f.apply_trns = 1; + else flags &= ~SPNG_DECODE_TRNS; + + if(ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE_ALPHA || + ihdr->color_type == SPNG_COLOR_TYPE_TRUECOLOR_ALPHA) flags &= ~SPNG_DECODE_TRNS; + + if(flags & SPNG_DECODE_GAMMA && ctx->stored.gama) f.apply_gamma = 1; + else flags &= ~SPNG_DECODE_GAMMA; + + if(flags & SPNG_DECODE_USE_SBIT && ctx->stored.sbit) f.use_sbit = 1; + else flags &= ~SPNG_DECODE_USE_SBIT; + + if(fmt & (SPNG_FMT_RGBA8 | SPNG_FMT_RGBA16)) + { + if(ihdr->color_type == SPNG_COLOR_TYPE_TRUECOLOR_ALPHA && + ihdr->bit_depth == depth_target) f.same_layout = 1; + } + else if(fmt == SPNG_FMT_RGB8) + { + if(ihdr->color_type == SPNG_COLOR_TYPE_TRUECOLOR && + ihdr->bit_depth == depth_target) f.same_layout = 1; + + f.apply_trns = 0; /* not applicable */ + } + else if(fmt & (SPNG_FMT_PNG | SPNG_FMT_RAW)) + { + f.same_layout = 1; + f.do_scaling = 0; + f.apply_gamma = 0; /* for now */ + f.apply_trns = 0; + } + else if(fmt == SPNG_FMT_G8 && ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE && ihdr->bit_depth <= 8) + { + if(ihdr->bit_depth == depth_target) f.same_layout = 1; + else if(ihdr->bit_depth < 8) f.unpack = 1; + + f.apply_trns = 0; + } + else if(fmt == SPNG_FMT_GA8 && ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE && ihdr->bit_depth <= 8) + { + if(ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE_ALPHA && + ihdr->bit_depth == depth_target) f.same_layout = 1; + else if(ihdr->bit_depth <= 8) f.unpack = 1; + } + else if(fmt == SPNG_FMT_GA16 && ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE && ihdr->bit_depth == 16) + { + if(ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE_ALPHA && + ihdr->bit_depth == depth_target) f.same_layout = 1; + else if(ihdr->bit_depth == 16) f.unpack = 1; + } + + /*if(f.same_layout && !flags && !f.interlaced) f.zerocopy = 1;*/ + + uint16_t *gamma_lut = NULL; + + if(f.apply_gamma) + { + float file_gamma = (float)ctx->gama / 100000.0f; + float max; + + unsigned lut_entries; + + if(fmt & (SPNG_FMT_RGBA8 | SPNG_FMT_RGB8)) + { + lut_entries = 256; + max = 255.0f; + + gamma_lut = ctx->gamma_lut8; + ctx->gamma_lut = ctx->gamma_lut8; + } + else /* SPNG_FMT_RGBA16 */ + { + lut_entries = 65536; + max = 65535.0f; + + ctx->gamma_lut16 = spng__malloc(ctx, lut_entries * sizeof(uint16_t)); + if(ctx->gamma_lut16 == NULL) return decode_err(ctx, SPNG_EMEM); + + gamma_lut = ctx->gamma_lut16; + ctx->gamma_lut = ctx->gamma_lut16; + } + + float screen_gamma = 2.2f; + float exponent = file_gamma * screen_gamma; + + if(FP_ZERO == fpclassify(exponent)) return decode_err(ctx, SPNG_EGAMA); + + exponent = 1.0f / exponent; + + unsigned i; + for(i=0; i < lut_entries; i++) + { + float c = pow((float)i / max, exponent) * max; + if(c > max) c = max; + + gamma_lut[i] = (uint16_t)c; + } + } + + struct spng_sbit *sb = &ctx->decode_sb; + + sb->red_bits = processing_depth; + sb->green_bits = processing_depth; + sb->blue_bits = processing_depth; + sb->alpha_bits = processing_depth; + sb->grayscale_bits = processing_depth; + + if(f.use_sbit) + { + if(ihdr->color_type == 0) + { + sb->grayscale_bits = ctx->sbit.grayscale_bits; + sb->alpha_bits = ihdr->bit_depth; + } + else if(ihdr->color_type == 2 || ihdr->color_type == 3) + { + sb->red_bits = ctx->sbit.red_bits; + sb->green_bits = ctx->sbit.green_bits; + sb->blue_bits = ctx->sbit.blue_bits; + sb->alpha_bits = ihdr->bit_depth; + } + else if(ihdr->color_type == 4) + { + sb->grayscale_bits = ctx->sbit.grayscale_bits; + sb->alpha_bits = ctx->sbit.alpha_bits; + } + else /* == 6 */ + { + sb->red_bits = ctx->sbit.red_bits; + sb->green_bits = ctx->sbit.green_bits; + sb->blue_bits = ctx->sbit.blue_bits; + sb->alpha_bits = ctx->sbit.alpha_bits; + } + } + + if(ihdr->bit_depth == 16 && fmt & (SPNG_FMT_RGBA8 | SPNG_FMT_RGB8)) + {/* samples are scaled down by 8 bits in the decode loop */ + sb->red_bits -= 8; + sb->green_bits -= 8; + sb->blue_bits -= 8; + sb->alpha_bits -= 8; + sb->grayscale_bits -= 8; + + processing_depth = 8; + } + + /* Prevent infinite loops in sample_to_target() */ + if(!depth_target || depth_target > 16 || + !processing_depth || processing_depth > 16 || + !sb->grayscale_bits || sb->grayscale_bits > processing_depth || + !sb->alpha_bits || sb->alpha_bits > processing_depth || + !sb->red_bits || sb->red_bits > processing_depth || + !sb->green_bits || sb->green_bits > processing_depth || + !sb->blue_bits || sb->blue_bits > processing_depth) + { + return decode_err(ctx, SPNG_ESBIT); + } + + if(sb->red_bits == sb->green_bits && + sb->green_bits == sb->blue_bits && + sb->blue_bits == sb->alpha_bits && + sb->alpha_bits == processing_depth && + processing_depth == depth_target) f.do_scaling = 0; + + struct spng_plte_entry *plte = ctx->decode_plte.rgba; + + /* Pre-process palette entries */ + if(f.indexed) + { + uint8_t red, green, blue, alpha; + + uint32_t i; + for(i=0; i < 256; i++) + { + if(f.apply_trns && i < ctx->trns.n_type3_entries) + ctx->plte.entries[i].alpha = ctx->trns.type3_alpha[i]; + else + ctx->plte.entries[i].alpha = 255; + + red = sample_to_target(ctx->plte.entries[i].red, 8, sb->red_bits, 8); + green = sample_to_target(ctx->plte.entries[i].green, 8, sb->green_bits, 8); + blue = sample_to_target(ctx->plte.entries[i].blue, 8, sb->blue_bits, 8); + alpha = sample_to_target(ctx->plte.entries[i].alpha, 8, sb->alpha_bits, 8); + +#if defined(SPNG_ARM) + if(fmt == SPNG_FMT_RGB8 && ihdr->bit_depth == 8) + {/* Working with 3 bytes at a time is more of an ARM thing */ + ctx->decode_plte.rgb[i * 3 + 0] = red; + ctx->decode_plte.rgb[i * 3 + 1] = green; + ctx->decode_plte.rgb[i * 3 + 2] = blue; + continue; + } +#endif + plte[i].red = red; + plte[i].green = green; + plte[i].blue = blue; + plte[i].alpha = alpha; + } + + f.apply_trns = 0; + } + + unsigned char *trns_px = ctx->trns_px; + + if(f.apply_trns) + { + uint16_t mask = ~0; + if(ctx->ihdr.bit_depth < 16) mask = (1 << ctx->ihdr.bit_depth) - 1; + + if(fmt & (SPNG_FMT_RGBA8 | SPNG_FMT_RGBA16)) + { + if(ihdr->color_type == SPNG_COLOR_TYPE_TRUECOLOR) + { + if(ihdr->bit_depth == 16) + { + memcpy(trns_px, &ctx->trns.red, 2); + memcpy(trns_px + 2, &ctx->trns.green, 2); + memcpy(trns_px + 4, &ctx->trns.blue, 2); + } + else + { + trns_px[0] = ctx->trns.red & mask; + trns_px[1] = ctx->trns.green & mask; + trns_px[2] = ctx->trns.blue & mask; + } + } + } + else if(ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE) // fmt == SPNG_FMT_GA8 && + { + if(ihdr->bit_depth == 16) + { + memcpy(trns_px, &ctx->trns.gray, 2); + } + else + { + trns_px[0] = ctx->trns.gray & mask; + } + } + } + + ctx->decode_flags = f; + + ctx->state = SPNG_STATE_DECODE_INIT; + + struct spng_row_info *ri = &ctx->row_info; + struct spng_subimage *sub = ctx->subimage; + + while(!sub[ri->pass].width || !sub[ri->pass].height) ri->pass++; + + if(f.interlaced) ri->row_num = adam7_y_start[ri->pass]; + + unsigned pixel_size = 4; /* SPNG_FMT_RGBA8 */ + + if(fmt == SPNG_FMT_RGBA16) pixel_size = 8; + else if(fmt == SPNG_FMT_RGB8) pixel_size = 3; + else if(fmt == SPNG_FMT_G8) pixel_size = 1; + else if(fmt == SPNG_FMT_GA8) pixel_size = 2; + + int i; + for(i=ri->pass; i <= ctx->last_pass; i++) + { + if(!sub[i].scanline_width) continue; + + if(fmt & (SPNG_FMT_PNG | SPNG_FMT_RAW)) sub[i].out_width = sub[i].scanline_width - 1; + else sub[i].out_width = (size_t)sub[i].width * pixel_size; + + if(sub[i].out_width > UINT32_MAX) return decode_err(ctx, SPNG_EOVERFLOW); + } + + /* Read the first filter byte, offsetting all reads by 1 byte. + The scanlines will be aligned with the start of the array with + the next scanline's filter byte at the end, + the last scanline will end up being 1 byte "shorter". */ + ret = read_scanline_bytes(ctx, &ri->filter, 1); + if(ret) return decode_err(ctx, ret); + + if(ri->filter > 4) return decode_err(ctx, SPNG_EFILTER); + + if(flags & SPNG_DECODE_PROGRESSIVE) + { + return 0; + } + + do + { + size_t ioffset = ri->row_num * ctx->image_width; + + ret = spng_decode_row(ctx, (unsigned char*)out + ioffset, ctx->image_width); + }while(!ret); + + if(ret != SPNG_EOI) return decode_err(ctx, ret); + + return 0; +} + +int spng_get_row_info(spng_ctx *ctx, struct spng_row_info *row_info) +{ + if(ctx == NULL || row_info == NULL || ctx->state < SPNG_STATE_DECODE_INIT) return 1; + + if(ctx->state >= SPNG_STATE_EOI) return SPNG_EOI; + + *row_info = ctx->row_info; + + return 0; +} + +static int write_chunks_before_idat(spng_ctx *ctx) +{ + if(ctx == NULL) return SPNG_EINTERNAL; + if(!ctx->encode_only) return SPNG_EINTERNAL; + if(!ctx->stored.ihdr) return SPNG_EINTERNAL; + + int ret; + uint32_t i; + size_t length; + const struct spng_ihdr *ihdr = &ctx->ihdr; + unsigned char *data = ctx->decode_plte.raw; + + ret = write_data(ctx, spng_signature, 8); + if(ret) return ret; + + write_u32(data, ihdr->width); + write_u32(data + 4, ihdr->height); + data[8] = ihdr->bit_depth; + data[9] = ihdr->color_type; + data[10] = ihdr->compression_method; + data[11] = ihdr->filter_method; + data[12] = ihdr->interlace_method; + + ret = write_chunk(ctx, type_ihdr, data, 13); + if(ret) return ret; + + if(ctx->stored.chrm) + { + write_u32(data, ctx->chrm_int.white_point_x); + write_u32(data + 4, ctx->chrm_int.white_point_y); + write_u32(data + 8, ctx->chrm_int.red_x); + write_u32(data + 12, ctx->chrm_int.red_y); + write_u32(data + 16, ctx->chrm_int.green_x); + write_u32(data + 20, ctx->chrm_int.green_y); + write_u32(data + 24, ctx->chrm_int.blue_x); + write_u32(data + 28, ctx->chrm_int.blue_y); + + ret = write_chunk(ctx, type_chrm, data, 32); + if(ret) return ret; + } + + if(ctx->stored.gama) + { + write_u32(data, ctx->gama); + + ret = write_chunk(ctx, type_gama, data, 4); + if(ret) return ret; + } + + if(ctx->stored.iccp) + { + uLongf dest_len = compressBound((uLong)ctx->iccp.profile_len); + + Bytef *buf = spng__malloc(ctx, dest_len); + if(buf == NULL) return SPNG_EMEM; + + ret = compress2(buf, &dest_len, (void*)ctx->iccp.profile, (uLong)ctx->iccp.profile_len, Z_DEFAULT_COMPRESSION); + + if(ret != Z_OK) + { + spng__free(ctx, buf); + return SPNG_EZLIB; + } + + size_t name_len = strlen(ctx->iccp.profile_name); + + length = name_len + 2; + length += dest_len; + + if(dest_len > length) return SPNG_EOVERFLOW; + + unsigned char *cdata = NULL; + + ret = write_header(ctx, type_iccp, length, &cdata); + + if(ret) + { + spng__free(ctx, buf); + return ret; + } + + memcpy(cdata, ctx->iccp.profile_name, name_len + 1); + cdata[name_len + 1] = 0; /* compression method */ + memcpy(cdata + name_len + 2, buf, dest_len); + + spng__free(ctx, buf); + + ret = finish_chunk(ctx); + if(ret) return ret; + } + + if(ctx->stored.sbit) + { + switch(ctx->ihdr.color_type) + { + case SPNG_COLOR_TYPE_GRAYSCALE: + { + length = 1; + + data[0] = ctx->sbit.grayscale_bits; + + break; + } + case SPNG_COLOR_TYPE_TRUECOLOR: + case SPNG_COLOR_TYPE_INDEXED: + { + length = 3; + + data[0] = ctx->sbit.red_bits; + data[1] = ctx->sbit.green_bits; + data[2] = ctx->sbit.blue_bits; + + break; + } + case SPNG_COLOR_TYPE_GRAYSCALE_ALPHA: + { + length = 2; + + data[0] = ctx->sbit.grayscale_bits; + data[1] = ctx->sbit.alpha_bits; + + break; + } + case SPNG_COLOR_TYPE_TRUECOLOR_ALPHA: + { + length = 4; + + data[0] = ctx->sbit.red_bits; + data[1] = ctx->sbit.green_bits; + data[2] = ctx->sbit.blue_bits; + data[3] = ctx->sbit.alpha_bits; + + break; + } + default: return SPNG_EINTERNAL; + } + + ret = write_chunk(ctx, type_sbit, data, length); + if(ret) return ret; + } + + if(ctx->stored.srgb) + { + ret = write_chunk(ctx, type_srgb, &ctx->srgb_rendering_intent, 1); + if(ret) return ret; + } + + ret = write_unknown_chunks(ctx, SPNG_AFTER_IHDR); + if(ret) return ret; + + if(ctx->stored.plte) + { + for(i=0; i < ctx->plte.n_entries; i++) + { + data[i * 3 + 0] = ctx->plte.entries[i].red; + data[i * 3 + 1] = ctx->plte.entries[i].green; + data[i * 3 + 2] = ctx->plte.entries[i].blue; + } + + ret = write_chunk(ctx, type_plte, data, ctx->plte.n_entries * 3); + if(ret) return ret; + } + + if(ctx->stored.bkgd) + { + switch(ctx->ihdr.color_type) + { + case SPNG_COLOR_TYPE_GRAYSCALE: + case SPNG_COLOR_TYPE_GRAYSCALE_ALPHA: + { + length = 2; + + write_u16(data, ctx->bkgd.gray); + + break; + } + case SPNG_COLOR_TYPE_TRUECOLOR: + case SPNG_COLOR_TYPE_TRUECOLOR_ALPHA: + { + length = 6; + + write_u16(data, ctx->bkgd.red); + write_u16(data + 2, ctx->bkgd.green); + write_u16(data + 4, ctx->bkgd.blue); + + break; + } + case SPNG_COLOR_TYPE_INDEXED: + { + length = 1; + + data[0] = ctx->bkgd.plte_index; + + break; + } + default: return SPNG_EINTERNAL; + } + + ret = write_chunk(ctx, type_bkgd, data, length); + if(ret) return ret; + } + + if(ctx->stored.hist) + { + length = ctx->plte.n_entries * 2; + + for(i=0; i < ctx->plte.n_entries; i++) + { + write_u16(data + i * 2, ctx->hist.frequency[i]); + } + + ret = write_chunk(ctx, type_hist, data, length); + if(ret) return ret; + } + + if(ctx->stored.trns) + { + if(ctx->ihdr.color_type == SPNG_COLOR_TYPE_GRAYSCALE) + { + write_u16(data, ctx->trns.gray); + + ret = write_chunk(ctx, type_trns, data, 2); + } + else if(ctx->ihdr.color_type == SPNG_COLOR_TYPE_TRUECOLOR) + { + write_u16(data, ctx->trns.red); + write_u16(data + 2, ctx->trns.green); + write_u16(data + 4, ctx->trns.blue); + + ret = write_chunk(ctx, type_trns, data, 6); + } + else if(ctx->ihdr.color_type == SPNG_COLOR_TYPE_INDEXED) + { + ret = write_chunk(ctx, type_trns, ctx->trns.type3_alpha, ctx->trns.n_type3_entries); + } + + if(ret) return ret; + } + + if(ctx->stored.phys) + { + write_u32(data, ctx->phys.ppu_x); + write_u32(data + 4, ctx->phys.ppu_y); + data[8] = ctx->phys.unit_specifier; + + ret = write_chunk(ctx, type_phys, data, 9); + if(ret) return ret; + } + + if(ctx->stored.splt) + { + const struct spng_splt *splt; + unsigned char *cdata = NULL; + + uint32_t k; + for(i=0; i < ctx->n_splt; i++) + { + splt = &ctx->splt_list[i]; + + size_t name_len = strlen(splt->name); + length = name_len + 1; + + if(splt->sample_depth == 8) length += splt->n_entries * 6 + 1; + else if(splt->sample_depth == 16) length += splt->n_entries * 10 + 1; + + ret = write_header(ctx, type_splt, length, &cdata); + if(ret) return ret; + + memcpy(cdata, splt->name, name_len + 1); + cdata += name_len + 2; + cdata[-1] = splt->sample_depth; + + if(splt->sample_depth == 8) + { + for(k=0; k < splt->n_entries; k++) + { + cdata[k * 6 + 0] = splt->entries[k].red; + cdata[k * 6 + 1] = splt->entries[k].green; + cdata[k * 6 + 2] = splt->entries[k].blue; + cdata[k * 6 + 3] = splt->entries[k].alpha; + write_u16(cdata + k * 6 + 4, splt->entries[k].frequency); + } + } + else if(splt->sample_depth == 16) + { + for(k=0; k < splt->n_entries; k++) + { + write_u16(cdata + k * 10 + 0, splt->entries[k].red); + write_u16(cdata + k * 10 + 2, splt->entries[k].green); + write_u16(cdata + k * 10 + 4, splt->entries[k].blue); + write_u16(cdata + k * 10 + 6, splt->entries[k].alpha); + write_u16(cdata + k * 10 + 8, splt->entries[k].frequency); + } + } + + ret = finish_chunk(ctx); + if(ret) return ret; + } + } + + if(ctx->stored.time) + { + write_u16(data, ctx->time.year); + data[2] = ctx->time.month; + data[3] = ctx->time.day; + data[4] = ctx->time.hour; + data[5] = ctx->time.minute; + data[6] = ctx->time.second; + + ret = write_chunk(ctx, type_time, data, 7); + if(ret) return ret; + } + + if(ctx->stored.text) + { + unsigned char *cdata = NULL; + const struct spng_text2 *text; + const uint8_t *text_type_array[4] = { 0, type_text, type_ztxt, type_itxt }; + + for(i=0; i < ctx->n_text; i++) + { + text = &ctx->text_list[i]; + + const uint8_t *text_chunk_type = text_type_array[text->type]; + Bytef *compressed_text = NULL; + size_t keyword_len = 0; + size_t language_tag_len = 0; + size_t translated_keyword_len = 0; + size_t compressed_length = 0; + size_t text_length = 0; + + keyword_len = strlen(text->keyword); + text_length = strlen(text->text); + + length = keyword_len + 1; + + if(text->type == SPNG_ZTXT) + { + length += 1; /* compression method */ + } + else if(text->type == SPNG_ITXT) + { + if(!text->language_tag || !text->translated_keyword) return SPNG_EINTERNAL; + + language_tag_len = strlen(text->language_tag); + translated_keyword_len = strlen(text->translated_keyword); + + length += language_tag_len; + if(length < language_tag_len) return SPNG_EOVERFLOW; + + length += translated_keyword_len; + if(length < translated_keyword_len) return SPNG_EOVERFLOW; + + length += 4; /* compression flag + method + nul for the two strings */ + if(length < 4) return SPNG_EOVERFLOW; + } + + if(text->compression_flag) + { + ret = spng__deflate_init(ctx, &ctx->text_options); + if(ret) return ret; + + z_stream *zstream = &ctx->zstream; + uLongf dest_len = deflateBound(zstream, (uLong)text_length); + + compressed_text = spng__malloc(ctx, dest_len); + + if(compressed_text == NULL) return SPNG_EMEM; + + zstream->next_in = (void*)text->text; + zstream->avail_in = (uInt)text_length; + + zstream->next_out = compressed_text; + zstream->avail_out = dest_len; + + ret = deflate(zstream, Z_FINISH); + + if(ret != Z_STREAM_END) + { + spng__free(ctx, compressed_text); + return SPNG_EZLIB; + } + + compressed_length = zstream->total_out; + + length += compressed_length; + if(length < compressed_length) return SPNG_EOVERFLOW; + } + else + { + text_length = strlen(text->text); + + length += text_length; + if(length < text_length) return SPNG_EOVERFLOW; + } + + ret = write_header(ctx, text_chunk_type, length, &cdata); + if(ret) + { + spng__free(ctx, compressed_text); + return ret; + } + + memcpy(cdata, text->keyword, keyword_len + 1); + cdata += keyword_len + 1; + + if(text->type == SPNG_ITXT) + { + cdata[0] = text->compression_flag; + cdata[1] = 0; /* compression method */ + cdata += 2; + + memcpy(cdata, text->language_tag, language_tag_len + 1); + cdata += language_tag_len + 1; + + memcpy(cdata, text->translated_keyword, translated_keyword_len + 1); + cdata += translated_keyword_len + 1; + } + else if(text->type == SPNG_ZTXT) + { + cdata[0] = 0; /* compression method */ + cdata++; + } + + if(text->compression_flag) memcpy(cdata, compressed_text, compressed_length); + else memcpy(cdata, text->text, text_length); + + spng__free(ctx, compressed_text); + + ret = finish_chunk(ctx); + if(ret) return ret; + } + } + + if(ctx->stored.offs) + { + write_s32(data, ctx->offs.x); + write_s32(data + 4, ctx->offs.y); + data[8] = ctx->offs.unit_specifier; + + ret = write_chunk(ctx, type_offs, data, 9); + if(ret) return ret; + } + + if(ctx->stored.exif) + { + ret = write_chunk(ctx, type_exif, ctx->exif.data, ctx->exif.length); + if(ret) return ret; + } + + ret = write_unknown_chunks(ctx, SPNG_AFTER_PLTE); + if(ret) return ret; + + return 0; +} + +static int write_chunks_after_idat(spng_ctx *ctx) +{ + if(ctx == NULL) return SPNG_EINTERNAL; + + int ret = write_unknown_chunks(ctx, SPNG_AFTER_IDAT); + if(ret) return ret; + + return write_iend(ctx); +} + +/* Compress and write scanline to IDAT stream */ +static int write_idat_bytes(spng_ctx *ctx, const void *scanline, size_t len, int flush) +{ + if(ctx == NULL || scanline == NULL) return SPNG_EINTERNAL; + if(len > UINT_MAX) return SPNG_EINTERNAL; + + int ret = 0; + unsigned char *data = NULL; + z_stream *zstream = &ctx->zstream; + uint32_t idat_length = SPNG_WRITE_SIZE; + + zstream->next_in = scanline; + zstream->avail_in = (uInt)len; + + do + { + ret = deflate(zstream, flush); + + if(zstream->avail_out == 0) + { + ret = finish_chunk(ctx); + if(ret) return encode_err(ctx, ret); + + ret = write_header(ctx, type_idat, idat_length, &data); + if(ret) return encode_err(ctx, ret); + + zstream->next_out = data; + zstream->avail_out = idat_length; + } + + }while(zstream->avail_in); + + if(ret != Z_OK) return SPNG_EZLIB; + + return 0; +} + +static int finish_idat(spng_ctx *ctx) +{ + int ret = 0; + unsigned char *data = NULL; + z_stream *zstream = &ctx->zstream; + uint32_t idat_length = SPNG_WRITE_SIZE; + + while(ret != Z_STREAM_END) + { + ret = deflate(zstream, Z_FINISH); + + if(ret) + { + if(ret == Z_STREAM_END) break; + + if(ret != Z_BUF_ERROR) return SPNG_EZLIB; + } + + if(zstream->avail_out == 0) + { + ret = finish_chunk(ctx); + if(ret) return encode_err(ctx, ret); + + ret = write_header(ctx, type_idat, idat_length, &data); + if(ret) return encode_err(ctx, ret); + + zstream->next_out = data; + zstream->avail_out = idat_length; + } + } + + uint32_t trimmed_length = idat_length - zstream->avail_out; + + ret = trim_chunk(ctx, trimmed_length); + if(ret) return ret; + + return finish_chunk(ctx); +} + +static int encode_scanline(spng_ctx *ctx, const void *scanline, size_t len) +{ + if(ctx == NULL || scanline == NULL) return SPNG_EINTERNAL; + + int ret, pass = ctx->row_info.pass; + uint8_t filter = 0; + struct spng_row_info *ri = &ctx->row_info; + const struct spng_subimage *sub = ctx->subimage; + struct encode_flags f = ctx->encode_flags; + unsigned char *filtered_scanline = ctx->filtered_scanline; + size_t scanline_width = sub[pass].scanline_width; + + if(len < scanline_width - 1) return SPNG_EINTERNAL; + + /* encode_row() interlaces directly to ctx->scanline */ + if(scanline != ctx->scanline) memcpy(ctx->scanline, scanline, scanline_width - 1); + + if(f.to_bigendian) u16_row_to_bigendian(ctx->scanline, scanline_width - 1); + const int requires_previous = f.filter_choice & (SPNG_FILTER_CHOICE_UP | SPNG_FILTER_CHOICE_AVG | SPNG_FILTER_CHOICE_PAETH); + + /* XXX: exclude 'requires_previous' filters by default for first scanline? */ + if(!ri->scanline_idx && requires_previous) + { + /* prev_scanline is all zeros for the first scanline */ + memset(ctx->prev_scanline, 0, scanline_width); + } + + filter = get_best_filter(ctx->prev_scanline, ctx->scanline, scanline_width, ctx->bytes_per_pixel, f.filter_choice); + + if(!filter) filtered_scanline = ctx->scanline; + + filtered_scanline[-1] = filter; + + if(filter) + { + ret = filter_scanline(filtered_scanline, ctx->prev_scanline, ctx->scanline, scanline_width, ctx->bytes_per_pixel, filter); + if(ret) return encode_err(ctx, ret); + } + + ret = write_idat_bytes(ctx, filtered_scanline - 1, scanline_width, Z_NO_FLUSH); + if(ret) return encode_err(ctx, ret); + + /* The previous scanline is always unfiltered */ + void *t = ctx->prev_scanline; + ctx->prev_scanline = ctx->scanline; + ctx->scanline = t; + + ret = update_row_info(ctx); + + if(ret == SPNG_EOI) + { + int error = finish_idat(ctx); + if(error) encode_err(ctx, error); + + if(f.finalize) + { + error = spng_encode_chunks(ctx); + if(error) return encode_err(ctx, error); + } + } + + return ret; +} + +static int encode_row(spng_ctx *ctx, const void *row, size_t len) +{ + if(ctx == NULL || row == NULL) return SPNG_EINTERNAL; + + const int pass = ctx->row_info.pass; + + if(!ctx->ihdr.interlace_method || pass == 6) return encode_scanline(ctx, row, len); + + uint32_t k; + const unsigned pixel_size = ctx->pixel_size; + const unsigned bit_depth = ctx->ihdr.bit_depth; + + if(bit_depth < 8) + { + const unsigned samples_per_byte = 8 / bit_depth; + const uint8_t mask = (1 << bit_depth) - 1; + const unsigned initial_shift = 8 - bit_depth; + unsigned shift_amount = initial_shift; + + unsigned char *scanline = ctx->scanline; + const unsigned char *row_uc = row; + uint8_t sample; + + memset(scanline, 0, ctx->subimage[pass].scanline_width); + + for(k=0; k < ctx->subimage[pass].width; k++) + { + size_t ioffset = adam7_x_start[pass] + k * adam7_x_delta[pass]; + + sample = row_uc[ioffset / samples_per_byte]; + + sample = sample >> (initial_shift - ioffset * bit_depth % 8); + sample = sample & mask; + sample = sample << shift_amount; + + scanline[0] |= sample; + + shift_amount -= bit_depth; + + if(shift_amount > 7) + { + shift_amount = initial_shift; + scanline++; + } + } + + return encode_scanline(ctx, ctx->scanline, len); + } + + for(k=0; k < ctx->subimage[pass].width; k++) + { + size_t ioffset = (adam7_x_start[pass] + (size_t) k * adam7_x_delta[pass]) * pixel_size; + + memcpy(ctx->scanline + k * pixel_size, (unsigned char*)row + ioffset, pixel_size); + } + + return encode_scanline(ctx, ctx->scanline, len); +} + +int spng_encode_scanline(spng_ctx *ctx, const void *scanline, size_t len) +{ + if(ctx == NULL || scanline == NULL) return SPNG_EINVAL; + if(ctx->state >= SPNG_STATE_EOI) return SPNG_EOI; + if(len < (ctx->subimage[ctx->row_info.pass].scanline_width -1) ) return SPNG_EBUFSIZ; + + return encode_scanline(ctx, scanline, len); +} + +int spng_encode_row(spng_ctx *ctx, const void *row, size_t len) +{ + if(ctx == NULL || row == NULL) return SPNG_EINVAL; + if(ctx->state >= SPNG_STATE_EOI) return SPNG_EOI; + if(len < ctx->image_width) return SPNG_EBUFSIZ; + + return encode_row(ctx, row, len); +} + +int spng_encode_chunks(spng_ctx *ctx) +{ + if(ctx == NULL) return 1; + if(!ctx->state) return SPNG_EBADSTATE; + if(ctx->state < SPNG_STATE_OUTPUT) return SPNG_ENODST; + if(!ctx->encode_only) return SPNG_ECTXTYPE; + + int ret = 0; + + if(ctx->state < SPNG_STATE_FIRST_IDAT) + { + if(!ctx->stored.ihdr) return SPNG_ENOIHDR; + + ret = write_chunks_before_idat(ctx); + if(ret) return encode_err(ctx, ret); + + ctx->state = SPNG_STATE_FIRST_IDAT; + } + else if(ctx->state == SPNG_STATE_FIRST_IDAT) + { + return 0; + } + else if(ctx->state == SPNG_STATE_EOI) + { + ret = write_chunks_after_idat(ctx); + if(ret) return encode_err(ctx, ret); + + ctx->state = SPNG_STATE_IEND; + } + else return SPNG_EOPSTATE; + + return 0; +} + +int spng_encode_image(spng_ctx *ctx, const void *img, size_t len, int fmt, int flags) +{ + if(ctx == NULL) return 1; + if(!ctx->state) return SPNG_EBADSTATE; + if(!ctx->encode_only) return SPNG_ECTXTYPE; + if(!ctx->stored.ihdr) return SPNG_ENOIHDR; + if( !(fmt == SPNG_FMT_PNG || fmt == SPNG_FMT_RAW) ) return SPNG_EFMT; + + int ret = 0; + const struct spng_ihdr *ihdr = &ctx->ihdr; + struct encode_flags *encode_flags = &ctx->encode_flags; + + if(ihdr->color_type == SPNG_COLOR_TYPE_INDEXED && !ctx->stored.plte) return SPNG_ENOPLTE; + + ret = calculate_image_width(ihdr, fmt, &ctx->image_width); + if(ret) return encode_err(ctx, ret); + + if(ctx->image_width > SIZE_MAX / ihdr->height) ctx->image_size = 0; /* overflow */ + else ctx->image_size = ctx->image_width * ihdr->height; + + if( !(flags & SPNG_ENCODE_PROGRESSIVE) ) + { + if(img == NULL) return 1; + if(!ctx->image_size) return SPNG_EOVERFLOW; + if(len != ctx->image_size) return SPNG_EBUFSIZ; + } + + ret = spng_encode_chunks(ctx); + if(ret) return encode_err(ctx, ret); + + ret = calculate_subimages(ctx); + if(ret) return encode_err(ctx, ret); + + if(ihdr->bit_depth < 8) ctx->bytes_per_pixel = 1; + else ctx->bytes_per_pixel = num_channels(ihdr) * (ihdr->bit_depth / 8); + + if(spng__optimize(SPNG_FILTER_CHOICE)) + { + /* Filtering would make no difference */ + if(!ctx->image_options.compression_level) + { + encode_flags->filter_choice = SPNG_DISABLE_FILTERING; + } + + /* Palette indices and low bit-depth images do not benefit from filtering */ + if(ihdr->color_type == SPNG_COLOR_TYPE_INDEXED || ihdr->bit_depth < 8) + { + encode_flags->filter_choice = SPNG_DISABLE_FILTERING; + } + } + + /* This is technically the same as disabling filtering */ + if(encode_flags->filter_choice == SPNG_FILTER_CHOICE_NONE) + { + encode_flags->filter_choice = SPNG_DISABLE_FILTERING; + } + + if(!encode_flags->filter_choice && spng__optimize(SPNG_IMG_COMPRESSION_STRATEGY)) + { + ctx->image_options.strategy = Z_DEFAULT_STRATEGY; + } + + ret = spng__deflate_init(ctx, &ctx->image_options); + if(ret) return encode_err(ctx, ret); + + size_t scanline_buf_size = ctx->subimage[ctx->widest_pass].scanline_width; + + scanline_buf_size += 32; + + if(scanline_buf_size < 32) return SPNG_EOVERFLOW; + + ctx->scanline_buf = spng__malloc(ctx, scanline_buf_size); + ctx->prev_scanline_buf = spng__malloc(ctx, scanline_buf_size); + + if(ctx->scanline_buf == NULL || ctx->prev_scanline_buf == NULL) return encode_err(ctx, SPNG_EMEM); + + /* Maintain alignment for pixels, filter at [-1] */ + ctx->scanline = ctx->scanline_buf + 16; + ctx->prev_scanline = ctx->prev_scanline_buf + 16; + + if(encode_flags->filter_choice) + { + ctx->filtered_scanline_buf = spng__malloc(ctx, scanline_buf_size); + if(ctx->filtered_scanline_buf == NULL) return encode_err(ctx, SPNG_EMEM); + + ctx->filtered_scanline = ctx->filtered_scanline_buf + 16; + } + + struct spng_subimage *sub = ctx->subimage; + struct spng_row_info *ri = &ctx->row_info; + + ctx->fmt = fmt; + + z_stream *zstream = &ctx->zstream; + zstream->avail_out = SPNG_WRITE_SIZE; + + ret = write_header(ctx, type_idat, zstream->avail_out, &zstream->next_out); + if(ret) return encode_err(ctx, ret); + + if(ihdr->interlace_method) encode_flags->interlace = 1; + + if(fmt & (SPNG_FMT_PNG | SPNG_FMT_RAW) ) encode_flags->same_layout = 1; + + if(ihdr->bit_depth == 16 && fmt != SPNG_FMT_RAW) encode_flags->to_bigendian = 1; + + if(flags & SPNG_ENCODE_FINALIZE) encode_flags->finalize = 1; + + while(!sub[ri->pass].width || !sub[ri->pass].height) ri->pass++; + + if(encode_flags->interlace) ri->row_num = adam7_y_start[ri->pass]; + + ctx->pixel_size = 4; /* SPNG_FMT_RGBA8 */ + + if(fmt == SPNG_FMT_RGBA16) ctx->pixel_size = 8; + else if(fmt == SPNG_FMT_RGB8) ctx->pixel_size = 3; + else if(fmt == SPNG_FMT_G8) ctx->pixel_size = 1; + else if(fmt == SPNG_FMT_GA8) ctx->pixel_size = 2; + else if(fmt & (SPNG_FMT_PNG | SPNG_FMT_RAW)) ctx->pixel_size = ctx->bytes_per_pixel; + + ctx->state = SPNG_STATE_ENCODE_INIT; + + if(flags & SPNG_ENCODE_PROGRESSIVE) + { + encode_flags->progressive = 1; + + return 0; + } + + do + { + size_t ioffset = ri->row_num * ctx->image_width; + + ret = encode_row(ctx, (unsigned char*)img + ioffset, ctx->image_width); + + }while(!ret); + + if(ret != SPNG_EOI) return encode_err(ctx, ret); + + return 0; +} + +spng_ctx *spng_ctx_new(int flags) +{ + struct spng_alloc alloc = + { + .malloc_fn = malloc, + .realloc_fn = realloc, + .calloc_fn = calloc, + .free_fn = free + }; + + return spng_ctx_new2(&alloc, flags); +} + +spng_ctx *spng_ctx_new2(struct spng_alloc *alloc, int flags) +{ + if(alloc == NULL) return NULL; + if(flags != (flags & SPNG__CTX_FLAGS_ALL)) return NULL; + + if(alloc->malloc_fn == NULL) return NULL; + if(alloc->realloc_fn == NULL) return NULL; + if(alloc->calloc_fn == NULL) return NULL; + if(alloc->free_fn == NULL) return NULL; + + spng_ctx *ctx = alloc->calloc_fn(1, sizeof(spng_ctx)); + if(ctx == NULL) return NULL; + + ctx->alloc = *alloc; + + ctx->max_width = spng_u32max; + ctx->max_height = spng_u32max; + + ctx->max_chunk_size = spng_u32max; + ctx->chunk_cache_limit = SIZE_MAX; + ctx->chunk_count_limit = SPNG_MAX_CHUNK_COUNT; + + ctx->state = SPNG_STATE_INIT; + + ctx->crc_action_critical = SPNG_CRC_ERROR; + ctx->crc_action_ancillary = SPNG_CRC_DISCARD; + + const struct spng__zlib_options image_defaults = + { + .compression_level = Z_DEFAULT_COMPRESSION, + .window_bits = 15, + .mem_level = 8, + .strategy = Z_FILTERED, + .data_type = 0 /* Z_BINARY */ + }; + + const struct spng__zlib_options text_defaults = + { + .compression_level = Z_DEFAULT_COMPRESSION, + .window_bits = 15, + .mem_level = 8, + .strategy = Z_DEFAULT_STRATEGY, + .data_type = 1 /* Z_TEXT */ + }; + + ctx->image_options = image_defaults; + ctx->text_options = text_defaults; + + ctx->optimize_option = ~0; + ctx->encode_flags.filter_choice = SPNG_FILTER_CHOICE_ALL; + + ctx->flags = flags; + + if(flags & SPNG_CTX_ENCODER) ctx->encode_only = 1; + + return ctx; +} + +void spng_ctx_free(spng_ctx *ctx) +{ + if(ctx == NULL) return; + + if(ctx->streaming && ctx->stream_buf != NULL) spng__free(ctx, ctx->stream_buf); + + if(!ctx->user.exif) spng__free(ctx, ctx->exif.data); + + if(!ctx->user.iccp) spng__free(ctx, ctx->iccp.profile); + + uint32_t i; + + if(ctx->splt_list != NULL && !ctx->user.splt) + { + for(i=0; i < ctx->n_splt; i++) + { + spng__free(ctx, ctx->splt_list[i].entries); + } + spng__free(ctx, ctx->splt_list); + } + + if(ctx->text_list != NULL) + { + for(i=0; i< ctx->n_text; i++) + { + if(ctx->user.text) break; + + spng__free(ctx, ctx->text_list[i].keyword); + if(ctx->text_list[i].compression_flag) spng__free(ctx, ctx->text_list[i].text); + } + spng__free(ctx, ctx->text_list); + } + + if(ctx->chunk_list != NULL && !ctx->user.unknown) + { + for(i=0; i< ctx->n_chunks; i++) + { + spng__free(ctx, ctx->chunk_list[i].data); + } + spng__free(ctx, ctx->chunk_list); + } + + if(ctx->deflate) deflateEnd(&ctx->zstream); + else inflateEnd(&ctx->zstream); + + if(!ctx->user_owns_out_png) spng__free(ctx, ctx->out_png); + + spng__free(ctx, ctx->gamma_lut16); + + spng__free(ctx, ctx->row_buf); + spng__free(ctx, ctx->scanline_buf); + spng__free(ctx, ctx->prev_scanline_buf); + spng__free(ctx, ctx->filtered_scanline_buf); + + spng_free_fn *free_fn = ctx->alloc.free_fn; + + memset(ctx, 0, sizeof(spng_ctx)); + + free_fn(ctx); +} + +static int buffer_read_fn(spng_ctx *ctx, void *user, void *data, size_t n) +{ + if(n > ctx->bytes_left) return SPNG_IO_EOF; + + (void)user; + (void)data; + ctx->data = ctx->data + ctx->last_read_size; + + ctx->last_read_size = n; + ctx->bytes_left -= n; + + return 0; +} + +static int file_read_fn(spng_ctx *ctx, void *user, void *data, size_t n) +{ + FILE *file = user; + (void)ctx; + + if(fread(data, n, 1, file) != 1) + { + if(feof(file)) return SPNG_IO_EOF; + else return SPNG_IO_ERROR; + } + + return 0; +} + +static int file_write_fn(spng_ctx *ctx, void *user, void *data, size_t n) +{ + FILE *file = user; + (void)ctx; + + if(fwrite(data, n, 1, file) != 1) return SPNG_IO_ERROR; + + return 0; +} + +int spng_set_png_buffer(spng_ctx *ctx, const void *buf, size_t size) +{ + if(ctx == NULL || buf == NULL) return 1; + if(!ctx->state) return SPNG_EBADSTATE; + if(ctx->encode_only) return SPNG_ECTXTYPE; /* not supported */ + + if(ctx->data != NULL) return SPNG_EBUF_SET; + + ctx->data = buf; + ctx->png_base = buf; + ctx->data_size = size; + ctx->bytes_left = size; + + ctx->read_fn = buffer_read_fn; + + ctx->state = SPNG_STATE_INPUT; + + return 0; +} + +int spng_set_png_stream(spng_ctx *ctx, spng_rw_fn *rw_func, void *user) +{ + if(ctx == NULL || rw_func == NULL) return 1; + if(!ctx->state) return SPNG_EBADSTATE; + + /* SPNG_STATE_OUTPUT shares the same value */ + if(ctx->state >= SPNG_STATE_INPUT) return SPNG_EBUF_SET; + + if(ctx->encode_only) + { + if(ctx->out_png != NULL) return SPNG_EBUF_SET; + + ctx->write_fn = rw_func; + ctx->write_ptr = ctx->stream_buf; + + ctx->state = SPNG_STATE_OUTPUT; + } + else + { + ctx->stream_buf = spng__malloc(ctx, SPNG_READ_SIZE); + if(ctx->stream_buf == NULL) return SPNG_EMEM; + + ctx->read_fn = rw_func; + ctx->data = ctx->stream_buf; + ctx->data_size = SPNG_READ_SIZE; + + ctx->state = SPNG_STATE_INPUT; + } + + ctx->stream_user_ptr = user; + + ctx->streaming = 1; + + return 0; +} + +int spng_set_png_file(spng_ctx *ctx, FILE *file) +{ + if(file == NULL) return 1; + + if(ctx->encode_only) return spng_set_png_stream(ctx, file_write_fn, file); + + return spng_set_png_stream(ctx, file_read_fn, file); +} + +void *spng_get_png_buffer(spng_ctx *ctx, size_t *len, int *error) +{ + int tmp = 0; + error = error ? error : &tmp; + *error = 0; + + if(ctx == NULL || !len) *error = SPNG_EINVAL; + + if(*error) return NULL; + + if(!ctx->encode_only) *error = SPNG_ECTXTYPE; + else if(!ctx->state) *error = SPNG_EBADSTATE; + else if(!ctx->internal_buffer) *error = SPNG_EOPSTATE; + else if(ctx->state < SPNG_STATE_EOI) *error = SPNG_EOPSTATE; + else if(ctx->state != SPNG_STATE_IEND) *error = SPNG_ENOTFINAL; + + if(*error) return NULL; + + ctx->user_owns_out_png = 1; + + *len = ctx->bytes_encoded; + + return ctx->out_png; +} + +int spng_set_image_limits(spng_ctx *ctx, uint32_t width, uint32_t height) +{ + if(ctx == NULL) return 1; + + if(width > spng_u32max || height > spng_u32max) return 1; + + ctx->max_width = width; + ctx->max_height = height; + + return 0; +} + +int spng_get_image_limits(spng_ctx *ctx, uint32_t *width, uint32_t *height) +{ + if(ctx == NULL || width == NULL || height == NULL) return 1; + + *width = ctx->max_width; + *height = ctx->max_height; + + return 0; +} + +int spng_set_chunk_limits(spng_ctx *ctx, size_t chunk_size, size_t cache_limit) +{ + if(ctx == NULL || chunk_size > spng_u32max || chunk_size > cache_limit) return 1; + + ctx->max_chunk_size = chunk_size; + + ctx->chunk_cache_limit = cache_limit; + + return 0; +} + +int spng_get_chunk_limits(spng_ctx *ctx, size_t *chunk_size, size_t *cache_limit) +{ + if(ctx == NULL || chunk_size == NULL || cache_limit == NULL) return 1; + + *chunk_size = ctx->max_chunk_size; + + *cache_limit = ctx->chunk_cache_limit; + + return 0; +} + +int spng_set_crc_action(spng_ctx *ctx, int critical, int ancillary) +{ + if(ctx == NULL) return 1; + if(ctx->encode_only) return SPNG_ECTXTYPE; + + if(critical > 2 || critical < 0) return 1; + if(ancillary > 2 || ancillary < 0) return 1; + + if(critical == SPNG_CRC_DISCARD) return 1; + + ctx->crc_action_critical = critical; + ctx->crc_action_ancillary = ancillary; + + return 0; +} + +int spng_set_option(spng_ctx *ctx, enum spng_option option, int value) +{ + if(ctx == NULL) return 1; + if(!ctx->state) return SPNG_EBADSTATE; + + switch(option) + { + case SPNG_KEEP_UNKNOWN_CHUNKS: + { + ctx->keep_unknown = value ? 1 : 0; + break; + } + case SPNG_IMG_COMPRESSION_LEVEL: + { + ctx->image_options.compression_level = value; + break; + } + case SPNG_IMG_WINDOW_BITS: + { + ctx->image_options.window_bits = value; + break; + } + case SPNG_IMG_MEM_LEVEL: + { + ctx->image_options.mem_level = value; + break; + } + case SPNG_IMG_COMPRESSION_STRATEGY: + { + ctx->image_options.strategy = value; + break; + } + case SPNG_TEXT_COMPRESSION_LEVEL: + { + ctx->text_options.compression_level = value; + break; + } + case SPNG_TEXT_WINDOW_BITS: + { + ctx->text_options.window_bits = value; + break; + } + case SPNG_TEXT_MEM_LEVEL: + { + ctx->text_options.mem_level = value; + break; + } + case SPNG_TEXT_COMPRESSION_STRATEGY: + { + ctx->text_options.strategy = value; + break; + } + case SPNG_FILTER_CHOICE: + { + if(value & ~SPNG_FILTER_CHOICE_ALL) return 1; + ctx->encode_flags.filter_choice = value; + break; + } + case SPNG_CHUNK_COUNT_LIMIT: + { + if(value < 0) return 1; + if(value > (int)ctx->chunk_count_total) return 1; + ctx->chunk_count_limit = value; + break; + } + case SPNG_ENCODE_TO_BUFFER: + { + if(value < 0) return 1; + if(!ctx->encode_only) return SPNG_ECTXTYPE; + if(ctx->state >= SPNG_STATE_OUTPUT) return SPNG_EOPSTATE; + + if(!value) break; + + ctx->internal_buffer = 1; + ctx->state = SPNG_STATE_OUTPUT; + + break; + } + default: return 1; + } + + /* Option can no longer be overriden by the library */ + if(option < 32) ctx->optimize_option &= ~(1 << option); + + return 0; +} + +int spng_get_option(spng_ctx *ctx, enum spng_option option, int *value) +{ + if(ctx == NULL || value == NULL) return 1; + if(!ctx->state) return SPNG_EBADSTATE; + + switch(option) + { + case SPNG_KEEP_UNKNOWN_CHUNKS: + { + *value = ctx->keep_unknown; + break; + } + case SPNG_IMG_COMPRESSION_LEVEL: + { + *value = ctx->image_options.compression_level; + break; + } + case SPNG_IMG_WINDOW_BITS: + { + *value = ctx->image_options.window_bits; + break; + } + case SPNG_IMG_MEM_LEVEL: + { + *value = ctx->image_options.mem_level; + break; + } + case SPNG_IMG_COMPRESSION_STRATEGY: + { + *value = ctx->image_options.strategy; + break; + } + case SPNG_TEXT_COMPRESSION_LEVEL: + { + *value = ctx->text_options.compression_level; + break; + } + case SPNG_TEXT_WINDOW_BITS: + { + *value = ctx->text_options.window_bits; + break; + } + case SPNG_TEXT_MEM_LEVEL: + { + *value = ctx->text_options.mem_level; + break; + } + case SPNG_TEXT_COMPRESSION_STRATEGY: + { + *value = ctx->text_options.strategy; + break; + } + case SPNG_FILTER_CHOICE: + { + *value = ctx->encode_flags.filter_choice; + break; + } + case SPNG_CHUNK_COUNT_LIMIT: + { + *value = ctx->chunk_count_limit; + break; + } + case SPNG_ENCODE_TO_BUFFER: + { + if(ctx->internal_buffer) *value = 1; + else *value = 0; + + break; + } + default: return 1; + } + + return 0; +} + +int spng_decoded_image_size(spng_ctx *ctx, int fmt, size_t *len) +{ + if(ctx == NULL || len == NULL) return 1; + + int ret = read_chunks(ctx, 1); + if(ret) return ret; + + ret = check_decode_fmt(&ctx->ihdr, fmt); + if(ret) return ret; + + return calculate_image_size(&ctx->ihdr, fmt, len); +} + +int spng_get_ihdr(spng_ctx *ctx, struct spng_ihdr *ihdr) +{ + if(ctx == NULL) return 1; + int ret = read_chunks(ctx, 1); + if(ret) return ret; + if(ihdr == NULL) return 1; + + *ihdr = ctx->ihdr; + + return 0; +} + +int spng_get_plte(spng_ctx *ctx, struct spng_plte *plte) +{ + SPNG_GET_CHUNK_BOILERPLATE(plte); + + *plte = ctx->plte; + + return 0; +} + +int spng_get_trns(spng_ctx *ctx, struct spng_trns *trns) +{ + SPNG_GET_CHUNK_BOILERPLATE(trns); + + *trns = ctx->trns; + + return 0; +} + +int spng_get_chrm(spng_ctx *ctx, struct spng_chrm *chrm) +{ + SPNG_GET_CHUNK_BOILERPLATE(chrm); + + chrm->white_point_x = (double)ctx->chrm_int.white_point_x / 100000.0; + chrm->white_point_y = (double)ctx->chrm_int.white_point_y / 100000.0; + chrm->red_x = (double)ctx->chrm_int.red_x / 100000.0; + chrm->red_y = (double)ctx->chrm_int.red_y / 100000.0; + chrm->blue_y = (double)ctx->chrm_int.blue_y / 100000.0; + chrm->blue_x = (double)ctx->chrm_int.blue_x / 100000.0; + chrm->green_x = (double)ctx->chrm_int.green_x / 100000.0; + chrm->green_y = (double)ctx->chrm_int.green_y / 100000.0; + + return 0; +} + +int spng_get_chrm_int(spng_ctx *ctx, struct spng_chrm_int *chrm) +{ + SPNG_GET_CHUNK_BOILERPLATE(chrm); + + *chrm = ctx->chrm_int; + + return 0; +} + +int spng_get_gama(spng_ctx *ctx, double *gamma) +{ + double *gama = gamma; + SPNG_GET_CHUNK_BOILERPLATE(gama); + + *gama = (double)ctx->gama / 100000.0; + + return 0; +} + +int spng_get_gama_int(spng_ctx *ctx, uint32_t *gama_int) +{ + uint32_t *gama = gama_int; + SPNG_GET_CHUNK_BOILERPLATE(gama); + + *gama_int = ctx->gama; + + return 0; +} + +int spng_get_iccp(spng_ctx *ctx, struct spng_iccp *iccp) +{ + SPNG_GET_CHUNK_BOILERPLATE(iccp); + + *iccp = ctx->iccp; + + return 0; +} + +int spng_get_sbit(spng_ctx *ctx, struct spng_sbit *sbit) +{ + SPNG_GET_CHUNK_BOILERPLATE(sbit); + + *sbit = ctx->sbit; + + return 0; +} + +int spng_get_srgb(spng_ctx *ctx, uint8_t *rendering_intent) +{ + uint8_t *srgb = rendering_intent; + SPNG_GET_CHUNK_BOILERPLATE(srgb); + + *srgb = ctx->srgb_rendering_intent; + + return 0; +} + +int spng_get_text(spng_ctx *ctx, struct spng_text *text, uint32_t *n_text) +{ + if(ctx == NULL) return 1; + int ret = read_chunks(ctx, 0); + if(ret) return ret; + if(!ctx->stored.text) return SPNG_ECHUNKAVAIL; + if(n_text == NULL) return 1; + + if(text == NULL) + { + *n_text = ctx->n_text; + return 0; + } + + if(*n_text < ctx->n_text) return 1; + + uint32_t i; + for(i=0; i< ctx->n_text; i++) + { + text[i].type = ctx->text_list[i].type; + memcpy(&text[i].keyword, ctx->text_list[i].keyword, strlen(ctx->text_list[i].keyword) + 1); + text[i].compression_method = 0; + text[i].compression_flag = ctx->text_list[i].compression_flag; + text[i].language_tag = ctx->text_list[i].language_tag; + text[i].translated_keyword = ctx->text_list[i].translated_keyword; + text[i].length = ctx->text_list[i].text_length; + text[i].text = ctx->text_list[i].text; + } + + return ret; +} + +int spng_get_bkgd(spng_ctx *ctx, struct spng_bkgd *bkgd) +{ + SPNG_GET_CHUNK_BOILERPLATE(bkgd); + + *bkgd = ctx->bkgd; + + return 0; +} + +int spng_get_hist(spng_ctx *ctx, struct spng_hist *hist) +{ + SPNG_GET_CHUNK_BOILERPLATE(hist); + + *hist = ctx->hist; + + return 0; +} + +int spng_get_phys(spng_ctx *ctx, struct spng_phys *phys) +{ + SPNG_GET_CHUNK_BOILERPLATE(phys); + + *phys = ctx->phys; + + return 0; +} + +int spng_get_splt(spng_ctx *ctx, struct spng_splt *splt, uint32_t *n_splt) +{ + if(ctx == NULL) return 1; + int ret = read_chunks(ctx, 0); + if(ret) return ret; + if(!ctx->stored.splt) return SPNG_ECHUNKAVAIL; + if(n_splt == NULL) return 1; + + if(splt == NULL) + { + *n_splt = ctx->n_splt; + return 0; + } + + if(*n_splt < ctx->n_splt) return 1; + + memcpy(splt, ctx->splt_list, ctx->n_splt * sizeof(struct spng_splt)); + + return 0; +} + +int spng_get_time(spng_ctx *ctx, struct spng_time *time) +{ + SPNG_GET_CHUNK_BOILERPLATE(time); + + *time = ctx->time; + + return 0; +} + +int spng_get_unknown_chunks(spng_ctx *ctx, struct spng_unknown_chunk *chunks, uint32_t *n_chunks) +{ + if(ctx == NULL) return 1; + int ret = read_chunks(ctx, 0); + if(ret) return ret; + if(!ctx->stored.unknown) return SPNG_ECHUNKAVAIL; + if(n_chunks == NULL) return 1; + + if(chunks == NULL) + { + *n_chunks = ctx->n_chunks; + return 0; + } + + if(*n_chunks < ctx->n_chunks) return 1; + + memcpy(chunks, ctx->chunk_list, sizeof(struct spng_unknown_chunk)); + + return 0; +} + +int spng_get_offs(spng_ctx *ctx, struct spng_offs *offs) +{ + SPNG_GET_CHUNK_BOILERPLATE(offs); + + *offs = ctx->offs; + + return 0; +} + +int spng_get_exif(spng_ctx *ctx, struct spng_exif *exif) +{ + SPNG_GET_CHUNK_BOILERPLATE(exif); + + *exif = ctx->exif; + + return 0; +} + +int spng_set_ihdr(spng_ctx *ctx, struct spng_ihdr *ihdr) +{ + SPNG_SET_CHUNK_BOILERPLATE(ihdr); + + if(ctx->stored.ihdr) return 1; + + ret = check_ihdr(ihdr, ctx->max_width, ctx->max_height); + if(ret) return ret; + + ctx->ihdr = *ihdr; + + ctx->stored.ihdr = 1; + ctx->user.ihdr = 1; + + return 0; +} + +int spng_set_plte(spng_ctx *ctx, struct spng_plte *plte) +{ + SPNG_SET_CHUNK_BOILERPLATE(plte); + + if(!ctx->stored.ihdr) return 1; + + if(check_plte(plte, &ctx->ihdr)) return 1; + + ctx->plte.n_entries = plte->n_entries; + + memcpy(ctx->plte.entries, plte->entries, plte->n_entries * sizeof(struct spng_plte_entry)); + + ctx->stored.plte = 1; + ctx->user.plte = 1; + + return 0; +} + +int spng_set_trns(spng_ctx *ctx, struct spng_trns *trns) +{ + SPNG_SET_CHUNK_BOILERPLATE(trns); + + if(!ctx->stored.ihdr) return SPNG_ENOIHDR; + + if(ctx->ihdr.color_type == SPNG_COLOR_TYPE_GRAYSCALE) + { + ctx->trns.gray = trns->gray; + } + else if(ctx->ihdr.color_type == SPNG_COLOR_TYPE_TRUECOLOR) + { + ctx->trns.red = trns->red; + ctx->trns.green = trns->green; + ctx->trns.blue = trns->blue; + } + else if(ctx->ihdr.color_type == SPNG_COLOR_TYPE_INDEXED) + { + if(!ctx->stored.plte) return SPNG_ETRNS_NO_PLTE; + if(trns->n_type3_entries > ctx->plte.n_entries) return 1; + + ctx->trns.n_type3_entries = trns->n_type3_entries; + memcpy(ctx->trns.type3_alpha, trns->type3_alpha, trns->n_type3_entries); + } + else return SPNG_ETRNS_COLOR_TYPE; + + ctx->stored.trns = 1; + ctx->user.trns = 1; + + return 0; +} + +int spng_set_chrm(spng_ctx *ctx, struct spng_chrm *chrm) +{ + SPNG_SET_CHUNK_BOILERPLATE(chrm); + + struct spng_chrm_int chrm_int; + + chrm_int.white_point_x = (uint32_t)(chrm->white_point_x * 100000.0); + chrm_int.white_point_y = (uint32_t)(chrm->white_point_y * 100000.0); + chrm_int.red_x = (uint32_t)(chrm->red_x * 100000.0); + chrm_int.red_y = (uint32_t)(chrm->red_y * 100000.0); + chrm_int.green_x = (uint32_t)(chrm->green_x * 100000.0); + chrm_int.green_y = (uint32_t)(chrm->green_y * 100000.0); + chrm_int.blue_x = (uint32_t)(chrm->blue_x * 100000.0); + chrm_int.blue_y = (uint32_t)(chrm->blue_y * 100000.0); + + if(check_chrm_int(&chrm_int)) return SPNG_ECHRM; + + ctx->chrm_int = chrm_int; + + ctx->stored.chrm = 1; + ctx->user.chrm = 1; + + return 0; +} + +int spng_set_chrm_int(spng_ctx *ctx, struct spng_chrm_int *chrm_int) +{ + SPNG_SET_CHUNK_BOILERPLATE(chrm_int); + + if(check_chrm_int(chrm_int)) return SPNG_ECHRM; + + ctx->chrm_int = *chrm_int; + + ctx->stored.chrm = 1; + ctx->user.chrm = 1; + + return 0; +} + +int spng_set_gama(spng_ctx *ctx, double gamma) +{ + SPNG_SET_CHUNK_BOILERPLATE(ctx); + + uint32_t gama = gamma * 100000.0; + + if(!gama) return 1; + if(gama > spng_u32max) return 1; + + ctx->gama = gama; + + ctx->stored.gama = 1; + ctx->user.gama = 1; + + return 0; +} + +int spng_set_gama_int(spng_ctx *ctx, uint32_t gamma) +{ + SPNG_SET_CHUNK_BOILERPLATE(ctx); + + if(!gamma) return 1; + if(gamma > spng_u32max) return 1; + + ctx->gama = gamma; + + ctx->stored.gama = 1; + ctx->user.gama = 1; + + return 0; +} + +int spng_set_iccp(spng_ctx *ctx, struct spng_iccp *iccp) +{ + SPNG_SET_CHUNK_BOILERPLATE(iccp); + + if(check_png_keyword(iccp->profile_name)) return SPNG_EICCP_NAME; + if(!iccp->profile_len) return SPNG_ECHUNK_SIZE; + if(iccp->profile_len > spng_u32max) return SPNG_ECHUNK_STDLEN; + + if(ctx->iccp.profile && !ctx->user.iccp) spng__free(ctx, ctx->iccp.profile); + + ctx->iccp = *iccp; + + ctx->stored.iccp = 1; + ctx->user.iccp = 1; + + return 0; +} + +int spng_set_sbit(spng_ctx *ctx, struct spng_sbit *sbit) +{ + SPNG_SET_CHUNK_BOILERPLATE(sbit); + + if(check_sbit(sbit, &ctx->ihdr)) return 1; + + if(!ctx->stored.ihdr) return 1; + + ctx->sbit = *sbit; + + ctx->stored.sbit = 1; + ctx->user.sbit = 1; + + return 0; +} + +int spng_set_srgb(spng_ctx *ctx, uint8_t rendering_intent) +{ + SPNG_SET_CHUNK_BOILERPLATE(ctx); + + if(rendering_intent > 3) return 1; + + ctx->srgb_rendering_intent = rendering_intent; + + ctx->stored.srgb = 1; + ctx->user.srgb = 1; + + return 0; +} + +int spng_set_text(spng_ctx *ctx, struct spng_text *text, uint32_t n_text) +{ + if(!n_text) return 1; + SPNG_SET_CHUNK_BOILERPLATE(text); + + uint32_t i; + for(i=0; i < n_text; i++) + { + if(check_png_keyword(text[i].keyword)) return SPNG_ETEXT_KEYWORD; + if(!text[i].length) return 1; + if(text[i].length > UINT_MAX) return 1; + if(text[i].text == NULL) return 1; + + if(text[i].type == SPNG_TEXT) + { + if(ctx->strict && check_png_text(text[i].text, text[i].length)) return 1; + } + else if(text[i].type == SPNG_ZTXT) + { + if(ctx->strict && check_png_text(text[i].text, text[i].length)) return 1; + + if(text[i].compression_method != 0) return SPNG_EZTXT_COMPRESSION_METHOD; + } + else if(text[i].type == SPNG_ITXT) + { + if(text[i].compression_flag > 1) return SPNG_EITXT_COMPRESSION_FLAG; + if(text[i].compression_method != 0) return SPNG_EITXT_COMPRESSION_METHOD; + if(text[i].language_tag == NULL) return SPNG_EITXT_LANG_TAG; + if(text[i].translated_keyword == NULL) return SPNG_EITXT_TRANSLATED_KEY; + } + else return 1; + + } + + struct spng_text2 *text_list = spng__calloc(ctx, sizeof(struct spng_text2), n_text); + + if(!text_list) return SPNG_EMEM; + + if(ctx->text_list != NULL) + { + for(i=0; i < ctx->n_text; i++) + { + if(ctx->user.text) break; + + spng__free(ctx, ctx->text_list[i].keyword); + if(ctx->text_list[i].compression_flag) spng__free(ctx, ctx->text_list[i].text); + } + spng__free(ctx, ctx->text_list); + } + + for(i=0; i < n_text; i++) + { + text_list[i].type = text[i].type; + /* Prevent issues with spng_text.keyword[80] going out of scope */ + text_list[i].keyword = text_list[i].user_keyword_storage; + memcpy(text_list[i].user_keyword_storage, text[i].keyword, strlen(text[i].keyword)); + text_list[i].text = text[i].text; + text_list[i].text_length = text[i].length; + + if(text[i].type == SPNG_ZTXT) + { + text_list[i].compression_flag = 1; + } + else if(text[i].type == SPNG_ITXT) + { + text_list[i].compression_flag = text[i].compression_flag; + text_list[i].language_tag = text[i].language_tag; + text_list[i].translated_keyword = text[i].translated_keyword; + } + } + + ctx->text_list = text_list; + ctx->n_text = n_text; + + ctx->stored.text = 1; + ctx->user.text = 1; + + return 0; +} + +int spng_set_bkgd(spng_ctx *ctx, struct spng_bkgd *bkgd) +{ + SPNG_SET_CHUNK_BOILERPLATE(bkgd); + + if(!ctx->stored.ihdr) return 1; + + if(ctx->ihdr.color_type == 0 || ctx->ihdr.color_type == 4) + { + ctx->bkgd.gray = bkgd->gray; + } + else if(ctx->ihdr.color_type == 2 || ctx->ihdr.color_type == 6) + { + ctx->bkgd.red = bkgd->red; + ctx->bkgd.green = bkgd->green; + ctx->bkgd.blue = bkgd->blue; + } + else if(ctx->ihdr.color_type == 3) + { + if(!ctx->stored.plte) return SPNG_EBKGD_NO_PLTE; + if(bkgd->plte_index >= ctx->plte.n_entries) return SPNG_EBKGD_PLTE_IDX; + + ctx->bkgd.plte_index = bkgd->plte_index; + } + + ctx->stored.bkgd = 1; + ctx->user.bkgd = 1; + + return 0; +} + +int spng_set_hist(spng_ctx *ctx, struct spng_hist *hist) +{ + SPNG_SET_CHUNK_BOILERPLATE(hist); + + if(!ctx->stored.plte) return SPNG_EHIST_NO_PLTE; + + ctx->hist = *hist; + + ctx->stored.hist = 1; + ctx->user.hist = 1; + + return 0; +} + +int spng_set_phys(spng_ctx *ctx, struct spng_phys *phys) +{ + SPNG_SET_CHUNK_BOILERPLATE(phys); + + if(check_phys(phys)) return SPNG_EPHYS; + + ctx->phys = *phys; + + ctx->stored.phys = 1; + ctx->user.phys = 1; + + return 0; +} + +int spng_set_splt(spng_ctx *ctx, struct spng_splt *splt, uint32_t n_splt) +{ + if(!n_splt) return 1; + SPNG_SET_CHUNK_BOILERPLATE(splt); + + uint32_t i; + for(i=0; i < n_splt; i++) + { + if(check_png_keyword(splt[i].name)) return SPNG_ESPLT_NAME; + if( !(splt[i].sample_depth == 8 || splt[i].sample_depth == 16) ) return SPNG_ESPLT_DEPTH; + } + + if(ctx->stored.splt && !ctx->user.splt) + { + for(i=0; i < ctx->n_splt; i++) + { + if(ctx->splt_list[i].entries != NULL) spng__free(ctx, ctx->splt_list[i].entries); + } + spng__free(ctx, ctx->splt_list); + } + + ctx->splt_list = splt; + ctx->n_splt = n_splt; + + ctx->stored.splt = 1; + ctx->user.splt = 1; + + return 0; +} + +int spng_set_time(spng_ctx *ctx, struct spng_time *time) +{ + SPNG_SET_CHUNK_BOILERPLATE(time); + + if(check_time(time)) return SPNG_ETIME; + + ctx->time = *time; + + ctx->stored.time = 1; + ctx->user.time = 1; + + return 0; +} + +int spng_set_unknown_chunks(spng_ctx *ctx, struct spng_unknown_chunk *chunks, uint32_t n_chunks) +{ + if(!n_chunks) return 1; + SPNG_SET_CHUNK_BOILERPLATE(chunks); + + uint32_t i; + for(i=0; i < n_chunks; i++) + { + if(chunks[i].length > spng_u32max) return SPNG_ECHUNK_STDLEN; + if(chunks[i].length && chunks[i].data == NULL) return 1; + + switch(chunks[i].location) + { + case SPNG_AFTER_IHDR: + case SPNG_AFTER_PLTE: + case SPNG_AFTER_IDAT: + break; + default: return SPNG_ECHUNK_POS; + } + } + + if(ctx->stored.unknown && !ctx->user.unknown) + { + for(i=0; i < ctx->n_chunks; i++) + { + spng__free(ctx, ctx->chunk_list[i].data); + } + spng__free(ctx, ctx->chunk_list); + } + + ctx->chunk_list = chunks; + ctx->n_chunks = n_chunks; + + ctx->stored.unknown = 1; + ctx->user.unknown = 1; + + return 0; +} + +int spng_set_offs(spng_ctx *ctx, struct spng_offs *offs) +{ + SPNG_SET_CHUNK_BOILERPLATE(offs); + + if(check_offs(offs)) return SPNG_EOFFS; + + ctx->offs = *offs; + + ctx->stored.offs = 1; + ctx->user.offs = 1; + + return 0; +} + +int spng_set_exif(spng_ctx *ctx, struct spng_exif *exif) +{ + SPNG_SET_CHUNK_BOILERPLATE(exif); + + if(check_exif(exif)) return SPNG_EEXIF; + + if(ctx->exif.data != NULL && !ctx->user.exif) spng__free(ctx, ctx->exif.data); + + ctx->exif = *exif; + + ctx->stored.exif = 1; + ctx->user.exif = 1; + + return 0; +} + +const char *spng_strerror(int err) +{ + switch(err) + { + case SPNG_IO_EOF: return "end of stream"; + case SPNG_IO_ERROR: return "stream error"; + case SPNG_OK: return "success"; + case SPNG_EINVAL: return "invalid argument"; + case SPNG_EMEM: return "out of memory"; + case SPNG_EOVERFLOW: return "arithmetic overflow"; + case SPNG_ESIGNATURE: return "invalid signature"; + case SPNG_EWIDTH: return "invalid image width"; + case SPNG_EHEIGHT: return "invalid image height"; + case SPNG_EUSER_WIDTH: return "image width exceeds user limit"; + case SPNG_EUSER_HEIGHT: return "image height exceeds user limit"; + case SPNG_EBIT_DEPTH: return "invalid bit depth"; + case SPNG_ECOLOR_TYPE: return "invalid color type"; + case SPNG_ECOMPRESSION_METHOD: return "invalid compression method"; + case SPNG_EFILTER_METHOD: return "invalid filter method"; + case SPNG_EINTERLACE_METHOD: return "invalid interlace method"; + case SPNG_EIHDR_SIZE: return "invalid IHDR chunk size"; + case SPNG_ENOIHDR: return "missing IHDR chunk"; + case SPNG_ECHUNK_POS: return "invalid chunk position"; + case SPNG_ECHUNK_SIZE: return "invalid chunk length"; + case SPNG_ECHUNK_CRC: return "invalid chunk checksum"; + case SPNG_ECHUNK_TYPE: return "invalid chunk type"; + case SPNG_ECHUNK_UNKNOWN_CRITICAL: return "unknown critical chunk"; + case SPNG_EDUP_PLTE: return "duplicate PLTE chunk"; + case SPNG_EDUP_CHRM: return "duplicate cHRM chunk"; + case SPNG_EDUP_GAMA: return "duplicate gAMA chunk"; + case SPNG_EDUP_ICCP: return "duplicate iCCP chunk"; + case SPNG_EDUP_SBIT: return "duplicate sBIT chunk"; + case SPNG_EDUP_SRGB: return "duplicate sRGB chunk"; + case SPNG_EDUP_BKGD: return "duplicate bKGD chunk"; + case SPNG_EDUP_HIST: return "duplicate hIST chunk"; + case SPNG_EDUP_TRNS: return "duplicate tRNS chunk"; + case SPNG_EDUP_PHYS: return "duplicate pHYs chunk"; + case SPNG_EDUP_TIME: return "duplicate tIME chunk"; + case SPNG_EDUP_OFFS: return "duplicate oFFs chunk"; + case SPNG_EDUP_EXIF: return "duplicate eXIf chunk"; + case SPNG_ECHRM: return "invalid cHRM chunk"; + case SPNG_EPLTE_IDX: return "invalid palette (PLTE) index"; + case SPNG_ETRNS_COLOR_TYPE: return "tRNS chunk with incompatible color type"; + case SPNG_ETRNS_NO_PLTE: return "missing palette (PLTE) for tRNS chunk"; + case SPNG_EGAMA: return "invalid gAMA chunk"; + case SPNG_EICCP_NAME: return "invalid iCCP profile name"; + case SPNG_EICCP_COMPRESSION_METHOD: return "invalid iCCP compression method"; + case SPNG_ESBIT: return "invalid sBIT chunk"; + case SPNG_ESRGB: return "invalid sRGB chunk"; + case SPNG_ETEXT: return "invalid tEXt chunk"; + case SPNG_ETEXT_KEYWORD: return "invalid tEXt keyword"; + case SPNG_EZTXT: return "invalid zTXt chunk"; + case SPNG_EZTXT_COMPRESSION_METHOD: return "invalid zTXt compression method"; + case SPNG_EITXT: return "invalid iTXt chunk"; + case SPNG_EITXT_COMPRESSION_FLAG: return "invalid iTXt compression flag"; + case SPNG_EITXT_COMPRESSION_METHOD: return "invalid iTXt compression method"; + case SPNG_EITXT_LANG_TAG: return "invalid iTXt language tag"; + case SPNG_EITXT_TRANSLATED_KEY: return "invalid iTXt translated key"; + case SPNG_EBKGD_NO_PLTE: return "missing palette for bKGD chunk"; + case SPNG_EBKGD_PLTE_IDX: return "invalid palette index for bKGD chunk"; + case SPNG_EHIST_NO_PLTE: return "missing palette for hIST chunk"; + case SPNG_EPHYS: return "invalid pHYs chunk"; + case SPNG_ESPLT_NAME: return "invalid suggested palette name"; + case SPNG_ESPLT_DUP_NAME: return "duplicate suggested palette (sPLT) name"; + case SPNG_ESPLT_DEPTH: return "invalid suggested palette (sPLT) sample depth"; + case SPNG_ETIME: return "invalid tIME chunk"; + case SPNG_EOFFS: return "invalid oFFs chunk"; + case SPNG_EEXIF: return "invalid eXIf chunk"; + case SPNG_EIDAT_TOO_SHORT: return "IDAT stream too short"; + case SPNG_EIDAT_STREAM: return "IDAT stream error"; + case SPNG_EZLIB: return "zlib error"; + case SPNG_EFILTER: return "invalid scanline filter"; + case SPNG_EBUFSIZ: return "invalid buffer size"; + case SPNG_EIO: return "i/o error"; + case SPNG_EOF: return "end of file"; + case SPNG_EBUF_SET: return "buffer already set"; + case SPNG_EBADSTATE: return "non-recoverable state"; + case SPNG_EFMT: return "invalid format"; + case SPNG_EFLAGS: return "invalid flags"; + case SPNG_ECHUNKAVAIL: return "chunk not available"; + case SPNG_ENCODE_ONLY: return "encode only context"; + case SPNG_EOI: return "reached end-of-image state"; + case SPNG_ENOPLTE: return "missing PLTE for indexed image"; + case SPNG_ECHUNK_LIMITS: return "reached chunk/cache limits"; + case SPNG_EZLIB_INIT: return "zlib init error"; + case SPNG_ECHUNK_STDLEN: return "chunk exceeds maximum standard length"; + case SPNG_EINTERNAL: return "internal error"; + case SPNG_ECTXTYPE: return "invalid operation for context type"; + case SPNG_ENOSRC: return "source PNG not set"; + case SPNG_ENODST: return "PNG output not set"; + case SPNG_EOPSTATE: return "invalid operation for state"; + case SPNG_ENOTFINAL: return "PNG not finalized"; + default: return "unknown error"; + } +} + +const char *spng_version_string(void) +{ + return SPNG_VERSION_STRING; +} + +#if defined(_MSC_VER) + #pragma warning(pop) +#endif + +/* The following SIMD optimizations are derived from libpng source code. */ + +/* +* PNG Reference Library License version 2 +* +* Copyright (c) 1995-2019 The PNG Reference Library Authors. +* Copyright (c) 2018-2019 Cosmin Truta. +* Copyright (c) 2000-2002, 2004, 2006-2018 Glenn Randers-Pehrson. +* Copyright (c) 1996-1997 Andreas Dilger. +* Copyright (c) 1995-1996 Guy Eric Schalnat, Group 42, Inc. +* +* The software is supplied "as is", without warranty of any kind, +* express or implied, including, without limitation, the warranties +* of merchantability, fitness for a particular purpose, title, and +* non-infringement. In no event shall the Copyright owners, or +* anyone distributing the software, be liable for any damages or +* other liability, whether in contract, tort or otherwise, arising +* from, out of, or in connection with the software, or the use or +* other dealings in the software, even if advised of the possibility +* of such damage. +* +* Permission is hereby granted to use, copy, modify, and distribute +* this software, or portions hereof, for any purpose, without fee, +* subject to the following restrictions: +* +* 1. The origin of this software must not be misrepresented; you +* must not claim that you wrote the original software. If you +* use this software in a product, an acknowledgment in the product +* documentation would be appreciated, but is not required. +* +* 2. Altered source versions must be plainly marked as such, and must +* not be misrepresented as being the original software. +* +* 3. This Copyright notice may not be removed or altered from any +* source or altered source distribution. +*/ + +#if defined(SPNG_X86) + +#ifndef SPNG_SSE + #define SPNG_SSE 1 +#endif + +#if defined(__GNUC__) && !defined(__clang__) + #if SPNG_SSE == 3 + #pragma GCC target("ssse3") + #elif SPNG_SSE == 4 + #pragma GCC target("sse4.1") + #else + #pragma GCC target("sse2") + #endif +#endif + +/* SSE2 optimised filter functions + * Derived from filter_neon_intrinsics.c + * + * Copyright (c) 2018 Cosmin Truta + * Copyright (c) 2016-2017 Glenn Randers-Pehrson + * Written by Mike Klein and Matt Sarett + * Derived from arm/filter_neon_intrinsics.c + * + * This code is derived from libpng source code. + * For conditions of distribution and use, see the disclaimer + * and license above. + */ + +#include +#include +#include + +/* Functions in this file look at most 3 pixels (a,b,c) to predict the 4th (d). + * They're positioned like this: + * prev: c b + * row: a d + * The Sub filter predicts d=a, Avg d=(a+b)/2, and Paeth predicts d to be + * whichever of a, b, or c is closest to p=a+b-c. + */ + +static __m128i load4(const void* p) +{ + int tmp; + memcpy(&tmp, p, sizeof(tmp)); + return _mm_cvtsi32_si128(tmp); +} + +static void store4(void* p, __m128i v) +{ + int tmp = _mm_cvtsi128_si32(v); + memcpy(p, &tmp, sizeof(int)); +} + +static __m128i load3(const void* p) +{ + uint32_t tmp = 0; + memcpy(&tmp, p, 3); + return _mm_cvtsi32_si128(tmp); +} + +static void store3(void* p, __m128i v) +{ + int tmp = _mm_cvtsi128_si32(v); + memcpy(p, &tmp, 3); +} + +static void defilter_sub3(size_t rowbytes, unsigned char *row) +{ + /* The Sub filter predicts each pixel as the previous pixel, a. + * There is no pixel to the left of the first pixel. It's encoded directly. + * That works with our main loop if we just say that left pixel was zero. + */ + size_t rb = rowbytes; + + __m128i a, d = _mm_setzero_si128(); + + while(rb >= 4) + { + a = d; d = load4(row); + d = _mm_add_epi8(d, a); + store3(row, d); + + row += 3; + rb -= 3; + } + + if(rb > 0) + { + a = d; d = load3(row); + d = _mm_add_epi8(d, a); + store3(row, d); + } +} + +static void defilter_sub4(size_t rowbytes, unsigned char *row) +{ + /* The Sub filter predicts each pixel as the previous pixel, a. + * There is no pixel to the left of the first pixel. It's encoded directly. + * That works with our main loop if we just say that left pixel was zero. + */ + size_t rb = rowbytes+4; + + __m128i a, d = _mm_setzero_si128(); + + while(rb > 4) + { + a = d; d = load4(row); + d = _mm_add_epi8(d, a); + store4(row, d); + + row += 4; + rb -= 4; + } +} + +static void defilter_avg3(size_t rowbytes, unsigned char *row, const unsigned char *prev) +{ + /* The Avg filter predicts each pixel as the (truncated) average of a and b. + * There's no pixel to the left of the first pixel. Luckily, it's + * predicted to be half of the pixel above it. So again, this works + * perfectly with our loop if we make sure a starts at zero. + */ + + size_t rb = rowbytes; + + const __m128i zero = _mm_setzero_si128(); + + __m128i b; + __m128i a, d = zero; + + while(rb >= 4) + { + __m128i avg; + b = load4(prev); + a = d; d = load4(row ); + + /* PNG requires a truncating average, so we can't just use _mm_avg_epu8 */ + avg = _mm_avg_epu8(a,b); + /* ...but we can fix it up by subtracting off 1 if it rounded up. */ + avg = _mm_sub_epi8(avg, _mm_and_si128(_mm_xor_si128(a, b), + _mm_set1_epi8(1))); + d = _mm_add_epi8(d, avg); + store3(row, d); + + prev += 3; + row += 3; + rb -= 3; + } + + if(rb > 0) + { + __m128i avg; + b = load3(prev); + a = d; d = load3(row ); + + /* PNG requires a truncating average, so we can't just use _mm_avg_epu8 */ + avg = _mm_avg_epu8(a, b); + /* ...but we can fix it up by subtracting off 1 if it rounded up. */ + avg = _mm_sub_epi8(avg, _mm_and_si128(_mm_xor_si128(a, b), + _mm_set1_epi8(1))); + + d = _mm_add_epi8(d, avg); + store3(row, d); + } +} + +static void defilter_avg4(size_t rowbytes, unsigned char *row, const unsigned char *prev) +{ + /* The Avg filter predicts each pixel as the (truncated) average of a and b. + * There's no pixel to the left of the first pixel. Luckily, it's + * predicted to be half of the pixel above it. So again, this works + * perfectly with our loop if we make sure a starts at zero. + */ + size_t rb = rowbytes+4; + + const __m128i zero = _mm_setzero_si128(); + __m128i b; + __m128i a, d = zero; + + while(rb > 4) + { + __m128i avg; + b = load4(prev); + a = d; d = load4(row ); + + /* PNG requires a truncating average, so we can't just use _mm_avg_epu8 */ + avg = _mm_avg_epu8(a,b); + /* ...but we can fix it up by subtracting off 1 if it rounded up. */ + avg = _mm_sub_epi8(avg, _mm_and_si128(_mm_xor_si128(a, b), + _mm_set1_epi8(1))); + + d = _mm_add_epi8(d, avg); + store4(row, d); + + prev += 4; + row += 4; + rb -= 4; + } +} + +/* Returns |x| for 16-bit lanes. */ +#if (SPNG_SSE >= 3) && !defined(_MSC_VER) +__attribute__((target("ssse3"))) +#endif +static __m128i abs_i16(__m128i x) +{ +#if SPNG_SSE >= 3 + return _mm_abs_epi16(x); +#else + /* Read this all as, return x<0 ? -x : x. + * To negate two's complement, you flip all the bits then add 1. + */ + __m128i is_negative = _mm_cmplt_epi16(x, _mm_setzero_si128()); + + /* Flip negative lanes. */ + x = _mm_xor_si128(x, is_negative); + + /* +1 to negative lanes, else +0. */ + x = _mm_sub_epi16(x, is_negative); + return x; +#endif +} + +/* Bytewise c ? t : e. */ +static __m128i if_then_else(__m128i c, __m128i t, __m128i e) +{ +#if SPNG_SSE >= 4 + return _mm_blendv_epi8(e, t, c); +#else + return _mm_or_si128(_mm_and_si128(c, t), _mm_andnot_si128(c, e)); +#endif +} + +static void defilter_paeth3(size_t rowbytes, unsigned char *row, const unsigned char *prev) +{ + /* Paeth tries to predict pixel d using the pixel to the left of it, a, + * and two pixels from the previous row, b and c: + * prev: c b + * row: a d + * The Paeth function predicts d to be whichever of a, b, or c is nearest to + * p=a+b-c. + * + * The first pixel has no left context, and so uses an Up filter, p = b. + * This works naturally with our main loop's p = a+b-c if we force a and c + * to zero. + * Here we zero b and d, which become c and a respectively at the start of + * the loop. + */ + size_t rb = rowbytes; + const __m128i zero = _mm_setzero_si128(); + __m128i c, b = zero, + a, d = zero; + + while(rb >= 4) + { + /* It's easiest to do this math (particularly, deal with pc) with 16-bit + * intermediates. + */ + __m128i pa,pb,pc,smallest,nearest; + c = b; b = _mm_unpacklo_epi8(load4(prev), zero); + a = d; d = _mm_unpacklo_epi8(load4(row ), zero); + + /* (p-a) == (a+b-c - a) == (b-c) */ + + pa = _mm_sub_epi16(b, c); + + /* (p-b) == (a+b-c - b) == (a-c) */ + pb = _mm_sub_epi16(a, c); + + /* (p-c) == (a+b-c - c) == (a+b-c-c) == (b-c)+(a-c) */ + pc = _mm_add_epi16(pa, pb); + + pa = abs_i16(pa); /* |p-a| */ + pb = abs_i16(pb); /* |p-b| */ + pc = abs_i16(pc); /* |p-c| */ + + smallest = _mm_min_epi16(pc, _mm_min_epi16(pa, pb)); + + /* Paeth breaks ties favoring a over b over c. */ + nearest = if_then_else(_mm_cmpeq_epi16(smallest, pa), a, + if_then_else(_mm_cmpeq_epi16(smallest, pb), b, c)); + + /* Note `_epi8`: we need addition to wrap modulo 255. */ + d = _mm_add_epi8(d, nearest); + store3(row, _mm_packus_epi16(d, d)); + + prev += 3; + row += 3; + rb -= 3; + } + + if(rb > 0) + { + /* It's easiest to do this math (particularly, deal with pc) with 16-bit + * intermediates. + */ + __m128i pa, pb, pc, smallest, nearest; + c = b; b = _mm_unpacklo_epi8(load3(prev), zero); + a = d; d = _mm_unpacklo_epi8(load3(row ), zero); + + /* (p-a) == (a+b-c - a) == (b-c) */ + pa = _mm_sub_epi16(b, c); + + /* (p-b) == (a+b-c - b) == (a-c) */ + pb = _mm_sub_epi16(a, c); + + /* (p-c) == (a+b-c - c) == (a+b-c-c) == (b-c)+(a-c) */ + pc = _mm_add_epi16(pa, pb); + + pa = abs_i16(pa); /* |p-a| */ + pb = abs_i16(pb); /* |p-b| */ + pc = abs_i16(pc); /* |p-c| */ + + smallest = _mm_min_epi16(pc, _mm_min_epi16(pa, pb)); + + /* Paeth breaks ties favoring a over b over c. */ + nearest = if_then_else(_mm_cmpeq_epi16(smallest, pa), a, + if_then_else(_mm_cmpeq_epi16(smallest, pb), b, c)); + + /* Note `_epi8`: we need addition to wrap modulo 255. */ + d = _mm_add_epi8(d, nearest); + store3(row, _mm_packus_epi16(d, d)); + } +} + +static void defilter_paeth4(size_t rowbytes, unsigned char *row, const unsigned char *prev) +{ + /* Paeth tries to predict pixel d using the pixel to the left of it, a, + * and two pixels from the previous row, b and c: + * prev: c b + * row: a d + * The Paeth function predicts d to be whichever of a, b, or c is nearest to + * p=a+b-c. + * + * The first pixel has no left context, and so uses an Up filter, p = b. + * This works naturally with our main loop's p = a+b-c if we force a and c + * to zero. + * Here we zero b and d, which become c and a respectively at the start of + * the loop. + */ + size_t rb = rowbytes+4; + + const __m128i zero = _mm_setzero_si128(); + __m128i pa, pb, pc, smallest, nearest; + __m128i c, b = zero, + a, d = zero; + + while(rb > 4) + { + /* It's easiest to do this math (particularly, deal with pc) with 16-bit + * intermediates. + */ + c = b; b = _mm_unpacklo_epi8(load4(prev), zero); + a = d; d = _mm_unpacklo_epi8(load4(row ), zero); + + /* (p-a) == (a+b-c - a) == (b-c) */ + pa = _mm_sub_epi16(b, c); + + /* (p-b) == (a+b-c - b) == (a-c) */ + pb = _mm_sub_epi16(a, c); + + /* (p-c) == (a+b-c - c) == (a+b-c-c) == (b-c)+(a-c) */ + pc = _mm_add_epi16(pa, pb); + + pa = abs_i16(pa); /* |p-a| */ + pb = abs_i16(pb); /* |p-b| */ + pc = abs_i16(pc); /* |p-c| */ + + smallest = _mm_min_epi16(pc, _mm_min_epi16(pa, pb)); + + /* Paeth breaks ties favoring a over b over c. */ + nearest = if_then_else(_mm_cmpeq_epi16(smallest, pa), a, + if_then_else(_mm_cmpeq_epi16(smallest, pb), b, c)); + + /* Note `_epi8`: we need addition to wrap modulo 255. */ + d = _mm_add_epi8(d, nearest); + store4(row, _mm_packus_epi16(d, d)); + + prev += 4; + row += 4; + rb -= 4; + } +} + +#endif /* SPNG_X86 */ + + +#if defined(SPNG_ARM) + +/* NEON optimised filter functions + * Derived from filter_neon_intrinsics.c + * + * Copyright (c) 2018 Cosmin Truta + * Copyright (c) 2014,2016 Glenn Randers-Pehrson + * Written by James Yu , October 2013. + * Based on filter_neon.S, written by Mans Rullgard, 2011. + * + * This code is derived from libpng source code. + * For conditions of distribution and use, see the disclaimer + * and license in this file. + */ + +#define png_aligncast(type, value) ((void*)(value)) +#define png_aligncastconst(type, value) ((const void*)(value)) + +/* libpng row pointers are not necessarily aligned to any particular boundary, + * however this code will only work with appropriate alignment. mips/mips_init.c + * checks for this (and will not compile unless it is done). This code uses + * variants of png_aligncast to avoid compiler warnings. + */ +#define png_ptr(type,pointer) png_aligncast(type *,pointer) +#define png_ptrc(type,pointer) png_aligncastconst(const type *,pointer) + +/* The following relies on a variable 'temp_pointer' being declared with type + * 'type'. This is written this way just to hide the GCC strict aliasing + * warning; note that the code is safe because there never is an alias between + * the input and output pointers. + */ +#define png_ldr(type,pointer)\ + (temp_pointer = png_ptr(type,pointer), *temp_pointer) + + +#if defined(_MSC_VER) && !defined(__clang__) && defined(_M_ARM64) + #include +#else + #include +#endif + +static void defilter_sub3(size_t rowbytes, unsigned char *row) +{ + unsigned char *rp = row; + unsigned char *rp_stop = row + rowbytes; + + uint8x16_t vtmp = vld1q_u8(rp); + uint8x8x2_t *vrpt = png_ptr(uint8x8x2_t, &vtmp); + uint8x8x2_t vrp = *vrpt; + + uint8x8x4_t vdest; + vdest.val[3] = vdup_n_u8(0); + + for (; rp < rp_stop;) + { + uint8x8_t vtmp1, vtmp2; + uint32x2_t *temp_pointer; + + vtmp1 = vext_u8(vrp.val[0], vrp.val[1], 3); + vdest.val[0] = vadd_u8(vdest.val[3], vrp.val[0]); + vtmp2 = vext_u8(vrp.val[0], vrp.val[1], 6); + vdest.val[1] = vadd_u8(vdest.val[0], vtmp1); + + vtmp1 = vext_u8(vrp.val[1], vrp.val[1], 1); + vdest.val[2] = vadd_u8(vdest.val[1], vtmp2); + vdest.val[3] = vadd_u8(vdest.val[2], vtmp1); + + vtmp = vld1q_u8(rp + 12); + vrpt = png_ptr(uint8x8x2_t, &vtmp); + vrp = *vrpt; + + vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[0]), 0); + rp += 3; + vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[1]), 0); + rp += 3; + vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[2]), 0); + rp += 3; + vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[3]), 0); + rp += 3; + } +} + +static void defilter_sub4(size_t rowbytes, unsigned char *row) +{ + unsigned char *rp = row; + unsigned char *rp_stop = row + rowbytes; + + uint8x8x4_t vdest; + vdest.val[3] = vdup_n_u8(0); + + for (; rp < rp_stop; rp += 16) + { + uint32x2x4_t vtmp = vld4_u32(png_ptr(uint32_t,rp)); + uint8x8x4_t *vrpt = png_ptr(uint8x8x4_t,&vtmp); + uint8x8x4_t vrp = *vrpt; + uint32x2x4_t *temp_pointer; + uint32x2x4_t vdest_val; + + vdest.val[0] = vadd_u8(vdest.val[3], vrp.val[0]); + vdest.val[1] = vadd_u8(vdest.val[0], vrp.val[1]); + vdest.val[2] = vadd_u8(vdest.val[1], vrp.val[2]); + vdest.val[3] = vadd_u8(vdest.val[2], vrp.val[3]); + + vdest_val = png_ldr(uint32x2x4_t, &vdest); + vst4_lane_u32(png_ptr(uint32_t,rp), vdest_val, 0); + } +} + +static void defilter_avg3(size_t rowbytes, unsigned char *row, const unsigned char *prev_row) +{ + unsigned char *rp = row; + const unsigned char *pp = prev_row; + unsigned char *rp_stop = row + rowbytes; + + uint8x16_t vtmp; + uint8x8x2_t *vrpt; + uint8x8x2_t vrp; + uint8x8x4_t vdest; + vdest.val[3] = vdup_n_u8(0); + + vtmp = vld1q_u8(rp); + vrpt = png_ptr(uint8x8x2_t,&vtmp); + vrp = *vrpt; + + for (; rp < rp_stop; pp += 12) + { + uint8x8_t vtmp1, vtmp2, vtmp3; + + uint8x8x2_t *vppt; + uint8x8x2_t vpp; + + uint32x2_t *temp_pointer; + + vtmp = vld1q_u8(pp); + vppt = png_ptr(uint8x8x2_t,&vtmp); + vpp = *vppt; + + vtmp1 = vext_u8(vrp.val[0], vrp.val[1], 3); + vdest.val[0] = vhadd_u8(vdest.val[3], vpp.val[0]); + vdest.val[0] = vadd_u8(vdest.val[0], vrp.val[0]); + + vtmp2 = vext_u8(vpp.val[0], vpp.val[1], 3); + vtmp3 = vext_u8(vrp.val[0], vrp.val[1], 6); + vdest.val[1] = vhadd_u8(vdest.val[0], vtmp2); + vdest.val[1] = vadd_u8(vdest.val[1], vtmp1); + + vtmp2 = vext_u8(vpp.val[0], vpp.val[1], 6); + vtmp1 = vext_u8(vrp.val[1], vrp.val[1], 1); + + vtmp = vld1q_u8(rp + 12); + vrpt = png_ptr(uint8x8x2_t,&vtmp); + vrp = *vrpt; + + vdest.val[2] = vhadd_u8(vdest.val[1], vtmp2); + vdest.val[2] = vadd_u8(vdest.val[2], vtmp3); + + vtmp2 = vext_u8(vpp.val[1], vpp.val[1], 1); + + vdest.val[3] = vhadd_u8(vdest.val[2], vtmp2); + vdest.val[3] = vadd_u8(vdest.val[3], vtmp1); + + vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[0]), 0); + rp += 3; + vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[1]), 0); + rp += 3; + vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[2]), 0); + rp += 3; + vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[3]), 0); + rp += 3; + } +} + +static void defilter_avg4(size_t rowbytes, unsigned char *row, const unsigned char *prev_row) +{ + unsigned char *rp = row; + unsigned char *rp_stop = row + rowbytes; + const unsigned char *pp = prev_row; + + uint8x8x4_t vdest; + vdest.val[3] = vdup_n_u8(0); + + for (; rp < rp_stop; rp += 16, pp += 16) + { + uint32x2x4_t vtmp; + uint8x8x4_t *vrpt, *vppt; + uint8x8x4_t vrp, vpp; + uint32x2x4_t *temp_pointer; + uint32x2x4_t vdest_val; + + vtmp = vld4_u32(png_ptr(uint32_t,rp)); + vrpt = png_ptr(uint8x8x4_t,&vtmp); + vrp = *vrpt; + vtmp = vld4_u32(png_ptrc(uint32_t,pp)); + vppt = png_ptr(uint8x8x4_t,&vtmp); + vpp = *vppt; + + vdest.val[0] = vhadd_u8(vdest.val[3], vpp.val[0]); + vdest.val[0] = vadd_u8(vdest.val[0], vrp.val[0]); + vdest.val[1] = vhadd_u8(vdest.val[0], vpp.val[1]); + vdest.val[1] = vadd_u8(vdest.val[1], vrp.val[1]); + vdest.val[2] = vhadd_u8(vdest.val[1], vpp.val[2]); + vdest.val[2] = vadd_u8(vdest.val[2], vrp.val[2]); + vdest.val[3] = vhadd_u8(vdest.val[2], vpp.val[3]); + vdest.val[3] = vadd_u8(vdest.val[3], vrp.val[3]); + + vdest_val = png_ldr(uint32x2x4_t, &vdest); + vst4_lane_u32(png_ptr(uint32_t,rp), vdest_val, 0); + } +} + +static uint8x8_t paeth_arm(uint8x8_t a, uint8x8_t b, uint8x8_t c) +{ + uint8x8_t d, e; + uint16x8_t p1, pa, pb, pc; + + p1 = vaddl_u8(a, b); /* a + b */ + pc = vaddl_u8(c, c); /* c * 2 */ + pa = vabdl_u8(b, c); /* pa */ + pb = vabdl_u8(a, c); /* pb */ + pc = vabdq_u16(p1, pc); /* pc */ + + p1 = vcleq_u16(pa, pb); /* pa <= pb */ + pa = vcleq_u16(pa, pc); /* pa <= pc */ + pb = vcleq_u16(pb, pc); /* pb <= pc */ + + p1 = vandq_u16(p1, pa); /* pa <= pb && pa <= pc */ + + d = vmovn_u16(pb); + e = vmovn_u16(p1); + + d = vbsl_u8(d, b, c); + e = vbsl_u8(e, a, d); + + return e; +} + +static void defilter_paeth3(size_t rowbytes, unsigned char *row, const unsigned char *prev_row) +{ + unsigned char *rp = row; + const unsigned char *pp = prev_row; + unsigned char *rp_stop = row + rowbytes; + + uint8x16_t vtmp; + uint8x8x2_t *vrpt; + uint8x8x2_t vrp; + uint8x8_t vlast = vdup_n_u8(0); + uint8x8x4_t vdest; + vdest.val[3] = vdup_n_u8(0); + + vtmp = vld1q_u8(rp); + vrpt = png_ptr(uint8x8x2_t,&vtmp); + vrp = *vrpt; + + for (; rp < rp_stop; pp += 12) + { + uint8x8x2_t *vppt; + uint8x8x2_t vpp; + uint8x8_t vtmp1, vtmp2, vtmp3; + uint32x2_t *temp_pointer; + + vtmp = vld1q_u8(pp); + vppt = png_ptr(uint8x8x2_t,&vtmp); + vpp = *vppt; + + vdest.val[0] = paeth_arm(vdest.val[3], vpp.val[0], vlast); + vdest.val[0] = vadd_u8(vdest.val[0], vrp.val[0]); + + vtmp1 = vext_u8(vrp.val[0], vrp.val[1], 3); + vtmp2 = vext_u8(vpp.val[0], vpp.val[1], 3); + vdest.val[1] = paeth_arm(vdest.val[0], vtmp2, vpp.val[0]); + vdest.val[1] = vadd_u8(vdest.val[1], vtmp1); + + vtmp1 = vext_u8(vrp.val[0], vrp.val[1], 6); + vtmp3 = vext_u8(vpp.val[0], vpp.val[1], 6); + vdest.val[2] = paeth_arm(vdest.val[1], vtmp3, vtmp2); + vdest.val[2] = vadd_u8(vdest.val[2], vtmp1); + + vtmp1 = vext_u8(vrp.val[1], vrp.val[1], 1); + vtmp2 = vext_u8(vpp.val[1], vpp.val[1], 1); + + vtmp = vld1q_u8(rp + 12); + vrpt = png_ptr(uint8x8x2_t,&vtmp); + vrp = *vrpt; + + vdest.val[3] = paeth_arm(vdest.val[2], vtmp2, vtmp3); + vdest.val[3] = vadd_u8(vdest.val[3], vtmp1); + + vlast = vtmp2; + + vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[0]), 0); + rp += 3; + vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[1]), 0); + rp += 3; + vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[2]), 0); + rp += 3; + vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[3]), 0); + rp += 3; + } +} + +static void defilter_paeth4(size_t rowbytes, unsigned char *row, const unsigned char *prev_row) +{ + unsigned char *rp = row; + unsigned char *rp_stop = row + rowbytes; + const unsigned char *pp = prev_row; + + uint8x8_t vlast = vdup_n_u8(0); + uint8x8x4_t vdest; + vdest.val[3] = vdup_n_u8(0); + + for (; rp < rp_stop; rp += 16, pp += 16) + { + uint32x2x4_t vtmp; + uint8x8x4_t *vrpt, *vppt; + uint8x8x4_t vrp, vpp; + uint32x2x4_t *temp_pointer; + uint32x2x4_t vdest_val; + + vtmp = vld4_u32(png_ptr(uint32_t,rp)); + vrpt = png_ptr(uint8x8x4_t,&vtmp); + vrp = *vrpt; + vtmp = vld4_u32(png_ptrc(uint32_t,pp)); + vppt = png_ptr(uint8x8x4_t,&vtmp); + vpp = *vppt; + + vdest.val[0] = paeth_arm(vdest.val[3], vpp.val[0], vlast); + vdest.val[0] = vadd_u8(vdest.val[0], vrp.val[0]); + vdest.val[1] = paeth_arm(vdest.val[0], vpp.val[1], vpp.val[0]); + vdest.val[1] = vadd_u8(vdest.val[1], vrp.val[1]); + vdest.val[2] = paeth_arm(vdest.val[1], vpp.val[2], vpp.val[1]); + vdest.val[2] = vadd_u8(vdest.val[2], vrp.val[2]); + vdest.val[3] = paeth_arm(vdest.val[2], vpp.val[3], vpp.val[2]); + vdest.val[3] = vadd_u8(vdest.val[3], vrp.val[3]); + + vlast = vpp.val[3]; + + vdest_val = png_ldr(uint32x2x4_t, &vdest); + vst4_lane_u32(png_ptr(uint32_t,rp), vdest_val, 0); + } +} + +/* NEON optimised palette expansion functions + * Derived from palette_neon_intrinsics.c + * + * Copyright (c) 2018-2019 Cosmin Truta + * Copyright (c) 2017-2018 Arm Holdings. All rights reserved. + * Written by Richard Townsend , February 2017. + * + * This code is derived from libpng source code. + * For conditions of distribution and use, see the disclaimer + * and license in this file. + * + * Related: https://developer.arm.com/documentation/101964/latest/Color-palette-expansion + * + * The functions were refactored to iterate forward. + * + */ + +/* Expands a palettized row into RGBA8. */ +static uint32_t expand_palette_rgba8_neon(unsigned char *row, const unsigned char *scanline, const unsigned char *plte, uint32_t width) +{ + const uint32_t scanline_stride = 4; + const uint32_t row_stride = scanline_stride * 4; + const uint32_t count = width / scanline_stride; + const uint32_t *palette = (const uint32_t*)plte; + + if(!count) return 0; + + uint32_t i; + uint32x4_t cur; + for(i=0; i < count; i++, scanline += scanline_stride) + { + cur = vld1q_dup_u32 (palette + scanline[0]); + cur = vld1q_lane_u32(palette + scanline[1], cur, 1); + cur = vld1q_lane_u32(palette + scanline[2], cur, 2); + cur = vld1q_lane_u32(palette + scanline[3], cur, 3); + vst1q_u32((uint32_t*)(row + i * row_stride), cur); + } + + return count * scanline_stride; +} + +/* Expands a palettized row into RGB8. */ +static uint32_t expand_palette_rgb8_neon(unsigned char *row, const unsigned char *scanline, const unsigned char *plte, uint32_t width) +{ + const uint32_t scanline_stride = 8; + const uint32_t row_stride = scanline_stride * 3; + const uint32_t count = width / scanline_stride; + + if(!count) return 0; + + uint32_t i; + uint8x8x3_t cur; + for(i=0; i < count; i++, scanline += scanline_stride) + { + cur = vld3_dup_u8 (plte + 3 * scanline[0]); + cur = vld3_lane_u8(plte + 3 * scanline[1], cur, 1); + cur = vld3_lane_u8(plte + 3 * scanline[2], cur, 2); + cur = vld3_lane_u8(plte + 3 * scanline[3], cur, 3); + cur = vld3_lane_u8(plte + 3 * scanline[4], cur, 4); + cur = vld3_lane_u8(plte + 3 * scanline[5], cur, 5); + cur = vld3_lane_u8(plte + 3 * scanline[6], cur, 6); + cur = vld3_lane_u8(plte + 3 * scanline[7], cur, 7); + vst3_u8(row + i * row_stride, cur); + } + + return count * scanline_stride; +} + +#endif /* SPNG_ARM */ diff --git a/cmake/externals/libspng/spng.h b/cmake/externals/libspng/spng.h new file mode 100644 index 000000000..8f946337b --- /dev/null +++ b/cmake/externals/libspng/spng.h @@ -0,0 +1,537 @@ +/* SPDX-License-Identifier: BSD-2-Clause */ +#ifndef SPNG_H +#define SPNG_H + +#ifdef __cplusplus +extern "C" { +#endif + +#if (defined(_WIN32) || defined(__CYGWIN__)) && !defined(SPNG_STATIC) + #if defined(SPNG__BUILD) + #define SPNG_API __declspec(dllexport) + #else + #define SPNG_API __declspec(dllimport) + #endif +#else + #define SPNG_API +#endif + +#if defined(_MSC_VER) + #define SPNG_CDECL __cdecl +#else + #define SPNG_CDECL +#endif + +#include +#include +#include + +#define SPNG_VERSION_MAJOR 0 +#define SPNG_VERSION_MINOR 7 +#define SPNG_VERSION_PATCH 4 + +enum spng_errno +{ + SPNG_IO_ERROR = -2, + SPNG_IO_EOF = -1, + SPNG_OK = 0, + SPNG_EINVAL, + SPNG_EMEM, + SPNG_EOVERFLOW, + SPNG_ESIGNATURE, + SPNG_EWIDTH, + SPNG_EHEIGHT, + SPNG_EUSER_WIDTH, + SPNG_EUSER_HEIGHT, + SPNG_EBIT_DEPTH, + SPNG_ECOLOR_TYPE, + SPNG_ECOMPRESSION_METHOD, + SPNG_EFILTER_METHOD, + SPNG_EINTERLACE_METHOD, + SPNG_EIHDR_SIZE, + SPNG_ENOIHDR, + SPNG_ECHUNK_POS, + SPNG_ECHUNK_SIZE, + SPNG_ECHUNK_CRC, + SPNG_ECHUNK_TYPE, + SPNG_ECHUNK_UNKNOWN_CRITICAL, + SPNG_EDUP_PLTE, + SPNG_EDUP_CHRM, + SPNG_EDUP_GAMA, + SPNG_EDUP_ICCP, + SPNG_EDUP_SBIT, + SPNG_EDUP_SRGB, + SPNG_EDUP_BKGD, + SPNG_EDUP_HIST, + SPNG_EDUP_TRNS, + SPNG_EDUP_PHYS, + SPNG_EDUP_TIME, + SPNG_EDUP_OFFS, + SPNG_EDUP_EXIF, + SPNG_ECHRM, + SPNG_EPLTE_IDX, + SPNG_ETRNS_COLOR_TYPE, + SPNG_ETRNS_NO_PLTE, + SPNG_EGAMA, + SPNG_EICCP_NAME, + SPNG_EICCP_COMPRESSION_METHOD, + SPNG_ESBIT, + SPNG_ESRGB, + SPNG_ETEXT, + SPNG_ETEXT_KEYWORD, + SPNG_EZTXT, + SPNG_EZTXT_COMPRESSION_METHOD, + SPNG_EITXT, + SPNG_EITXT_COMPRESSION_FLAG, + SPNG_EITXT_COMPRESSION_METHOD, + SPNG_EITXT_LANG_TAG, + SPNG_EITXT_TRANSLATED_KEY, + SPNG_EBKGD_NO_PLTE, + SPNG_EBKGD_PLTE_IDX, + SPNG_EHIST_NO_PLTE, + SPNG_EPHYS, + SPNG_ESPLT_NAME, + SPNG_ESPLT_DUP_NAME, + SPNG_ESPLT_DEPTH, + SPNG_ETIME, + SPNG_EOFFS, + SPNG_EEXIF, + SPNG_EIDAT_TOO_SHORT, + SPNG_EIDAT_STREAM, + SPNG_EZLIB, + SPNG_EFILTER, + SPNG_EBUFSIZ, + SPNG_EIO, + SPNG_EOF, + SPNG_EBUF_SET, + SPNG_EBADSTATE, + SPNG_EFMT, + SPNG_EFLAGS, + SPNG_ECHUNKAVAIL, + SPNG_ENCODE_ONLY, + SPNG_EOI, + SPNG_ENOPLTE, + SPNG_ECHUNK_LIMITS, + SPNG_EZLIB_INIT, + SPNG_ECHUNK_STDLEN, + SPNG_EINTERNAL, + SPNG_ECTXTYPE, + SPNG_ENOSRC, + SPNG_ENODST, + SPNG_EOPSTATE, + SPNG_ENOTFINAL, +}; + +enum spng_text_type +{ + SPNG_TEXT = 1, + SPNG_ZTXT = 2, + SPNG_ITXT = 3 +}; + +enum spng_color_type +{ + SPNG_COLOR_TYPE_GRAYSCALE = 0, + SPNG_COLOR_TYPE_TRUECOLOR = 2, + SPNG_COLOR_TYPE_INDEXED = 3, + SPNG_COLOR_TYPE_GRAYSCALE_ALPHA = 4, + SPNG_COLOR_TYPE_TRUECOLOR_ALPHA = 6 +}; + +enum spng_filter +{ + SPNG_FILTER_NONE = 0, + SPNG_FILTER_SUB = 1, + SPNG_FILTER_UP = 2, + SPNG_FILTER_AVERAGE = 3, + SPNG_FILTER_PAETH = 4 +}; + +enum spng_filter_choice +{ + SPNG_DISABLE_FILTERING = 0, + SPNG_FILTER_CHOICE_NONE = 8, + SPNG_FILTER_CHOICE_SUB = 16, + SPNG_FILTER_CHOICE_UP = 32, + SPNG_FILTER_CHOICE_AVG = 64, + SPNG_FILTER_CHOICE_PAETH = 128, + SPNG_FILTER_CHOICE_ALL = (8|16|32|64|128) +}; + +enum spng_interlace_method +{ + SPNG_INTERLACE_NONE = 0, + SPNG_INTERLACE_ADAM7 = 1 +}; + +/* Channels are always in byte-order */ +enum spng_format +{ + SPNG_FMT_RGBA8 = 1, + SPNG_FMT_RGBA16 = 2, + SPNG_FMT_RGB8 = 4, + + /* Partially implemented, see documentation */ + SPNG_FMT_GA8 = 16, + SPNG_FMT_GA16 = 32, + SPNG_FMT_G8 = 64, + + /* No conversion or scaling */ + SPNG_FMT_PNG = 256, + SPNG_FMT_RAW = 512 /* big-endian (everything else is host-endian) */ +}; + +enum spng_ctx_flags +{ + SPNG_CTX_IGNORE_ADLER32 = 1, /* Ignore checksum in DEFLATE streams */ + SPNG_CTX_ENCODER = 2 /* Create an encoder context */ +}; + +enum spng_decode_flags +{ + SPNG_DECODE_USE_TRNS = 1, /* Deprecated */ + SPNG_DECODE_USE_GAMA = 2, /* Deprecated */ + SPNG_DECODE_USE_SBIT = 8, /* Undocumented */ + + SPNG_DECODE_TRNS = 1, /* Apply transparency */ + SPNG_DECODE_GAMMA = 2, /* Apply gamma correction */ + SPNG_DECODE_PROGRESSIVE = 256 /* Initialize for progressive reads */ +}; + +enum spng_crc_action +{ + /* Default for critical chunks */ + SPNG_CRC_ERROR = 0, + + /* Discard chunk, invalid for critical chunks. + Since v0.6.2: default for ancillary chunks */ + SPNG_CRC_DISCARD = 1, + + /* Ignore and don't calculate checksum. + Since v0.6.2: also ignores checksums in DEFLATE streams */ + SPNG_CRC_USE = 2 +}; + +enum spng_encode_flags +{ + SPNG_ENCODE_PROGRESSIVE = 1, /* Initialize for progressive writes */ + SPNG_ENCODE_FINALIZE = 2, /* Finalize PNG after encoding image */ +}; + +struct spng_ihdr +{ + uint32_t width; + uint32_t height; + uint8_t bit_depth; + uint8_t color_type; + uint8_t compression_method; + uint8_t filter_method; + uint8_t interlace_method; +}; + +struct spng_plte_entry +{ + uint8_t red; + uint8_t green; + uint8_t blue; + + uint8_t alpha; /* Reserved for internal use */ +}; + +struct spng_plte +{ + uint32_t n_entries; + struct spng_plte_entry entries[256]; +}; + +struct spng_trns +{ + uint16_t gray; + + uint16_t red; + uint16_t green; + uint16_t blue; + + uint32_t n_type3_entries; + uint8_t type3_alpha[256]; +}; + +struct spng_chrm_int +{ + uint32_t white_point_x; + uint32_t white_point_y; + uint32_t red_x; + uint32_t red_y; + uint32_t green_x; + uint32_t green_y; + uint32_t blue_x; + uint32_t blue_y; +}; + +struct spng_chrm +{ + double white_point_x; + double white_point_y; + double red_x; + double red_y; + double green_x; + double green_y; + double blue_x; + double blue_y; +}; + +struct spng_iccp +{ + char profile_name[80]; + size_t profile_len; + char *profile; +}; + +struct spng_sbit +{ + uint8_t grayscale_bits; + uint8_t red_bits; + uint8_t green_bits; + uint8_t blue_bits; + uint8_t alpha_bits; +}; + +struct spng_text +{ + char keyword[80]; + int type; + + size_t length; + char *text; + + uint8_t compression_flag; /* iTXt only */ + uint8_t compression_method; /* iTXt, ztXt only */ + char *language_tag; /* iTXt only */ + char *translated_keyword; /* iTXt only */ +}; + +struct spng_bkgd +{ + uint16_t gray; /* Only for gray/gray alpha */ + uint16_t red; + uint16_t green; + uint16_t blue; + uint16_t plte_index; /* Only for indexed color */ +}; + +struct spng_hist +{ + uint16_t frequency[256]; +}; + +struct spng_phys +{ + uint32_t ppu_x, ppu_y; + uint8_t unit_specifier; +}; + +struct spng_splt_entry +{ + uint16_t red; + uint16_t green; + uint16_t blue; + uint16_t alpha; + uint16_t frequency; +}; + +struct spng_splt +{ + char name[80]; + uint8_t sample_depth; + uint32_t n_entries; + struct spng_splt_entry *entries; +}; + +struct spng_time +{ + uint16_t year; + uint8_t month; + uint8_t day; + uint8_t hour; + uint8_t minute; + uint8_t second; +}; + +struct spng_offs +{ + int32_t x, y; + uint8_t unit_specifier; +}; + +struct spng_exif +{ + size_t length; + char *data; +}; + +struct spng_chunk +{ + size_t offset; + uint32_t length; + uint8_t type[4]; + uint32_t crc; +}; + +enum spng_location +{ + SPNG_AFTER_IHDR = 1, + SPNG_AFTER_PLTE = 2, + SPNG_AFTER_IDAT = 8, +}; + +struct spng_unknown_chunk +{ + uint8_t type[4]; + size_t length; + void *data; + enum spng_location location; +}; + +enum spng_option +{ + SPNG_KEEP_UNKNOWN_CHUNKS = 1, + + SPNG_IMG_COMPRESSION_LEVEL, + SPNG_IMG_WINDOW_BITS, + SPNG_IMG_MEM_LEVEL, + SPNG_IMG_COMPRESSION_STRATEGY, + + SPNG_TEXT_COMPRESSION_LEVEL, + SPNG_TEXT_WINDOW_BITS, + SPNG_TEXT_MEM_LEVEL, + SPNG_TEXT_COMPRESSION_STRATEGY, + + SPNG_FILTER_CHOICE, + SPNG_CHUNK_COUNT_LIMIT, + SPNG_ENCODE_TO_BUFFER, +}; + +typedef void* SPNG_CDECL spng_malloc_fn(size_t size); +typedef void* SPNG_CDECL spng_realloc_fn(void* ptr, size_t size); +typedef void* SPNG_CDECL spng_calloc_fn(size_t count, size_t size); +typedef void SPNG_CDECL spng_free_fn(void* ptr); + +struct spng_alloc +{ + spng_malloc_fn *malloc_fn; + spng_realloc_fn *realloc_fn; + spng_calloc_fn *calloc_fn; + spng_free_fn *free_fn; +}; + +struct spng_row_info +{ + uint32_t scanline_idx; + uint32_t row_num; /* deinterlaced row index */ + int pass; + uint8_t filter; +}; + +typedef struct spng_ctx spng_ctx; + +typedef int spng_read_fn(spng_ctx *ctx, void *user, void *dest, size_t length); +typedef int spng_write_fn(spng_ctx *ctx, void *user, void *src, size_t length); + +typedef int spng_rw_fn(spng_ctx *ctx, void *user, void *dst_src, size_t length); + +SPNG_API spng_ctx *spng_ctx_new(int flags); +SPNG_API spng_ctx *spng_ctx_new2(struct spng_alloc *alloc, int flags); +SPNG_API void spng_ctx_free(spng_ctx *ctx); + +SPNG_API int spng_set_png_buffer(spng_ctx *ctx, const void *buf, size_t size); +SPNG_API int spng_set_png_stream(spng_ctx *ctx, spng_rw_fn *rw_func, void *user); +SPNG_API int spng_set_png_file(spng_ctx *ctx, FILE *file); + +SPNG_API void *spng_get_png_buffer(spng_ctx *ctx, size_t *len, int *error); + +SPNG_API int spng_set_image_limits(spng_ctx *ctx, uint32_t width, uint32_t height); +SPNG_API int spng_get_image_limits(spng_ctx *ctx, uint32_t *width, uint32_t *height); + +SPNG_API int spng_set_chunk_limits(spng_ctx *ctx, size_t chunk_size, size_t cache_size); +SPNG_API int spng_get_chunk_limits(spng_ctx *ctx, size_t *chunk_size, size_t *cache_size); + +SPNG_API int spng_set_crc_action(spng_ctx *ctx, int critical, int ancillary); + +SPNG_API int spng_set_option(spng_ctx *ctx, enum spng_option option, int value); +SPNG_API int spng_get_option(spng_ctx *ctx, enum spng_option option, int *value); + +SPNG_API int spng_decoded_image_size(spng_ctx *ctx, int fmt, size_t *len); + +/* Decode */ +SPNG_API int spng_decode_image(spng_ctx *ctx, void *out, size_t len, int fmt, int flags); + +/* Progressive decode */ +SPNG_API int spng_decode_scanline(spng_ctx *ctx, void *out, size_t len); +SPNG_API int spng_decode_row(spng_ctx *ctx, void *out, size_t len); +SPNG_API int spng_decode_chunks(spng_ctx *ctx); + +/* Encode/decode */ +SPNG_API int spng_get_row_info(spng_ctx *ctx, struct spng_row_info *row_info); + +/* Encode */ +SPNG_API int spng_encode_image(spng_ctx *ctx, const void *img, size_t len, int fmt, int flags); + +/* Progressive encode */ +SPNG_API int spng_encode_scanline(spng_ctx *ctx, const void *scanline, size_t len); +SPNG_API int spng_encode_row(spng_ctx *ctx, const void *row, size_t len); +SPNG_API int spng_encode_chunks(spng_ctx *ctx); + +SPNG_API int spng_get_ihdr(spng_ctx *ctx, struct spng_ihdr *ihdr); +SPNG_API int spng_get_plte(spng_ctx *ctx, struct spng_plte *plte); +SPNG_API int spng_get_trns(spng_ctx *ctx, struct spng_trns *trns); +SPNG_API int spng_get_chrm(spng_ctx *ctx, struct spng_chrm *chrm); +SPNG_API int spng_get_chrm_int(spng_ctx *ctx, struct spng_chrm_int *chrm_int); +SPNG_API int spng_get_gama(spng_ctx *ctx, double *gamma); +SPNG_API int spng_get_gama_int(spng_ctx *ctx, uint32_t *gama_int); +SPNG_API int spng_get_iccp(spng_ctx *ctx, struct spng_iccp *iccp); +SPNG_API int spng_get_sbit(spng_ctx *ctx, struct spng_sbit *sbit); +SPNG_API int spng_get_srgb(spng_ctx *ctx, uint8_t *rendering_intent); +SPNG_API int spng_get_text(spng_ctx *ctx, struct spng_text *text, uint32_t *n_text); +SPNG_API int spng_get_bkgd(spng_ctx *ctx, struct spng_bkgd *bkgd); +SPNG_API int spng_get_hist(spng_ctx *ctx, struct spng_hist *hist); +SPNG_API int spng_get_phys(spng_ctx *ctx, struct spng_phys *phys); +SPNG_API int spng_get_splt(spng_ctx *ctx, struct spng_splt *splt, uint32_t *n_splt); +SPNG_API int spng_get_time(spng_ctx *ctx, struct spng_time *time); +SPNG_API int spng_get_unknown_chunks(spng_ctx *ctx, struct spng_unknown_chunk *chunks, uint32_t *n_chunks); + +/* Official extensions */ +SPNG_API int spng_get_offs(spng_ctx *ctx, struct spng_offs *offs); +SPNG_API int spng_get_exif(spng_ctx *ctx, struct spng_exif *exif); + + +SPNG_API int spng_set_ihdr(spng_ctx *ctx, struct spng_ihdr *ihdr); +SPNG_API int spng_set_plte(spng_ctx *ctx, struct spng_plte *plte); +SPNG_API int spng_set_trns(spng_ctx *ctx, struct spng_trns *trns); +SPNG_API int spng_set_chrm(spng_ctx *ctx, struct spng_chrm *chrm); +SPNG_API int spng_set_chrm_int(spng_ctx *ctx, struct spng_chrm_int *chrm_int); +SPNG_API int spng_set_gama(spng_ctx *ctx, double gamma); +SPNG_API int spng_set_gama_int(spng_ctx *ctx, uint32_t gamma); +SPNG_API int spng_set_iccp(spng_ctx *ctx, struct spng_iccp *iccp); +SPNG_API int spng_set_sbit(spng_ctx *ctx, struct spng_sbit *sbit); +SPNG_API int spng_set_srgb(spng_ctx *ctx, uint8_t rendering_intent); +SPNG_API int spng_set_text(spng_ctx *ctx, struct spng_text *text, uint32_t n_text); +SPNG_API int spng_set_bkgd(spng_ctx *ctx, struct spng_bkgd *bkgd); +SPNG_API int spng_set_hist(spng_ctx *ctx, struct spng_hist *hist); +SPNG_API int spng_set_phys(spng_ctx *ctx, struct spng_phys *phys); +SPNG_API int spng_set_splt(spng_ctx *ctx, struct spng_splt *splt, uint32_t n_splt); +SPNG_API int spng_set_time(spng_ctx *ctx, struct spng_time *time); +SPNG_API int spng_set_unknown_chunks(spng_ctx *ctx, struct spng_unknown_chunk *chunks, uint32_t n_chunks); + +/* Official extensions */ +SPNG_API int spng_set_offs(spng_ctx *ctx, struct spng_offs *offs); +SPNG_API int spng_set_exif(spng_ctx *ctx, struct spng_exif *exif); + + +SPNG_API const char *spng_strerror(int err); +SPNG_API const char *spng_version_string(void); + +#ifdef __cplusplus +} +#endif + +#endif /* SPNG_H */ diff --git a/cmake/externals/opencv.cmake b/cmake/externals/opencv.cmake index b972266ac..e3fa7e904 100644 --- a/cmake/externals/opencv.cmake +++ b/cmake/externals/opencv.cmake @@ -1,167 +1,16 @@ -set(BUILD_ANDROID_EXAMPLES OFF CACHE INTERNAL "") -set(BUILD_ANDROID_PROJECTS OFF CACHE INTERNAL "") -set(BUILD_ANDROID_SERVICE OFF CACHE INTERNAL "") -set(BUILD_DOCS OFF CACHE INTERNAL "") -set(BUILD_FAT_JAVA_LIB OFF CACHE INTERNAL "") -set(BUILD_IPP_IW OFF CACHE INTERNAL "") -set(BUILD_ITT OFF CACHE INTERNAL "") -set(BUILD_JASPER OFF CACHE INTERNAL "") -set(BUILD_JAVA OFF CACHE INTERNAL "") -set(BUILD_JPEG OFF CACHE INTERNAL "") -set(BUILD_OBJC OFF CACHE INTERNAL "") -set(BUILD_OPENJPEG OFF CACHE INTERNAL "") -set(BUILD_PNG OFF CACHE INTERNAL "") -set(BUILD_opencv_apps OFF CACHE INTERNAL "") -set(BUILD_opencv_calib3d OFF CACHE INTERNAL "") -set(BUILD_opencv_dnn OFF CACHE INTERNAL "") -set(BUILD_opencv_features2d OFF CACHE INTERNAL "") -set(BUILD_opencv_flann OFF CACHE INTERNAL "") -set(BUILD_opencv_gapi OFF CACHE INTERNAL "") -set(BUILD_opencv_highgui OFF CACHE INTERNAL "") -set(BUILD_opencv_imgcodecs OFF CACHE INTERNAL "") -set(BUILD_opencv_java OFF CACHE INTERNAL "") -set(BUILD_opencv_js OFF CACHE INTERNAL "") -set(BUILD_opencv_ml OFF CACHE INTERNAL "") -set(BUILD_opencv_objc OFF CACHE INTERNAL "") -set(BUILD_opencv_objdetect OFF CACHE INTERNAL "") -set(BUILD_opencv_photo OFF CACHE INTERNAL "") -set(BUILD_opencv_python2 OFF CACHE INTERNAL "") -set(BUILD_opencv_python3 OFF CACHE INTERNAL "") -set(BUILD_opencv_stitching OFF CACHE INTERNAL "") -set(BUILD_opencv_ts OFF CACHE INTERNAL "") -set(BUILD_opencv_video OFF CACHE INTERNAL "") -set(BUILD_opencv_videoio OFF CACHE INTERNAL "") -set(BUILD_opencv_world OFF CACHE INTERNAL "") -set(BUILD_OPENEXR OFF CACHE INTERNAL "") -set(BUILD_TBB OFF CACHE INTERNAL "") -set(BUILD_TIFF OFF CACHE INTERNAL "") -set(BUILD_WEBP OFF CACHE INTERNAL "") -set(BUILD_WITH_STATIC_CRT OFF CACHE INTERNAL "") -if(OCOS_BUILD_APPLE_FRAMEWORK) - # tell OpenCV to build zlib so we can link to the static library - set(BUILD_ZLIB ON CACHE INTERNAL "") -else() - set(BUILD_ZLIB OFF CACHE INTERNAL "") -endif() -set(ENABLE_FAST_MATH OFF CACHE INTERNAL "") -set(ENABLE_PRECOMPILED_HEADERS OFF CACHE INTERNAL "") -set(WITH_ANDROID_MEDIANDK OFF CACHE INTERNAL "") -set(WITH_AVFOUNDATION OFF CACHE INTERNAL "") -set(WITH_CAP_IOS OFF CACHE INTERNAL "") -set(WITH_CAROTENE OFF CACHE INTERNAL "") -set(WITH_CLP OFF CACHE INTERNAL "") -set(WITH_CPUFEATURES OFF CACHE INTERNAL "") -set(WITH_DIRECTX OFF CACHE INTERNAL "") -set(WITH_DSHOW OFF CACHE INTERNAL "") -set(WITH_EIGEN OFF CACHE INTERNAL "") -set(WITH_FFMPEG OFF CACHE INTERNAL "") -set(WITH_GDCM OFF CACHE INTERNAL "") -set(WITH_GSTREAMER OFF CACHE INTERNAL "") -set(WITH_GTK OFF CACHE INTERNAL "") -set(WITH_HALIDE OFF CACHE INTERNAL "") -set(WITH_HPX OFF CACHE INTERNAL "") -set(WITH_IMGCODEC_HDR OFF CACHE INTERNAL "") -set(WITH_IMGCODEC_PFM OFF CACHE INTERNAL "") -set(WITH_IMGCODEC_PXM OFF CACHE INTERNAL "") -set(WITH_IMGCODEC_SUNRASTER OFF CACHE INTERNAL "") -set(WITH_INF_ENGINE OFF CACHE INTERNAL "") -set(WITH_IPP OFF CACHE INTERNAL "") -set(WITH_ITT OFF CACHE INTERNAL "") -set(WITH_JASPER OFF CACHE INTERNAL "") -set(WITH_JPEG OFF CACHE INTERNAL "") -set(WITH_MSMF OFF CACHE INTERNAL "") -set(WITH_NGRAPH OFF CACHE INTERNAL "") -set(WITH_ONNX OFF CACHE INTERNAL "") -set(WITH_OPENCL OFF CACHE INTERNAL "") -set(WITH_OPENCL_SVM OFF CACHE INTERNAL "") -set(WITH_OPENEXR OFF CACHE INTERNAL "") -set(WITH_OPENJPEG OFF CACHE INTERNAL "") -set(WITH_OPENMP OFF CACHE INTERNAL "") -set(WITH_OPENVX OFF CACHE INTERNAL "") -set(WITH_PNG OFF CACHE INTERNAL "") -set(WITH_PROTOBUF OFF CACHE INTERNAL "") -set(WITH_PTHREADS_PF OFF CACHE INTERNAL "") -set(WITH_QUIRC OFF CACHE INTERNAL "") -set(WITH_TBB OFF CACHE INTERNAL "") -set(WITH_TENGINE OFF CACHE INTERNAL "") -set(WITH_TIFF OFF CACHE INTERNAL "") -set(WITH_V4L OFF CACHE INTERNAL "") -set(WITH_VULKAN OFF CACHE INTERNAL "") -set(WITH_WEBP OFF CACHE INTERNAL "") -set(WITH_WIN32UI OFF CACHE INTERNAL "") - if (OCOS_ENABLE_OPENCV_CODECS) - set(BUILD_opencv_imgcodecs ON CACHE INTERNAL "") - - set(BUILD_JPEG ON CACHE INTERNAL "") - set(BUILD_PNG ON CACHE INTERNAL "") - - set(WITH_JPEG ON CACHE INTERNAL "") - set(WITH_PNG ON CACHE INTERNAL "") -endif() - -set(BUILD_SHARED_LIBS OFF CACHE INTERNAL "") -set(BUILD_DOCS OFF CACHE INTERNAL "") -set(BUILD_EXAMPLES OFF CACHE INTERNAL "") -set(BUILD_TESTS OFF CACHE INTERNAL "") - -if(IOS) - # copy what OpenCV's platforms/ios/build_framework.py does and set CPU_BASELINE=DETECT - # https://github.com/opencv/opencv/blob/4223495e6cd67011f86b8ecd9be1fa105018f3b1/platforms/ios/build_framework.py#L253 - set(CPU_BASELINE DETECT) -endif() + # Include the subdirectories for libspng and libjpeg + add_subdirectory(${PROJECT_SOURCE_DIR}/cmake/externals/libspng) + add_subdirectory(${PROJECT_SOURCE_DIR}/cmake/externals/libjpeg) -if (CMAKE_SYSTEM_NAME MATCHES "Darwin") - # error _png_do_read_transformations in liblibpng.a(pngrtran.c.o) not found for architecture arm64 - # workaround to disable NEON optimizations - add_definitions(-DPNG_ARM_NEON_IMPLEMENTATION=0) - add_definitions(-DPNG_ARM_NEON_OPT=0) -endif() - -if (MSVC AND CMAKE_GENERATOR_PLATFORM) - string(TOLOWER ${CMAKE_GENERATOR_PLATFORM} _GEN_PLATFORM) - if (${_GEN_PLATFORM} MATCHES "arm|arm64") - set(OPENCV_SKIP_SYSTEM_PROCESSOR_DETECTION ON) - endif() -endif() - -FetchContent_Declare( - opencv - GIT_REPOSITORY https://github.com/opencv/opencv.git - GIT_TAG 4.5.4 - GIT_SHALLOW TRUE - -DBUILD_DOCS:BOOL=FALSE - -DBUILD_EXAMPLES:BOOL=FALSE - -DBUILD_TESTS:BOOL=FALSE - -DBUILD_SHARED_LIBS:BOOL=FALSE - -DCMAKE_INSTALL_PREFIX:PATH=${CMAKE_CURRENT_BINARY_DIR}/opencv - -DCV_TRACE:BOOL=FALSE - PATCH_COMMAND git checkout . && git apply --whitespace=fix --ignore-space-change --ignore-whitespace ${CMAKE_CURRENT_SOURCE_DIR}/cmake/externals/opencv-no-rtti.patch -) - -FetchContent_MakeAvailable(opencv) -set(opencv_INCLUDE_DIRS "") -list(APPEND opencv_INCLUDE_DIRS ${OPENCV_CONFIG_FILE_INCLUDE_DIR}) -list(APPEND opencv_INCLUDE_DIRS - ${OPENCV_MODULE_opencv_core_LOCATION}/include - ${OPENCV_MODULE_opencv_imgproc_LOCATION}/include) -set(opencv_LIBS "") -list(APPEND opencv_LIBS opencv_core opencv_imgproc) - -if (OCOS_ENABLE_OPENCV_CODECS) - list(APPEND opencv_INCLUDE_DIRS ${OPENCV_MODULE_opencv_imgcodecs_LOCATION}/include) - list(APPEND opencv_LIBS opencv_imgcodecs) -endif() + # Specify where the static libraries are built + set(LIBRARY_OUTPUT_PATH ${CMAKE_CURRENT_BINARY_DIR}/cvout) -# unset it to avoid affecting other projects. -unset(EXECUTABLE_OUTPUT_PATH CACHE) + # Add the libraries + add_library(libspng STATIC IMPORTED) + add_library(libjpeg STATIC IMPORTED) -if (CMAKE_SYSTEM_NAME MATCHES "Windows") - set(opencv_projs gen_opencv_java_source gen_opencv_js_source gen_opencv_python_source) - list(APPEND opencv_projs gen_opencv_objc_source gen_opencv_objc_source_ios gen_opencv_objc_source_osx) - list(APPEND opencv_projs opencv_highgui_plugins opencv_videoio_plugins) - foreach(p ${opencv_projs}) - set_target_properties(${p} PROPERTIES FOLDER "externals/opencv") - set_target_properties(${p} PROPERTIES EXCLUDE_FROM_ALL TRUE EXCLUDE_FROM_DEFAULT_BUILD TRUE) - endforeach() -endif() + # Set the properties for the libraries + set_property(TARGET libspng PROPERTY IMPORTED_LOCATION ${CMAKE_CURRENT_BINARY_DIR}/cmake/externals/cvout/libspng.a) + set_property(TARGET libjpeg PROPERTY IMPORTED_LOCATION ${CMAKE_CURRENT_BINARY_DIR}/cmake/externals/cvout/libjpeg.a) +endif() \ No newline at end of file From 181dc854091b06266469d620a7a8cf3b713c09cd Mon Sep 17 00:00:00 2001 From: Sayan Shaw Date: Wed, 15 May 2024 13:26:53 -0700 Subject: [PATCH 2/3] try with fetch content populate --- cmake/externals/cvout/.gitkeep | 0 cmake/externals/libjpeg/CMakeLists.txt | 30 - cmake/externals/libjpeg/README | 371 -- cmake/externals/libjpeg/change.log | 502 -- cmake/externals/libjpeg/jaricom.c | 153 - cmake/externals/libjpeg/jcapimin.c | 288 - cmake/externals/libjpeg/jcapistd.c | 162 - cmake/externals/libjpeg/jcarith.c | 945 ---- cmake/externals/libjpeg/jccoefct.c | 456 -- cmake/externals/libjpeg/jccolor.c | 598 -- cmake/externals/libjpeg/jcdctmgr.c | 466 -- cmake/externals/libjpeg/jchuff.c | 1656 ------ cmake/externals/libjpeg/jcinit.c | 249 - cmake/externals/libjpeg/jcmainct.c | 297 - cmake/externals/libjpeg/jcmarker.c | 717 --- cmake/externals/libjpeg/jcmaster.c | 675 --- cmake/externals/libjpeg/jcomapi.c | 244 - cmake/externals/libjpeg/jconfig.h | 64 - cmake/externals/libjpeg/jcparam.c | 591 -- cmake/externals/libjpeg/jcprepct.c | 358 -- cmake/externals/libjpeg/jcsample.c | 545 -- cmake/externals/libjpeg/jctrans.c | 399 -- cmake/externals/libjpeg/jdapimin.c | 412 -- cmake/externals/libjpeg/jdapistd.c | 276 - cmake/externals/libjpeg/jdarith.c | 796 --- cmake/externals/libjpeg/jdatadst.c | 263 - cmake/externals/libjpeg/jdatasrc.c | 271 - cmake/externals/libjpeg/jdcoefct.c | 744 --- cmake/externals/libjpeg/jdcolor.c | 769 --- cmake/externals/libjpeg/jdct.h | 409 -- cmake/externals/libjpeg/jddctmgr.c | 384 -- cmake/externals/libjpeg/jdhuff.c | 1559 ------ cmake/externals/libjpeg/jdinput.c | 657 --- cmake/externals/libjpeg/jdmainct.c | 511 -- cmake/externals/libjpeg/jdmarker.c | 1505 ----- cmake/externals/libjpeg/jdmaster.c | 532 -- cmake/externals/libjpeg/jdmerge.c | 437 -- cmake/externals/libjpeg/jdpostct.c | 290 - cmake/externals/libjpeg/jdsample.c | 341 -- cmake/externals/libjpeg/jdtrans.c | 140 - cmake/externals/libjpeg/jerror.c | 253 - cmake/externals/libjpeg/jerror.h | 304 -- cmake/externals/libjpeg/jfdctflt.c | 176 - cmake/externals/libjpeg/jfdctfst.c | 232 - cmake/externals/libjpeg/jfdctint.c | 4415 --------------- cmake/externals/libjpeg/jidctflt.c | 238 - cmake/externals/libjpeg/jidctfst.c | 351 -- cmake/externals/libjpeg/jidctint.c | 5240 ------------------ cmake/externals/libjpeg/jinclude.h | 157 - cmake/externals/libjpeg/jmemansi.c | 167 - cmake/externals/libjpeg/jmemmgr.c | 1115 ---- cmake/externals/libjpeg/jmemnobs.c | 113 - cmake/externals/libjpeg/jmemsys.h | 198 - cmake/externals/libjpeg/jmorecfg.h | 457 -- cmake/externals/libjpeg/jpegint.h | 445 -- cmake/externals/libjpeg/jpeglib.h | 1183 ---- cmake/externals/libjpeg/jquant1.c | 851 --- cmake/externals/libjpeg/jquant2.c | 1311 ----- cmake/externals/libjpeg/jutils.c | 224 - cmake/externals/libjpeg/jversion.h | 14 - cmake/externals/libspng/CMakeLists.txt | 39 - cmake/externals/libspng/LICENSE | 25 - cmake/externals/libspng/spng.c | 6980 ------------------------ cmake/externals/libspng/spng.h | 537 -- cmake/externals/opencv.cmake | 179 +- 65 files changed, 167 insertions(+), 45099 deletions(-) delete mode 100644 cmake/externals/cvout/.gitkeep delete mode 100644 cmake/externals/libjpeg/CMakeLists.txt delete mode 100644 cmake/externals/libjpeg/README delete mode 100644 cmake/externals/libjpeg/change.log delete mode 100644 cmake/externals/libjpeg/jaricom.c delete mode 100644 cmake/externals/libjpeg/jcapimin.c delete mode 100644 cmake/externals/libjpeg/jcapistd.c delete mode 100644 cmake/externals/libjpeg/jcarith.c delete mode 100644 cmake/externals/libjpeg/jccoefct.c delete mode 100644 cmake/externals/libjpeg/jccolor.c delete mode 100644 cmake/externals/libjpeg/jcdctmgr.c delete mode 100644 cmake/externals/libjpeg/jchuff.c delete mode 100644 cmake/externals/libjpeg/jcinit.c delete mode 100644 cmake/externals/libjpeg/jcmainct.c delete mode 100644 cmake/externals/libjpeg/jcmarker.c delete mode 100644 cmake/externals/libjpeg/jcmaster.c delete mode 100644 cmake/externals/libjpeg/jcomapi.c delete mode 100644 cmake/externals/libjpeg/jconfig.h delete mode 100644 cmake/externals/libjpeg/jcparam.c delete mode 100644 cmake/externals/libjpeg/jcprepct.c delete mode 100644 cmake/externals/libjpeg/jcsample.c delete mode 100644 cmake/externals/libjpeg/jctrans.c delete mode 100644 cmake/externals/libjpeg/jdapimin.c delete mode 100644 cmake/externals/libjpeg/jdapistd.c delete mode 100644 cmake/externals/libjpeg/jdarith.c delete mode 100644 cmake/externals/libjpeg/jdatadst.c delete mode 100644 cmake/externals/libjpeg/jdatasrc.c delete mode 100644 cmake/externals/libjpeg/jdcoefct.c delete mode 100644 cmake/externals/libjpeg/jdcolor.c delete mode 100644 cmake/externals/libjpeg/jdct.h delete mode 100644 cmake/externals/libjpeg/jddctmgr.c delete mode 100644 cmake/externals/libjpeg/jdhuff.c delete mode 100644 cmake/externals/libjpeg/jdinput.c delete mode 100644 cmake/externals/libjpeg/jdmainct.c delete mode 100644 cmake/externals/libjpeg/jdmarker.c delete mode 100644 cmake/externals/libjpeg/jdmaster.c delete mode 100644 cmake/externals/libjpeg/jdmerge.c delete mode 100644 cmake/externals/libjpeg/jdpostct.c delete mode 100644 cmake/externals/libjpeg/jdsample.c delete mode 100644 cmake/externals/libjpeg/jdtrans.c delete mode 100644 cmake/externals/libjpeg/jerror.c delete mode 100644 cmake/externals/libjpeg/jerror.h delete mode 100644 cmake/externals/libjpeg/jfdctflt.c delete mode 100644 cmake/externals/libjpeg/jfdctfst.c delete mode 100644 cmake/externals/libjpeg/jfdctint.c delete mode 100644 cmake/externals/libjpeg/jidctflt.c delete mode 100644 cmake/externals/libjpeg/jidctfst.c delete mode 100644 cmake/externals/libjpeg/jidctint.c delete mode 100644 cmake/externals/libjpeg/jinclude.h delete mode 100644 cmake/externals/libjpeg/jmemansi.c delete mode 100644 cmake/externals/libjpeg/jmemmgr.c delete mode 100644 cmake/externals/libjpeg/jmemnobs.c delete mode 100644 cmake/externals/libjpeg/jmemsys.h delete mode 100644 cmake/externals/libjpeg/jmorecfg.h delete mode 100644 cmake/externals/libjpeg/jpegint.h delete mode 100644 cmake/externals/libjpeg/jpeglib.h delete mode 100644 cmake/externals/libjpeg/jquant1.c delete mode 100644 cmake/externals/libjpeg/jquant2.c delete mode 100644 cmake/externals/libjpeg/jutils.c delete mode 100644 cmake/externals/libjpeg/jversion.h delete mode 100644 cmake/externals/libspng/CMakeLists.txt delete mode 100644 cmake/externals/libspng/LICENSE delete mode 100644 cmake/externals/libspng/spng.c delete mode 100644 cmake/externals/libspng/spng.h diff --git a/cmake/externals/cvout/.gitkeep b/cmake/externals/cvout/.gitkeep deleted file mode 100644 index e69de29bb..000000000 diff --git a/cmake/externals/libjpeg/CMakeLists.txt b/cmake/externals/libjpeg/CMakeLists.txt deleted file mode 100644 index 4e88164cc..000000000 --- a/cmake/externals/libjpeg/CMakeLists.txt +++ /dev/null @@ -1,30 +0,0 @@ -# ---------------------------------------------------------------------------- -# CMake file for libjpeg. See root CMakeLists.txt -# -# ---------------------------------------------------------------------------- -project(JPEG_LIBRARY) - -file(GLOB lib_srcs *.c) -file(GLOB lib_hdrs *.h) - -# ---------------------------------------------------------------------------------- -# Define the library target: -# ---------------------------------------------------------------------------------- - -add_library(JPEG_LIBRARY STATIC ${OPENCV_3RDPARTY_EXCLUDE_FROM_ALL} ${lib_srcs} ${lib_hdrs}) - -if(CV_GCC OR CV_CLANG) - set_source_files_properties(jcdctmgr.c PROPERTIES COMPILE_FLAGS "-O1") -endif() - -set_target_properties(JPEG_LIBRARY - PROPERTIES OUTPUT_NAME JPEG_LIBRARY - DEBUG_POSTFIX "${OPENCV_DEBUG_POSTFIX}" - COMPILE_PDB_NAME JPEG_LIBRARY - COMPILE_PDB_NAME_DEBUG "JPEG_LIBRARY${OPENCV_DEBUG_POSTFIX}" - ARCHIVE_OUTPUT_DIRECTORY "${3P_LIBRARY_OUTPUT_PATH}" - ) - -if(ENABLE_SOLUTION_FOLDERS) - set_target_properties(JPEG_LIBRARY PROPERTIES FOLDER "3rdparty") -endif() diff --git a/cmake/externals/libjpeg/README b/cmake/externals/libjpeg/README deleted file mode 100644 index 13b245c5c..000000000 --- a/cmake/externals/libjpeg/README +++ /dev/null @@ -1,371 +0,0 @@ -The Independent JPEG Group's JPEG software -========================================== - -README for release 9d of 12-Jan-2020 -==================================== - -This distribution contains the ninth public release of the Independent JPEG -Group's free JPEG software. You are welcome to redistribute this software and -to use it for any purpose, subject to the conditions under LEGAL ISSUES, below. - -This software is the work of Tom Lane, Guido Vollbeding, Philip Gladstone, -Bill Allombert, Jim Boucher, Lee Crocker, Bob Friesenhahn, Ben Jackson, -John Korejwa, Julian Minguillon, Luis Ortiz, George Phillips, Davide Rossi, -Ge' Weijers, and other members of the Independent JPEG Group. - -IJG is not affiliated with the ISO/IEC JTC1/SC29/WG1 standards committee -(previously known as JPEG, together with ITU-T SG16). - - -DOCUMENTATION ROADMAP -===================== - -This file contains the following sections: - -OVERVIEW General description of JPEG and the IJG software. -LEGAL ISSUES Copyright, lack of warranty, terms of distribution. -REFERENCES Where to learn more about JPEG. -ARCHIVE LOCATIONS Where to find newer versions of this software. -ACKNOWLEDGMENTS Special thanks. -FILE FORMAT WARS Software *not* to get. -TO DO Plans for future IJG releases. - -Other documentation files in the distribution are: - -User documentation: - install.txt How to configure and install the IJG software. - usage.txt Usage instructions for cjpeg, djpeg, jpegtran, - rdjpgcom, and wrjpgcom. - *.1 Unix-style man pages for programs (same info as usage.txt). - wizard.txt Advanced usage instructions for JPEG wizards only. - change.log Version-to-version change highlights. -Programmer and internal documentation: - libjpeg.txt How to use the JPEG library in your own programs. - example.c Sample code for calling the JPEG library. - structure.txt Overview of the JPEG library's internal structure. - filelist.txt Road map of IJG files. - coderules.txt Coding style rules --- please read if you contribute code. - -Please read at least the files install.txt and usage.txt. Some information -can also be found in the JPEG FAQ (Frequently Asked Questions) article. See -ARCHIVE LOCATIONS below to find out where to obtain the FAQ article. - -If you want to understand how the JPEG code works, we suggest reading one or -more of the REFERENCES, then looking at the documentation files (in roughly -the order listed) before diving into the code. - - -OVERVIEW -======== - -This package contains C software to implement JPEG image encoding, decoding, -and transcoding. JPEG (pronounced "jay-peg") is a standardized compression -method for full-color and grayscale images. - -This software implements JPEG baseline, extended-sequential, and progressive -compression processes. Provision is made for supporting all variants of these -processes, although some uncommon parameter settings aren't implemented yet. -We have made no provision for supporting the hierarchical or lossless -processes defined in the standard. - -We provide a set of library routines for reading and writing JPEG image files, -plus two sample applications "cjpeg" and "djpeg", which use the library to -perform conversion between JPEG and some other popular image file formats. -The library is intended to be reused in other applications. - -In order to support file conversion and viewing software, we have included -considerable functionality beyond the bare JPEG coding/decoding capability; -for example, the color quantization modules are not strictly part of JPEG -decoding, but they are essential for output to colormapped file formats or -colormapped displays. These extra functions can be compiled out of the -library if not required for a particular application. - -We have also included "jpegtran", a utility for lossless transcoding between -different JPEG processes, and "rdjpgcom" and "wrjpgcom", two simple -applications for inserting and extracting textual comments in JFIF files. - -The emphasis in designing this software has been on achieving portability and -flexibility, while also making it fast enough to be useful. In particular, -the software is not intended to be read as a tutorial on JPEG. (See the -REFERENCES section for introductory material.) Rather, it is intended to -be reliable, portable, industrial-strength code. We do not claim to have -achieved that goal in every aspect of the software, but we strive for it. - -We welcome the use of this software as a component of commercial products. -No royalty is required, but we do ask for an acknowledgement in product -documentation, as described under LEGAL ISSUES. - - -LEGAL ISSUES -============ - -In plain English: - -1. We don't promise that this software works. (But if you find any bugs, - please let us know!) -2. You can use this software for whatever you want. You don't have to pay us. -3. You may not pretend that you wrote this software. If you use it in a - program, you must acknowledge somewhere in your documentation that - you've used the IJG code. - -In legalese: - -The authors make NO WARRANTY or representation, either express or implied, -with respect to this software, its quality, accuracy, merchantability, or -fitness for a particular purpose. This software is provided "AS IS", and you, -its user, assume the entire risk as to its quality and accuracy. - -This software is copyright (C) 1991-2020, Thomas G. Lane, Guido Vollbeding. -All Rights Reserved except as specified below. - -Permission is hereby granted to use, copy, modify, and distribute this -software (or portions thereof) for any purpose, without fee, subject to these -conditions: -(1) If any part of the source code for this software is distributed, then this -README file must be included, with this copyright and no-warranty notice -unaltered; and any additions, deletions, or changes to the original files -must be clearly indicated in accompanying documentation. -(2) If only executable code is distributed, then the accompanying -documentation must state that "this software is based in part on the work of -the Independent JPEG Group". -(3) Permission for use of this software is granted only if the user accepts -full responsibility for any undesirable consequences; the authors accept -NO LIABILITY for damages of any kind. - -These conditions apply to any software derived from or based on the IJG code, -not just to the unmodified library. If you use our work, you ought to -acknowledge us. - -Permission is NOT granted for the use of any IJG author's name or company name -in advertising or publicity relating to this software or products derived from -it. This software may be referred to only as "the Independent JPEG Group's -software". - -We specifically permit and encourage the use of this software as the basis of -commercial products, provided that all warranty or liability claims are -assumed by the product vendor. - - -The Unix configuration script "configure" was produced with GNU Autoconf. -It is copyright by the Free Software Foundation but is freely distributable. -The same holds for its supporting scripts (config.guess, config.sub, -ltmain.sh). Another support script, install-sh, is copyright by X Consortium -but is also freely distributable. - - -REFERENCES -========== - -We recommend reading one or more of these references before trying to -understand the innards of the JPEG software. - -The best short technical introduction to the JPEG compression algorithm is - Wallace, Gregory K. "The JPEG Still Picture Compression Standard", - Communications of the ACM, April 1991 (vol. 34 no. 4), pp. 30-44. -(Adjacent articles in that issue discuss MPEG motion picture compression, -applications of JPEG, and related topics.) If you don't have the CACM issue -handy, a PDF file containing a revised version of Wallace's article is -available at http://www.ijg.org/files/Wallace.JPEG.pdf. The file (actually -a preprint for an article that appeared in IEEE Trans. Consumer Electronics) -omits the sample images that appeared in CACM, but it includes corrections -and some added material. Note: the Wallace article is copyright ACM and IEEE, -and it may not be used for commercial purposes. - -A somewhat less technical, more leisurely introduction to JPEG can be found in -"The Data Compression Book" by Mark Nelson and Jean-loup Gailly, published by -M&T Books (New York), 2nd ed. 1996, ISBN 1-55851-434-1. This book provides -good explanations and example C code for a multitude of compression methods -including JPEG. It is an excellent source if you are comfortable reading C -code but don't know much about data compression in general. The book's JPEG -sample code is far from industrial-strength, but when you are ready to look -at a full implementation, you've got one here... - -The best currently available description of JPEG is the textbook "JPEG Still -Image Data Compression Standard" by William B. Pennebaker and Joan L. -Mitchell, published by Van Nostrand Reinhold, 1993, ISBN 0-442-01272-1. -Price US$59.95, 638 pp. The book includes the complete text of the ISO JPEG -standards (DIS 10918-1 and draft DIS 10918-2). -Although this is by far the most detailed and comprehensive exposition of -JPEG publicly available, we point out that it is still missing an explanation -of the most essential properties and algorithms of the underlying DCT -technology. -If you think that you know about DCT-based JPEG after reading this book, -then you are in delusion. The real fundamentals and corresponding potential -of DCT-based JPEG are not publicly known so far, and that is the reason for -all the mistaken developments taking place in the image coding domain. - -The original JPEG standard is divided into two parts, Part 1 being the actual -specification, while Part 2 covers compliance testing methods. Part 1 is -titled "Digital Compression and Coding of Continuous-tone Still Images, -Part 1: Requirements and guidelines" and has document numbers ISO/IEC IS -10918-1, ITU-T T.81. Part 2 is titled "Digital Compression and Coding of -Continuous-tone Still Images, Part 2: Compliance testing" and has document -numbers ISO/IEC IS 10918-2, ITU-T T.83. -IJG JPEG 8 introduced an implementation of the JPEG SmartScale extension -which is specified in two documents: A contributed document at ITU and ISO -with title "ITU-T JPEG-Plus Proposal for Extending ITU-T T.81 for Advanced -Image Coding", April 2006, Geneva, Switzerland. The latest version of this -document is Revision 3. And a contributed document ISO/IEC JTC1/SC29/WG1 N -5799 with title "Evolution of JPEG", June/July 2011, Berlin, Germany. -IJG JPEG 9 introduces a reversible color transform for improved lossless -compression which is described in a contributed document ISO/IEC JTC1/SC29/ -WG1 N 6080 with title "JPEG 9 Lossless Coding", June/July 2012, Paris, -France. - -The JPEG standard does not specify all details of an interchangeable file -format. For the omitted details we follow the "JFIF" conventions, version 2. -JFIF version 1 has been adopted as Recommendation ITU-T T.871 (05/2011) : -Information technology - Digital compression and coding of continuous-tone -still images: JPEG File Interchange Format (JFIF). It is available as a -free download in PDF file format from http://www.itu.int/rec/T-REC-T.871. -A PDF file of the older JFIF document is available at -http://www.w3.org/Graphics/JPEG/jfif3.pdf. - -The TIFF 6.0 file format specification can be obtained by FTP from -ftp://ftp.sgi.com/graphics/tiff/TIFF6.ps.gz. The JPEG incorporation scheme -found in the TIFF 6.0 spec of 3-June-92 has a number of serious problems. -IJG does not recommend use of the TIFF 6.0 design (TIFF Compression tag 6). -Instead, we recommend the JPEG design proposed by TIFF Technical Note #2 -(Compression tag 7). Copies of this Note can be obtained from -http://www.ijg.org/files/. It is expected that the next revision -of the TIFF spec will replace the 6.0 JPEG design with the Note's design. -Although IJG's own code does not support TIFF/JPEG, the free libtiff library -uses our library to implement TIFF/JPEG per the Note. - - -ARCHIVE LOCATIONS -================= - -The "official" archive site for this software is www.ijg.org. -The most recent released version can always be found there in -directory "files". This particular version will be archived as -http://www.ijg.org/files/jpegsrc.v9d.tar.gz, and in Windows-compatible -"zip" archive format as http://www.ijg.org/files/jpegsr9d.zip. - -The JPEG FAQ (Frequently Asked Questions) article is a source of some -general information about JPEG. -It is available on the World Wide Web at http://www.faqs.org/faqs/jpeg-faq/ -and other news.answers archive sites, including the official news.answers -archive at rtfm.mit.edu: ftp://rtfm.mit.edu/pub/usenet/news.answers/jpeg-faq/. -If you don't have Web or FTP access, send e-mail to mail-server@rtfm.mit.edu -with body - send usenet/news.answers/jpeg-faq/part1 - send usenet/news.answers/jpeg-faq/part2 - - -ACKNOWLEDGMENTS -=============== - -Thank to Juergen Bruder for providing me with a copy of the common DCT -algorithm article, only to find out that I had come to the same result -in a more direct and comprehensible way with a more generative approach. - -Thank to Istvan Sebestyen and Joan L. Mitchell for inviting me to the -ITU JPEG (Study Group 16) meeting in Geneva, Switzerland. - -Thank to Thomas Wiegand and Gary Sullivan for inviting me to the -Joint Video Team (MPEG & ITU) meeting in Geneva, Switzerland. - -Thank to Thomas Richter and Daniel Lee for inviting me to the -ISO/IEC JTC1/SC29/WG1 (previously known as JPEG, together with ITU-T SG16) -meeting in Berlin, Germany. - -Thank to John Korejwa and Massimo Ballerini for inviting me to -fruitful consultations in Boston, MA and Milan, Italy. - -Thank to Hendrik Elstner, Roland Fassauer, Simone Zuck, Guenther -Maier-Gerber, Walter Stoeber, Fred Schmitz, and Norbert Braunagel -for corresponding business development. - -Thank to Nico Zschach and Dirk Stelling of the technical support team -at the Digital Images company in Halle for providing me with extra -equipment for configuration tests. - -Thank to Richard F. Lyon (then of Foveon Inc.) for fruitful -communication about JPEG configuration in Sigma Photo Pro software. - -Thank to Andrew Finkenstadt for hosting the ijg.org site. - -Thank to Thomas G. Lane for the original design and development of -this singular software package. - -Thank to Lars Goehler, Andreas Heinecke, Sebastian Fuss, Yvonne Roebert, -Andrej Werner, and Ulf-Dietrich Braumann for support and public relations. - - -FILE FORMAT WARS -================ - -The ISO/IEC JTC1/SC29/WG1 standards committee (previously known as JPEG, -together with ITU-T SG16) currently promotes different formats containing -the name "JPEG" which is misleading because these formats are incompatible -with original DCT-based JPEG and are based on faulty technologies. -IJG therefore does not and will not support such momentary mistakes -(see REFERENCES). -There exist also distributions under the name "OpenJPEG" promoting such -kind of formats which is misleading because they don't support original -JPEG images. -We have no sympathy for the promotion of inferior formats. Indeed, one of -the original reasons for developing this free software was to help force -convergence on common, interoperable format standards for JPEG files. -Don't use an incompatible file format! -(In any case, our decoder will remain capable of reading existing JPEG -image files indefinitely.) - -The ISO committee pretends to be "responsible for the popular JPEG" in their -public reports which is not true because they don't respond to actual -requirements for the maintenance of the original JPEG specification. -Furthermore, the ISO committee pretends to "ensure interoperability" with -their standards which is not true because their "standards" support only -application-specific and proprietary use cases and contain mathematically -incorrect code. - -There are currently different distributions in circulation containing the -name "libjpeg" which is misleading because they don't have the features and -are incompatible with formats supported by actual IJG libjpeg distributions. -One of those fakes is released by members of the ISO committee and just uses -the name of libjpeg for misdirection of people, similar to the abuse of the -name JPEG as described above, while having nothing in common with actual IJG -libjpeg distributions and containing mathematically incorrect code. -The other one claims to be a "derivative" or "fork" of the original libjpeg, -but violates the license conditions as described under LEGAL ISSUES above -and violates basic C programming properties. -We have no sympathy for the release of misleading, incorrect and illegal -distributions derived from obsolete code bases. -Don't use an obsolete code base! - -According to the UCC (Uniform Commercial Code) law, IJG has the lawful and -legal right to foreclose on certain standardization bodies and other -institutions or corporations that knowingly perform substantial and -systematic deceptive acts and practices, fraud, theft, and damaging of the -value of the people of this planet without their knowing, willing and -intentional consent. -The titles, ownership, and rights of these institutions and all their assets -are now duly secured and held in trust for the free people of this planet. -People of the planet, on every country, may have a financial interest in -the assets of these former principals, agents, and beneficiaries of the -foreclosed institutions and corporations. -IJG asserts what is: that each man, woman, and child has unalienable value -and rights granted and deposited in them by the Creator and not any one of -the people is subordinate to any artificial principality, corporate fiction -or the special interest of another without their appropriate knowing, -willing and intentional consent made by contract or accommodation agreement. -IJG expresses that which already was. -The people have already determined and demanded that public administration -entities, national governments, and their supporting judicial systems must -be fully transparent, accountable, and liable. -IJG has secured the value for all concerned free people of the planet. - -A partial list of foreclosed institutions and corporations ("Hall of Shame") -is currently prepared and will be published later. - - -TO DO -===== - -Version 9 is the second release of a new generation JPEG standard -to overcome the limitations of the original JPEG specification, -and is the first true source reference JPEG codec. -More features are being prepared for coming releases... - -Please send bug reports, offers of help, etc. to jpeg-info@jpegclub.org. diff --git a/cmake/externals/libjpeg/change.log b/cmake/externals/libjpeg/change.log deleted file mode 100644 index ed8ec6afc..000000000 --- a/cmake/externals/libjpeg/change.log +++ /dev/null @@ -1,502 +0,0 @@ -CHANGE LOG for Independent JPEG Group's JPEG software - - -Version 9d 12-Jan-2020 ------------------------ - -Optimize the optimal Huffman code table generation to produce -slightly smaller files. Thank to John Korejwa for suggestion. -Note: Requires rebuild of testimgp.jpg. - -Decoding Huffman: Use default tables if tables are not defined. -Thank to Simone Azzalin for report (Motion JPEG), -and to Martin Strunz for hint. - -Add sanity check in optimal Huffman code table generation. -Thank to Adam Farley for suggestion. - -rdtarga.c: use read_byte(), with EOF check, instead of getc() -in read_*_pixel(). -Thank to Chijin Zhou for cjpeg potential vulnerability report. - -jmemnobs.c: respect the max_memory_to_use setting in -jpeg_mem_available() computation. Thank to Sheng Shu and -Dongdong She for djpeg potential vulnerability report. - -jdarith.c, jdhuff.c: avoid left shift of negative value -compiler warning in decode_mcu_AC_refine(). -Thank to Indu Bhagat for suggestion. - -Add x64 (64-bit) platform support, avoid compiler warnings. -Thank to Jonathan Potter, Feiyun Wang, and Sheng Shu for suggestion. - -Adjust libjpeg version specification for pkg-config file. -Thank to Chen Chen for suggestion. - -Restore GIF read and write support from libjpeg version 6a. -Thank to Wolfgang Werner (W.W.) Heinz for suggestion. - -Improve consistency in raw (downsampled) image data processing mode. -Thank to Zhongyuan Zhou for hint. - -Avoid out of bounds array read (AC derived table pointers) -in start pass in jdhuff.c. Thank to Peng Li for report. - -Improve code sanity (jdhuff.c). -Thank to Reza Mirzazade farkhani for reports. - -Add jpegtran -drop option; add options to the crop extension and wipe -to fill the extra area with content from the source image region, -instead of gray out. - - -Version 9c 14-Jan-2018 ------------------------ - -jpegtran: add an option to the -wipe switch to fill the region -with the average of adjacent blocks, instead of gray out. -Thank to Caitlyn Feddock and Maddie Ziegler for inspiration. - -Make range extension bits adjustable (in jpegint.h). -Thank to Robin Watts for suggestion. - -Provide macros for fflush() and ferror() in jinclude.h in order -to facilitate adaption by applications using an own FILE class. -Thank to Gerhard Huber for suggestion. - -Add libjpeg pkg-config file. Thank to Mark Lavi, Vincent Torri, -Patrick McMunn, and Huw Davies for suggestion. - -Add sanity checks in cjpeg image reader modules. -Thank to Bingchang, Liu for reports. - - -Version 9b 17-Jan-2016 ------------------------ - -Improvements and optimizations in DCT and color calculations. -Normalize range limit array composition and access pattern. -Thank to Sia Furler and Maddie Ziegler for inspiration. - -Use merged upsample with scaled DCT sizes larger than 8. -Thank to Taylor Hatala for inspiration. - -Check for excessive comment lengths in argument parsing in wrjpgcom.c. -Thank to Julian Cohen for hint. - -Add makefile.b32 for use with Borland C++ 32-bit (bcc32). -Thank to Joe Slater for contribution. - -Document 'f' specifier for jpegtran -crop specification. -Thank to Michele Martone for suggestion. - -Use defined value from header instead of hardwired number in rdswitch.c. -Thank to Robert Sprowson for hint. - - -Version 9a 19-Jan-2014 ------------------------ - -Add support for wide gamut color spaces (JFIF version 2). -Improve clarity and accuracy in color conversion modules. -Note: Requires rebuild of test images. - -Extend the bit depth support to all values from 8 to 12 -(BITS_IN_JSAMPLE configuration option in jmorecfg.h). -jpegtran now supports N bits sample data precision with all N from 8 to 12 -in a single instance. Thank to Roland Fassauer for inspiration. - -Try to resolve issues with new boolean type definition. -Thank also to v4hn for suggestion. - -Enable option to use default Huffman tables for lossless compression -(for hardware solution), and in this case improve lossless RGB compression -with reversible color transform. Thank to Benny Alexandar for hint. - -Extend the entropy decoding structure, so that extraneous bytes between -compressed scan data and following marker can be reported correctly. -Thank to Nigel Tao for hint. - -Add jpegtran -wipe option and extension for -crop. -Thank to Andrew Senior, David Clunie, and Josef Schmid for suggestion. - - -Version 9 13-Jan-2013 ----------------------- - -Add cjpeg -rgb1 option to create an RGB JPEG file, and insert -a simple reversible color transform into the processing which -significantly improves the compression. -The recommended command for lossless coding of RGB images is now -cjpeg -rgb1 -block 1 -arithmetic. -As said, this option improves the compression significantly, but -the files are not compatible with JPEG decoders prior to IJG v9 -due to the included color transform. -The used color transform and marker signaling is compatible with -other JPEG standards (e.g., JPEG-LS part 2). - -Remove the automatic de-ANSI-fication support (Automake 1.12). -Thank also to Nitin A Kamble for suggestion. - -Add remark for jpeg_mem_dest() in jdatadst.c. -Thank to Elie-Gregoire Khoury for the hint. - -Support files with invalid component identifiers (created -by Adobe PDF). Thank to Robin Watts for the suggestion. - -Adapt full buffer case in jcmainct.c for use with scaled DCT. -Thank to Sergii Biloshytskyi for the suggestion. - -Add type identifier for declaration of noreturn functions. -Thank to Brett L. Moore for the suggestion. - -Correct argument type in format string, avoid compiler warnings. -Thank to Vincent Torri for hint. - -Add missing #include directives in configuration checks, avoid -configuration errors. Thank to John Spencer for the hint. - - -Version 8d 15-Jan-2012 ------------------------ - -Add cjpeg -rgb option to create RGB JPEG files. -Using this switch suppresses the conversion from RGB -colorspace input to the default YCbCr JPEG colorspace. -This feature allows true lossless JPEG coding of RGB color images. -The recommended command for this purpose is currently -cjpeg -rgb -block 1 -arithmetic. -SmartScale capable decoder (introduced with IJG JPEG 8) required. -Thank to Michael Koch for the initial suggestion. - -Add option to disable the region adjustment in the transupp crop code. -Thank to Jeffrey Friedl for the suggestion. - -Thank to Richard Jones and Edd Dawson for various minor corrections. - -Thank to Akim Demaille for configure.ac cleanup. - - -Version 8c 16-Jan-2011 ------------------------ - -Add option to compression library and cjpeg (-block N) to use -different DCT block size. -All N from 1 to 16 are possible. Default is 8 (baseline format). -Larger values produce higher compression, -smaller values produce higher quality. -SmartScale capable decoder (introduced with IJG JPEG 8) required. - - -Version 8b 16-May-2010 ------------------------ - -Repair problem in new memory source manager with corrupt JPEG data. -Thank to Ted Campbell and Samuel Chun for the report. - -Repair problem in Makefile.am test target. -Thank to anonymous user for the report. - -Support MinGW installation with automatic configure. -Thank to Volker Grabsch for the suggestion. - - -Version 8a 28-Feb-2010 ------------------------ - -Writing tables-only datastreams via jpeg_write_tables works again. - -Support 32-bit BMPs (RGB image with Alpha channel) for read in cjpeg. -Thank to Brett Blackham for the suggestion. - -Improve accuracy in floating point IDCT calculation. -Thank to Robert Hooke for the hint. - - -Version 8 10-Jan-2010 ----------------------- - -jpegtran now supports the same -scale option as djpeg for "lossless" resize. -An implementation of the JPEG SmartScale extension is required for this -feature. A (draft) specification of the JPEG SmartScale extension is -available as a contributed document at ITU and ISO. Revision 2 or later -of the document is required (latest document version is Revision 3). -The SmartScale extension will enable more features beside lossless resize -in future implementations, as described in the document (new compression -options). - -Add sanity check in BMP reader module to avoid cjpeg crash for empty input -image (thank to Isaev Ildar of ISP RAS, Moscow, RU for reporting this error). - -Add data source and destination managers for read from and write to -memory buffers. New API functions jpeg_mem_src and jpeg_mem_dest. -Thank to Roberto Boni from Italy for the suggestion. - - -Version 7 27-Jun-2009 ----------------------- - -New scaled DCTs implemented. -djpeg now supports scalings N/8 with all N from 1 to 16. -cjpeg now supports scalings 8/N with all N from 1 to 16. -Scaled DCTs with size larger than 8 are now also used for resolving the -common 2x2 chroma subsampling case without additional spatial resampling. -Separate spatial resampling for those kind of files is now only necessary -for N>8 scaling cases. -Furthermore, separate scaled DCT functions are provided for direct resolving -of the common asymmetric subsampling cases (2x1 and 1x2) without additional -spatial resampling. - -cjpeg -quality option has been extended for support of separate quality -settings for luminance and chrominance (or in general, for every provided -quantization table slot). -New API function jpeg_default_qtables() and q_scale_factor array in library. - -Added -nosmooth option to cjpeg, complementary to djpeg. -New variable "do_fancy_downsampling" in library, complement to fancy -upsampling. Fancy upsampling now uses direct DCT scaling with sizes -larger than 8. The old method is not reversible and has been removed. - -Support arithmetic entropy encoding and decoding. -Added files jaricom.c, jcarith.c, jdarith.c. - -Straighten the file structure: -Removed files jidctred.c, jcphuff.c, jchuff.h, jdphuff.c, jdhuff.h. - -jpegtran has a new "lossless" cropping feature. - -Implement -perfect option in jpegtran, new API function -jtransform_perfect_transform() in transupp. (DP 204_perfect.dpatch) - -Better error messages for jpegtran fopen failure. -(DP 203_jpegtran_errmsg.dpatch) - -Fix byte order issue with 16bit PPM/PGM files in rdppm.c/wrppm.c: -according to Netpbm, the de facto standard implementation of the PNM formats, -the most significant byte is first. (DP 203_rdppm.dpatch) - -Add -raw option to rdjpgcom not to mangle the output. -(DP 205_rdjpgcom_raw.dpatch) - -Make rdjpgcom locale aware. (DP 201_rdjpgcom_locale.dpatch) - -Add extern "C" to jpeglib.h. -This avoids the need to put extern "C" { ... } around #include "jpeglib.h" -in your C++ application. Defining the symbol DONT_USE_EXTERN_C in the -configuration prevents this. (DP 202_jpeglib.h_c++.dpatch) - - -Version 6b 27-Mar-1998 ------------------------ - -jpegtran has new features for lossless image transformations (rotation -and flipping) as well as "lossless" reduction to grayscale. - -jpegtran now copies comments by default; it has a -copy switch to enable -copying all APPn blocks as well, or to suppress comments. (Formerly it -always suppressed comments and APPn blocks.) jpegtran now also preserves -JFIF version and resolution information. - -New decompressor library feature: COM and APPn markers found in the input -file can be saved in memory for later use by the application. (Before, -you had to code this up yourself with a custom marker processor.) - -There is an unused field "void * client_data" now in compress and decompress -parameter structs; this may be useful in some applications. - -JFIF version number information is now saved by the decoder and accepted by -the encoder. jpegtran uses this to copy the source file's version number, -to ensure "jpegtran -copy all" won't create bogus files that contain JFXX -extensions but claim to be version 1.01. Applications that generate their -own JFXX extension markers also (finally) have a supported way to cause the -encoder to emit JFIF version number 1.02. - -djpeg's trace mode reports JFIF 1.02 thumbnail images as such, rather -than as unknown APP0 markers. - -In -verbose mode, djpeg and rdjpgcom will try to print the contents of -APP12 markers as text. Some digital cameras store useful text information -in APP12 markers. - -Handling of truncated data streams is more robust: blocks beyond the one in -which the error occurs will be output as uniform gray, or left unchanged -if decoding a progressive JPEG. The appearance no longer depends on the -Huffman tables being used. - -Huffman tables are checked for validity much more carefully than before. - -To avoid the Unisys LZW patent, djpeg's GIF output capability has been -changed to produce "uncompressed GIFs", and cjpeg's GIF input capability -has been removed altogether. We're not happy about it either, but there -seems to be no good alternative. - -The configure script now supports building libjpeg as a shared library -on many flavors of Unix (all the ones that GNU libtool knows how to -build shared libraries for). Use "./configure --enable-shared" to -try this out. - -New jconfig file and makefiles for Microsoft Visual C++ and Developer Studio. -Also, a jconfig file and a build script for Metrowerks CodeWarrior -on Apple Macintosh. makefile.dj has been updated for DJGPP v2, and there -are miscellaneous other minor improvements in the makefiles. - -jmemmac.c now knows how to create temporary files following Mac System 7 -conventions. - -djpeg's -map switch is now able to read raw-format PPM files reliably. - -cjpeg -progressive -restart no longer generates any unnecessary DRI markers. - -Multiple calls to jpeg_simple_progression for a single JPEG object -no longer leak memory. - - -Version 6a 7-Feb-96 --------------------- - -Library initialization sequence modified to detect version mismatches -and struct field packing mismatches between library and calling application. -This change requires applications to be recompiled, but does not require -any application source code change. - -All routine declarations changed to the style "GLOBAL(type) name ...", -that is, GLOBAL, LOCAL, METHODDEF, EXTERN are now macros taking the -routine's return type as an argument. This makes it possible to add -Microsoft-style linkage keywords to all the routines by changing just -these macros. Note that any application code that was using these macros -will have to be changed. - -DCT coefficient quantization tables are now stored in normal array order -rather than zigzag order. Application code that calls jpeg_add_quant_table, -or otherwise manipulates quantization tables directly, will need to be -changed. If you need to make such code work with either older or newer -versions of the library, a test like "#if JPEG_LIB_VERSION >= 61" is -recommended. - -djpeg's trace capability now dumps DQT tables in natural order, not zigzag -order. This allows the trace output to be made into a "-qtables" file -more easily. - -New system-dependent memory manager module for use on Apple Macintosh. - -Fix bug in cjpeg's -smooth option: last one or two scanlines would be -duplicates of the prior line unless the image height mod 16 was 1 or 2. - -Repair minor problems in VMS, BCC, MC6 makefiles. - -New configure script based on latest GNU Autoconf. - -Correct the list of include files needed by MetroWerks C for ccommand(). - -Numerous small documentation updates. - - -Version 6 2-Aug-95 -------------------- - -Progressive JPEG support: library can read and write full progressive JPEG -files. A "buffered image" mode supports incremental decoding for on-the-fly -display of progressive images. Simply recompiling an existing IJG-v5-based -decoder with v6 should allow it to read progressive files, though of course -without any special progressive display. - -New "jpegtran" application performs lossless transcoding between different -JPEG formats; primarily, it can be used to convert baseline to progressive -JPEG and vice versa. In support of jpegtran, the library now allows lossless -reading and writing of JPEG files as DCT coefficient arrays. This ability -may be of use in other applications. - -Notes for programmers: -* We changed jpeg_start_decompress() to be able to suspend; this makes all -decoding modes available to suspending-input applications. However, -existing applications that use suspending input will need to be changed -to check the return value from jpeg_start_decompress(). You don't need to -do anything if you don't use a suspending data source. -* We changed the interface to the virtual array routines: access_virt_array -routines now take a count of the number of rows to access this time. The -last parameter to request_virt_array routines is now interpreted as the -maximum number of rows that may be accessed at once, but not necessarily -the height of every access. - - -Version 5b 15-Mar-95 ---------------------- - -Correct bugs with grayscale images having v_samp_factor > 1. - -jpeg_write_raw_data() now supports output suspension. - -Correct bugs in "configure" script for case of compiling in -a directory other than the one containing the source files. - -Repair bug in jquant1.c: sometimes didn't use as many colors as it could. - -Borland C makefile and jconfig file work under either MS-DOS or OS/2. - -Miscellaneous improvements to documentation. - - -Version 5a 7-Dec-94 --------------------- - -Changed color conversion roundoff behavior so that grayscale values are -represented exactly. (This causes test image files to change.) - -Make ordered dither use 16x16 instead of 4x4 pattern for a small quality -improvement. - -New configure script based on latest GNU Autoconf. -Fix configure script to handle CFLAGS correctly. -Rename *.auto files to *.cfg, so that configure script still works if -file names have been truncated for DOS. - -Fix bug in rdbmp.c: didn't allow for extra data between header and image. - -Modify rdppm.c/wrppm.c to handle 2-byte raw PPM/PGM formats for 12-bit data. - -Fix several bugs in rdrle.c. - -NEED_SHORT_EXTERNAL_NAMES option was broken. - -Revise jerror.h/jerror.c for more flexibility in message table. - -Repair oversight in jmemname.c NO_MKTEMP case: file could be there -but unreadable. - - -Version 5 24-Sep-94 --------------------- - -Version 5 represents a nearly complete redesign and rewrite of the IJG -software. Major user-visible changes include: - * Automatic configuration simplifies installation for most Unix systems. - * A range of speed vs. image quality tradeoffs are supported. - This includes resizing of an image during decompression: scaling down - by a factor of 1/2, 1/4, or 1/8 is handled very efficiently. - * New programs rdjpgcom and wrjpgcom allow insertion and extraction - of text comments in a JPEG file. - -The application programmer's interface to the library has changed completely. -Notable improvements include: - * We have eliminated the use of callback routines for handling the - uncompressed image data. The application now sees the library as a - set of routines that it calls to read or write image data on a - scanline-by-scanline basis. - * The application image data is represented in a conventional interleaved- - pixel format, rather than as a separate array for each color channel. - This can save a copying step in many programs. - * The handling of compressed data has been cleaned up: the application can - supply routines to source or sink the compressed data. It is possible to - suspend processing on source/sink buffer overrun, although this is not - supported in all operating modes. - * All static state has been eliminated from the library, so that multiple - instances of compression or decompression can be active concurrently. - * JPEG abbreviated datastream formats are supported, ie, quantization and - Huffman tables can be stored separately from the image data. - * And not only that, but the documentation of the library has improved - considerably! - - -The last widely used release before the version 5 rewrite was version 4A of -18-Feb-93. Change logs before that point have been discarded, since they -are not of much interest after the rewrite. diff --git a/cmake/externals/libjpeg/jaricom.c b/cmake/externals/libjpeg/jaricom.c deleted file mode 100644 index 690068861..000000000 --- a/cmake/externals/libjpeg/jaricom.c +++ /dev/null @@ -1,153 +0,0 @@ -/* - * jaricom.c - * - * Developed 1997-2011 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains probability estimation tables for common use in - * arithmetic entropy encoding and decoding routines. - * - * This data represents Table D.3 in the JPEG spec (D.2 in the draft), - * ISO/IEC IS 10918-1 and CCITT Recommendation ITU-T T.81, and Table 24 - * in the JBIG spec, ISO/IEC IS 11544 and CCITT Recommendation ITU-T T.82. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - -/* The following #define specifies the packing of the four components - * into the compact INT32 representation. - * Note that this formula must match the actual arithmetic encoder - * and decoder implementation. The implementation has to be changed - * if this formula is changed. - * The current organization is leaned on Markus Kuhn's JBIG - * implementation (jbig_tab.c). - */ - -#define V(i,a,b,c,d) (((INT32)a << 16) | ((INT32)c << 8) | ((INT32)d << 7) | b) - -const INT32 jpeg_aritab[113+1] = { -/* - * Index, Qe_Value, Next_Index_LPS, Next_Index_MPS, Switch_MPS - */ - V( 0, 0x5a1d, 1, 1, 1 ), - V( 1, 0x2586, 14, 2, 0 ), - V( 2, 0x1114, 16, 3, 0 ), - V( 3, 0x080b, 18, 4, 0 ), - V( 4, 0x03d8, 20, 5, 0 ), - V( 5, 0x01da, 23, 6, 0 ), - V( 6, 0x00e5, 25, 7, 0 ), - V( 7, 0x006f, 28, 8, 0 ), - V( 8, 0x0036, 30, 9, 0 ), - V( 9, 0x001a, 33, 10, 0 ), - V( 10, 0x000d, 35, 11, 0 ), - V( 11, 0x0006, 9, 12, 0 ), - V( 12, 0x0003, 10, 13, 0 ), - V( 13, 0x0001, 12, 13, 0 ), - V( 14, 0x5a7f, 15, 15, 1 ), - V( 15, 0x3f25, 36, 16, 0 ), - V( 16, 0x2cf2, 38, 17, 0 ), - V( 17, 0x207c, 39, 18, 0 ), - V( 18, 0x17b9, 40, 19, 0 ), - V( 19, 0x1182, 42, 20, 0 ), - V( 20, 0x0cef, 43, 21, 0 ), - V( 21, 0x09a1, 45, 22, 0 ), - V( 22, 0x072f, 46, 23, 0 ), - V( 23, 0x055c, 48, 24, 0 ), - V( 24, 0x0406, 49, 25, 0 ), - V( 25, 0x0303, 51, 26, 0 ), - V( 26, 0x0240, 52, 27, 0 ), - V( 27, 0x01b1, 54, 28, 0 ), - V( 28, 0x0144, 56, 29, 0 ), - V( 29, 0x00f5, 57, 30, 0 ), - V( 30, 0x00b7, 59, 31, 0 ), - V( 31, 0x008a, 60, 32, 0 ), - V( 32, 0x0068, 62, 33, 0 ), - V( 33, 0x004e, 63, 34, 0 ), - V( 34, 0x003b, 32, 35, 0 ), - V( 35, 0x002c, 33, 9, 0 ), - V( 36, 0x5ae1, 37, 37, 1 ), - V( 37, 0x484c, 64, 38, 0 ), - V( 38, 0x3a0d, 65, 39, 0 ), - V( 39, 0x2ef1, 67, 40, 0 ), - V( 40, 0x261f, 68, 41, 0 ), - V( 41, 0x1f33, 69, 42, 0 ), - V( 42, 0x19a8, 70, 43, 0 ), - V( 43, 0x1518, 72, 44, 0 ), - V( 44, 0x1177, 73, 45, 0 ), - V( 45, 0x0e74, 74, 46, 0 ), - V( 46, 0x0bfb, 75, 47, 0 ), - V( 47, 0x09f8, 77, 48, 0 ), - V( 48, 0x0861, 78, 49, 0 ), - V( 49, 0x0706, 79, 50, 0 ), - V( 50, 0x05cd, 48, 51, 0 ), - V( 51, 0x04de, 50, 52, 0 ), - V( 52, 0x040f, 50, 53, 0 ), - V( 53, 0x0363, 51, 54, 0 ), - V( 54, 0x02d4, 52, 55, 0 ), - V( 55, 0x025c, 53, 56, 0 ), - V( 56, 0x01f8, 54, 57, 0 ), - V( 57, 0x01a4, 55, 58, 0 ), - V( 58, 0x0160, 56, 59, 0 ), - V( 59, 0x0125, 57, 60, 0 ), - V( 60, 0x00f6, 58, 61, 0 ), - V( 61, 0x00cb, 59, 62, 0 ), - V( 62, 0x00ab, 61, 63, 0 ), - V( 63, 0x008f, 61, 32, 0 ), - V( 64, 0x5b12, 65, 65, 1 ), - V( 65, 0x4d04, 80, 66, 0 ), - V( 66, 0x412c, 81, 67, 0 ), - V( 67, 0x37d8, 82, 68, 0 ), - V( 68, 0x2fe8, 83, 69, 0 ), - V( 69, 0x293c, 84, 70, 0 ), - V( 70, 0x2379, 86, 71, 0 ), - V( 71, 0x1edf, 87, 72, 0 ), - V( 72, 0x1aa9, 87, 73, 0 ), - V( 73, 0x174e, 72, 74, 0 ), - V( 74, 0x1424, 72, 75, 0 ), - V( 75, 0x119c, 74, 76, 0 ), - V( 76, 0x0f6b, 74, 77, 0 ), - V( 77, 0x0d51, 75, 78, 0 ), - V( 78, 0x0bb6, 77, 79, 0 ), - V( 79, 0x0a40, 77, 48, 0 ), - V( 80, 0x5832, 80, 81, 1 ), - V( 81, 0x4d1c, 88, 82, 0 ), - V( 82, 0x438e, 89, 83, 0 ), - V( 83, 0x3bdd, 90, 84, 0 ), - V( 84, 0x34ee, 91, 85, 0 ), - V( 85, 0x2eae, 92, 86, 0 ), - V( 86, 0x299a, 93, 87, 0 ), - V( 87, 0x2516, 86, 71, 0 ), - V( 88, 0x5570, 88, 89, 1 ), - V( 89, 0x4ca9, 95, 90, 0 ), - V( 90, 0x44d9, 96, 91, 0 ), - V( 91, 0x3e22, 97, 92, 0 ), - V( 92, 0x3824, 99, 93, 0 ), - V( 93, 0x32b4, 99, 94, 0 ), - V( 94, 0x2e17, 93, 86, 0 ), - V( 95, 0x56a8, 95, 96, 1 ), - V( 96, 0x4f46, 101, 97, 0 ), - V( 97, 0x47e5, 102, 98, 0 ), - V( 98, 0x41cf, 103, 99, 0 ), - V( 99, 0x3c3d, 104, 100, 0 ), - V( 100, 0x375e, 99, 93, 0 ), - V( 101, 0x5231, 105, 102, 0 ), - V( 102, 0x4c0f, 106, 103, 0 ), - V( 103, 0x4639, 107, 104, 0 ), - V( 104, 0x415e, 103, 99, 0 ), - V( 105, 0x5627, 105, 106, 1 ), - V( 106, 0x50e7, 108, 107, 0 ), - V( 107, 0x4b85, 109, 103, 0 ), - V( 108, 0x5597, 110, 109, 0 ), - V( 109, 0x504f, 111, 107, 0 ), - V( 110, 0x5a10, 110, 111, 1 ), - V( 111, 0x5522, 112, 109, 0 ), - V( 112, 0x59eb, 112, 111, 1 ), -/* - * This last entry is used for fixed probability estimate of 0.5 - * as suggested in Section 10.3 Table 5 of ITU-T Rec. T.851. - */ - V( 113, 0x5a1d, 113, 113, 0 ) -}; diff --git a/cmake/externals/libjpeg/jcapimin.c b/cmake/externals/libjpeg/jcapimin.c deleted file mode 100644 index 639ce86f4..000000000 --- a/cmake/externals/libjpeg/jcapimin.c +++ /dev/null @@ -1,288 +0,0 @@ -/* - * jcapimin.c - * - * Copyright (C) 1994-1998, Thomas G. Lane. - * Modified 2003-2010 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains application interface code for the compression half - * of the JPEG library. These are the "minimum" API routines that may be - * needed in either the normal full-compression case or the transcoding-only - * case. - * - * Most of the routines intended to be called directly by an application - * are in this file or in jcapistd.c. But also see jcparam.c for - * parameter-setup helper routines, jcomapi.c for routines shared by - * compression and decompression, and jctrans.c for the transcoding case. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* - * Initialization of a JPEG compression object. - * The error manager must already be set up (in case memory manager fails). - */ - -GLOBAL(void) -jpeg_CreateCompress (j_compress_ptr cinfo, int version, size_t structsize) -{ - int i; - - /* Guard against version mismatches between library and caller. */ - cinfo->mem = NULL; /* so jpeg_destroy knows mem mgr not called */ - if (version != JPEG_LIB_VERSION) - ERREXIT2(cinfo, JERR_BAD_LIB_VERSION, JPEG_LIB_VERSION, version); - if (structsize != SIZEOF(struct jpeg_compress_struct)) - ERREXIT2(cinfo, JERR_BAD_STRUCT_SIZE, - (int) SIZEOF(struct jpeg_compress_struct), (int) structsize); - - /* For debugging purposes, we zero the whole master structure. - * But the application has already set the err pointer, and may have set - * client_data, so we have to save and restore those fields. - * Note: if application hasn't set client_data, tools like Purify may - * complain here. - */ - { - struct jpeg_error_mgr * err = cinfo->err; - void * client_data = cinfo->client_data; /* ignore Purify complaint here */ - MEMZERO(cinfo, SIZEOF(struct jpeg_compress_struct)); - cinfo->err = err; - cinfo->client_data = client_data; - } - cinfo->is_decompressor = FALSE; - - /* Initialize a memory manager instance for this object */ - jinit_memory_mgr((j_common_ptr) cinfo); - - /* Zero out pointers to permanent structures. */ - cinfo->progress = NULL; - cinfo->dest = NULL; - - cinfo->comp_info = NULL; - - for (i = 0; i < NUM_QUANT_TBLS; i++) { - cinfo->quant_tbl_ptrs[i] = NULL; - cinfo->q_scale_factor[i] = 100; - } - - for (i = 0; i < NUM_HUFF_TBLS; i++) { - cinfo->dc_huff_tbl_ptrs[i] = NULL; - cinfo->ac_huff_tbl_ptrs[i] = NULL; - } - - /* Must do it here for emit_dqt in case jpeg_write_tables is used */ - cinfo->block_size = DCTSIZE; - cinfo->natural_order = jpeg_natural_order; - cinfo->lim_Se = DCTSIZE2-1; - - cinfo->script_space = NULL; - - cinfo->input_gamma = 1.0; /* in case application forgets */ - - /* OK, I'm ready */ - cinfo->global_state = CSTATE_START; -} - - -/* - * Destruction of a JPEG compression object - */ - -GLOBAL(void) -jpeg_destroy_compress (j_compress_ptr cinfo) -{ - jpeg_destroy((j_common_ptr) cinfo); /* use common routine */ -} - - -/* - * Abort processing of a JPEG compression operation, - * but don't destroy the object itself. - */ - -GLOBAL(void) -jpeg_abort_compress (j_compress_ptr cinfo) -{ - jpeg_abort((j_common_ptr) cinfo); /* use common routine */ -} - - -/* - * Forcibly suppress or un-suppress all quantization and Huffman tables. - * Marks all currently defined tables as already written (if suppress) - * or not written (if !suppress). This will control whether they get emitted - * by a subsequent jpeg_start_compress call. - * - * This routine is exported for use by applications that want to produce - * abbreviated JPEG datastreams. It logically belongs in jcparam.c, but - * since it is called by jpeg_start_compress, we put it here --- otherwise - * jcparam.o would be linked whether the application used it or not. - */ - -GLOBAL(void) -jpeg_suppress_tables (j_compress_ptr cinfo, boolean suppress) -{ - int i; - JQUANT_TBL * qtbl; - JHUFF_TBL * htbl; - - for (i = 0; i < NUM_QUANT_TBLS; i++) { - if ((qtbl = cinfo->quant_tbl_ptrs[i]) != NULL) - qtbl->sent_table = suppress; - } - - for (i = 0; i < NUM_HUFF_TBLS; i++) { - if ((htbl = cinfo->dc_huff_tbl_ptrs[i]) != NULL) - htbl->sent_table = suppress; - if ((htbl = cinfo->ac_huff_tbl_ptrs[i]) != NULL) - htbl->sent_table = suppress; - } -} - - -/* - * Finish JPEG compression. - * - * If a multipass operating mode was selected, this may do a great deal of - * work including most of the actual output. - */ - -GLOBAL(void) -jpeg_finish_compress (j_compress_ptr cinfo) -{ - JDIMENSION iMCU_row; - - if (cinfo->global_state == CSTATE_SCANNING || - cinfo->global_state == CSTATE_RAW_OK) { - /* Terminate first pass */ - if (cinfo->next_scanline < cinfo->image_height) - ERREXIT(cinfo, JERR_TOO_LITTLE_DATA); - (*cinfo->master->finish_pass) (cinfo); - } else if (cinfo->global_state != CSTATE_WRCOEFS) - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - /* Perform any remaining passes */ - while (! cinfo->master->is_last_pass) { - (*cinfo->master->prepare_for_pass) (cinfo); - for (iMCU_row = 0; iMCU_row < cinfo->total_iMCU_rows; iMCU_row++) { - if (cinfo->progress != NULL) { - cinfo->progress->pass_counter = (long) iMCU_row; - cinfo->progress->pass_limit = (long) cinfo->total_iMCU_rows; - (*cinfo->progress->progress_monitor) ((j_common_ptr) cinfo); - } - /* We bypass the main controller and invoke coef controller directly; - * all work is being done from the coefficient buffer. - */ - if (! (*cinfo->coef->compress_data) (cinfo, (JSAMPIMAGE) NULL)) - ERREXIT(cinfo, JERR_CANT_SUSPEND); - } - (*cinfo->master->finish_pass) (cinfo); - } - /* Write EOI, do final cleanup */ - (*cinfo->marker->write_file_trailer) (cinfo); - (*cinfo->dest->term_destination) (cinfo); - /* We can use jpeg_abort to release memory and reset global_state */ - jpeg_abort((j_common_ptr) cinfo); -} - - -/* - * Write a special marker. - * This is only recommended for writing COM or APPn markers. - * Must be called after jpeg_start_compress() and before - * first call to jpeg_write_scanlines() or jpeg_write_raw_data(). - */ - -GLOBAL(void) -jpeg_write_marker (j_compress_ptr cinfo, int marker, - const JOCTET *dataptr, unsigned int datalen) -{ - JMETHOD(void, write_marker_byte, (j_compress_ptr info, int val)); - - if (cinfo->next_scanline != 0 || - (cinfo->global_state != CSTATE_SCANNING && - cinfo->global_state != CSTATE_RAW_OK && - cinfo->global_state != CSTATE_WRCOEFS)) - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - - (*cinfo->marker->write_marker_header) (cinfo, marker, datalen); - write_marker_byte = cinfo->marker->write_marker_byte; /* copy for speed */ - while (datalen--) { - (*write_marker_byte) (cinfo, *dataptr); - dataptr++; - } -} - -/* Same, but piecemeal. */ - -GLOBAL(void) -jpeg_write_m_header (j_compress_ptr cinfo, int marker, unsigned int datalen) -{ - if (cinfo->next_scanline != 0 || - (cinfo->global_state != CSTATE_SCANNING && - cinfo->global_state != CSTATE_RAW_OK && - cinfo->global_state != CSTATE_WRCOEFS)) - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - - (*cinfo->marker->write_marker_header) (cinfo, marker, datalen); -} - -GLOBAL(void) -jpeg_write_m_byte (j_compress_ptr cinfo, int val) -{ - (*cinfo->marker->write_marker_byte) (cinfo, val); -} - - -/* - * Alternate compression function: just write an abbreviated table file. - * Before calling this, all parameters and a data destination must be set up. - * - * To produce a pair of files containing abbreviated tables and abbreviated - * image data, one would proceed as follows: - * - * initialize JPEG object - * set JPEG parameters - * set destination to table file - * jpeg_write_tables(cinfo); - * set destination to image file - * jpeg_start_compress(cinfo, FALSE); - * write data... - * jpeg_finish_compress(cinfo); - * - * jpeg_write_tables has the side effect of marking all tables written - * (same as jpeg_suppress_tables(..., TRUE)). Thus a subsequent start_compress - * will not re-emit the tables unless it is passed write_all_tables=TRUE. - */ - -GLOBAL(void) -jpeg_write_tables (j_compress_ptr cinfo) -{ - if (cinfo->global_state != CSTATE_START) - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - - /* (Re)initialize error mgr and destination modules */ - (*cinfo->err->reset_error_mgr) ((j_common_ptr) cinfo); - (*cinfo->dest->init_destination) (cinfo); - /* Initialize the marker writer ... bit of a crock to do it here. */ - jinit_marker_writer(cinfo); - /* Write them tables! */ - (*cinfo->marker->write_tables_only) (cinfo); - /* And clean up. */ - (*cinfo->dest->term_destination) (cinfo); - /* - * In library releases up through v6a, we called jpeg_abort() here to free - * any working memory allocated by the destination manager and marker - * writer. Some applications had a problem with that: they allocated space - * of their own from the library memory manager, and didn't want it to go - * away during write_tables. So now we do nothing. This will cause a - * memory leak if an app calls write_tables repeatedly without doing a full - * compression cycle or otherwise resetting the JPEG object. However, that - * seems less bad than unexpectedly freeing memory in the normal case. - * An app that prefers the old behavior can call jpeg_abort for itself after - * each call to jpeg_write_tables(). - */ -} diff --git a/cmake/externals/libjpeg/jcapistd.c b/cmake/externals/libjpeg/jcapistd.c deleted file mode 100644 index 0917afa97..000000000 --- a/cmake/externals/libjpeg/jcapistd.c +++ /dev/null @@ -1,162 +0,0 @@ -/* - * jcapistd.c - * - * Copyright (C) 1994-1996, Thomas G. Lane. - * Modified 2013 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains application interface code for the compression half - * of the JPEG library. These are the "standard" API routines that are - * used in the normal full-compression case. They are not used by a - * transcoding-only application. Note that if an application links in - * jpeg_start_compress, it will end up linking in the entire compressor. - * We thus must separate this file from jcapimin.c to avoid linking the - * whole compression library into a transcoder. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* - * Compression initialization. - * Before calling this, all parameters and a data destination must be set up. - * - * We require a write_all_tables parameter as a failsafe check when writing - * multiple datastreams from the same compression object. Since prior runs - * will have left all the tables marked sent_table=TRUE, a subsequent run - * would emit an abbreviated stream (no tables) by default. This may be what - * is wanted, but for safety's sake it should not be the default behavior: - * programmers should have to make a deliberate choice to emit abbreviated - * images. Therefore the documentation and examples should encourage people - * to pass write_all_tables=TRUE; then it will take active thought to do the - * wrong thing. - */ - -GLOBAL(void) -jpeg_start_compress (j_compress_ptr cinfo, boolean write_all_tables) -{ - if (cinfo->global_state != CSTATE_START) - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - - if (write_all_tables) - jpeg_suppress_tables(cinfo, FALSE); /* mark all tables to be written */ - - /* (Re)initialize error mgr and destination modules */ - (*cinfo->err->reset_error_mgr) ((j_common_ptr) cinfo); - (*cinfo->dest->init_destination) (cinfo); - /* Perform master selection of active modules */ - jinit_compress_master(cinfo); - /* Set up for the first pass */ - (*cinfo->master->prepare_for_pass) (cinfo); - /* Ready for application to drive first pass through jpeg_write_scanlines - * or jpeg_write_raw_data. - */ - cinfo->next_scanline = 0; - cinfo->global_state = (cinfo->raw_data_in ? CSTATE_RAW_OK : CSTATE_SCANNING); -} - - -/* - * Write some scanlines of data to the JPEG compressor. - * - * The return value will be the number of lines actually written. - * This should be less than the supplied num_lines only in case that - * the data destination module has requested suspension of the compressor, - * or if more than image_height scanlines are passed in. - * - * Note: we warn about excess calls to jpeg_write_scanlines() since - * this likely signals an application programmer error. However, - * excess scanlines passed in the last valid call are *silently* ignored, - * so that the application need not adjust num_lines for end-of-image - * when using a multiple-scanline buffer. - */ - -GLOBAL(JDIMENSION) -jpeg_write_scanlines (j_compress_ptr cinfo, JSAMPARRAY scanlines, - JDIMENSION num_lines) -{ - JDIMENSION row_ctr, rows_left; - - if (cinfo->global_state != CSTATE_SCANNING) - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - if (cinfo->next_scanline >= cinfo->image_height) - WARNMS(cinfo, JWRN_TOO_MUCH_DATA); - - /* Call progress monitor hook if present */ - if (cinfo->progress != NULL) { - cinfo->progress->pass_counter = (long) cinfo->next_scanline; - cinfo->progress->pass_limit = (long) cinfo->image_height; - (*cinfo->progress->progress_monitor) ((j_common_ptr) cinfo); - } - - /* Give master control module another chance if this is first call to - * jpeg_write_scanlines. This lets output of the frame/scan headers be - * delayed so that application can write COM, etc, markers between - * jpeg_start_compress and jpeg_write_scanlines. - */ - if (cinfo->master->call_pass_startup) - (*cinfo->master->pass_startup) (cinfo); - - /* Ignore any extra scanlines at bottom of image. */ - rows_left = cinfo->image_height - cinfo->next_scanline; - if (num_lines > rows_left) - num_lines = rows_left; - - row_ctr = 0; - (*cinfo->main->process_data) (cinfo, scanlines, &row_ctr, num_lines); - cinfo->next_scanline += row_ctr; - return row_ctr; -} - - -/* - * Alternate entry point to write raw data. - * Processes exactly one iMCU row per call, unless suspended. - */ - -GLOBAL(JDIMENSION) -jpeg_write_raw_data (j_compress_ptr cinfo, JSAMPIMAGE data, - JDIMENSION num_lines) -{ - JDIMENSION lines_per_iMCU_row; - - if (cinfo->global_state != CSTATE_RAW_OK) - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - if (cinfo->next_scanline >= cinfo->image_height) { - WARNMS(cinfo, JWRN_TOO_MUCH_DATA); - return 0; - } - - /* Call progress monitor hook if present */ - if (cinfo->progress != NULL) { - cinfo->progress->pass_counter = (long) cinfo->next_scanline; - cinfo->progress->pass_limit = (long) cinfo->image_height; - (*cinfo->progress->progress_monitor) ((j_common_ptr) cinfo); - } - - /* Give master control module another chance if this is first call to - * jpeg_write_raw_data. This lets output of the frame/scan headers be - * delayed so that application can write COM, etc, markers between - * jpeg_start_compress and jpeg_write_raw_data. - */ - if (cinfo->master->call_pass_startup) - (*cinfo->master->pass_startup) (cinfo); - - /* Verify that at least one iMCU row has been passed. */ - lines_per_iMCU_row = cinfo->max_v_samp_factor * cinfo->min_DCT_v_scaled_size; - if (num_lines < lines_per_iMCU_row) - ERREXIT(cinfo, JERR_BUFFER_SIZE); - - /* Directly compress the row. */ - if (! (*cinfo->coef->compress_data) (cinfo, data)) { - /* If compressor did not consume the whole row, suspend processing. */ - return 0; - } - - /* OK, we processed one iMCU row. */ - cinfo->next_scanline += lines_per_iMCU_row; - return lines_per_iMCU_row; -} diff --git a/cmake/externals/libjpeg/jcarith.c b/cmake/externals/libjpeg/jcarith.c deleted file mode 100644 index 1b45089a3..000000000 --- a/cmake/externals/libjpeg/jcarith.c +++ /dev/null @@ -1,945 +0,0 @@ -/* - * jcarith.c - * - * Developed 1997-2020 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains portable arithmetic entropy encoding routines for JPEG - * (implementing the ISO/IEC IS 10918-1 and CCITT Recommendation ITU-T T.81). - * - * Both sequential and progressive modes are supported in this single module. - * - * Suspension is not currently supported in this module. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* Expanded entropy encoder object for arithmetic encoding. */ - -typedef struct { - struct jpeg_entropy_encoder pub; /* public fields */ - - INT32 c; /* C register, base of coding interval, layout as in sec. D.1.3 */ - INT32 a; /* A register, normalized size of coding interval */ - INT32 sc; /* counter for stacked 0xFF values which might overflow */ - INT32 zc; /* counter for pending 0x00 output values which might * - * be discarded at the end ("Pacman" termination) */ - int ct; /* bit shift counter, determines when next byte will be written */ - int buffer; /* buffer for most recent output byte != 0xFF */ - - int last_dc_val[MAX_COMPS_IN_SCAN]; /* last DC coef for each component */ - int dc_context[MAX_COMPS_IN_SCAN]; /* context index for DC conditioning */ - - unsigned int restarts_to_go; /* MCUs left in this restart interval */ - int next_restart_num; /* next restart number to write (0-7) */ - - /* Pointers to statistics areas (these workspaces have image lifespan) */ - unsigned char * dc_stats[NUM_ARITH_TBLS]; - unsigned char * ac_stats[NUM_ARITH_TBLS]; - - /* Statistics bin for coding with fixed probability 0.5 */ - unsigned char fixed_bin[4]; -} arith_entropy_encoder; - -typedef arith_entropy_encoder * arith_entropy_ptr; - -/* The following two definitions specify the allocation chunk size - * for the statistics area. - * According to sections F.1.4.4.1.3 and F.1.4.4.2, we need at least - * 49 statistics bins for DC, and 245 statistics bins for AC coding. - * - * We use a compact representation with 1 byte per statistics bin, - * thus the numbers directly represent byte sizes. - * This 1 byte per statistics bin contains the meaning of the MPS - * (more probable symbol) in the highest bit (mask 0x80), and the - * index into the probability estimation state machine table - * in the lower bits (mask 0x7F). - */ - -#define DC_STAT_BINS 64 -#define AC_STAT_BINS 256 - -/* NOTE: Uncomment the following #define if you want to use the - * given formula for calculating the AC conditioning parameter Kx - * for spectral selection progressive coding in section G.1.3.2 - * of the spec (Kx = Kmin + SRL (8 + Se - Kmin) 4). - * Although the spec and P&M authors claim that this "has proven - * to give good results for 8 bit precision samples", I'm not - * convinced yet that this is really beneficial. - * Early tests gave only very marginal compression enhancements - * (a few - around 5 or so - bytes even for very large files), - * which would turn out rather negative if we'd suppress the - * DAC (Define Arithmetic Conditioning) marker segments for - * the default parameters in the future. - * Note that currently the marker writing module emits 12-byte - * DAC segments for a full-component scan in a color image. - * This is not worth worrying about IMHO. However, since the - * spec defines the default values to be used if the tables - * are omitted (unlike Huffman tables, which are required - * anyway), one might optimize this behaviour in the future, - * and then it would be disadvantageous to use custom tables if - * they don't provide sufficient gain to exceed the DAC size. - * - * On the other hand, I'd consider it as a reasonable result - * that the conditioning has no significant influence on the - * compression performance. This means that the basic - * statistical model is already rather stable. - * - * Thus, at the moment, we use the default conditioning values - * anyway, and do not use the custom formula. - * -#define CALCULATE_SPECTRAL_CONDITIONING - */ - -/* IRIGHT_SHIFT is like RIGHT_SHIFT, but works on int rather than INT32. - * We assume that int right shift is unsigned if INT32 right shift is, - * which should be safe. - */ - -#ifdef RIGHT_SHIFT_IS_UNSIGNED -#define ISHIFT_TEMPS int ishift_temp; -#define IRIGHT_SHIFT(x,shft) \ - ((ishift_temp = (x)) < 0 ? \ - (ishift_temp >> (shft)) | ((~0) << (16-(shft))) : \ - (ishift_temp >> (shft))) -#else -#define ISHIFT_TEMPS -#define IRIGHT_SHIFT(x,shft) ((x) >> (shft)) -#endif - - -LOCAL(void) -emit_byte (int val, j_compress_ptr cinfo) -/* Write next output byte; we do not support suspension in this module. */ -{ - struct jpeg_destination_mgr * dest = cinfo->dest; - - *dest->next_output_byte++ = (JOCTET) val; - if (--dest->free_in_buffer == 0) - if (! (*dest->empty_output_buffer) (cinfo)) - ERREXIT(cinfo, JERR_CANT_SUSPEND); -} - - -/* - * Finish up at the end of an arithmetic-compressed scan. - */ - -METHODDEF(void) -finish_pass (j_compress_ptr cinfo) -{ - arith_entropy_ptr e = (arith_entropy_ptr) cinfo->entropy; - INT32 temp; - - /* Section D.1.8: Termination of encoding */ - - /* Find the e->c in the coding interval with the largest - * number of trailing zero bits */ - if ((temp = (e->a - 1 + e->c) & 0xFFFF0000L) < e->c) - e->c = temp + 0x8000L; - else - e->c = temp; - /* Send remaining bytes to output */ - e->c <<= e->ct; - if (e->c & 0xF8000000L) { - /* One final overflow has to be handled */ - if (e->buffer >= 0) { - if (e->zc) - do emit_byte(0x00, cinfo); - while (--e->zc); - emit_byte(e->buffer + 1, cinfo); - if (e->buffer + 1 == 0xFF) - emit_byte(0x00, cinfo); - } - e->zc += e->sc; /* carry-over converts stacked 0xFF bytes to 0x00 */ - e->sc = 0; - } else { - if (e->buffer == 0) - ++e->zc; - else if (e->buffer >= 0) { - if (e->zc) - do emit_byte(0x00, cinfo); - while (--e->zc); - emit_byte(e->buffer, cinfo); - } - if (e->sc) { - if (e->zc) - do emit_byte(0x00, cinfo); - while (--e->zc); - do { - emit_byte(0xFF, cinfo); - emit_byte(0x00, cinfo); - } while (--e->sc); - } - } - /* Output final bytes only if they are not 0x00 */ - if (e->c & 0x7FFF800L) { - if (e->zc) /* output final pending zero bytes */ - do emit_byte(0x00, cinfo); - while (--e->zc); - emit_byte((int) ((e->c >> 19) & 0xFF), cinfo); - if (((e->c >> 19) & 0xFF) == 0xFF) - emit_byte(0x00, cinfo); - if (e->c & 0x7F800L) { - emit_byte((int) ((e->c >> 11) & 0xFF), cinfo); - if (((e->c >> 11) & 0xFF) == 0xFF) - emit_byte(0x00, cinfo); - } - } -} - - -/* - * The core arithmetic encoding routine (common in JPEG and JBIG). - * This needs to go as fast as possible. - * Machine-dependent optimization facilities - * are not utilized in this portable implementation. - * However, this code should be fairly efficient and - * may be a good base for further optimizations anyway. - * - * Parameter 'val' to be encoded may be 0 or 1 (binary decision). - * - * Note: I've added full "Pacman" termination support to the - * byte output routines, which is equivalent to the optional - * Discard_final_zeros procedure (Figure D.15) in the spec. - * Thus, we always produce the shortest possible output - * stream compliant to the spec (no trailing zero bytes, - * except for FF stuffing). - * - * I've also introduced a new scheme for accessing - * the probability estimation state machine table, - * derived from Markus Kuhn's JBIG implementation. - */ - -LOCAL(void) -arith_encode (j_compress_ptr cinfo, unsigned char *st, int val) -{ - register arith_entropy_ptr e = (arith_entropy_ptr) cinfo->entropy; - register unsigned char nl, nm; - register INT32 qe, temp; - register int sv; - - /* Fetch values from our compact representation of Table D.3(D.2): - * Qe values and probability estimation state machine - */ - sv = *st; - qe = jpeg_aritab[sv & 0x7F]; /* => Qe_Value */ - nl = qe & 0xFF; qe >>= 8; /* Next_Index_LPS + Switch_MPS */ - nm = qe & 0xFF; qe >>= 8; /* Next_Index_MPS */ - - /* Encode & estimation procedures per sections D.1.4 & D.1.5 */ - e->a -= qe; - if (val != (sv >> 7)) { - /* Encode the less probable symbol */ - if (e->a >= qe) { - /* If the interval size (qe) for the less probable symbol (LPS) - * is larger than the interval size for the MPS, then exchange - * the two symbols for coding efficiency, otherwise code the LPS - * as usual: */ - e->c += e->a; - e->a = qe; - } - *st = (sv & 0x80) ^ nl; /* Estimate_after_LPS */ - } else { - /* Encode the more probable symbol */ - if (e->a >= 0x8000L) - return; /* A >= 0x8000 -> ready, no renormalization required */ - if (e->a < qe) { - /* If the interval size (qe) for the less probable symbol (LPS) - * is larger than the interval size for the MPS, then exchange - * the two symbols for coding efficiency: */ - e->c += e->a; - e->a = qe; - } - *st = (sv & 0x80) ^ nm; /* Estimate_after_MPS */ - } - - /* Renormalization & data output per section D.1.6 */ - do { - e->a <<= 1; - e->c <<= 1; - if (--e->ct == 0) { - /* Another byte is ready for output */ - temp = e->c >> 19; - if (temp > 0xFF) { - /* Handle overflow over all stacked 0xFF bytes */ - if (e->buffer >= 0) { - if (e->zc) - do emit_byte(0x00, cinfo); - while (--e->zc); - emit_byte(e->buffer + 1, cinfo); - if (e->buffer + 1 == 0xFF) - emit_byte(0x00, cinfo); - } - e->zc += e->sc; /* carry-over converts stacked 0xFF bytes to 0x00 */ - e->sc = 0; - /* Note: The 3 spacer bits in the C register guarantee - * that the new buffer byte can't be 0xFF here - * (see page 160 in the P&M JPEG book). */ - /* New output byte, might overflow later */ - e->buffer = (int) (temp & 0xFF); - } else if (temp == 0xFF) { - ++e->sc; /* stack 0xFF byte (which might overflow later) */ - } else { - /* Output all stacked 0xFF bytes, they will not overflow any more */ - if (e->buffer == 0) - ++e->zc; - else if (e->buffer >= 0) { - if (e->zc) - do emit_byte(0x00, cinfo); - while (--e->zc); - emit_byte(e->buffer, cinfo); - } - if (e->sc) { - if (e->zc) - do emit_byte(0x00, cinfo); - while (--e->zc); - do { - emit_byte(0xFF, cinfo); - emit_byte(0x00, cinfo); - } while (--e->sc); - } - /* New output byte (can still overflow) */ - e->buffer = (int) (temp & 0xFF); - } - e->c &= 0x7FFFFL; - e->ct += 8; - } - } while (e->a < 0x8000L); -} - - -/* - * Emit a restart marker & resynchronize predictions. - */ - -LOCAL(void) -emit_restart (j_compress_ptr cinfo, int restart_num) -{ - arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; - int ci; - jpeg_component_info * compptr; - - finish_pass(cinfo); - - emit_byte(0xFF, cinfo); - emit_byte(JPEG_RST0 + restart_num, cinfo); - - /* Re-initialize statistics areas */ - for (ci = 0; ci < cinfo->comps_in_scan; ci++) { - compptr = cinfo->cur_comp_info[ci]; - /* DC needs no table for refinement scan */ - if (cinfo->Ss == 0 && cinfo->Ah == 0) { - MEMZERO(entropy->dc_stats[compptr->dc_tbl_no], DC_STAT_BINS); - /* Reset DC predictions to 0 */ - entropy->last_dc_val[ci] = 0; - entropy->dc_context[ci] = 0; - } - /* AC needs no table when not present */ - if (cinfo->Se) { - MEMZERO(entropy->ac_stats[compptr->ac_tbl_no], AC_STAT_BINS); - } - } - - /* Reset arithmetic encoding variables */ - entropy->c = 0; - entropy->a = 0x10000L; - entropy->sc = 0; - entropy->zc = 0; - entropy->ct = 11; - entropy->buffer = -1; /* empty */ -} - - -/* - * MCU encoding for DC initial scan (either spectral selection, - * or first pass of successive approximation). - */ - -METHODDEF(boolean) -encode_mcu_DC_first (j_compress_ptr cinfo, JBLOCKARRAY MCU_data) -{ - arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; - unsigned char *st; - int blkn, ci, tbl; - int v, v2, m; - ISHIFT_TEMPS - - /* Emit restart marker if needed */ - if (cinfo->restart_interval) { - if (entropy->restarts_to_go == 0) { - emit_restart(cinfo, entropy->next_restart_num); - entropy->restarts_to_go = cinfo->restart_interval; - entropy->next_restart_num++; - entropy->next_restart_num &= 7; - } - entropy->restarts_to_go--; - } - - /* Encode the MCU data blocks */ - for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { - ci = cinfo->MCU_membership[blkn]; - tbl = cinfo->cur_comp_info[ci]->dc_tbl_no; - - /* Compute the DC value after the required point transform by Al. - * This is simply an arithmetic right shift. - */ - m = IRIGHT_SHIFT((int) (MCU_data[blkn][0][0]), cinfo->Al); - - /* Sections F.1.4.1 & F.1.4.4.1: Encoding of DC coefficients */ - - /* Table F.4: Point to statistics bin S0 for DC coefficient coding */ - st = entropy->dc_stats[tbl] + entropy->dc_context[ci]; - - /* Figure F.4: Encode_DC_DIFF */ - if ((v = m - entropy->last_dc_val[ci]) == 0) { - arith_encode(cinfo, st, 0); - entropy->dc_context[ci] = 0; /* zero diff category */ - } else { - entropy->last_dc_val[ci] = m; - arith_encode(cinfo, st, 1); - /* Figure F.6: Encoding nonzero value v */ - /* Figure F.7: Encoding the sign of v */ - if (v > 0) { - arith_encode(cinfo, st + 1, 0); /* Table F.4: SS = S0 + 1 */ - st += 2; /* Table F.4: SP = S0 + 2 */ - entropy->dc_context[ci] = 4; /* small positive diff category */ - } else { - v = -v; - arith_encode(cinfo, st + 1, 1); /* Table F.4: SS = S0 + 1 */ - st += 3; /* Table F.4: SN = S0 + 3 */ - entropy->dc_context[ci] = 8; /* small negative diff category */ - } - /* Figure F.8: Encoding the magnitude category of v */ - m = 0; - if (v -= 1) { - arith_encode(cinfo, st, 1); - m = 1; - v2 = v; - st = entropy->dc_stats[tbl] + 20; /* Table F.4: X1 = 20 */ - while (v2 >>= 1) { - arith_encode(cinfo, st, 1); - m <<= 1; - st += 1; - } - } - arith_encode(cinfo, st, 0); - /* Section F.1.4.4.1.2: Establish dc_context conditioning category */ - if (m < (int) ((1L << cinfo->arith_dc_L[tbl]) >> 1)) - entropy->dc_context[ci] = 0; /* zero diff category */ - else if (m > (int) ((1L << cinfo->arith_dc_U[tbl]) >> 1)) - entropy->dc_context[ci] += 8; /* large diff category */ - /* Figure F.9: Encoding the magnitude bit pattern of v */ - st += 14; - while (m >>= 1) - arith_encode(cinfo, st, (m & v) ? 1 : 0); - } - } - - return TRUE; -} - - -/* - * MCU encoding for AC initial scan (either spectral selection, - * or first pass of successive approximation). - */ - -METHODDEF(boolean) -encode_mcu_AC_first (j_compress_ptr cinfo, JBLOCKARRAY MCU_data) -{ - arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; - const int * natural_order; - JBLOCKROW block; - unsigned char *st; - int tbl, k, ke; - int v, v2, m; - - /* Emit restart marker if needed */ - if (cinfo->restart_interval) { - if (entropy->restarts_to_go == 0) { - emit_restart(cinfo, entropy->next_restart_num); - entropy->restarts_to_go = cinfo->restart_interval; - entropy->next_restart_num++; - entropy->next_restart_num &= 7; - } - entropy->restarts_to_go--; - } - - natural_order = cinfo->natural_order; - - /* Encode the MCU data block */ - block = MCU_data[0]; - tbl = cinfo->cur_comp_info[0]->ac_tbl_no; - - /* Sections F.1.4.2 & F.1.4.4.2: Encoding of AC coefficients */ - - /* Establish EOB (end-of-block) index */ - ke = cinfo->Se; - do { - /* We must apply the point transform by Al. For AC coefficients this - * is an integer division with rounding towards 0. To do this portably - * in C, we shift after obtaining the absolute value. - */ - if ((v = (*block)[natural_order[ke]]) >= 0) { - if (v >>= cinfo->Al) break; - } else { - v = -v; - if (v >>= cinfo->Al) break; - } - } while (--ke); - - /* Figure F.5: Encode_AC_Coefficients */ - for (k = cinfo->Ss - 1; k < ke;) { - st = entropy->ac_stats[tbl] + 3 * k; - arith_encode(cinfo, st, 0); /* EOB decision */ - for (;;) { - if ((v = (*block)[natural_order[++k]]) >= 0) { - if (v >>= cinfo->Al) { - arith_encode(cinfo, st + 1, 1); - arith_encode(cinfo, entropy->fixed_bin, 0); - break; - } - } else { - v = -v; - if (v >>= cinfo->Al) { - arith_encode(cinfo, st + 1, 1); - arith_encode(cinfo, entropy->fixed_bin, 1); - break; - } - } - arith_encode(cinfo, st + 1, 0); - st += 3; - } - st += 2; - /* Figure F.8: Encoding the magnitude category of v */ - m = 0; - if (v -= 1) { - arith_encode(cinfo, st, 1); - m = 1; - v2 = v; - if (v2 >>= 1) { - arith_encode(cinfo, st, 1); - m <<= 1; - st = entropy->ac_stats[tbl] + - (k <= cinfo->arith_ac_K[tbl] ? 189 : 217); - while (v2 >>= 1) { - arith_encode(cinfo, st, 1); - m <<= 1; - st += 1; - } - } - } - arith_encode(cinfo, st, 0); - /* Figure F.9: Encoding the magnitude bit pattern of v */ - st += 14; - while (m >>= 1) - arith_encode(cinfo, st, (m & v) ? 1 : 0); - } - /* Encode EOB decision only if k < cinfo->Se */ - if (k < cinfo->Se) { - st = entropy->ac_stats[tbl] + 3 * k; - arith_encode(cinfo, st, 1); - } - - return TRUE; -} - - -/* - * MCU encoding for DC successive approximation refinement scan. - * Note: we assume such scans can be multi-component, - * although the spec is not very clear on the point. - */ - -METHODDEF(boolean) -encode_mcu_DC_refine (j_compress_ptr cinfo, JBLOCKARRAY MCU_data) -{ - arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; - unsigned char *st; - int Al, blkn; - - /* Emit restart marker if needed */ - if (cinfo->restart_interval) { - if (entropy->restarts_to_go == 0) { - emit_restart(cinfo, entropy->next_restart_num); - entropy->restarts_to_go = cinfo->restart_interval; - entropy->next_restart_num++; - entropy->next_restart_num &= 7; - } - entropy->restarts_to_go--; - } - - st = entropy->fixed_bin; /* use fixed probability estimation */ - Al = cinfo->Al; - - /* Encode the MCU data blocks */ - for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { - /* We simply emit the Al'th bit of the DC coefficient value. */ - arith_encode(cinfo, st, (MCU_data[blkn][0][0] >> Al) & 1); - } - - return TRUE; -} - - -/* - * MCU encoding for AC successive approximation refinement scan. - */ - -METHODDEF(boolean) -encode_mcu_AC_refine (j_compress_ptr cinfo, JBLOCKARRAY MCU_data) -{ - arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; - const int * natural_order; - JBLOCKROW block; - unsigned char *st; - int tbl, k, ke, kex; - int v; - - /* Emit restart marker if needed */ - if (cinfo->restart_interval) { - if (entropy->restarts_to_go == 0) { - emit_restart(cinfo, entropy->next_restart_num); - entropy->restarts_to_go = cinfo->restart_interval; - entropy->next_restart_num++; - entropy->next_restart_num &= 7; - } - entropy->restarts_to_go--; - } - - natural_order = cinfo->natural_order; - - /* Encode the MCU data block */ - block = MCU_data[0]; - tbl = cinfo->cur_comp_info[0]->ac_tbl_no; - - /* Section G.1.3.3: Encoding of AC coefficients */ - - /* Establish EOB (end-of-block) index */ - ke = cinfo->Se; - do { - /* We must apply the point transform by Al. For AC coefficients this - * is an integer division with rounding towards 0. To do this portably - * in C, we shift after obtaining the absolute value. - */ - if ((v = (*block)[natural_order[ke]]) >= 0) { - if (v >>= cinfo->Al) break; - } else { - v = -v; - if (v >>= cinfo->Al) break; - } - } while (--ke); - - /* Establish EOBx (previous stage end-of-block) index */ - for (kex = ke; kex > 0; kex--) - if ((v = (*block)[natural_order[kex]]) >= 0) { - if (v >>= cinfo->Ah) break; - } else { - v = -v; - if (v >>= cinfo->Ah) break; - } - - /* Figure G.10: Encode_AC_Coefficients_SA */ - for (k = cinfo->Ss - 1; k < ke;) { - st = entropy->ac_stats[tbl] + 3 * k; - if (k >= kex) - arith_encode(cinfo, st, 0); /* EOB decision */ - for (;;) { - if ((v = (*block)[natural_order[++k]]) >= 0) { - if (v >>= cinfo->Al) { - if (v >> 1) /* previously nonzero coef */ - arith_encode(cinfo, st + 2, (v & 1)); - else { /* newly nonzero coef */ - arith_encode(cinfo, st + 1, 1); - arith_encode(cinfo, entropy->fixed_bin, 0); - } - break; - } - } else { - v = -v; - if (v >>= cinfo->Al) { - if (v >> 1) /* previously nonzero coef */ - arith_encode(cinfo, st + 2, (v & 1)); - else { /* newly nonzero coef */ - arith_encode(cinfo, st + 1, 1); - arith_encode(cinfo, entropy->fixed_bin, 1); - } - break; - } - } - arith_encode(cinfo, st + 1, 0); - st += 3; - } - } - /* Encode EOB decision only if k < cinfo->Se */ - if (k < cinfo->Se) { - st = entropy->ac_stats[tbl] + 3 * k; - arith_encode(cinfo, st, 1); - } - - return TRUE; -} - - -/* - * Encode and output one MCU's worth of arithmetic-compressed coefficients. - */ - -METHODDEF(boolean) -encode_mcu (j_compress_ptr cinfo, JBLOCKARRAY MCU_data) -{ - arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; - const int * natural_order; - JBLOCKROW block; - unsigned char *st; - int tbl, k, ke; - int v, v2, m; - int blkn, ci; - jpeg_component_info * compptr; - - /* Emit restart marker if needed */ - if (cinfo->restart_interval) { - if (entropy->restarts_to_go == 0) { - emit_restart(cinfo, entropy->next_restart_num); - entropy->restarts_to_go = cinfo->restart_interval; - entropy->next_restart_num++; - entropy->next_restart_num &= 7; - } - entropy->restarts_to_go--; - } - - natural_order = cinfo->natural_order; - - /* Encode the MCU data blocks */ - for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { - block = MCU_data[blkn]; - ci = cinfo->MCU_membership[blkn]; - compptr = cinfo->cur_comp_info[ci]; - - /* Sections F.1.4.1 & F.1.4.4.1: Encoding of DC coefficients */ - - tbl = compptr->dc_tbl_no; - - /* Table F.4: Point to statistics bin S0 for DC coefficient coding */ - st = entropy->dc_stats[tbl] + entropy->dc_context[ci]; - - /* Figure F.4: Encode_DC_DIFF */ - if ((v = (*block)[0] - entropy->last_dc_val[ci]) == 0) { - arith_encode(cinfo, st, 0); - entropy->dc_context[ci] = 0; /* zero diff category */ - } else { - entropy->last_dc_val[ci] = (*block)[0]; - arith_encode(cinfo, st, 1); - /* Figure F.6: Encoding nonzero value v */ - /* Figure F.7: Encoding the sign of v */ - if (v > 0) { - arith_encode(cinfo, st + 1, 0); /* Table F.4: SS = S0 + 1 */ - st += 2; /* Table F.4: SP = S0 + 2 */ - entropy->dc_context[ci] = 4; /* small positive diff category */ - } else { - v = -v; - arith_encode(cinfo, st + 1, 1); /* Table F.4: SS = S0 + 1 */ - st += 3; /* Table F.4: SN = S0 + 3 */ - entropy->dc_context[ci] = 8; /* small negative diff category */ - } - /* Figure F.8: Encoding the magnitude category of v */ - m = 0; - if (v -= 1) { - arith_encode(cinfo, st, 1); - m = 1; - v2 = v; - st = entropy->dc_stats[tbl] + 20; /* Table F.4: X1 = 20 */ - while (v2 >>= 1) { - arith_encode(cinfo, st, 1); - m <<= 1; - st += 1; - } - } - arith_encode(cinfo, st, 0); - /* Section F.1.4.4.1.2: Establish dc_context conditioning category */ - if (m < (int) ((1L << cinfo->arith_dc_L[tbl]) >> 1)) - entropy->dc_context[ci] = 0; /* zero diff category */ - else if (m > (int) ((1L << cinfo->arith_dc_U[tbl]) >> 1)) - entropy->dc_context[ci] += 8; /* large diff category */ - /* Figure F.9: Encoding the magnitude bit pattern of v */ - st += 14; - while (m >>= 1) - arith_encode(cinfo, st, (m & v) ? 1 : 0); - } - - /* Sections F.1.4.2 & F.1.4.4.2: Encoding of AC coefficients */ - - if ((ke = cinfo->lim_Se) == 0) continue; - tbl = compptr->ac_tbl_no; - - /* Establish EOB (end-of-block) index */ - do { - if ((*block)[natural_order[ke]]) break; - } while (--ke); - - /* Figure F.5: Encode_AC_Coefficients */ - for (k = 0; k < ke;) { - st = entropy->ac_stats[tbl] + 3 * k; - arith_encode(cinfo, st, 0); /* EOB decision */ - while ((v = (*block)[natural_order[++k]]) == 0) { - arith_encode(cinfo, st + 1, 0); - st += 3; - } - arith_encode(cinfo, st + 1, 1); - /* Figure F.6: Encoding nonzero value v */ - /* Figure F.7: Encoding the sign of v */ - if (v > 0) { - arith_encode(cinfo, entropy->fixed_bin, 0); - } else { - v = -v; - arith_encode(cinfo, entropy->fixed_bin, 1); - } - st += 2; - /* Figure F.8: Encoding the magnitude category of v */ - m = 0; - if (v -= 1) { - arith_encode(cinfo, st, 1); - m = 1; - v2 = v; - if (v2 >>= 1) { - arith_encode(cinfo, st, 1); - m <<= 1; - st = entropy->ac_stats[tbl] + - (k <= cinfo->arith_ac_K[tbl] ? 189 : 217); - while (v2 >>= 1) { - arith_encode(cinfo, st, 1); - m <<= 1; - st += 1; - } - } - } - arith_encode(cinfo, st, 0); - /* Figure F.9: Encoding the magnitude bit pattern of v */ - st += 14; - while (m >>= 1) - arith_encode(cinfo, st, (m & v) ? 1 : 0); - } - /* Encode EOB decision only if k < cinfo->lim_Se */ - if (k < cinfo->lim_Se) { - st = entropy->ac_stats[tbl] + 3 * k; - arith_encode(cinfo, st, 1); - } - } - - return TRUE; -} - - -/* - * Initialize for an arithmetic-compressed scan. - */ - -METHODDEF(void) -start_pass (j_compress_ptr cinfo, boolean gather_statistics) -{ - arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; - int ci, tbl; - jpeg_component_info * compptr; - - if (gather_statistics) - /* Make sure to avoid that in the master control logic! - * We are fully adaptive here and need no extra - * statistics gathering pass! - */ - ERREXIT(cinfo, JERR_NOT_COMPILED); - - /* We assume jcmaster.c already validated the progressive scan parameters. */ - - /* Select execution routines */ - if (cinfo->progressive_mode) { - if (cinfo->Ah == 0) { - if (cinfo->Ss == 0) - entropy->pub.encode_mcu = encode_mcu_DC_first; - else - entropy->pub.encode_mcu = encode_mcu_AC_first; - } else { - if (cinfo->Ss == 0) - entropy->pub.encode_mcu = encode_mcu_DC_refine; - else - entropy->pub.encode_mcu = encode_mcu_AC_refine; - } - } else - entropy->pub.encode_mcu = encode_mcu; - - /* Allocate & initialize requested statistics areas */ - for (ci = 0; ci < cinfo->comps_in_scan; ci++) { - compptr = cinfo->cur_comp_info[ci]; - /* DC needs no table for refinement scan */ - if (cinfo->Ss == 0 && cinfo->Ah == 0) { - tbl = compptr->dc_tbl_no; - if (tbl < 0 || tbl >= NUM_ARITH_TBLS) - ERREXIT1(cinfo, JERR_NO_ARITH_TABLE, tbl); - if (entropy->dc_stats[tbl] == NULL) - entropy->dc_stats[tbl] = (unsigned char *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, DC_STAT_BINS); - MEMZERO(entropy->dc_stats[tbl], DC_STAT_BINS); - /* Initialize DC predictions to 0 */ - entropy->last_dc_val[ci] = 0; - entropy->dc_context[ci] = 0; - } - /* AC needs no table when not present */ - if (cinfo->Se) { - tbl = compptr->ac_tbl_no; - if (tbl < 0 || tbl >= NUM_ARITH_TBLS) - ERREXIT1(cinfo, JERR_NO_ARITH_TABLE, tbl); - if (entropy->ac_stats[tbl] == NULL) - entropy->ac_stats[tbl] = (unsigned char *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, AC_STAT_BINS); - MEMZERO(entropy->ac_stats[tbl], AC_STAT_BINS); -#ifdef CALCULATE_SPECTRAL_CONDITIONING - if (cinfo->progressive_mode) - /* Section G.1.3.2: Set appropriate arithmetic conditioning value Kx */ - cinfo->arith_ac_K[tbl] = cinfo->Ss + ((8 + cinfo->Se - cinfo->Ss) >> 4); -#endif - } - } - - /* Initialize arithmetic encoding variables */ - entropy->c = 0; - entropy->a = 0x10000L; - entropy->sc = 0; - entropy->zc = 0; - entropy->ct = 11; - entropy->buffer = -1; /* empty */ - - /* Initialize restart stuff */ - entropy->restarts_to_go = cinfo->restart_interval; - entropy->next_restart_num = 0; -} - - -/* - * Module initialization routine for arithmetic entropy encoding. - */ - -GLOBAL(void) -jinit_arith_encoder (j_compress_ptr cinfo) -{ - arith_entropy_ptr entropy; - int i; - - entropy = (arith_entropy_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(arith_entropy_encoder)); - cinfo->entropy = &entropy->pub; - entropy->pub.start_pass = start_pass; - entropy->pub.finish_pass = finish_pass; - - /* Mark tables unallocated */ - for (i = 0; i < NUM_ARITH_TBLS; i++) { - entropy->dc_stats[i] = NULL; - entropy->ac_stats[i] = NULL; - } - - /* Initialize index for fixed probability estimation */ - entropy->fixed_bin[0] = 113; -} diff --git a/cmake/externals/libjpeg/jccoefct.c b/cmake/externals/libjpeg/jccoefct.c deleted file mode 100644 index 494aa2298..000000000 --- a/cmake/externals/libjpeg/jccoefct.c +++ /dev/null @@ -1,456 +0,0 @@ -/* - * jccoefct.c - * - * Copyright (C) 1994-1997, Thomas G. Lane. - * Modified 2003-2022 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains the coefficient buffer controller for compression. - * This controller is the top level of the JPEG compressor proper. - * The coefficient buffer lies between forward-DCT and entropy encoding steps. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* We use a full-image coefficient buffer when doing Huffman optimization, - * and also for writing multiple-scan JPEG files. In all cases, the DCT - * step is run during the first pass, and subsequent passes need only read - * the buffered coefficients. - */ -#ifdef ENTROPY_OPT_SUPPORTED -#define FULL_COEF_BUFFER_SUPPORTED -#else -#ifdef C_MULTISCAN_FILES_SUPPORTED -#define FULL_COEF_BUFFER_SUPPORTED -#endif -#endif - - -/* Private buffer controller object */ - -typedef struct { - struct jpeg_c_coef_controller pub; /* public fields */ - - JDIMENSION iMCU_row_num; /* iMCU row # within image */ - JDIMENSION MCU_ctr; /* counts MCUs processed in current row */ - int MCU_vert_offset; /* counts MCU rows within iMCU row */ - int MCU_rows_per_iMCU_row; /* number of such rows needed */ - - /* For single-pass compression, it's sufficient to buffer just one MCU - * (although this may prove a bit slow in practice). - * We append a workspace of C_MAX_BLOCKS_IN_MCU coefficient blocks, - * and reuse it for each MCU constructed and sent. - * In multi-pass modes, this array points to the current MCU's blocks - * within the virtual arrays. - */ - JBLOCKROW MCU_buffer[C_MAX_BLOCKS_IN_MCU]; - - /* In multi-pass modes, we need a virtual block array for each component. */ - jvirt_barray_ptr whole_image[MAX_COMPONENTS]; - - /* Workspace for single-pass compression (omitted otherwise). */ - JBLOCK blk_buffer[C_MAX_BLOCKS_IN_MCU]; -} my_coef_controller; - -typedef my_coef_controller * my_coef_ptr; - - -/* Forward declarations */ -METHODDEF(boolean) compress_data - JPP((j_compress_ptr cinfo, JSAMPIMAGE input_buf)); -#ifdef FULL_COEF_BUFFER_SUPPORTED -METHODDEF(boolean) compress_first_pass - JPP((j_compress_ptr cinfo, JSAMPIMAGE input_buf)); -METHODDEF(boolean) compress_output - JPP((j_compress_ptr cinfo, JSAMPIMAGE input_buf)); -#endif - - -LOCAL(void) -start_iMCU_row (j_compress_ptr cinfo) -/* Reset within-iMCU-row counters for a new row */ -{ - my_coef_ptr coef = (my_coef_ptr) cinfo->coef; - - /* In an interleaved scan, an MCU row is the same as an iMCU row. - * In a noninterleaved scan, an iMCU row has v_samp_factor MCU rows. - * But at the bottom of the image, process only what's left. - */ - if (cinfo->comps_in_scan > 1) { - coef->MCU_rows_per_iMCU_row = 1; - } else { - if (coef->iMCU_row_num < (cinfo->total_iMCU_rows-1)) - coef->MCU_rows_per_iMCU_row = cinfo->cur_comp_info[0]->v_samp_factor; - else - coef->MCU_rows_per_iMCU_row = cinfo->cur_comp_info[0]->last_row_height; - } - - coef->MCU_ctr = 0; - coef->MCU_vert_offset = 0; -} - - -/* - * Initialize for a processing pass. - */ - -METHODDEF(void) -start_pass_coef (j_compress_ptr cinfo, J_BUF_MODE pass_mode) -{ - my_coef_ptr coef = (my_coef_ptr) cinfo->coef; - - coef->iMCU_row_num = 0; - start_iMCU_row(cinfo); - - switch (pass_mode) { - case JBUF_PASS_THRU: - if (coef->whole_image[0] != NULL) - ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); - coef->pub.compress_data = compress_data; - break; -#ifdef FULL_COEF_BUFFER_SUPPORTED - case JBUF_SAVE_AND_PASS: - if (coef->whole_image[0] == NULL) - ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); - coef->pub.compress_data = compress_first_pass; - break; - case JBUF_CRANK_DEST: - if (coef->whole_image[0] == NULL) - ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); - coef->pub.compress_data = compress_output; - break; -#endif - default: - ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); - } -} - - -/* - * Process some data in the single-pass case. - * We process the equivalent of one fully interleaved MCU row ("iMCU" row) - * per call, ie, v_samp_factor block rows for each component in the image. - * Returns TRUE if the iMCU row is completed, FALSE if suspended. - * - * NB: input_buf contains a plane for each component in image, - * which we index according to the component's SOF position. - */ - -METHODDEF(boolean) -compress_data (j_compress_ptr cinfo, JSAMPIMAGE input_buf) -{ - my_coef_ptr coef = (my_coef_ptr) cinfo->coef; - JDIMENSION MCU_col_num; /* index of current MCU within row */ - JDIMENSION last_MCU_col = cinfo->MCUs_per_row - 1; - JDIMENSION last_iMCU_row = cinfo->total_iMCU_rows - 1; - int ci, xindex, yindex, yoffset, blockcnt; - JBLOCKROW blkp; - JSAMPARRAY input_ptr; - JDIMENSION xpos; - jpeg_component_info *compptr; - forward_DCT_ptr forward_DCT; - - /* Loop to write as much as one whole iMCU row */ - for (yoffset = coef->MCU_vert_offset; yoffset < coef->MCU_rows_per_iMCU_row; - yoffset++) { - for (MCU_col_num = coef->MCU_ctr; MCU_col_num <= last_MCU_col; - MCU_col_num++) { - /* Determine where data comes from in input_buf and do the DCT thing. - * Each call on forward_DCT processes a horizontal row of DCT blocks as - * wide as an MCU. Dummy blocks at the right or bottom edge are filled in - * specially. The data in them does not matter for image reconstruction, - * so we fill them with values that will encode to the smallest amount of - * data, viz: all zeroes in the AC entries, DC entries equal to previous - * block's DC value. (Thanks to Thomas Kinsman for this idea.) - */ - blkp = coef->blk_buffer; /* pointer to current DCT block within MCU */ - for (ci = 0; ci < cinfo->comps_in_scan; ci++) { - compptr = cinfo->cur_comp_info[ci]; - forward_DCT = cinfo->fdct->forward_DCT[compptr->component_index]; - input_ptr = input_buf[compptr->component_index] + - yoffset * compptr->DCT_v_scaled_size; - /* ypos == (yoffset + yindex) * compptr->DCT_v_scaled_size */ - blockcnt = (MCU_col_num < last_MCU_col) ? compptr->MCU_width - : compptr->last_col_width; - xpos = MCU_col_num * compptr->MCU_sample_width; - for (yindex = 0; yindex < compptr->MCU_height; yindex++) { - if (coef->iMCU_row_num < last_iMCU_row || - yoffset + yindex < compptr->last_row_height) { - (*forward_DCT) (cinfo, compptr, input_ptr, blkp, - xpos, (JDIMENSION) blockcnt); - input_ptr += compptr->DCT_v_scaled_size; - blkp += blockcnt; - /* Dummy blocks at right edge */ - if ((xindex = compptr->MCU_width - blockcnt) == 0) - continue; - } else { - /* At bottom of image, need a whole row of dummy blocks */ - xindex = compptr->MCU_width; - } - /* Fill in any dummy blocks needed in this row */ - MEMZERO(blkp, xindex * SIZEOF(JBLOCK)); - do { - blkp[0][0] = blkp[-1][0]; - blkp++; - } while (--xindex); - } - } - /* Try to write the MCU. In event of a suspension failure, we will - * re-DCT the MCU on restart (a bit inefficient, could be fixed...) - */ - if (! (*cinfo->entropy->encode_mcu) (cinfo, coef->MCU_buffer)) { - /* Suspension forced; update state counters and exit */ - coef->MCU_vert_offset = yoffset; - coef->MCU_ctr = MCU_col_num; - return FALSE; - } - } - /* Completed an MCU row, but perhaps not an iMCU row */ - coef->MCU_ctr = 0; - } - /* Completed the iMCU row, advance counters for next one */ - coef->iMCU_row_num++; - start_iMCU_row(cinfo); - return TRUE; -} - - -#ifdef FULL_COEF_BUFFER_SUPPORTED - -/* - * Process some data in the first pass of a multi-pass case. - * We process the equivalent of one fully interleaved MCU row ("iMCU" row) - * per call, ie, v_samp_factor block rows for each component in the image. - * This amount of data is read from the source buffer, DCT'd and quantized, - * and saved into the virtual arrays. We also generate suitable dummy blocks - * as needed at the right and lower edges. (The dummy blocks are constructed - * in the virtual arrays, which have been padded appropriately.) This makes - * it possible for subsequent passes not to worry about real vs. dummy blocks. - * - * We must also emit the data to the entropy encoder. This is conveniently - * done by calling compress_output() after we've loaded the current strip - * of the virtual arrays. - * - * NB: input_buf contains a plane for each component in image. All - * components are DCT'd and loaded into the virtual arrays in this pass. - * However, it may be that only a subset of the components are emitted to - * the entropy encoder during this first pass; be careful about looking - * at the scan-dependent variables (MCU dimensions, etc). - */ - -METHODDEF(boolean) -compress_first_pass (j_compress_ptr cinfo, JSAMPIMAGE input_buf) -{ - my_coef_ptr coef = (my_coef_ptr) cinfo->coef; - JDIMENSION last_iMCU_row = cinfo->total_iMCU_rows - 1; - JDIMENSION blocks_across, MCUs_across, MCUindex; - int bi, ci, h_samp_factor, block_row, block_rows, ndummy; - JCOEF lastDC; - jpeg_component_info *compptr; - JBLOCKARRAY buffer; - JBLOCKROW thisblockrow, lastblockrow; - JSAMPARRAY input_ptr; - forward_DCT_ptr forward_DCT; - - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - /* Align the virtual buffer for this component. */ - buffer = (*cinfo->mem->access_virt_barray) - ((j_common_ptr) cinfo, coef->whole_image[ci], - coef->iMCU_row_num * compptr->v_samp_factor, - (JDIMENSION) compptr->v_samp_factor, TRUE); - /* Count non-dummy DCT block rows in this iMCU row. */ - if (coef->iMCU_row_num < last_iMCU_row) - block_rows = compptr->v_samp_factor; - else { - /* NB: can't use last_row_height here, since may not be set! */ - block_rows = (int) (compptr->height_in_blocks % compptr->v_samp_factor); - if (block_rows == 0) block_rows = compptr->v_samp_factor; - } - blocks_across = compptr->width_in_blocks; - h_samp_factor = compptr->h_samp_factor; - /* Count number of dummy blocks to be added at the right margin. */ - ndummy = (int) (blocks_across % h_samp_factor); - if (ndummy > 0) - ndummy = h_samp_factor - ndummy; - forward_DCT = cinfo->fdct->forward_DCT[ci]; - input_ptr = input_buf[ci]; - /* Perform DCT for all non-dummy blocks in this iMCU row. Each call - * on forward_DCT processes a complete horizontal row of DCT blocks. - */ - for (block_row = 0; block_row < block_rows; block_row++) { - thisblockrow = buffer[block_row]; - (*forward_DCT) (cinfo, compptr, input_ptr, thisblockrow, - (JDIMENSION) 0, blocks_across); - input_ptr += compptr->DCT_v_scaled_size; - if (ndummy > 0) { - /* Create dummy blocks at the right edge of the image. */ - thisblockrow += blocks_across; /* => first dummy block */ - FMEMZERO((void FAR *) thisblockrow, ndummy * SIZEOF(JBLOCK)); - lastDC = thisblockrow[-1][0]; - for (bi = 0; bi < ndummy; bi++) { - thisblockrow[bi][0] = lastDC; - } - } - } - /* If at end of image, create dummy block rows as needed. - * The tricky part here is that within each MCU, we want the DC values - * of the dummy blocks to match the last real block's DC value. - * This squeezes a few more bytes out of the resulting file... - */ - if (block_row < compptr->v_samp_factor) { - blocks_across += ndummy; /* include lower right corner */ - MCUs_across = blocks_across / h_samp_factor; - do { - thisblockrow = buffer[block_row]; - lastblockrow = buffer[block_row-1]; - FMEMZERO((void FAR *) thisblockrow, - (size_t) blocks_across * SIZEOF(JBLOCK)); - for (MCUindex = 0; MCUindex < MCUs_across; MCUindex++) { - lastDC = lastblockrow[h_samp_factor-1][0]; - for (bi = 0; bi < h_samp_factor; bi++) { - thisblockrow[bi][0] = lastDC; - } - thisblockrow += h_samp_factor; /* advance to next MCU in row */ - lastblockrow += h_samp_factor; - } - } while (++block_row < compptr->v_samp_factor); - } - } - /* NB: compress_output will increment iMCU_row_num if successful. - * A suspension return will result in redoing all the work above next time. - */ - - /* Emit data to the entropy encoder, sharing code with subsequent passes */ - return compress_output(cinfo, input_buf); -} - - -/* - * Process some data in subsequent passes of a multi-pass case. - * We process the equivalent of one fully interleaved MCU row ("iMCU" row) - * per call, ie, v_samp_factor block rows for each component in the scan. - * The data is obtained from the virtual arrays and fed to the entropy coder. - * Returns TRUE if the iMCU row is completed, FALSE if suspended. - * - * NB: input_buf is ignored; it is likely to be a NULL pointer. - */ - -METHODDEF(boolean) -compress_output (j_compress_ptr cinfo, JSAMPIMAGE input_buf) -{ - my_coef_ptr coef = (my_coef_ptr) cinfo->coef; - JDIMENSION MCU_col_num; /* index of current MCU within row */ - int ci, xindex, yindex, yoffset; - JDIMENSION start_col; - JBLOCKARRAY blkp; - JBLOCKARRAY buffer[MAX_COMPS_IN_SCAN]; - JBLOCKROW buffer_ptr; - jpeg_component_info *compptr; - - /* Align the virtual buffers for the components used in this scan. - * NB: during first pass, this is safe only because the buffers will - * already be aligned properly, so jmemmgr.c won't need to do any I/O. - */ - for (ci = 0; ci < cinfo->comps_in_scan; ci++) { - compptr = cinfo->cur_comp_info[ci]; - buffer[ci] = (*cinfo->mem->access_virt_barray) - ((j_common_ptr) cinfo, coef->whole_image[compptr->component_index], - coef->iMCU_row_num * compptr->v_samp_factor, - (JDIMENSION) compptr->v_samp_factor, FALSE); - } - - /* Loop to process one whole iMCU row */ - for (yoffset = coef->MCU_vert_offset; yoffset < coef->MCU_rows_per_iMCU_row; - yoffset++) { - for (MCU_col_num = coef->MCU_ctr; MCU_col_num < cinfo->MCUs_per_row; - MCU_col_num++) { - /* Construct list of pointers to DCT blocks belonging to this MCU */ - blkp = coef->MCU_buffer; /* pointer to current DCT block within MCU */ - for (ci = 0; ci < cinfo->comps_in_scan; ci++) { - compptr = cinfo->cur_comp_info[ci]; - start_col = MCU_col_num * compptr->MCU_width; - for (yindex = 0; yindex < compptr->MCU_height; yindex++) { - buffer_ptr = buffer[ci][yoffset + yindex] + start_col; - xindex = compptr->MCU_width; - do { - *blkp++ = buffer_ptr++; - } while (--xindex); - } - } - /* Try to write the MCU. */ - if (! (*cinfo->entropy->encode_mcu) (cinfo, coef->MCU_buffer)) { - /* Suspension forced; update state counters and exit */ - coef->MCU_vert_offset = yoffset; - coef->MCU_ctr = MCU_col_num; - return FALSE; - } - } - /* Completed an MCU row, but perhaps not an iMCU row */ - coef->MCU_ctr = 0; - } - /* Completed the iMCU row, advance counters for next one */ - coef->iMCU_row_num++; - start_iMCU_row(cinfo); - return TRUE; -} - -#endif /* FULL_COEF_BUFFER_SUPPORTED */ - - -/* - * Initialize coefficient buffer controller. - */ - -GLOBAL(void) -jinit_c_coef_controller (j_compress_ptr cinfo, boolean need_full_buffer) -{ - my_coef_ptr coef; - - if (need_full_buffer) { -#ifdef FULL_COEF_BUFFER_SUPPORTED - /* Allocate a full-image virtual array for each component, */ - /* padded to a multiple of samp_factor DCT blocks in each direction. */ - int ci; - jpeg_component_info *compptr; - - coef = (my_coef_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, - SIZEOF(my_coef_controller) - SIZEOF(coef->blk_buffer)); - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - coef->whole_image[ci] = (*cinfo->mem->request_virt_barray) - ((j_common_ptr) cinfo, JPOOL_IMAGE, FALSE, - (JDIMENSION) jround_up((long) compptr->width_in_blocks, - (long) compptr->h_samp_factor), - (JDIMENSION) jround_up((long) compptr->height_in_blocks, - (long) compptr->v_samp_factor), - (JDIMENSION) compptr->v_samp_factor); - } -#else - ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); -#endif - } else { - /* We only need a single-MCU buffer. */ - JBLOCKARRAY blkp; - JBLOCKROW buffer_ptr; - int bi; - - coef = (my_coef_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_coef_controller)); - blkp = coef->MCU_buffer; - buffer_ptr = coef->blk_buffer; - bi = C_MAX_BLOCKS_IN_MCU; - do { - *blkp++ = buffer_ptr++; - } while (--bi); - coef->whole_image[0] = NULL; /* flag for no virtual arrays */ - } - - coef->pub.start_pass = start_pass_coef; - cinfo->coef = &coef->pub; -} diff --git a/cmake/externals/libjpeg/jccolor.c b/cmake/externals/libjpeg/jccolor.c deleted file mode 100644 index c028dd9db..000000000 --- a/cmake/externals/libjpeg/jccolor.c +++ /dev/null @@ -1,598 +0,0 @@ -/* - * jccolor.c - * - * Copyright (C) 1991-1996, Thomas G. Lane. - * Modified 2011-2023 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains input colorspace conversion routines. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* Private subobject */ - -typedef struct { - struct jpeg_color_converter pub; /* public fields */ - - /* Private state for RGB->YCC conversion */ - INT32 * rgb_ycc_tab; /* => table for RGB to YCbCr conversion */ -} my_color_converter; - -typedef my_color_converter * my_cconvert_ptr; - - -/**************** RGB -> YCbCr conversion: most common case **************/ - -/* - * YCbCr is defined per Recommendation ITU-R BT.601-7 (03/2011), - * previously known as Recommendation CCIR 601-1, except that Cb and Cr - * are normalized to the range 0..MAXJSAMPLE rather than -0.5 .. 0.5. - * sRGB (standard RGB color space) is defined per IEC 61966-2-1:1999. - * sYCC (standard luma-chroma-chroma color space with extended gamut) - * is defined per IEC 61966-2-1:1999 Amendment A1:2003 Annex F. - * bg-sRGB and bg-sYCC (big gamut standard color spaces) - * are defined per IEC 61966-2-1:1999 Amendment A1:2003 Annex G. - * Note that the derived conversion coefficients given in some of these - * documents are imprecise. The general conversion equations are - * Y = Kr * R + (1 - Kr - Kb) * G + Kb * B - * Cb = (B - Y) / (1 - Kb) / K - * Cr = (R - Y) / (1 - Kr) / K - * With Kr = 0.299 and Kb = 0.114 (derived according to SMPTE RP 177-1993 - * from the 1953 FCC NTSC primaries and CIE Illuminant C), K = 2 for sYCC, - * the conversion equations to be implemented are therefore - * Y = 0.299 * R + 0.587 * G + 0.114 * B - * Cb = -0.168735892 * R - 0.331264108 * G + 0.5 * B + CENTERJSAMPLE - * Cr = 0.5 * R - 0.418687589 * G - 0.081312411 * B + CENTERJSAMPLE - * Note: older versions of the IJG code used a zero offset of MAXJSAMPLE/2, - * rather than CENTERJSAMPLE, for Cb and Cr. This gave equal positive and - * negative swings for Cb/Cr, but meant that grayscale values (Cb=Cr=0) - * were not represented exactly. Now we sacrifice exact representation of - * maximum red and maximum blue in order to get exact grayscales. - * - * To avoid floating-point arithmetic, we represent the fractional constants - * as integers scaled up by 2^16 (about 4 digits precision); we have to divide - * the products by 2^16, with appropriate rounding, to get the correct answer. - * - * For even more speed, we avoid doing any multiplications in the inner loop - * by precalculating the constants times R,G,B for all possible values. - * For 8-bit JSAMPLEs this is very reasonable (only 256 entries per table); - * for 9-bit to 12-bit samples it is still acceptable. It's not very - * reasonable for 16-bit samples, but if you want lossless storage - * you shouldn't be changing colorspace anyway. - * The CENTERJSAMPLE offsets and the rounding fudge-factor of 0.5 are included - * in the tables to save adding them separately in the inner loop. - */ - -#define SCALEBITS 16 /* speediest right-shift on some machines */ -#define CBCR_OFFSET ((INT32) CENTERJSAMPLE << SCALEBITS) -#define ONE_HALF ((INT32) 1 << (SCALEBITS-1)) -#define FIX(x) ((INT32) ((x) * (1L< Y section */ -#define G_Y_OFF (1*(MAXJSAMPLE+1)) /* offset to G => Y section */ -#define B_Y_OFF (2*(MAXJSAMPLE+1)) /* etc. */ -#define R_CB_OFF (3*(MAXJSAMPLE+1)) -#define G_CB_OFF (4*(MAXJSAMPLE+1)) -#define B_CB_OFF (5*(MAXJSAMPLE+1)) -#define R_CR_OFF B_CB_OFF /* B=>Cb, R=>Cr are the same */ -#define G_CR_OFF (6*(MAXJSAMPLE+1)) -#define B_CR_OFF (7*(MAXJSAMPLE+1)) -#define TABLE_SIZE (8*(MAXJSAMPLE+1)) - - -/* - * Initialize for RGB->YCC colorspace conversion. - */ - -METHODDEF(void) -rgb_ycc_start (j_compress_ptr cinfo) -{ - my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; - INT32 * rgb_ycc_tab; - INT32 i; - - /* Allocate and fill in the conversion tables. */ - cconvert->rgb_ycc_tab = rgb_ycc_tab = (INT32 *) - (*cinfo->mem->alloc_small) ((j_common_ptr) cinfo, JPOOL_IMAGE, - TABLE_SIZE * SIZEOF(INT32)); - - for (i = 0; i <= MAXJSAMPLE; i++) { - rgb_ycc_tab[i+R_Y_OFF] = FIX(0.299) * i; - rgb_ycc_tab[i+G_Y_OFF] = FIX(0.587) * i; - rgb_ycc_tab[i+B_Y_OFF] = FIX(0.114) * i + ONE_HALF; - rgb_ycc_tab[i+R_CB_OFF] = (- FIX(0.168735892)) * i; - rgb_ycc_tab[i+G_CB_OFF] = (- FIX(0.331264108)) * i; - /* We use a rounding fudge-factor of 0.5-epsilon for Cb and Cr. - * This ensures that the maximum output will round to MAXJSAMPLE - * not MAXJSAMPLE+1, and thus that we don't have to range-limit. - */ - rgb_ycc_tab[i+B_CB_OFF] = (i << (SCALEBITS-1)) + CBCR_OFFSET + ONE_HALF-1; -/* B=>Cb and R=>Cr tables are the same - rgb_ycc_tab[i+R_CR_OFF] = (i << (SCALEBITS-1)) + CBCR_OFFSET + ONE_HALF-1; -*/ - rgb_ycc_tab[i+G_CR_OFF] = (- FIX(0.418687589)) * i; - rgb_ycc_tab[i+B_CR_OFF] = (- FIX(0.081312411)) * i; - } -} - - -/* - * Convert some rows of samples to the JPEG colorspace. - * - * Note that we change from the application's interleaved-pixel format - * to our internal noninterleaved, one-plane-per-component format. The - * input buffer is therefore three times as wide as the output buffer. - * - * A starting row offset is provided only for the output buffer. The - * caller can easily adjust the passed input_buf value to accommodate - * any row offset required on that side. - */ - -METHODDEF(void) -rgb_ycc_convert (j_compress_ptr cinfo, - JSAMPARRAY input_buf, JSAMPIMAGE output_buf, - JDIMENSION output_row, int num_rows) -{ - my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; - register int r, g, b; - register INT32 * ctab = cconvert->rgb_ycc_tab; - register JSAMPROW inptr; - register JSAMPROW outptr0, outptr1, outptr2; - register JDIMENSION col; - JDIMENSION num_cols = cinfo->image_width; - - while (--num_rows >= 0) { - inptr = *input_buf++; - outptr0 = output_buf[0][output_row]; - outptr1 = output_buf[1][output_row]; - outptr2 = output_buf[2][output_row]; - output_row++; - for (col = 0; col < num_cols; col++) { - r = GETJSAMPLE(inptr[RGB_RED]); - g = GETJSAMPLE(inptr[RGB_GREEN]); - b = GETJSAMPLE(inptr[RGB_BLUE]); - inptr += RGB_PIXELSIZE; - /* If the inputs are 0..MAXJSAMPLE, the outputs of these equations - * must be too; we do not need an explicit range-limiting operation. - * Hence the value being shifted is never negative, and we don't - * need the general RIGHT_SHIFT macro. - */ - /* Y */ - outptr0[col] = (JSAMPLE) - ((ctab[r+R_Y_OFF] + ctab[g+G_Y_OFF] + ctab[b+B_Y_OFF]) - >> SCALEBITS); - /* Cb */ - outptr1[col] = (JSAMPLE) - ((ctab[r+R_CB_OFF] + ctab[g+G_CB_OFF] + ctab[b+B_CB_OFF]) - >> SCALEBITS); - /* Cr */ - outptr2[col] = (JSAMPLE) - ((ctab[r+R_CR_OFF] + ctab[g+G_CR_OFF] + ctab[b+B_CR_OFF]) - >> SCALEBITS); - } - } -} - - -/**************** Cases other than RGB -> YCbCr **************/ - - -/* - * Convert some rows of samples to the JPEG colorspace. - * This version handles RGB->grayscale conversion, - * which is the same as the RGB->Y portion of RGB->YCbCr. - * We assume rgb_ycc_start has been called (we only use the Y tables). - */ - -METHODDEF(void) -rgb_gray_convert (j_compress_ptr cinfo, - JSAMPARRAY input_buf, JSAMPIMAGE output_buf, - JDIMENSION output_row, int num_rows) -{ - my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; - register INT32 y; - register INT32 * ctab = cconvert->rgb_ycc_tab; - register JSAMPROW inptr; - register JSAMPROW outptr; - register JDIMENSION col; - JDIMENSION num_cols = cinfo->image_width; - - while (--num_rows >= 0) { - inptr = *input_buf++; - outptr = output_buf[0][output_row++]; - for (col = 0; col < num_cols; col++) { - y = ctab[R_Y_OFF + GETJSAMPLE(inptr[RGB_RED])]; - y += ctab[G_Y_OFF + GETJSAMPLE(inptr[RGB_GREEN])]; - y += ctab[B_Y_OFF + GETJSAMPLE(inptr[RGB_BLUE])]; - inptr += RGB_PIXELSIZE; - outptr[col] = (JSAMPLE) (y >> SCALEBITS); - } - } -} - - -/* - * Convert some rows of samples to the JPEG colorspace. - * This version handles Adobe-style CMYK->YCCK conversion, - * where we convert R=1-C, G=1-M, and B=1-Y to YCbCr using the - * same conversion as above, while passing K (black) unchanged. - * We assume rgb_ycc_start has been called. - */ - -METHODDEF(void) -cmyk_ycck_convert (j_compress_ptr cinfo, - JSAMPARRAY input_buf, JSAMPIMAGE output_buf, - JDIMENSION output_row, int num_rows) -{ - my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; - register int r, g, b; - register INT32 * ctab = cconvert->rgb_ycc_tab; - register JSAMPROW inptr; - register JSAMPROW outptr0, outptr1, outptr2, outptr3; - register JDIMENSION col; - JDIMENSION num_cols = cinfo->image_width; - - while (--num_rows >= 0) { - inptr = *input_buf++; - outptr0 = output_buf[0][output_row]; - outptr1 = output_buf[1][output_row]; - outptr2 = output_buf[2][output_row]; - outptr3 = output_buf[3][output_row]; - output_row++; - for (col = 0; col < num_cols; col++) { - r = MAXJSAMPLE - GETJSAMPLE(inptr[0]); - g = MAXJSAMPLE - GETJSAMPLE(inptr[1]); - b = MAXJSAMPLE - GETJSAMPLE(inptr[2]); - /* K passes through as-is */ - outptr3[col] = inptr[3]; /* don't need GETJSAMPLE here */ - inptr += 4; - /* If the inputs are 0..MAXJSAMPLE, the outputs of these equations - * must be too; we do not need an explicit range-limiting operation. - * Hence the value being shifted is never negative, and we don't - * need the general RIGHT_SHIFT macro. - */ - /* Y */ - outptr0[col] = (JSAMPLE) - ((ctab[r+R_Y_OFF] + ctab[g+G_Y_OFF] + ctab[b+B_Y_OFF]) - >> SCALEBITS); - /* Cb */ - outptr1[col] = (JSAMPLE) - ((ctab[r+R_CB_OFF] + ctab[g+G_CB_OFF] + ctab[b+B_CB_OFF]) - >> SCALEBITS); - /* Cr */ - outptr2[col] = (JSAMPLE) - ((ctab[r+R_CR_OFF] + ctab[g+G_CR_OFF] + ctab[b+B_CR_OFF]) - >> SCALEBITS); - } - } -} - - -/* - * Convert some rows of samples to the JPEG colorspace. - * [R,G,B] to [R-G,G,B-G] conversion with modulo calculation - * (forward reversible color transform). - * This can be seen as an adaption of the general RGB->YCbCr - * conversion equation with Kr = Kb = 0, while replacing the - * normalization by modulo calculation. - */ - -METHODDEF(void) -rgb_rgb1_convert (j_compress_ptr cinfo, - JSAMPARRAY input_buf, JSAMPIMAGE output_buf, - JDIMENSION output_row, int num_rows) -{ - register int r, g, b; - register JSAMPROW inptr; - register JSAMPROW outptr0, outptr1, outptr2; - register JDIMENSION col; - JDIMENSION num_cols = cinfo->image_width; - - while (--num_rows >= 0) { - inptr = *input_buf++; - outptr0 = output_buf[0][output_row]; - outptr1 = output_buf[1][output_row]; - outptr2 = output_buf[2][output_row]; - output_row++; - for (col = 0; col < num_cols; col++) { - r = GETJSAMPLE(inptr[RGB_RED]); - g = GETJSAMPLE(inptr[RGB_GREEN]); - b = GETJSAMPLE(inptr[RGB_BLUE]); - inptr += RGB_PIXELSIZE; - /* Assume that MAXJSAMPLE+1 is a power of 2, so that the MOD - * (modulo) operator is equivalent to the bitmask operator AND. - */ - outptr0[col] = (JSAMPLE) ((r - g + CENTERJSAMPLE) & MAXJSAMPLE); - outptr1[col] = (JSAMPLE) g; - outptr2[col] = (JSAMPLE) ((b - g + CENTERJSAMPLE) & MAXJSAMPLE); - } - } -} - - -/* - * Convert some rows of samples to the JPEG colorspace. - * This version handles grayscale output with no conversion. - * The source can be either plain grayscale or YCC (since Y == gray). - */ - -METHODDEF(void) -grayscale_convert (j_compress_ptr cinfo, - JSAMPARRAY input_buf, JSAMPIMAGE output_buf, - JDIMENSION output_row, int num_rows) -{ - register JSAMPROW inptr; - register JSAMPROW outptr; - register JDIMENSION count; - register int instride = cinfo->input_components; - JDIMENSION num_cols = cinfo->image_width; - - while (--num_rows >= 0) { - inptr = *input_buf++; - outptr = output_buf[0][output_row++]; - for (count = num_cols; count > 0; count--) { - *outptr++ = *inptr; /* don't need GETJSAMPLE() here */ - inptr += instride; - } - } -} - - -/* - * Convert some rows of samples to the JPEG colorspace. - * No colorspace conversion, but change from interleaved - * to separate-planes representation. - */ - -METHODDEF(void) -rgb_convert (j_compress_ptr cinfo, - JSAMPARRAY input_buf, JSAMPIMAGE output_buf, - JDIMENSION output_row, int num_rows) -{ - register JSAMPROW inptr; - register JSAMPROW outptr0, outptr1, outptr2; - register JDIMENSION col; - JDIMENSION num_cols = cinfo->image_width; - - while (--num_rows >= 0) { - inptr = *input_buf++; - outptr0 = output_buf[0][output_row]; - outptr1 = output_buf[1][output_row]; - outptr2 = output_buf[2][output_row]; - output_row++; - for (col = 0; col < num_cols; col++) { - /* We can dispense with GETJSAMPLE() here */ - outptr0[col] = inptr[RGB_RED]; - outptr1[col] = inptr[RGB_GREEN]; - outptr2[col] = inptr[RGB_BLUE]; - inptr += RGB_PIXELSIZE; - } - } -} - - -/* - * Convert some rows of samples to the JPEG colorspace. - * This version handles multi-component colorspaces without conversion. - * We assume input_components == num_components. - */ - -METHODDEF(void) -null_convert (j_compress_ptr cinfo, - JSAMPARRAY input_buf, JSAMPIMAGE output_buf, - JDIMENSION output_row, int num_rows) -{ - register JSAMPROW inptr; - register JSAMPROW outptr; - register JDIMENSION count; - register int num_comps = cinfo->num_components; - JDIMENSION num_cols = cinfo->image_width; - int ci; - - while (--num_rows >= 0) { - /* It seems fastest to make a separate pass for each component. */ - for (ci = 0; ci < num_comps; ci++) { - inptr = input_buf[0] + ci; - outptr = output_buf[ci][output_row]; - for (count = num_cols; count > 0; count--) { - *outptr++ = *inptr; /* don't need GETJSAMPLE() here */ - inptr += num_comps; - } - } - input_buf++; - output_row++; - } -} - - -/* - * Empty method for start_pass. - */ - -METHODDEF(void) -null_method (j_compress_ptr cinfo) -{ - /* no work needed */ -} - - -/* - * Module initialization routine for input colorspace conversion. - */ - -GLOBAL(void) -jinit_color_converter (j_compress_ptr cinfo) -{ - my_cconvert_ptr cconvert; - - cconvert = (my_cconvert_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_color_converter)); - cinfo->cconvert = &cconvert->pub; - /* set start_pass to null method until we find out differently */ - cconvert->pub.start_pass = null_method; - - /* Make sure input_components agrees with in_color_space */ - switch (cinfo->in_color_space) { - case JCS_GRAYSCALE: - if (cinfo->input_components != 1) - ERREXIT(cinfo, JERR_BAD_IN_COLORSPACE); - break; - - case JCS_RGB: - case JCS_BG_RGB: -#if RGB_PIXELSIZE != 3 - if (cinfo->input_components != RGB_PIXELSIZE) - ERREXIT(cinfo, JERR_BAD_IN_COLORSPACE); - break; -#endif /* else share code with YCbCr */ - - case JCS_YCbCr: - case JCS_BG_YCC: - if (cinfo->input_components != 3) - ERREXIT(cinfo, JERR_BAD_IN_COLORSPACE); - break; - - case JCS_CMYK: - case JCS_YCCK: - if (cinfo->input_components != 4) - ERREXIT(cinfo, JERR_BAD_IN_COLORSPACE); - break; - - default: /* JCS_UNKNOWN can be anything */ - if (cinfo->input_components < 1) - ERREXIT(cinfo, JERR_BAD_IN_COLORSPACE); - } - - /* Support color transform only for RGB colorspaces */ - if (cinfo->color_transform && - cinfo->jpeg_color_space != JCS_RGB && - cinfo->jpeg_color_space != JCS_BG_RGB) - ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); - - /* Check num_components, set conversion method based on requested space */ - switch (cinfo->jpeg_color_space) { - case JCS_GRAYSCALE: - if (cinfo->num_components != 1) - ERREXIT(cinfo, JERR_BAD_J_COLORSPACE); - switch (cinfo->in_color_space) { - case JCS_GRAYSCALE: - case JCS_YCbCr: - case JCS_BG_YCC: - cconvert->pub.color_convert = grayscale_convert; - break; - case JCS_RGB: - cconvert->pub.start_pass = rgb_ycc_start; - cconvert->pub.color_convert = rgb_gray_convert; - break; - default: - ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); - } - break; - - case JCS_RGB: - case JCS_BG_RGB: - if (cinfo->num_components != 3) - ERREXIT(cinfo, JERR_BAD_J_COLORSPACE); - if (cinfo->in_color_space != cinfo->jpeg_color_space) - ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); - switch (cinfo->color_transform) { - case JCT_NONE: - cconvert->pub.color_convert = rgb_convert; - break; - case JCT_SUBTRACT_GREEN: - cconvert->pub.color_convert = rgb_rgb1_convert; - break; - default: - ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); - } - break; - - case JCS_YCbCr: - if (cinfo->num_components != 3) - ERREXIT(cinfo, JERR_BAD_J_COLORSPACE); - switch (cinfo->in_color_space) { - case JCS_RGB: - cconvert->pub.start_pass = rgb_ycc_start; - cconvert->pub.color_convert = rgb_ycc_convert; - break; - case JCS_YCbCr: - cconvert->pub.color_convert = null_convert; - break; - default: - ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); - } - break; - - case JCS_BG_YCC: - if (cinfo->num_components != 3) - ERREXIT(cinfo, JERR_BAD_J_COLORSPACE); - switch (cinfo->in_color_space) { - case JCS_RGB: - /* For conversion from normal RGB input to BG_YCC representation, - * the Cb/Cr values are first computed as usual, and then - * quantized further after DCT processing by a factor of - * 2 in reference to the nominal quantization factor. - */ - /* need quantization scale by factor of 2 after DCT */ - cinfo->comp_info[1].component_needed = TRUE; - cinfo->comp_info[2].component_needed = TRUE; - /* compute normal YCC first */ - cconvert->pub.start_pass = rgb_ycc_start; - cconvert->pub.color_convert = rgb_ycc_convert; - break; - case JCS_YCbCr: - /* need quantization scale by factor of 2 after DCT */ - cinfo->comp_info[1].component_needed = TRUE; - cinfo->comp_info[2].component_needed = TRUE; - /*FALLTHROUGH*/ - case JCS_BG_YCC: - /* Pass through for BG_YCC input */ - cconvert->pub.color_convert = null_convert; - break; - default: - ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); - } - break; - - case JCS_CMYK: - if (cinfo->num_components != 4) - ERREXIT(cinfo, JERR_BAD_J_COLORSPACE); - if (cinfo->in_color_space != JCS_CMYK) - ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); - cconvert->pub.color_convert = null_convert; - break; - - case JCS_YCCK: - if (cinfo->num_components != 4) - ERREXIT(cinfo, JERR_BAD_J_COLORSPACE); - switch (cinfo->in_color_space) { - case JCS_CMYK: - cconvert->pub.start_pass = rgb_ycc_start; - cconvert->pub.color_convert = cmyk_ycck_convert; - break; - case JCS_YCCK: - cconvert->pub.color_convert = null_convert; - break; - default: - ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); - } - break; - - default: /* allow null conversion of JCS_UNKNOWN */ - if (cinfo->jpeg_color_space != cinfo->in_color_space || - cinfo->num_components != cinfo->input_components) - ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); - cconvert->pub.color_convert = null_convert; - } -} diff --git a/cmake/externals/libjpeg/jcdctmgr.c b/cmake/externals/libjpeg/jcdctmgr.c deleted file mode 100644 index a48ccd814..000000000 --- a/cmake/externals/libjpeg/jcdctmgr.c +++ /dev/null @@ -1,466 +0,0 @@ -/* - * jcdctmgr.c - * - * Copyright (C) 1994-1996, Thomas G. Lane. - * Modified 2003-2020 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains the forward-DCT management logic. - * This code selects a particular DCT implementation to be used, - * and it performs related housekeeping chores including coefficient - * quantization. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" -#include "jdct.h" /* Private declarations for DCT subsystem */ - - -/* Private subobject for this module */ - -typedef struct { - struct jpeg_forward_dct pub; /* public fields */ - - /* Pointer to the DCT routine actually in use */ - forward_DCT_method_ptr do_dct[MAX_COMPONENTS]; - -#ifdef DCT_FLOAT_SUPPORTED - /* Same as above for the floating-point case. */ - float_DCT_method_ptr do_float_dct[MAX_COMPONENTS]; -#endif -} my_fdct_controller; - -typedef my_fdct_controller * my_fdct_ptr; - - -/* The allocated post-DCT divisor tables -- big enough for any - * supported variant and not identical to the quant table entries, - * because of scaling (especially for an unnormalized DCT) -- - * are pointed to by dct_table in the per-component comp_info - * structures. Each table is given in normal array order. - */ - -typedef union { - DCTELEM int_array[DCTSIZE2]; -#ifdef DCT_FLOAT_SUPPORTED - FAST_FLOAT float_array[DCTSIZE2]; -#endif -} divisor_table; - - -/* The current scaled-DCT routines require ISLOW-style divisor tables, - * so be sure to compile that code if either ISLOW or SCALING is requested. - */ -#ifdef DCT_ISLOW_SUPPORTED -#define PROVIDE_ISLOW_TABLES -#else -#ifdef DCT_SCALING_SUPPORTED -#define PROVIDE_ISLOW_TABLES -#endif -#endif - - -/* - * Perform forward DCT on one or more blocks of a component. - * - * The input samples are taken from the sample_data[] array starting at - * position start_col, and moving to the right for any additional blocks. - * The quantized coefficients are returned in coef_blocks[]. - */ - -METHODDEF(void) -forward_DCT (j_compress_ptr cinfo, jpeg_component_info * compptr, - JSAMPARRAY sample_data, JBLOCKROW coef_blocks, - JDIMENSION start_col, JDIMENSION num_blocks) -/* This version is used for integer DCT implementations. */ -{ - /* This routine is heavily used, so it's worth coding it tightly. */ - my_fdct_ptr fdct = (my_fdct_ptr) cinfo->fdct; - forward_DCT_method_ptr do_dct = fdct->do_dct[compptr->component_index]; - DCTELEM * divisors = (DCTELEM *) compptr->dct_table; - DCTELEM workspace[DCTSIZE2]; /* work area for FDCT subroutine */ - JDIMENSION bi; - - for (bi = 0; bi < num_blocks; bi++, start_col += compptr->DCT_h_scaled_size) { - /* Perform the DCT */ - (*do_dct) (workspace, sample_data, start_col); - - /* Quantize/descale the coefficients, and store into coef_blocks[] */ - { register DCTELEM temp, qval; - register int i; - register JCOEFPTR output_ptr = coef_blocks[bi]; - - for (i = 0; i < DCTSIZE2; i++) { - qval = divisors[i]; - temp = workspace[i]; - /* Divide the coefficient value by qval, ensuring proper rounding. - * Since C does not specify the direction of rounding for negative - * quotients, we have to force the dividend positive for portability. - * - * In most files, at least half of the output values will be zero - * (at default quantization settings, more like three-quarters...) - * so we should ensure that this case is fast. On many machines, - * a comparison is enough cheaper than a divide to make a special test - * a win. Since both inputs will be nonnegative, we need only test - * for a < b to discover whether a/b is 0. - * If your machine's division is fast enough, define FAST_DIVIDE. - */ -#ifdef FAST_DIVIDE -#define DIVIDE_BY(a,b) a /= b -#else -#define DIVIDE_BY(a,b) if (a >= b) a /= b; else a = 0 -#endif - if (temp < 0) { - temp = -temp; - temp += qval>>1; /* for rounding */ - DIVIDE_BY(temp, qval); - temp = -temp; - } else { - temp += qval>>1; /* for rounding */ - DIVIDE_BY(temp, qval); - } - output_ptr[i] = (JCOEF) temp; - } - } - } -} - - -#ifdef DCT_FLOAT_SUPPORTED - -METHODDEF(void) -forward_DCT_float (j_compress_ptr cinfo, jpeg_component_info * compptr, - JSAMPARRAY sample_data, JBLOCKROW coef_blocks, - JDIMENSION start_col, JDIMENSION num_blocks) -/* This version is used for floating-point DCT implementations. */ -{ - /* This routine is heavily used, so it's worth coding it tightly. */ - my_fdct_ptr fdct = (my_fdct_ptr) cinfo->fdct; - float_DCT_method_ptr do_dct = fdct->do_float_dct[compptr->component_index]; - FAST_FLOAT * divisors = (FAST_FLOAT *) compptr->dct_table; - FAST_FLOAT workspace[DCTSIZE2]; /* work area for FDCT subroutine */ - JDIMENSION bi; - - for (bi = 0; bi < num_blocks; bi++, start_col += compptr->DCT_h_scaled_size) { - /* Perform the DCT */ - (*do_dct) (workspace, sample_data, start_col); - - /* Quantize/descale the coefficients, and store into coef_blocks[] */ - { register FAST_FLOAT temp; - register int i; - register JCOEFPTR output_ptr = coef_blocks[bi]; - - for (i = 0; i < DCTSIZE2; i++) { - /* Apply the quantization and scaling factor */ - temp = workspace[i] * divisors[i]; - /* Round to nearest integer. - * Since C does not specify the direction of rounding for negative - * quotients, we have to force the dividend positive for portability. - * The maximum coefficient size is +-16K (for 12-bit data), so this - * code should work for either 16-bit or 32-bit ints. - */ - output_ptr[i] = (JCOEF) ((int) (temp + (FAST_FLOAT) 16384.5) - 16384); - } - } - } -} - -#endif /* DCT_FLOAT_SUPPORTED */ - - -/* - * Initialize for a processing pass. - * Verify that all referenced Q-tables are present, and set up - * the divisor table for each one. - * In the current implementation, DCT of all components is done during - * the first pass, even if only some components will be output in the - * first scan. Hence all components should be examined here. - */ - -METHODDEF(void) -start_pass_fdctmgr (j_compress_ptr cinfo) -{ - my_fdct_ptr fdct = (my_fdct_ptr) cinfo->fdct; - int ci, qtblno, i; - jpeg_component_info *compptr; - int method = 0; - JQUANT_TBL * qtbl; - DCTELEM * dtbl; - - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - /* Select the proper DCT routine for this component's scaling */ - switch ((compptr->DCT_h_scaled_size << 8) + compptr->DCT_v_scaled_size) { -#ifdef DCT_SCALING_SUPPORTED - case ((1 << 8) + 1): - fdct->do_dct[ci] = jpeg_fdct_1x1; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((2 << 8) + 2): - fdct->do_dct[ci] = jpeg_fdct_2x2; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((3 << 8) + 3): - fdct->do_dct[ci] = jpeg_fdct_3x3; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((4 << 8) + 4): - fdct->do_dct[ci] = jpeg_fdct_4x4; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((5 << 8) + 5): - fdct->do_dct[ci] = jpeg_fdct_5x5; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((6 << 8) + 6): - fdct->do_dct[ci] = jpeg_fdct_6x6; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((7 << 8) + 7): - fdct->do_dct[ci] = jpeg_fdct_7x7; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((9 << 8) + 9): - fdct->do_dct[ci] = jpeg_fdct_9x9; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((10 << 8) + 10): - fdct->do_dct[ci] = jpeg_fdct_10x10; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((11 << 8) + 11): - fdct->do_dct[ci] = jpeg_fdct_11x11; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((12 << 8) + 12): - fdct->do_dct[ci] = jpeg_fdct_12x12; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((13 << 8) + 13): - fdct->do_dct[ci] = jpeg_fdct_13x13; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((14 << 8) + 14): - fdct->do_dct[ci] = jpeg_fdct_14x14; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((15 << 8) + 15): - fdct->do_dct[ci] = jpeg_fdct_15x15; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((16 << 8) + 16): - fdct->do_dct[ci] = jpeg_fdct_16x16; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((16 << 8) + 8): - fdct->do_dct[ci] = jpeg_fdct_16x8; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((14 << 8) + 7): - fdct->do_dct[ci] = jpeg_fdct_14x7; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((12 << 8) + 6): - fdct->do_dct[ci] = jpeg_fdct_12x6; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((10 << 8) + 5): - fdct->do_dct[ci] = jpeg_fdct_10x5; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((8 << 8) + 4): - fdct->do_dct[ci] = jpeg_fdct_8x4; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((6 << 8) + 3): - fdct->do_dct[ci] = jpeg_fdct_6x3; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((4 << 8) + 2): - fdct->do_dct[ci] = jpeg_fdct_4x2; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((2 << 8) + 1): - fdct->do_dct[ci] = jpeg_fdct_2x1; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((8 << 8) + 16): - fdct->do_dct[ci] = jpeg_fdct_8x16; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((7 << 8) + 14): - fdct->do_dct[ci] = jpeg_fdct_7x14; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((6 << 8) + 12): - fdct->do_dct[ci] = jpeg_fdct_6x12; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((5 << 8) + 10): - fdct->do_dct[ci] = jpeg_fdct_5x10; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((4 << 8) + 8): - fdct->do_dct[ci] = jpeg_fdct_4x8; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((3 << 8) + 6): - fdct->do_dct[ci] = jpeg_fdct_3x6; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((2 << 8) + 4): - fdct->do_dct[ci] = jpeg_fdct_2x4; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; - case ((1 << 8) + 2): - fdct->do_dct[ci] = jpeg_fdct_1x2; - method = JDCT_ISLOW; /* jfdctint uses islow-style table */ - break; -#endif - case ((DCTSIZE << 8) + DCTSIZE): - switch (cinfo->dct_method) { -#ifdef DCT_ISLOW_SUPPORTED - case JDCT_ISLOW: - fdct->do_dct[ci] = jpeg_fdct_islow; - method = JDCT_ISLOW; - break; -#endif -#ifdef DCT_IFAST_SUPPORTED - case JDCT_IFAST: - fdct->do_dct[ci] = jpeg_fdct_ifast; - method = JDCT_IFAST; - break; -#endif -#ifdef DCT_FLOAT_SUPPORTED - case JDCT_FLOAT: - fdct->do_float_dct[ci] = jpeg_fdct_float; - method = JDCT_FLOAT; - break; -#endif - default: - ERREXIT(cinfo, JERR_NOT_COMPILED); - } - break; - default: - ERREXIT2(cinfo, JERR_BAD_DCTSIZE, - compptr->DCT_h_scaled_size, compptr->DCT_v_scaled_size); - } - qtblno = compptr->quant_tbl_no; - /* Make sure specified quantization table is present */ - if (qtblno < 0 || qtblno >= NUM_QUANT_TBLS || - cinfo->quant_tbl_ptrs[qtblno] == NULL) - ERREXIT1(cinfo, JERR_NO_QUANT_TABLE, qtblno); - qtbl = cinfo->quant_tbl_ptrs[qtblno]; - /* Create divisor table from quant table */ - switch (method) { -#ifdef PROVIDE_ISLOW_TABLES - case JDCT_ISLOW: - /* For LL&M IDCT method, divisors are equal to raw quantization - * coefficients multiplied by 8 (to counteract scaling). - */ - dtbl = (DCTELEM *) compptr->dct_table; - for (i = 0; i < DCTSIZE2; i++) { - dtbl[i] = - ((DCTELEM) qtbl->quantval[i]) << (compptr->component_needed ? 4 : 3); - } - fdct->pub.forward_DCT[ci] = forward_DCT; - break; -#endif -#ifdef DCT_IFAST_SUPPORTED - case JDCT_IFAST: - { - /* For AA&N IDCT method, divisors are equal to quantization - * coefficients scaled by scalefactor[row]*scalefactor[col], where - * scalefactor[0] = 1 - * scalefactor[k] = cos(k*PI/16) * sqrt(2) for k=1..7 - * We apply a further scale factor of 8. - */ -#define CONST_BITS 14 - static const INT16 aanscales[DCTSIZE2] = { - /* precomputed values scaled up by 14 bits */ - 16384, 22725, 21407, 19266, 16384, 12873, 8867, 4520, - 22725, 31521, 29692, 26722, 22725, 17855, 12299, 6270, - 21407, 29692, 27969, 25172, 21407, 16819, 11585, 5906, - 19266, 26722, 25172, 22654, 19266, 15137, 10426, 5315, - 16384, 22725, 21407, 19266, 16384, 12873, 8867, 4520, - 12873, 17855, 16819, 15137, 12873, 10114, 6967, 3552, - 8867, 12299, 11585, 10426, 8867, 6967, 4799, 2446, - 4520, 6270, 5906, 5315, 4520, 3552, 2446, 1247 - }; - SHIFT_TEMPS - - dtbl = (DCTELEM *) compptr->dct_table; - for (i = 0; i < DCTSIZE2; i++) { - dtbl[i] = (DCTELEM) - DESCALE(MULTIPLY16V16((INT32) qtbl->quantval[i], - (INT32) aanscales[i]), - compptr->component_needed ? CONST_BITS-4 : CONST_BITS-3); - } - } - fdct->pub.forward_DCT[ci] = forward_DCT; - break; -#endif -#ifdef DCT_FLOAT_SUPPORTED - case JDCT_FLOAT: - { - /* For float AA&N IDCT method, divisors are equal to quantization - * coefficients scaled by scalefactor[row]*scalefactor[col], where - * scalefactor[0] = 1 - * scalefactor[k] = cos(k*PI/16) * sqrt(2) for k=1..7 - * We apply a further scale factor of 8. - * What's actually stored is 1/divisor so that the inner loop can - * use a multiplication rather than a division. - */ - FAST_FLOAT * fdtbl = (FAST_FLOAT *) compptr->dct_table; - int row, col; - static const double aanscalefactor[DCTSIZE] = { - 1.0, 1.387039845, 1.306562965, 1.175875602, - 1.0, 0.785694958, 0.541196100, 0.275899379 - }; - - i = 0; - for (row = 0; row < DCTSIZE; row++) { - for (col = 0; col < DCTSIZE; col++) { - fdtbl[i] = (FAST_FLOAT) - (1.0 / ((double) qtbl->quantval[i] * - aanscalefactor[row] * aanscalefactor[col] * - (compptr->component_needed ? 16.0 : 8.0))); - i++; - } - } - } - fdct->pub.forward_DCT[ci] = forward_DCT_float; - break; -#endif - default: - ERREXIT(cinfo, JERR_NOT_COMPILED); - } - } -} - - -/* - * Initialize FDCT manager. - */ - -GLOBAL(void) -jinit_forward_dct (j_compress_ptr cinfo) -{ - my_fdct_ptr fdct; - int ci; - jpeg_component_info *compptr; - - fdct = (my_fdct_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_fdct_controller)); - cinfo->fdct = &fdct->pub; - fdct->pub.start_pass = start_pass_fdctmgr; - - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - /* Allocate a divisor table for each component */ - compptr->dct_table = (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(divisor_table)); - } -} diff --git a/cmake/externals/libjpeg/jchuff.c b/cmake/externals/libjpeg/jchuff.c deleted file mode 100644 index 1f527b218..000000000 --- a/cmake/externals/libjpeg/jchuff.c +++ /dev/null @@ -1,1656 +0,0 @@ -/* - * jchuff.c - * - * Copyright (C) 1991-1997, Thomas G. Lane. - * Modified 2006-2023 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains Huffman entropy encoding routines. - * Both sequential and progressive modes are supported in this single module. - * - * Much of the complexity here has to do with supporting output suspension. - * If the data destination module demands suspension, we want to be able to - * back up to the start of the current MCU. To do this, we copy state - * variables into local working storage, and update them back to the - * permanent JPEG objects only upon successful completion of an MCU. - * - * We do not support output suspension for the progressive JPEG mode, since - * the library currently does not allow multiple-scan files to be written - * with output suspension. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* The legal range of a DCT coefficient is - * -1024 .. +1023 for 8-bit sample data precision; - * -16384 .. +16383 for 12-bit sample data precision. - * Hence the magnitude should always fit in sample data precision + 2 bits. - */ - -/* Derived data constructed for each Huffman table */ - -typedef struct { - unsigned int ehufco[256]; /* code for each symbol */ - char ehufsi[256]; /* length of code for each symbol */ - /* If no code has been allocated for a symbol S, ehufsi[S] contains 0 */ -} c_derived_tbl; - - -/* Expanded entropy encoder object for Huffman encoding. - * - * The savable_state subrecord contains fields that change within an MCU, - * but must not be updated permanently until we complete the MCU. - */ - -typedef struct { - INT32 put_buffer; /* current bit-accumulation buffer */ - int put_bits; /* # of bits now in it */ - int last_dc_val[MAX_COMPS_IN_SCAN]; /* last DC coef for each component */ -} savable_state; - -/* This macro is to work around compilers with missing or broken - * structure assignment. You'll need to fix this code if you have - * such a compiler and you change MAX_COMPS_IN_SCAN. - */ - -#ifndef NO_STRUCT_ASSIGN -#define ASSIGN_STATE(dest,src) ((dest) = (src)) -#else -#if MAX_COMPS_IN_SCAN == 4 -#define ASSIGN_STATE(dest,src) \ - ((dest).put_buffer = (src).put_buffer, \ - (dest).put_bits = (src).put_bits, \ - (dest).last_dc_val[0] = (src).last_dc_val[0], \ - (dest).last_dc_val[1] = (src).last_dc_val[1], \ - (dest).last_dc_val[2] = (src).last_dc_val[2], \ - (dest).last_dc_val[3] = (src).last_dc_val[3]) -#endif -#endif - - -typedef struct { - struct jpeg_entropy_encoder pub; /* public fields */ - - savable_state saved; /* Bit buffer & DC state at start of MCU */ - - /* These fields are NOT loaded into local working state. */ - unsigned int restarts_to_go; /* MCUs left in this restart interval */ - int next_restart_num; /* next restart number to write (0-7) */ - - /* Pointers to derived tables (these workspaces have image lifespan) */ - c_derived_tbl * dc_derived_tbls[NUM_HUFF_TBLS]; - c_derived_tbl * ac_derived_tbls[NUM_HUFF_TBLS]; - - /* Statistics tables for optimization */ - long * dc_count_ptrs[NUM_HUFF_TBLS]; - long * ac_count_ptrs[NUM_HUFF_TBLS]; - - /* Following fields used only in progressive mode */ - - /* Mode flag: TRUE for optimization, FALSE for actual data output */ - boolean gather_statistics; - - /* next_output_byte/free_in_buffer are local copies of cinfo->dest fields. - */ - JOCTET * next_output_byte; /* => next byte to write in buffer */ - size_t free_in_buffer; /* # of byte spaces remaining in buffer */ - j_compress_ptr cinfo; /* link to cinfo (needed for dump_buffer) */ - - /* Coding status for AC components */ - int ac_tbl_no; /* the table number of the single component */ - unsigned int EOBRUN; /* run length of EOBs */ - unsigned int BE; /* # of buffered correction bits before MCU */ - char * bit_buffer; /* buffer for correction bits (1 per char) */ - /* packing correction bits tightly would save some space but cost time... */ -} huff_entropy_encoder; - -typedef huff_entropy_encoder * huff_entropy_ptr; - -/* Working state while writing an MCU (sequential mode). - * This struct contains all the fields that are needed by subroutines. - */ - -typedef struct { - JOCTET * next_output_byte; /* => next byte to write in buffer */ - size_t free_in_buffer; /* # of byte spaces remaining in buffer */ - savable_state cur; /* Current bit buffer & DC state */ - j_compress_ptr cinfo; /* dump_buffer needs access to this */ -} working_state; - -/* MAX_CORR_BITS is the number of bits the AC refinement correction-bit - * buffer can hold. Larger sizes may slightly improve compression, but - * 1000 is already well into the realm of overkill. - * The minimum safe size is 64 bits. - */ - -#define MAX_CORR_BITS 1000 /* Max # of correction bits I can buffer */ - -/* IRIGHT_SHIFT is like RIGHT_SHIFT, but works on int rather than INT32. - * We assume that int right shift is unsigned if INT32 right shift is, - * which should be safe. - */ - -#ifdef RIGHT_SHIFT_IS_UNSIGNED -#define ISHIFT_TEMPS int ishift_temp; -#define IRIGHT_SHIFT(x,shft) \ - ((ishift_temp = (x)) < 0 ? \ - (ishift_temp >> (shft)) | ((~0) << (16-(shft))) : \ - (ishift_temp >> (shft))) -#else -#define ISHIFT_TEMPS -#define IRIGHT_SHIFT(x,shft) ((x) >> (shft)) -#endif - - -/* - * Compute the derived values for a Huffman table. - * This routine also performs some validation checks on the table. - */ - -LOCAL(void) -jpeg_make_c_derived_tbl (j_compress_ptr cinfo, boolean isDC, int tblno, - c_derived_tbl ** pdtbl) -{ - JHUFF_TBL *htbl; - c_derived_tbl *dtbl; - int p, i, l, lastp, si, maxsymbol; - char huffsize[257]; - unsigned int huffcode[257]; - unsigned int code; - - /* Note that huffsize[] and huffcode[] are filled in code-length order, - * paralleling the order of the symbols themselves in htbl->huffval[]. - */ - - /* Find the input Huffman table */ - if (tblno < 0 || tblno >= NUM_HUFF_TBLS) - ERREXIT1(cinfo, JERR_NO_HUFF_TABLE, tblno); - htbl = - isDC ? cinfo->dc_huff_tbl_ptrs[tblno] : cinfo->ac_huff_tbl_ptrs[tblno]; - if (htbl == NULL) - htbl = jpeg_std_huff_table((j_common_ptr) cinfo, isDC, tblno); - - /* Allocate a workspace if we haven't already done so. */ - if (*pdtbl == NULL) - *pdtbl = (c_derived_tbl *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(c_derived_tbl)); - dtbl = *pdtbl; - - /* Figure C.1: make table of Huffman code length for each symbol */ - - p = 0; - for (l = 1; l <= 16; l++) { - i = (int) htbl->bits[l]; - if (i < 0 || p + i > 256) /* protect against table overrun */ - ERREXIT(cinfo, JERR_BAD_HUFF_TABLE); - while (i--) - huffsize[p++] = (char) l; - } - huffsize[p] = 0; - lastp = p; - - /* Figure C.2: generate the codes themselves */ - /* We also validate that the counts represent a legal Huffman code tree. */ - - code = 0; - si = huffsize[0]; - p = 0; - while (huffsize[p]) { - while (((int) huffsize[p]) == si) { - huffcode[p++] = code; - code++; - } - /* code is now 1 more than the last code used for codelength si; but - * it must still fit in si bits, since no code is allowed to be all ones. - */ - if (((INT32) code) >= (((INT32) 1) << si)) - ERREXIT(cinfo, JERR_BAD_HUFF_TABLE); - code <<= 1; - si++; - } - - /* Figure C.3: generate encoding tables */ - /* These are code and size indexed by symbol value */ - - /* Set all codeless symbols to have code length 0; - * this lets us detect duplicate VAL entries here, and later - * allows emit_bits to detect any attempt to emit such symbols. - */ - MEMZERO(dtbl->ehufsi, SIZEOF(dtbl->ehufsi)); - - /* This is also a convenient place to check for out-of-range - * and duplicated VAL entries. We allow 0..255 for AC symbols - * but only 0..15 for DC. (We could constrain them further - * based on data depth and mode, but this seems enough.) - */ - maxsymbol = isDC ? 15 : 255; - - for (p = 0; p < lastp; p++) { - i = htbl->huffval[p]; - if (i < 0 || i > maxsymbol || dtbl->ehufsi[i]) - ERREXIT(cinfo, JERR_BAD_HUFF_TABLE); - dtbl->ehufco[i] = huffcode[p]; - dtbl->ehufsi[i] = huffsize[p]; - } -} - - -/* Outputting bytes to the file. - * NB: these must be called only when actually outputting, - * that is, entropy->gather_statistics == FALSE. - */ - -/* Emit a byte, taking 'action' if must suspend. */ -#define emit_byte_s(state,val,action) \ - { *(state)->next_output_byte++ = (JOCTET) (val); \ - if (--(state)->free_in_buffer == 0) \ - if (! dump_buffer_s(state)) \ - { action; } } - -/* Emit a byte */ -#define emit_byte_e(entropy,val) \ - { *(entropy)->next_output_byte++ = (JOCTET) (val); \ - if (--(entropy)->free_in_buffer == 0) \ - dump_buffer_e(entropy); } - - -LOCAL(boolean) -dump_buffer_s (working_state * state) -/* Empty the output buffer; return TRUE if successful, FALSE if must suspend */ -{ - struct jpeg_destination_mgr * dest = state->cinfo->dest; - - if (! (*dest->empty_output_buffer) (state->cinfo)) - return FALSE; - /* After a successful buffer dump, must reset buffer pointers */ - state->next_output_byte = dest->next_output_byte; - state->free_in_buffer = dest->free_in_buffer; - return TRUE; -} - - -LOCAL(void) -dump_buffer_e (huff_entropy_ptr entropy) -/* Empty the output buffer; we do not support suspension in this case. */ -{ - struct jpeg_destination_mgr * dest = entropy->cinfo->dest; - - if (! (*dest->empty_output_buffer) (entropy->cinfo)) - ERREXIT(entropy->cinfo, JERR_CANT_SUSPEND); - /* After a successful buffer dump, must reset buffer pointers */ - entropy->next_output_byte = dest->next_output_byte; - entropy->free_in_buffer = dest->free_in_buffer; -} - - -/* Outputting bits to the file */ - -/* Only the right 24 bits of put_buffer are used; the valid bits are - * left-justified in this part. At most 16 bits can be passed to emit_bits - * in one call, and we never retain more than 7 bits in put_buffer - * between calls, so 24 bits are sufficient. - */ - -INLINE -LOCAL(boolean) -emit_bits_s (working_state * state, unsigned int code, int size) -/* Emit some bits; return TRUE if successful, FALSE if must suspend */ -{ - /* This routine is heavily used, so it's worth coding tightly. */ - register INT32 put_buffer; - register int put_bits; - - /* if size is 0, caller used an invalid Huffman table entry */ - if (size == 0) - ERREXIT(state->cinfo, JERR_HUFF_MISSING_CODE); - - /* mask off any extra bits in code */ - put_buffer = ((INT32) code) & ((((INT32) 1) << size) - 1); - - /* new number of bits in buffer */ - put_bits = size + state->cur.put_bits; - - put_buffer <<= 24 - put_bits; /* align incoming bits */ - - /* and merge with old buffer contents */ - put_buffer |= state->cur.put_buffer; - - while (put_bits >= 8) { - int c = (int) ((put_buffer >> 16) & 0xFF); - - emit_byte_s(state, c, return FALSE); - if (c == 0xFF) { /* need to stuff a zero byte? */ - emit_byte_s(state, 0, return FALSE); - } - put_buffer <<= 8; - put_bits -= 8; - } - - state->cur.put_buffer = put_buffer; /* update state variables */ - state->cur.put_bits = put_bits; - - return TRUE; -} - - -INLINE -LOCAL(void) -emit_bits_e (huff_entropy_ptr entropy, unsigned int code, int size) -/* Emit some bits, unless we are in gather mode */ -{ - /* This routine is heavily used, so it's worth coding tightly. */ - register INT32 put_buffer; - register int put_bits; - - /* if size is 0, caller used an invalid Huffman table entry */ - if (size == 0) - ERREXIT(entropy->cinfo, JERR_HUFF_MISSING_CODE); - - if (entropy->gather_statistics) - return; /* do nothing if we're only getting stats */ - - /* mask off any extra bits in code */ - put_buffer = ((INT32) code) & ((((INT32) 1) << size) - 1); - - /* new number of bits in buffer */ - put_bits = size + entropy->saved.put_bits; - - put_buffer <<= 24 - put_bits; /* align incoming bits */ - - /* and merge with old buffer contents */ - put_buffer |= entropy->saved.put_buffer; - - while (put_bits >= 8) { - int c = (int) ((put_buffer >> 16) & 0xFF); - - emit_byte_e(entropy, c); - if (c == 0xFF) { /* need to stuff a zero byte? */ - emit_byte_e(entropy, 0); - } - put_buffer <<= 8; - put_bits -= 8; - } - - entropy->saved.put_buffer = put_buffer; /* update variables */ - entropy->saved.put_bits = put_bits; -} - - -LOCAL(boolean) -flush_bits_s (working_state * state) -{ - if (! emit_bits_s(state, 0x7F, 7)) /* fill any partial byte with ones */ - return FALSE; - state->cur.put_buffer = 0; /* and reset bit-buffer to empty */ - state->cur.put_bits = 0; - return TRUE; -} - - -LOCAL(void) -flush_bits_e (huff_entropy_ptr entropy) -{ - emit_bits_e(entropy, 0x7F, 7); /* fill any partial byte with ones */ - entropy->saved.put_buffer = 0; /* and reset bit-buffer to empty */ - entropy->saved.put_bits = 0; -} - - -/* - * Emit (or just count) a Huffman symbol. - */ - -INLINE -LOCAL(void) -emit_dc_symbol (huff_entropy_ptr entropy, int tbl_no, int symbol) -{ - if (entropy->gather_statistics) - entropy->dc_count_ptrs[tbl_no][symbol]++; - else { - c_derived_tbl * tbl = entropy->dc_derived_tbls[tbl_no]; - emit_bits_e(entropy, tbl->ehufco[symbol], tbl->ehufsi[symbol]); - } -} - - -INLINE -LOCAL(void) -emit_ac_symbol (huff_entropy_ptr entropy, int tbl_no, int symbol) -{ - if (entropy->gather_statistics) - entropy->ac_count_ptrs[tbl_no][symbol]++; - else { - c_derived_tbl * tbl = entropy->ac_derived_tbls[tbl_no]; - emit_bits_e(entropy, tbl->ehufco[symbol], tbl->ehufsi[symbol]); - } -} - - -/* - * Emit bits from a correction bit buffer. - */ - -LOCAL(void) -emit_buffered_bits (huff_entropy_ptr entropy, char * bufstart, - unsigned int nbits) -{ - if (entropy->gather_statistics) - return; /* no real work */ - - while (nbits > 0) { - emit_bits_e(entropy, (unsigned int) (*bufstart), 1); - bufstart++; - nbits--; - } -} - - -/* - * Emit any pending EOBRUN symbol. - */ - -LOCAL(void) -emit_eobrun (huff_entropy_ptr entropy) -{ - register int temp, nbits; - - if (entropy->EOBRUN > 0) { /* if there is any pending EOBRUN */ - temp = entropy->EOBRUN; - nbits = 0; - while ((temp >>= 1)) - nbits++; - /* safety check: shouldn't happen given limited correction-bit buffer */ - if (nbits > 14) - ERREXIT(entropy->cinfo, JERR_HUFF_MISSING_CODE); - - emit_ac_symbol(entropy, entropy->ac_tbl_no, nbits << 4); - if (nbits) - emit_bits_e(entropy, entropy->EOBRUN, nbits); - - entropy->EOBRUN = 0; - - /* Emit any buffered correction bits */ - emit_buffered_bits(entropy, entropy->bit_buffer, entropy->BE); - entropy->BE = 0; - } -} - - -/* - * Emit a restart marker & resynchronize predictions. - */ - -LOCAL(boolean) -emit_restart_s (working_state * state, int restart_num) -{ - int ci; - - if (! flush_bits_s(state)) - return FALSE; - - emit_byte_s(state, 0xFF, return FALSE); - emit_byte_s(state, JPEG_RST0 + restart_num, return FALSE); - - /* Re-initialize DC predictions to 0 */ - for (ci = 0; ci < state->cinfo->comps_in_scan; ci++) - state->cur.last_dc_val[ci] = 0; - - /* The restart counter is not updated until we successfully write the MCU. */ - - return TRUE; -} - - -LOCAL(void) -emit_restart_e (huff_entropy_ptr entropy, int restart_num) -{ - int ci; - - emit_eobrun(entropy); - - if (! entropy->gather_statistics) { - flush_bits_e(entropy); - emit_byte_e(entropy, 0xFF); - emit_byte_e(entropy, JPEG_RST0 + restart_num); - } - - if (entropy->cinfo->Ss == 0) { - /* Re-initialize DC predictions to 0 */ - for (ci = 0; ci < entropy->cinfo->comps_in_scan; ci++) - entropy->saved.last_dc_val[ci] = 0; - } else { - /* Re-initialize all AC-related fields to 0 */ - entropy->EOBRUN = 0; - entropy->BE = 0; - } -} - - -/* - * MCU encoding for DC initial scan (either spectral selection, - * or first pass of successive approximation). - */ - -METHODDEF(boolean) -encode_mcu_DC_first (j_compress_ptr cinfo, JBLOCKARRAY MCU_data) -{ - huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; - register int temp, temp2; - register int nbits; - int max_coef_bits; - int blkn, ci, tbl; - ISHIFT_TEMPS - - entropy->next_output_byte = cinfo->dest->next_output_byte; - entropy->free_in_buffer = cinfo->dest->free_in_buffer; - - /* Emit restart marker if needed */ - if (cinfo->restart_interval) - if (entropy->restarts_to_go == 0) - emit_restart_e(entropy, entropy->next_restart_num); - - /* Since we're encoding a difference, the range limit is twice as much. */ - max_coef_bits = cinfo->data_precision + 3; - - /* Encode the MCU data blocks */ - for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { - ci = cinfo->MCU_membership[blkn]; - tbl = cinfo->cur_comp_info[ci]->dc_tbl_no; - - /* Compute the DC value after the required point transform by Al. - * This is simply an arithmetic right shift. - */ - temp = IRIGHT_SHIFT((int) (MCU_data[blkn][0][0]), cinfo->Al); - - /* DC differences are figured on the point-transformed values. */ - if ((temp2 = temp - entropy->saved.last_dc_val[ci]) == 0) { - /* Count/emit the Huffman-coded symbol for the number of bits */ - emit_dc_symbol(entropy, tbl, 0); - - continue; - } - - entropy->saved.last_dc_val[ci] = temp; - - /* Encode the DC coefficient difference per section G.1.2.1 */ - if ((temp = temp2) < 0) { - temp = -temp; /* temp is abs value of input */ - /* For a negative input, want temp2 = bitwise complement of abs(input) */ - /* This code assumes we are on a two's complement machine */ - temp2--; - } - - /* Find the number of bits needed for the magnitude of the coefficient */ - nbits = 0; - do nbits++; /* there must be at least one 1 bit */ - while ((temp >>= 1)); - /* Check for out-of-range coefficient values */ - if (nbits > max_coef_bits) - ERREXIT(cinfo, JERR_BAD_DCT_COEF); - - /* Count/emit the Huffman-coded symbol for the number of bits */ - emit_dc_symbol(entropy, tbl, nbits); - - /* Emit that number of bits of the value, if positive, */ - /* or the complement of its magnitude, if negative. */ - emit_bits_e(entropy, (unsigned int) temp2, nbits); - } - - cinfo->dest->next_output_byte = entropy->next_output_byte; - cinfo->dest->free_in_buffer = entropy->free_in_buffer; - - /* Update restart-interval state too */ - if (cinfo->restart_interval) { - if (entropy->restarts_to_go == 0) { - entropy->restarts_to_go = cinfo->restart_interval; - entropy->next_restart_num++; - entropy->next_restart_num &= 7; - } - entropy->restarts_to_go--; - } - - return TRUE; -} - - -/* - * MCU encoding for AC initial scan (either spectral selection, - * or first pass of successive approximation). - */ - -METHODDEF(boolean) -encode_mcu_AC_first (j_compress_ptr cinfo, JBLOCKARRAY MCU_data) -{ - huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; - const int * natural_order; - JBLOCKROW block; - register int temp, temp2; - register int nbits; - register int r, k; - int Se, Al, max_coef_bits; - - entropy->next_output_byte = cinfo->dest->next_output_byte; - entropy->free_in_buffer = cinfo->dest->free_in_buffer; - - /* Emit restart marker if needed */ - if (cinfo->restart_interval) - if (entropy->restarts_to_go == 0) - emit_restart_e(entropy, entropy->next_restart_num); - - Se = cinfo->Se; - Al = cinfo->Al; - natural_order = cinfo->natural_order; - max_coef_bits = cinfo->data_precision + 2; - - /* Encode the MCU data block */ - block = MCU_data[0]; - - /* Encode the AC coefficients per section G.1.2.2, fig. G.3 */ - - r = 0; /* r = run length of zeros */ - - for (k = cinfo->Ss; k <= Se; k++) { - if ((temp = (*block)[natural_order[k]]) == 0) { - r++; - continue; - } - /* We must apply the point transform by Al. For AC coefficients this - * is an integer division with rounding towards 0. To do this portably - * in C, we shift after obtaining the absolute value; so the code is - * interwoven with finding the abs value (temp) and output bits (temp2). - */ - if (temp < 0) { - temp = -temp; /* temp is abs value of input */ - /* Apply the point transform, and watch out for case */ - /* that nonzero coef is zero after point transform. */ - if ((temp >>= Al) == 0) { - r++; - continue; - } - /* For a negative coef, want temp2 = bitwise complement of abs(coef) */ - temp2 = ~temp; - } else { - /* Apply the point transform, and watch out for case */ - /* that nonzero coef is zero after point transform. */ - if ((temp >>= Al) == 0) { - r++; - continue; - } - temp2 = temp; - } - - /* Emit any pending EOBRUN */ - if (entropy->EOBRUN > 0) - emit_eobrun(entropy); - /* if run length > 15, must emit special run-length-16 codes (0xF0) */ - while (r > 15) { - emit_ac_symbol(entropy, entropy->ac_tbl_no, 0xF0); - r -= 16; - } - - /* Find the number of bits needed for the magnitude of the coefficient */ - nbits = 0; - do nbits++; /* there must be at least one 1 bit */ - while ((temp >>= 1)); - /* Check for out-of-range coefficient values */ - if (nbits > max_coef_bits) - ERREXIT(cinfo, JERR_BAD_DCT_COEF); - - /* Count/emit Huffman symbol for run length / number of bits */ - emit_ac_symbol(entropy, entropy->ac_tbl_no, (r << 4) + nbits); - - /* Emit that number of bits of the value, if positive, */ - /* or the complement of its magnitude, if negative. */ - emit_bits_e(entropy, (unsigned int) temp2, nbits); - - r = 0; /* reset zero run length */ - } - - if (r > 0) { /* If there are trailing zeroes, */ - entropy->EOBRUN++; /* count an EOB */ - if (entropy->EOBRUN == 0x7FFF) - emit_eobrun(entropy); /* force it out to avoid overflow */ - } - - cinfo->dest->next_output_byte = entropy->next_output_byte; - cinfo->dest->free_in_buffer = entropy->free_in_buffer; - - /* Update restart-interval state too */ - if (cinfo->restart_interval) { - if (entropy->restarts_to_go == 0) { - entropy->restarts_to_go = cinfo->restart_interval; - entropy->next_restart_num++; - entropy->next_restart_num &= 7; - } - entropy->restarts_to_go--; - } - - return TRUE; -} - - -/* - * MCU encoding for DC successive approximation refinement scan. - * Note: we assume such scans can be multi-component, - * although the spec is not very clear on the point. - */ - -METHODDEF(boolean) -encode_mcu_DC_refine (j_compress_ptr cinfo, JBLOCKARRAY MCU_data) -{ - huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; - int Al, blkn; - - entropy->next_output_byte = cinfo->dest->next_output_byte; - entropy->free_in_buffer = cinfo->dest->free_in_buffer; - - /* Emit restart marker if needed */ - if (cinfo->restart_interval) - if (entropy->restarts_to_go == 0) - emit_restart_e(entropy, entropy->next_restart_num); - - Al = cinfo->Al; - - /* Encode the MCU data blocks */ - for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { - /* We simply emit the Al'th bit of the DC coefficient value. */ - emit_bits_e(entropy, (unsigned int) (MCU_data[blkn][0][0] >> Al), 1); - } - - cinfo->dest->next_output_byte = entropy->next_output_byte; - cinfo->dest->free_in_buffer = entropy->free_in_buffer; - - /* Update restart-interval state too */ - if (cinfo->restart_interval) { - if (entropy->restarts_to_go == 0) { - entropy->restarts_to_go = cinfo->restart_interval; - entropy->next_restart_num++; - entropy->next_restart_num &= 7; - } - entropy->restarts_to_go--; - } - - return TRUE; -} - - -/* - * MCU encoding for AC successive approximation refinement scan. - */ - -METHODDEF(boolean) -encode_mcu_AC_refine (j_compress_ptr cinfo, JBLOCKARRAY MCU_data) -{ - huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; - const int * natural_order; - JBLOCKROW block; - register int temp; - register int r, k; - int Se, Al; - int EOB; - char *BR_buffer; - unsigned int BR; - int absvalues[DCTSIZE2]; - - entropy->next_output_byte = cinfo->dest->next_output_byte; - entropy->free_in_buffer = cinfo->dest->free_in_buffer; - - /* Emit restart marker if needed */ - if (cinfo->restart_interval) - if (entropy->restarts_to_go == 0) - emit_restart_e(entropy, entropy->next_restart_num); - - Se = cinfo->Se; - Al = cinfo->Al; - natural_order = cinfo->natural_order; - - /* Encode the MCU data block */ - block = MCU_data[0]; - - /* It is convenient to make a pre-pass to determine the transformed - * coefficients' absolute values and the EOB position. - */ - EOB = 0; - for (k = cinfo->Ss; k <= Se; k++) { - temp = (*block)[natural_order[k]]; - /* We must apply the point transform by Al. For AC coefficients this - * is an integer division with rounding towards 0. To do this portably - * in C, we shift after obtaining the absolute value. - */ - if (temp < 0) - temp = -temp; /* temp is abs value of input */ - temp >>= Al; /* apply the point transform */ - absvalues[k] = temp; /* save abs value for main pass */ - if (temp == 1) - EOB = k; /* EOB = index of last newly-nonzero coef */ - } - - /* Encode the AC coefficients per section G.1.2.3, fig. G.7 */ - - r = 0; /* r = run length of zeros */ - BR = 0; /* BR = count of buffered bits added now */ - BR_buffer = entropy->bit_buffer + entropy->BE; /* Append bits to buffer */ - - for (k = cinfo->Ss; k <= Se; k++) { - if ((temp = absvalues[k]) == 0) { - r++; - continue; - } - - /* Emit any required ZRLs, but not if they can be folded into EOB */ - while (r > 15 && k <= EOB) { - /* emit any pending EOBRUN and the BE correction bits */ - emit_eobrun(entropy); - /* Emit ZRL */ - emit_ac_symbol(entropy, entropy->ac_tbl_no, 0xF0); - r -= 16; - /* Emit buffered correction bits that must be associated with ZRL */ - emit_buffered_bits(entropy, BR_buffer, BR); - BR_buffer = entropy->bit_buffer; /* BE bits are gone now */ - BR = 0; - } - - /* If the coef was previously nonzero, it only needs a correction bit. - * NOTE: a straight translation of the spec's figure G.7 would suggest - * that we also need to test r > 15. But if r > 15, we can only get here - * if k > EOB, which implies that this coefficient is not 1. - */ - if (temp > 1) { - /* The correction bit is the next bit of the absolute value. */ - BR_buffer[BR++] = (char) (temp & 1); - continue; - } - - /* Emit any pending EOBRUN and the BE correction bits */ - emit_eobrun(entropy); - - /* Count/emit Huffman symbol for run length / number of bits */ - emit_ac_symbol(entropy, entropy->ac_tbl_no, (r << 4) + 1); - - /* Emit output bit for newly-nonzero coef */ - temp = ((*block)[natural_order[k]] < 0) ? 0 : 1; - emit_bits_e(entropy, (unsigned int) temp, 1); - - /* Emit buffered correction bits that must be associated with this code */ - emit_buffered_bits(entropy, BR_buffer, BR); - BR_buffer = entropy->bit_buffer; /* BE bits are gone now */ - BR = 0; - r = 0; /* reset zero run length */ - } - - if (r > 0 || BR > 0) { /* If there are trailing zeroes, */ - entropy->EOBRUN++; /* count an EOB */ - entropy->BE += BR; /* concat my correction bits to older ones */ - /* We force out the EOB if we risk either: - * 1. overflow of the EOB counter; - * 2. overflow of the correction bit buffer during the next MCU. - */ - if (entropy->EOBRUN == 0x7FFF || entropy->BE > (MAX_CORR_BITS-DCTSIZE2+1)) - emit_eobrun(entropy); - } - - cinfo->dest->next_output_byte = entropy->next_output_byte; - cinfo->dest->free_in_buffer = entropy->free_in_buffer; - - /* Update restart-interval state too */ - if (cinfo->restart_interval) { - if (entropy->restarts_to_go == 0) { - entropy->restarts_to_go = cinfo->restart_interval; - entropy->next_restart_num++; - entropy->next_restart_num &= 7; - } - entropy->restarts_to_go--; - } - - return TRUE; -} - - -/* Encode a single block's worth of coefficients */ - -LOCAL(boolean) -encode_one_block (working_state * state, JCOEFPTR block, int last_dc_val, - c_derived_tbl *dctbl, c_derived_tbl *actbl) -{ - register int temp, temp2; - register int nbits; - register int r, k; - int Se = state->cinfo->lim_Se; - int max_coef_bits = state->cinfo->data_precision + 3; - const int * natural_order = state->cinfo->natural_order; - - /* Encode the DC coefficient difference per section F.1.2.1 */ - - if ((temp = block[0] - last_dc_val) == 0) { - /* Emit the Huffman-coded symbol for the number of bits */ - if (! emit_bits_s(state, dctbl->ehufco[0], dctbl->ehufsi[0])) - return FALSE; - } else { - if ((temp2 = temp) < 0) { - temp = -temp; /* temp is abs value of input */ - /* For a negative input, want temp2 = bitwise complement of abs(input) */ - /* This code assumes we are on a two's complement machine */ - temp2--; - } - - /* Find the number of bits needed for the magnitude of the coefficient */ - nbits = 0; - do nbits++; /* there must be at least one 1 bit */ - while ((temp >>= 1)); - /* Check for out-of-range coefficient values. - * Since we're encoding a difference, the range limit is twice as much. - */ - if (nbits > max_coef_bits) - ERREXIT(state->cinfo, JERR_BAD_DCT_COEF); - - /* Emit the Huffman-coded symbol for the number of bits */ - if (! emit_bits_s(state, dctbl->ehufco[nbits], dctbl->ehufsi[nbits])) - return FALSE; - - /* Emit that number of bits of the value, if positive, */ - /* or the complement of its magnitude, if negative. */ - if (! emit_bits_s(state, (unsigned int) temp2, nbits)) - return FALSE; - } - - /* Encode the AC coefficients per section F.1.2.2 */ - - r = 0; /* r = run length of zeros */ - - for (k = 1; k <= Se; k++) { - if ((temp = block[natural_order[k]]) == 0) { - r++; - continue; - } - - /* if run length > 15, must emit special run-length-16 codes (0xF0) */ - while (r > 15) { - if (! emit_bits_s(state, actbl->ehufco[0xF0], actbl->ehufsi[0xF0])) - return FALSE; - r -= 16; - } - - if ((temp2 = temp) < 0) { - temp = -temp; /* temp is abs value of input */ - /* For a negative coef, want temp2 = bitwise complement of abs(coef) */ - /* This code assumes we are on a two's complement machine */ - temp2--; - } - - /* Find the number of bits needed for the magnitude of the coefficient */ - nbits = 0; - do nbits++; /* there must be at least one 1 bit */ - while ((temp >>= 1)); - /* Check for out-of-range coefficient values. - * Use ">=" instead of ">" so can use the - * same one larger limit from DC check here. - */ - if (nbits >= max_coef_bits) - ERREXIT(state->cinfo, JERR_BAD_DCT_COEF); - - /* Emit Huffman symbol for run length / number of bits */ - temp = (r << 4) + nbits; - if (! emit_bits_s(state, actbl->ehufco[temp], actbl->ehufsi[temp])) - return FALSE; - - /* Emit that number of bits of the value, if positive, */ - /* or the complement of its magnitude, if negative. */ - if (! emit_bits_s(state, (unsigned int) temp2, nbits)) - return FALSE; - - r = 0; /* reset zero run length */ - } - - /* If the last coef(s) were zero, emit an end-of-block code */ - if (r > 0) - if (! emit_bits_s(state, actbl->ehufco[0], actbl->ehufsi[0])) - return FALSE; - - return TRUE; -} - - -/* - * Encode and output one MCU's worth of Huffman-compressed coefficients. - */ - -METHODDEF(boolean) -encode_mcu_huff (j_compress_ptr cinfo, JBLOCKARRAY MCU_data) -{ - huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; - working_state state; - int blkn, ci; - jpeg_component_info * compptr; - - /* Load up working state */ - state.next_output_byte = cinfo->dest->next_output_byte; - state.free_in_buffer = cinfo->dest->free_in_buffer; - ASSIGN_STATE(state.cur, entropy->saved); - state.cinfo = cinfo; - - /* Emit restart marker if needed */ - if (cinfo->restart_interval) { - if (entropy->restarts_to_go == 0) - if (! emit_restart_s(&state, entropy->next_restart_num)) - return FALSE; - } - - /* Encode the MCU data blocks */ - for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { - ci = cinfo->MCU_membership[blkn]; - compptr = cinfo->cur_comp_info[ci]; - if (! encode_one_block(&state, - MCU_data[blkn][0], state.cur.last_dc_val[ci], - entropy->dc_derived_tbls[compptr->dc_tbl_no], - entropy->ac_derived_tbls[compptr->ac_tbl_no])) - return FALSE; - /* Update last_dc_val */ - state.cur.last_dc_val[ci] = MCU_data[blkn][0][0]; - } - - /* Completed MCU, so update state */ - cinfo->dest->next_output_byte = state.next_output_byte; - cinfo->dest->free_in_buffer = state.free_in_buffer; - ASSIGN_STATE(entropy->saved, state.cur); - - /* Update restart-interval state too */ - if (cinfo->restart_interval) { - if (entropy->restarts_to_go == 0) { - entropy->restarts_to_go = cinfo->restart_interval; - entropy->next_restart_num++; - entropy->next_restart_num &= 7; - } - entropy->restarts_to_go--; - } - - return TRUE; -} - - -/* - * Finish up at the end of a Huffman-compressed scan. - */ - -METHODDEF(void) -finish_pass_huff (j_compress_ptr cinfo) -{ - huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; - working_state state; - - if (cinfo->progressive_mode) { - entropy->next_output_byte = cinfo->dest->next_output_byte; - entropy->free_in_buffer = cinfo->dest->free_in_buffer; - - /* Flush out any buffered data */ - emit_eobrun(entropy); - flush_bits_e(entropy); - - cinfo->dest->next_output_byte = entropy->next_output_byte; - cinfo->dest->free_in_buffer = entropy->free_in_buffer; - } else { - /* Load up working state ... flush_bits needs it */ - state.next_output_byte = cinfo->dest->next_output_byte; - state.free_in_buffer = cinfo->dest->free_in_buffer; - ASSIGN_STATE(state.cur, entropy->saved); - state.cinfo = cinfo; - - /* Flush out the last data */ - if (! flush_bits_s(&state)) - ERREXIT(cinfo, JERR_CANT_SUSPEND); - - /* Update state */ - cinfo->dest->next_output_byte = state.next_output_byte; - cinfo->dest->free_in_buffer = state.free_in_buffer; - ASSIGN_STATE(entropy->saved, state.cur); - } -} - - -/* - * Huffman coding optimization. - * - * We first scan the supplied data and count the number of uses of each symbol - * that is to be Huffman-coded. (This process MUST agree with the code above.) - * Then we build a Huffman coding tree for the observed counts. - * Symbols which are not needed at all for the particular image are not - * assigned any code, which saves space in the DHT marker as well as in - * the compressed data. - */ - - -/* Process a single block's worth of coefficients */ - -LOCAL(void) -htest_one_block (j_compress_ptr cinfo, JCOEFPTR block, int last_dc_val, - long dc_counts[], long ac_counts[]) -{ - register int temp; - register int nbits; - register int r, k; - int Se = cinfo->lim_Se; - int max_coef_bits = cinfo->data_precision + 3; - const int * natural_order = cinfo->natural_order; - - /* Encode the DC coefficient difference per section F.1.2.1 */ - - if ((temp = block[0] - last_dc_val) == 0) { - /* Count the Huffman symbol for the number of bits */ - dc_counts[0]++; - } else { - if (temp < 0) - temp = -temp; /* temp is abs value of input */ - - /* Find the number of bits needed for the magnitude of the coefficient */ - nbits = 0; - do nbits++; /* there must be at least one 1 bit */ - while ((temp >>= 1)); - /* Check for out-of-range coefficient values. - * Since we're encoding a difference, the range limit is twice as much. - */ - if (nbits > max_coef_bits) - ERREXIT(cinfo, JERR_BAD_DCT_COEF); - - /* Count the Huffman symbol for the number of bits */ - dc_counts[nbits]++; - } - - /* Encode the AC coefficients per section F.1.2.2 */ - - r = 0; /* r = run length of zeros */ - - for (k = 1; k <= Se; k++) { - if ((temp = block[natural_order[k]]) == 0) { - r++; - continue; - } - - /* if run length > 15, must emit special run-length-16 codes (0xF0) */ - while (r > 15) { - ac_counts[0xF0]++; - r -= 16; - } - - if (temp < 0) - temp = -temp; /* temp is abs value of input */ - - /* Find the number of bits needed for the magnitude of the coefficient */ - nbits = 0; - do nbits++; /* there must be at least one 1 bit */ - while ((temp >>= 1)); - /* Check for out-of-range coefficient values. - * Use ">=" instead of ">" so can use the - * same one larger limit from DC check here. - */ - if (nbits >= max_coef_bits) - ERREXIT(cinfo, JERR_BAD_DCT_COEF); - - /* Count Huffman symbol for run length / number of bits */ - ac_counts[(r << 4) + nbits]++; - - r = 0; /* reset zero run length */ - } - - /* If the last coef(s) were zero, emit an end-of-block code */ - if (r > 0) - ac_counts[0]++; -} - - -/* - * Trial-encode one MCU's worth of Huffman-compressed coefficients. - * No data is actually output, so no suspension return is possible. - */ - -METHODDEF(boolean) -encode_mcu_gather (j_compress_ptr cinfo, JBLOCKARRAY MCU_data) -{ - huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; - int blkn, ci; - jpeg_component_info * compptr; - - /* Take care of restart intervals if needed */ - if (cinfo->restart_interval) { - if (entropy->restarts_to_go == 0) { - /* Re-initialize DC predictions to 0 */ - for (ci = 0; ci < cinfo->comps_in_scan; ci++) - entropy->saved.last_dc_val[ci] = 0; - /* Update restart state */ - entropy->restarts_to_go = cinfo->restart_interval; - } - entropy->restarts_to_go--; - } - - for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { - ci = cinfo->MCU_membership[blkn]; - compptr = cinfo->cur_comp_info[ci]; - htest_one_block(cinfo, MCU_data[blkn][0], entropy->saved.last_dc_val[ci], - entropy->dc_count_ptrs[compptr->dc_tbl_no], - entropy->ac_count_ptrs[compptr->ac_tbl_no]); - entropy->saved.last_dc_val[ci] = MCU_data[blkn][0][0]; - } - - return TRUE; -} - - -/* - * Generate the best Huffman code table for the given counts, fill htbl. - * - * The JPEG standard requires that no symbol be assigned a codeword of all - * one bits (so that padding bits added at the end of a compressed segment - * can't look like a valid code). Because of the canonical ordering of - * codewords, this just means that there must be an unused slot in the - * longest codeword length category. Section K.2 of the JPEG spec suggests - * reserving such a slot by pretending that symbol 256 is a valid symbol - * with count 1. In theory that's not optimal; giving it count zero but - * including it in the symbol set anyway should give a better Huffman code. - * But the theoretically better code actually seems to come out worse in - * practice, because it produces more all-ones bytes (which incur stuffed - * zero bytes in the final file). In any case the difference is tiny. - * - * The JPEG standard requires Huffman codes to be no more than 16 bits long. - * If some symbols have a very small but nonzero probability, the Huffman tree - * must be adjusted to meet the code length restriction. We currently use - * the adjustment method suggested in JPEG section K.2. This method is *not* - * optimal; it may not choose the best possible limited-length code. But - * typically only very-low-frequency symbols will be given less-than-optimal - * lengths, so the code is almost optimal. Experimental comparisons against - * an optimal limited-length-code algorithm indicate that the difference is - * microscopic --- usually less than a hundredth of a percent of total size. - * So the extra complexity of an optimal algorithm doesn't seem worthwhile. - */ - -LOCAL(void) -jpeg_gen_optimal_table (j_compress_ptr cinfo, JHUFF_TBL * htbl, long freq[]) -{ -#define MAX_CLEN 32 /* assumed maximum initial code length */ - UINT8 bits[MAX_CLEN+1]; /* bits[k] = # of symbols with code length k */ - int codesize[257]; /* codesize[k] = code length of symbol k */ - int others[257]; /* next symbol in current branch of tree */ - int c1, c2, i, j; - UINT8 *p; - long v; - - freq[256] = 1; /* make sure 256 has a nonzero count */ - /* Including the pseudo-symbol 256 in the Huffman procedure guarantees - * that no real symbol is given code-value of all ones, because 256 - * will be placed last in the largest codeword category. - * In the symbol list build procedure this element serves as sentinel - * for the zero run loop. - */ - -#ifndef DONT_USE_FANCY_HUFF_OPT - - /* Build list of symbols sorted in order of descending frequency */ - /* This approach has several benefits (thank to John Korejwa for the idea): - * 1. - * If a codelength category is split during the length limiting procedure - * below, the feature that more frequent symbols are assigned shorter - * codewords remains valid for the adjusted code. - * 2. - * To reduce consecutive ones in a Huffman data stream (thus reducing the - * number of stuff bytes in JPEG) it is preferable to follow 0 branches - * (and avoid 1 branches) as much as possible. This is easily done by - * assigning symbols to leaves of the Huffman tree in order of decreasing - * frequency, with no secondary sort based on codelengths. - * 3. - * The symbol list can be built independently from the assignment of code - * lengths by the Huffman procedure below. - * Note: The symbol list build procedure must be performed first, because - * the Huffman procedure assigning the codelengths clobbers the frequency - * counts! - */ - - /* Here we use the others array as a linked list of nonzero frequencies - * to be sorted. Already sorted elements are removed from the list. - */ - - /* Building list */ - - /* This item does not correspond to a valid symbol frequency and is used - * as starting index. - */ - j = 256; - - for (i = 0;; i++) { - if (freq[i] == 0) /* skip zero frequencies */ - continue; - if (i > 255) - break; - others[j] = i; /* this symbol value */ - j = i; /* previous symbol value */ - } - others[j] = -1; /* mark end of list */ - - /* Sorting list */ - - p = htbl->huffval; - while ((c1 = others[256]) >= 0) { - v = freq[c1]; - i = c1; /* first symbol value */ - j = 256; /* pseudo symbol value for starting index */ - while ((c2 = others[c1]) >= 0) { - if (freq[c2] > v) { - v = freq[c2]; - i = c2; /* this symbol value */ - j = c1; /* previous symbol value */ - } - c1 = c2; - } - others[j] = others[i]; /* remove this symbol i from list */ - *p++ = (UINT8) i; - } - -#endif /* DONT_USE_FANCY_HUFF_OPT */ - - /* This algorithm is explained in section K.2 of the JPEG standard */ - - MEMZERO(bits, SIZEOF(bits)); - MEMZERO(codesize, SIZEOF(codesize)); - for (i = 0; i < 257; i++) - others[i] = -1; /* init links to empty */ - - /* Huffman's basic algorithm to assign optimal code lengths to symbols */ - - for (;;) { - /* Find the smallest nonzero frequency, set c1 = its symbol */ - /* In case of ties, take the larger symbol number */ - c1 = -1; - v = 1000000000L; - for (i = 0; i <= 256; i++) { - if (freq[i] && freq[i] <= v) { - v = freq[i]; - c1 = i; - } - } - - /* Find the next smallest nonzero frequency, set c2 = its symbol */ - /* In case of ties, take the larger symbol number */ - c2 = -1; - v = 1000000000L; - for (i = 0; i <= 256; i++) { - if (freq[i] && freq[i] <= v && i != c1) { - v = freq[i]; - c2 = i; - } - } - - /* Done if we've merged everything into one frequency */ - if (c2 < 0) - break; - - /* Else merge the two counts/trees */ - freq[c1] += freq[c2]; - freq[c2] = 0; - - /* Increment the codesize of everything in c1's tree branch */ - codesize[c1]++; - while (others[c1] >= 0) { - c1 = others[c1]; - codesize[c1]++; - } - - others[c1] = c2; /* chain c2 onto c1's tree branch */ - - /* Increment the codesize of everything in c2's tree branch */ - codesize[c2]++; - while (others[c2] >= 0) { - c2 = others[c2]; - codesize[c2]++; - } - } - - /* Now count the number of symbols of each code length */ - for (i = 0; i <= 256; i++) { - if (codesize[i]) { - /* The JPEG standard seems to think that this can't happen, */ - /* but I'm paranoid... */ - if (codesize[i] > MAX_CLEN) - ERREXIT(cinfo, JERR_HUFF_CLEN_OUTOFBOUNDS); - - bits[codesize[i]]++; - } - } - - /* JPEG doesn't allow symbols with code lengths over 16 bits, so if the pure - * Huffman procedure assigned any such lengths, we must adjust the coding. - * Here is what the JPEG spec says about how this next bit works: - * Since symbols are paired for the longest Huffman code, the symbols are - * removed from this length category two at a time. The prefix for the pair - * (which is one bit shorter) is allocated to one of the pair; then, - * skipping the BITS entry for that prefix length, a code word from the next - * shortest nonzero BITS entry is converted into a prefix for two code words - * one bit longer. - */ - - for (i = MAX_CLEN; i > 16; i--) { - while (bits[i] > 0) { - j = i - 2; /* find length of new prefix to be used */ - while (bits[j] == 0) { - if (j == 0) - ERREXIT(cinfo, JERR_HUFF_CLEN_OUTOFBOUNDS); - j--; - } - - bits[i] -= 2; /* remove two symbols */ - bits[i-1]++; /* one goes in this length */ - bits[j+1] += 2; /* two new symbols in this length */ - bits[j]--; /* symbol of this length is now a prefix */ - } - } - - /* Remove the count for the pseudo-symbol 256 from the largest codelength */ - while (bits[i] == 0) /* find largest codelength still in use */ - i--; - bits[i]--; - - /* Return final symbol counts (only for lengths 0..16) */ - MEMCOPY(htbl->bits, bits, SIZEOF(htbl->bits)); - -#ifdef DONT_USE_FANCY_HUFF_OPT - - /* Return a list of the symbols sorted by code length */ - /* Note: Due to the codelength changes made above, it can happen - * that more frequent symbols are assigned longer codewords. - */ - p = htbl->huffval; - for (i = 1; i <= MAX_CLEN; i++) { - for (j = 0; j <= 255; j++) { - if (codesize[j] == i) { - *p++ = (UINT8) j; - } - } - } - -#endif /* DONT_USE_FANCY_HUFF_OPT */ - - /* Set sent_table FALSE so updated table will be written to JPEG file. */ - htbl->sent_table = FALSE; -} - - -/* - * Finish up a statistics-gathering pass and create the new Huffman tables. - */ - -METHODDEF(void) -finish_pass_gather (j_compress_ptr cinfo) -{ - huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; - int ci, tbl; - jpeg_component_info * compptr; - JHUFF_TBL **htblptr; - boolean did_dc[NUM_HUFF_TBLS]; - boolean did_ac[NUM_HUFF_TBLS]; - - if (cinfo->progressive_mode) - /* Flush out buffered data (all we care about is counting the EOB symbol) */ - emit_eobrun(entropy); - - /* It's important not to apply jpeg_gen_optimal_table more than once - * per table, because it clobbers the input frequency counts! - */ - MEMZERO(did_dc, SIZEOF(did_dc)); - MEMZERO(did_ac, SIZEOF(did_ac)); - - for (ci = 0; ci < cinfo->comps_in_scan; ci++) { - compptr = cinfo->cur_comp_info[ci]; - /* DC needs no table for refinement scan */ - if (cinfo->Ss == 0 && cinfo->Ah == 0) { - tbl = compptr->dc_tbl_no; - if (! did_dc[tbl]) { - htblptr = & cinfo->dc_huff_tbl_ptrs[tbl]; - if (*htblptr == NULL) - *htblptr = jpeg_alloc_huff_table((j_common_ptr) cinfo); - jpeg_gen_optimal_table(cinfo, *htblptr, entropy->dc_count_ptrs[tbl]); - did_dc[tbl] = TRUE; - } - } - /* AC needs no table when not present */ - if (cinfo->Se) { - tbl = compptr->ac_tbl_no; - if (! did_ac[tbl]) { - htblptr = & cinfo->ac_huff_tbl_ptrs[tbl]; - if (*htblptr == NULL) - *htblptr = jpeg_alloc_huff_table((j_common_ptr) cinfo); - jpeg_gen_optimal_table(cinfo, *htblptr, entropy->ac_count_ptrs[tbl]); - did_ac[tbl] = TRUE; - } - } - } -} - - -/* - * Initialize for a Huffman-compressed scan. - * If gather_statistics is TRUE, we do not output anything during the scan, - * just count the Huffman symbols used and generate Huffman code tables. - */ - -METHODDEF(void) -start_pass_huff (j_compress_ptr cinfo, boolean gather_statistics) -{ - huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; - int ci, tbl; - jpeg_component_info * compptr; - - if (gather_statistics) - entropy->pub.finish_pass = finish_pass_gather; - else - entropy->pub.finish_pass = finish_pass_huff; - - if (cinfo->progressive_mode) { - entropy->cinfo = cinfo; - entropy->gather_statistics = gather_statistics; - - /* We assume jcmaster.c already validated the scan parameters. */ - - /* Select execution routine */ - if (cinfo->Ah == 0) { - if (cinfo->Ss == 0) - entropy->pub.encode_mcu = encode_mcu_DC_first; - else - entropy->pub.encode_mcu = encode_mcu_AC_first; - } else { - if (cinfo->Ss == 0) - entropy->pub.encode_mcu = encode_mcu_DC_refine; - else { - entropy->pub.encode_mcu = encode_mcu_AC_refine; - /* AC refinement needs a correction bit buffer */ - if (entropy->bit_buffer == NULL) - entropy->bit_buffer = (char *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, MAX_CORR_BITS * SIZEOF(char)); - } - } - - /* Initialize AC stuff */ - entropy->ac_tbl_no = cinfo->cur_comp_info[0]->ac_tbl_no; - entropy->EOBRUN = 0; - entropy->BE = 0; - } else { - if (gather_statistics) - entropy->pub.encode_mcu = encode_mcu_gather; - else - entropy->pub.encode_mcu = encode_mcu_huff; - } - - for (ci = 0; ci < cinfo->comps_in_scan; ci++) { - compptr = cinfo->cur_comp_info[ci]; - /* DC needs no table for refinement scan */ - if (cinfo->Ss == 0 && cinfo->Ah == 0) { - tbl = compptr->dc_tbl_no; - if (gather_statistics) { - /* Check for invalid table index */ - /* (make_c_derived_tbl does this in the other path) */ - if (tbl < 0 || tbl >= NUM_HUFF_TBLS) - ERREXIT1(cinfo, JERR_NO_HUFF_TABLE, tbl); - /* Allocate and zero the statistics tables */ - /* Note that jpeg_gen_optimal_table expects 257 entries in each table! */ - if (entropy->dc_count_ptrs[tbl] == NULL) - entropy->dc_count_ptrs[tbl] = (long *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, 257 * SIZEOF(long)); - MEMZERO(entropy->dc_count_ptrs[tbl], 257 * SIZEOF(long)); - } else { - /* Compute derived values for Huffman tables */ - /* We may do this more than once for a table, but it's not expensive */ - jpeg_make_c_derived_tbl(cinfo, TRUE, tbl, - & entropy->dc_derived_tbls[tbl]); - } - /* Initialize DC predictions to 0 */ - entropy->saved.last_dc_val[ci] = 0; - } - /* AC needs no table when not present */ - if (cinfo->Se) { - tbl = compptr->ac_tbl_no; - if (gather_statistics) { - if (tbl < 0 || tbl >= NUM_HUFF_TBLS) - ERREXIT1(cinfo, JERR_NO_HUFF_TABLE, tbl); - if (entropy->ac_count_ptrs[tbl] == NULL) - entropy->ac_count_ptrs[tbl] = (long *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, 257 * SIZEOF(long)); - MEMZERO(entropy->ac_count_ptrs[tbl], 257 * SIZEOF(long)); - } else { - jpeg_make_c_derived_tbl(cinfo, FALSE, tbl, - & entropy->ac_derived_tbls[tbl]); - } - } - } - - /* Initialize bit buffer to empty */ - entropy->saved.put_buffer = 0; - entropy->saved.put_bits = 0; - - /* Initialize restart stuff */ - entropy->restarts_to_go = cinfo->restart_interval; - entropy->next_restart_num = 0; -} - - -/* - * Module initialization routine for Huffman entropy encoding. - */ - -GLOBAL(void) -jinit_huff_encoder (j_compress_ptr cinfo) -{ - huff_entropy_ptr entropy; - int i; - - entropy = (huff_entropy_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(huff_entropy_encoder)); - cinfo->entropy = &entropy->pub; - entropy->pub.start_pass = start_pass_huff; - - /* Mark tables unallocated */ - for (i = 0; i < NUM_HUFF_TBLS; i++) { - entropy->dc_derived_tbls[i] = entropy->ac_derived_tbls[i] = NULL; - entropy->dc_count_ptrs[i] = entropy->ac_count_ptrs[i] = NULL; - } - - if (cinfo->progressive_mode) - entropy->bit_buffer = NULL; /* needed only in AC refinement scan */ -} diff --git a/cmake/externals/libjpeg/jcinit.c b/cmake/externals/libjpeg/jcinit.c deleted file mode 100644 index 2aea7ca2a..000000000 --- a/cmake/externals/libjpeg/jcinit.c +++ /dev/null @@ -1,249 +0,0 @@ -/* - * jcinit.c - * - * Copyright (C) 1991-1997, Thomas G. Lane. - * Modified 2003-2017 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains initialization logic for the JPEG compressor. - * This routine is in charge of selecting the modules to be executed and - * making an initialization call to each one. - * - * Logically, this code belongs in jcmaster.c. It's split out because - * linking this routine implies linking the entire compression library. - * For a transcoding-only application, we want to be able to use jcmaster.c - * without linking in the whole library. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* - * Compute JPEG image dimensions and related values. - * NOTE: this is exported for possible use by application. - * Hence it mustn't do anything that can't be done twice. - */ - -GLOBAL(void) -jpeg_calc_jpeg_dimensions (j_compress_ptr cinfo) -/* Do computations that are needed before master selection phase */ -{ - /* Sanity check on input image dimensions to prevent overflow in - * following calculations. - * We do check jpeg_width and jpeg_height in initial_setup in jcmaster.c, - * but image_width and image_height can come from arbitrary data, - * and we need some space for multiplication by block_size. - */ - if (((long) cinfo->image_width >> 24) || ((long) cinfo->image_height >> 24)) - ERREXIT1(cinfo, JERR_IMAGE_TOO_BIG, (unsigned int) JPEG_MAX_DIMENSION); - -#ifdef DCT_SCALING_SUPPORTED - - /* Compute actual JPEG image dimensions and DCT scaling choices. */ - if (cinfo->scale_num >= cinfo->scale_denom * cinfo->block_size) { - /* Provide block_size/1 scaling */ - cinfo->jpeg_width = cinfo->image_width * cinfo->block_size; - cinfo->jpeg_height = cinfo->image_height * cinfo->block_size; - cinfo->min_DCT_h_scaled_size = 1; - cinfo->min_DCT_v_scaled_size = 1; - } else if (cinfo->scale_num * 2 >= cinfo->scale_denom * cinfo->block_size) { - /* Provide block_size/2 scaling */ - cinfo->jpeg_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 2L); - cinfo->jpeg_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 2L); - cinfo->min_DCT_h_scaled_size = 2; - cinfo->min_DCT_v_scaled_size = 2; - } else if (cinfo->scale_num * 3 >= cinfo->scale_denom * cinfo->block_size) { - /* Provide block_size/3 scaling */ - cinfo->jpeg_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 3L); - cinfo->jpeg_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 3L); - cinfo->min_DCT_h_scaled_size = 3; - cinfo->min_DCT_v_scaled_size = 3; - } else if (cinfo->scale_num * 4 >= cinfo->scale_denom * cinfo->block_size) { - /* Provide block_size/4 scaling */ - cinfo->jpeg_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 4L); - cinfo->jpeg_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 4L); - cinfo->min_DCT_h_scaled_size = 4; - cinfo->min_DCT_v_scaled_size = 4; - } else if (cinfo->scale_num * 5 >= cinfo->scale_denom * cinfo->block_size) { - /* Provide block_size/5 scaling */ - cinfo->jpeg_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 5L); - cinfo->jpeg_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 5L); - cinfo->min_DCT_h_scaled_size = 5; - cinfo->min_DCT_v_scaled_size = 5; - } else if (cinfo->scale_num * 6 >= cinfo->scale_denom * cinfo->block_size) { - /* Provide block_size/6 scaling */ - cinfo->jpeg_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 6L); - cinfo->jpeg_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 6L); - cinfo->min_DCT_h_scaled_size = 6; - cinfo->min_DCT_v_scaled_size = 6; - } else if (cinfo->scale_num * 7 >= cinfo->scale_denom * cinfo->block_size) { - /* Provide block_size/7 scaling */ - cinfo->jpeg_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 7L); - cinfo->jpeg_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 7L); - cinfo->min_DCT_h_scaled_size = 7; - cinfo->min_DCT_v_scaled_size = 7; - } else if (cinfo->scale_num * 8 >= cinfo->scale_denom * cinfo->block_size) { - /* Provide block_size/8 scaling */ - cinfo->jpeg_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 8L); - cinfo->jpeg_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 8L); - cinfo->min_DCT_h_scaled_size = 8; - cinfo->min_DCT_v_scaled_size = 8; - } else if (cinfo->scale_num * 9 >= cinfo->scale_denom * cinfo->block_size) { - /* Provide block_size/9 scaling */ - cinfo->jpeg_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 9L); - cinfo->jpeg_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 9L); - cinfo->min_DCT_h_scaled_size = 9; - cinfo->min_DCT_v_scaled_size = 9; - } else if (cinfo->scale_num * 10 >= cinfo->scale_denom * cinfo->block_size) { - /* Provide block_size/10 scaling */ - cinfo->jpeg_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 10L); - cinfo->jpeg_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 10L); - cinfo->min_DCT_h_scaled_size = 10; - cinfo->min_DCT_v_scaled_size = 10; - } else if (cinfo->scale_num * 11 >= cinfo->scale_denom * cinfo->block_size) { - /* Provide block_size/11 scaling */ - cinfo->jpeg_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 11L); - cinfo->jpeg_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 11L); - cinfo->min_DCT_h_scaled_size = 11; - cinfo->min_DCT_v_scaled_size = 11; - } else if (cinfo->scale_num * 12 >= cinfo->scale_denom * cinfo->block_size) { - /* Provide block_size/12 scaling */ - cinfo->jpeg_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 12L); - cinfo->jpeg_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 12L); - cinfo->min_DCT_h_scaled_size = 12; - cinfo->min_DCT_v_scaled_size = 12; - } else if (cinfo->scale_num * 13 >= cinfo->scale_denom * cinfo->block_size) { - /* Provide block_size/13 scaling */ - cinfo->jpeg_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 13L); - cinfo->jpeg_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 13L); - cinfo->min_DCT_h_scaled_size = 13; - cinfo->min_DCT_v_scaled_size = 13; - } else if (cinfo->scale_num * 14 >= cinfo->scale_denom * cinfo->block_size) { - /* Provide block_size/14 scaling */ - cinfo->jpeg_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 14L); - cinfo->jpeg_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 14L); - cinfo->min_DCT_h_scaled_size = 14; - cinfo->min_DCT_v_scaled_size = 14; - } else if (cinfo->scale_num * 15 >= cinfo->scale_denom * cinfo->block_size) { - /* Provide block_size/15 scaling */ - cinfo->jpeg_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 15L); - cinfo->jpeg_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 15L); - cinfo->min_DCT_h_scaled_size = 15; - cinfo->min_DCT_v_scaled_size = 15; - } else { - /* Provide block_size/16 scaling */ - cinfo->jpeg_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * cinfo->block_size, 16L); - cinfo->jpeg_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * cinfo->block_size, 16L); - cinfo->min_DCT_h_scaled_size = 16; - cinfo->min_DCT_v_scaled_size = 16; - } - -#else /* !DCT_SCALING_SUPPORTED */ - - /* Hardwire it to "no scaling" */ - cinfo->jpeg_width = cinfo->image_width; - cinfo->jpeg_height = cinfo->image_height; - cinfo->min_DCT_h_scaled_size = DCTSIZE; - cinfo->min_DCT_v_scaled_size = DCTSIZE; - -#endif /* DCT_SCALING_SUPPORTED */ -} - - -/* - * Master selection of compression modules. - * This is done once at the start of processing an image. We determine - * which modules will be used and give them appropriate initialization calls. - */ - -GLOBAL(void) -jinit_compress_master (j_compress_ptr cinfo) -{ - long samplesperrow; - JDIMENSION jd_samplesperrow; - - /* For now, precision must match compiled-in value... */ - if (cinfo->data_precision != BITS_IN_JSAMPLE) - ERREXIT1(cinfo, JERR_BAD_PRECISION, cinfo->data_precision); - - /* Sanity check on input image dimensions */ - if (cinfo->image_height <= 0 || cinfo->image_width <= 0 || - cinfo->input_components <= 0) - ERREXIT(cinfo, JERR_EMPTY_IMAGE); - - /* Width of an input scanline must be representable as JDIMENSION. */ - samplesperrow = (long) cinfo->image_width * (long) cinfo->input_components; - jd_samplesperrow = (JDIMENSION) samplesperrow; - if ((long) jd_samplesperrow != samplesperrow) - ERREXIT(cinfo, JERR_WIDTH_OVERFLOW); - - /* Compute JPEG image dimensions and related values. */ - jpeg_calc_jpeg_dimensions(cinfo); - - /* Initialize master control (includes parameter checking/processing) */ - jinit_c_master_control(cinfo, FALSE /* full compression */); - - /* Preprocessing */ - if (! cinfo->raw_data_in) { - jinit_color_converter(cinfo); - jinit_downsampler(cinfo); - jinit_c_prep_controller(cinfo, FALSE /* never need full buffer here */); - } - /* Forward DCT */ - jinit_forward_dct(cinfo); - /* Entropy encoding: either Huffman or arithmetic coding. */ - if (cinfo->arith_code) - jinit_arith_encoder(cinfo); - else { - jinit_huff_encoder(cinfo); - } - - /* Need a full-image coefficient buffer in any multi-pass mode. */ - jinit_c_coef_controller(cinfo, - (boolean) (cinfo->num_scans > 1 || cinfo->optimize_coding)); - jinit_c_main_controller(cinfo, FALSE /* never need full buffer here */); - - jinit_marker_writer(cinfo); - - /* We can now tell the memory manager to allocate virtual arrays. */ - (*cinfo->mem->realize_virt_arrays) ((j_common_ptr) cinfo); - - /* Write the datastream header (SOI) immediately. - * Frame and scan headers are postponed till later. - * This lets application insert special markers after the SOI. - */ - (*cinfo->marker->write_file_header) (cinfo); -} diff --git a/cmake/externals/libjpeg/jcmainct.c b/cmake/externals/libjpeg/jcmainct.c deleted file mode 100644 index 39b97902e..000000000 --- a/cmake/externals/libjpeg/jcmainct.c +++ /dev/null @@ -1,297 +0,0 @@ -/* - * jcmainct.c - * - * Copyright (C) 1994-1996, Thomas G. Lane. - * Modified 2003-2012 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains the main buffer controller for compression. - * The main buffer lies between the pre-processor and the JPEG - * compressor proper; it holds downsampled data in the JPEG colorspace. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* Note: currently, there is no operating mode in which a full-image buffer - * is needed at this step. If there were, that mode could not be used with - * "raw data" input, since this module is bypassed in that case. However, - * we've left the code here for possible use in special applications. - */ -#undef FULL_MAIN_BUFFER_SUPPORTED - - -/* Private buffer controller object */ - -typedef struct { - struct jpeg_c_main_controller pub; /* public fields */ - - JDIMENSION cur_iMCU_row; /* number of current iMCU row */ - JDIMENSION rowgroup_ctr; /* counts row groups received in iMCU row */ - boolean suspended; /* remember if we suspended output */ - J_BUF_MODE pass_mode; /* current operating mode */ - - /* If using just a strip buffer, this points to the entire set of buffers - * (we allocate one for each component). In the full-image case, this - * points to the currently accessible strips of the virtual arrays. - */ - JSAMPARRAY buffer[MAX_COMPONENTS]; - -#ifdef FULL_MAIN_BUFFER_SUPPORTED - /* If using full-image storage, this array holds pointers to virtual-array - * control blocks for each component. Unused if not full-image storage. - */ - jvirt_sarray_ptr whole_image[MAX_COMPONENTS]; -#endif -} my_main_controller; - -typedef my_main_controller * my_main_ptr; - - -/* Forward declarations */ -METHODDEF(void) process_data_simple_main - JPP((j_compress_ptr cinfo, JSAMPARRAY input_buf, - JDIMENSION *in_row_ctr, JDIMENSION in_rows_avail)); -#ifdef FULL_MAIN_BUFFER_SUPPORTED -METHODDEF(void) process_data_buffer_main - JPP((j_compress_ptr cinfo, JSAMPARRAY input_buf, - JDIMENSION *in_row_ctr, JDIMENSION in_rows_avail)); -#endif - - -/* - * Initialize for a processing pass. - */ - -METHODDEF(void) -start_pass_main (j_compress_ptr cinfo, J_BUF_MODE pass_mode) -{ - my_main_ptr mainp = (my_main_ptr) cinfo->main; - - /* Do nothing in raw-data mode. */ - if (cinfo->raw_data_in) - return; - - mainp->cur_iMCU_row = 0; /* initialize counters */ - mainp->rowgroup_ctr = 0; - mainp->suspended = FALSE; - mainp->pass_mode = pass_mode; /* save mode for use by process_data */ - - switch (pass_mode) { - case JBUF_PASS_THRU: -#ifdef FULL_MAIN_BUFFER_SUPPORTED - if (mainp->whole_image[0] != NULL) - ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); -#endif - mainp->pub.process_data = process_data_simple_main; - break; -#ifdef FULL_MAIN_BUFFER_SUPPORTED - case JBUF_SAVE_SOURCE: - case JBUF_CRANK_DEST: - case JBUF_SAVE_AND_PASS: - if (mainp->whole_image[0] == NULL) - ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); - mainp->pub.process_data = process_data_buffer_main; - break; -#endif - default: - ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); - break; - } -} - - -/* - * Process some data. - * This routine handles the simple pass-through mode, - * where we have only a strip buffer. - */ - -METHODDEF(void) -process_data_simple_main (j_compress_ptr cinfo, - JSAMPARRAY input_buf, JDIMENSION *in_row_ctr, - JDIMENSION in_rows_avail) -{ - my_main_ptr mainp = (my_main_ptr) cinfo->main; - - while (mainp->cur_iMCU_row < cinfo->total_iMCU_rows) { - /* Read input data if we haven't filled the main buffer yet */ - if (mainp->rowgroup_ctr < (JDIMENSION) cinfo->min_DCT_v_scaled_size) - (*cinfo->prep->pre_process_data) (cinfo, - input_buf, in_row_ctr, in_rows_avail, - mainp->buffer, &mainp->rowgroup_ctr, - (JDIMENSION) cinfo->min_DCT_v_scaled_size); - - /* If we don't have a full iMCU row buffered, return to application for - * more data. Note that preprocessor will always pad to fill the iMCU row - * at the bottom of the image. - */ - if (mainp->rowgroup_ctr != (JDIMENSION) cinfo->min_DCT_v_scaled_size) - return; - - /* Send the completed row to the compressor */ - if (! (*cinfo->coef->compress_data) (cinfo, mainp->buffer)) { - /* If compressor did not consume the whole row, then we must need to - * suspend processing and return to the application. In this situation - * we pretend we didn't yet consume the last input row; otherwise, if - * it happened to be the last row of the image, the application would - * think we were done. - */ - if (! mainp->suspended) { - (*in_row_ctr)--; - mainp->suspended = TRUE; - } - return; - } - /* We did finish the row. Undo our little suspension hack if a previous - * call suspended; then mark the main buffer empty. - */ - if (mainp->suspended) { - (*in_row_ctr)++; - mainp->suspended = FALSE; - } - mainp->rowgroup_ctr = 0; - mainp->cur_iMCU_row++; - } -} - - -#ifdef FULL_MAIN_BUFFER_SUPPORTED - -/* - * Process some data. - * This routine handles all of the modes that use a full-size buffer. - */ - -METHODDEF(void) -process_data_buffer_main (j_compress_ptr cinfo, - JSAMPARRAY input_buf, JDIMENSION *in_row_ctr, - JDIMENSION in_rows_avail) -{ - my_main_ptr mainp = (my_main_ptr) cinfo->main; - int ci; - jpeg_component_info *compptr; - boolean writing = (mainp->pass_mode != JBUF_CRANK_DEST); - - while (mainp->cur_iMCU_row < cinfo->total_iMCU_rows) { - /* Realign the virtual buffers if at the start of an iMCU row. */ - if (mainp->rowgroup_ctr == 0) { - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - mainp->buffer[ci] = (*cinfo->mem->access_virt_sarray) - ((j_common_ptr) cinfo, mainp->whole_image[ci], mainp->cur_iMCU_row * - ((JDIMENSION) (compptr->v_samp_factor * cinfo->min_DCT_v_scaled_size)), - (JDIMENSION) (compptr->v_samp_factor * cinfo->min_DCT_v_scaled_size), - writing); - } - /* In a read pass, pretend we just read some source data. */ - if (! writing) { - *in_row_ctr += (JDIMENSION) - (cinfo->max_v_samp_factor * cinfo->min_DCT_v_scaled_size); - mainp->rowgroup_ctr = (JDIMENSION) cinfo->min_DCT_v_scaled_size; - } - } - - /* If a write pass, read input data until the current iMCU row is full. */ - /* Note: preprocessor will pad if necessary to fill the last iMCU row. */ - if (writing) { - (*cinfo->prep->pre_process_data) (cinfo, - input_buf, in_row_ctr, in_rows_avail, - mainp->buffer, &mainp->rowgroup_ctr, - (JDIMENSION) cinfo->min_DCT_v_scaled_size); - /* Return to application if we need more data to fill the iMCU row. */ - if (mainp->rowgroup_ctr < (JDIMENSION) cinfo->min_DCT_v_scaled_size) - return; - } - - /* Emit data, unless this is a sink-only pass. */ - if (mainp->pass_mode != JBUF_SAVE_SOURCE) { - if (! (*cinfo->coef->compress_data) (cinfo, mainp->buffer)) { - /* If compressor did not consume the whole row, then we must need to - * suspend processing and return to the application. In this situation - * we pretend we didn't yet consume the last input row; otherwise, if - * it happened to be the last row of the image, the application would - * think we were done. - */ - if (! mainp->suspended) { - (*in_row_ctr)--; - mainp->suspended = TRUE; - } - return; - } - /* We did finish the row. Undo our little suspension hack if a previous - * call suspended; then mark the main buffer empty. - */ - if (mainp->suspended) { - (*in_row_ctr)++; - mainp->suspended = FALSE; - } - } - - /* If get here, we are done with this iMCU row. Mark buffer empty. */ - mainp->rowgroup_ctr = 0; - mainp->cur_iMCU_row++; - } -} - -#endif /* FULL_MAIN_BUFFER_SUPPORTED */ - - -/* - * Initialize main buffer controller. - */ - -GLOBAL(void) -jinit_c_main_controller (j_compress_ptr cinfo, boolean need_full_buffer) -{ - my_main_ptr mainp; - int ci; - jpeg_component_info *compptr; - - mainp = (my_main_ptr) - (*cinfo->mem->alloc_small) ((j_common_ptr) cinfo, JPOOL_IMAGE, - SIZEOF(my_main_controller)); - cinfo->main = &mainp->pub; - mainp->pub.start_pass = start_pass_main; - - /* We don't need to create a buffer in raw-data mode. */ - if (cinfo->raw_data_in) - return; - - /* Create the buffer. It holds downsampled data, so each component - * may be of a different size. - */ - if (need_full_buffer) { -#ifdef FULL_MAIN_BUFFER_SUPPORTED - /* Allocate a full-image virtual array for each component */ - /* Note we pad the bottom to a multiple of the iMCU height */ - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - mainp->whole_image[ci] = (*cinfo->mem->request_virt_sarray) - ((j_common_ptr) cinfo, JPOOL_IMAGE, FALSE, - compptr->width_in_blocks * ((JDIMENSION) compptr->DCT_h_scaled_size), - ((JDIMENSION) jround_up((long) compptr->height_in_blocks, - (long) compptr->v_samp_factor)) * - ((JDIMENSION) cinfo->min_DCT_v_scaled_size), - (JDIMENSION) (compptr->v_samp_factor * compptr->DCT_v_scaled_size)); - } -#else - ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); -#endif - } else { -#ifdef FULL_MAIN_BUFFER_SUPPORTED - mainp->whole_image[0] = NULL; /* flag for no virtual arrays */ -#endif - /* Allocate a strip buffer for each component */ - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - mainp->buffer[ci] = (*cinfo->mem->alloc_sarray) - ((j_common_ptr) cinfo, JPOOL_IMAGE, - compptr->width_in_blocks * ((JDIMENSION) compptr->DCT_h_scaled_size), - (JDIMENSION) (compptr->v_samp_factor * compptr->DCT_v_scaled_size)); - } - } -} diff --git a/cmake/externals/libjpeg/jcmarker.c b/cmake/externals/libjpeg/jcmarker.c deleted file mode 100644 index 8874cd867..000000000 --- a/cmake/externals/libjpeg/jcmarker.c +++ /dev/null @@ -1,717 +0,0 @@ -/* - * jcmarker.c - * - * Copyright (C) 1991-1998, Thomas G. Lane. - * Modified 2003-2019 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains routines to write JPEG datastream markers. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -typedef enum { /* JPEG marker codes */ - M_SOF0 = 0xc0, - M_SOF1 = 0xc1, - M_SOF2 = 0xc2, - M_SOF3 = 0xc3, - - M_SOF5 = 0xc5, - M_SOF6 = 0xc6, - M_SOF7 = 0xc7, - - M_JPG = 0xc8, - M_SOF9 = 0xc9, - M_SOF10 = 0xca, - M_SOF11 = 0xcb, - - M_SOF13 = 0xcd, - M_SOF14 = 0xce, - M_SOF15 = 0xcf, - - M_DHT = 0xc4, - - M_DAC = 0xcc, - - M_RST0 = 0xd0, - M_RST1 = 0xd1, - M_RST2 = 0xd2, - M_RST3 = 0xd3, - M_RST4 = 0xd4, - M_RST5 = 0xd5, - M_RST6 = 0xd6, - M_RST7 = 0xd7, - - M_SOI = 0xd8, - M_EOI = 0xd9, - M_SOS = 0xda, - M_DQT = 0xdb, - M_DNL = 0xdc, - M_DRI = 0xdd, - M_DHP = 0xde, - M_EXP = 0xdf, - - M_APP0 = 0xe0, - M_APP1 = 0xe1, - M_APP2 = 0xe2, - M_APP3 = 0xe3, - M_APP4 = 0xe4, - M_APP5 = 0xe5, - M_APP6 = 0xe6, - M_APP7 = 0xe7, - M_APP8 = 0xe8, - M_APP9 = 0xe9, - M_APP10 = 0xea, - M_APP11 = 0xeb, - M_APP12 = 0xec, - M_APP13 = 0xed, - M_APP14 = 0xee, - M_APP15 = 0xef, - - M_JPG0 = 0xf0, - M_JPG8 = 0xf8, - M_JPG13 = 0xfd, - M_COM = 0xfe, - - M_TEM = 0x01, - - M_ERROR = 0x100 -} JPEG_MARKER; - - -/* Private state */ - -typedef struct { - struct jpeg_marker_writer pub; /* public fields */ - - unsigned int last_restart_interval; /* last DRI value emitted; 0 after SOI */ -} my_marker_writer; - -typedef my_marker_writer * my_marker_ptr; - - -/* - * Basic output routines. - * - * Note that we do not support suspension while writing a marker. - * Therefore, an application using suspension must ensure that there is - * enough buffer space for the initial markers (typ. 600-700 bytes) before - * calling jpeg_start_compress, and enough space to write the trailing EOI - * (a few bytes) before calling jpeg_finish_compress. Multipass compression - * modes are not supported at all with suspension, so those two are the only - * points where markers will be written. - */ - -LOCAL(void) -emit_byte (j_compress_ptr cinfo, int val) -/* Emit a byte */ -{ - struct jpeg_destination_mgr * dest = cinfo->dest; - - *(dest->next_output_byte)++ = (JOCTET) val; - if (--dest->free_in_buffer == 0) { - if (! (*dest->empty_output_buffer) (cinfo)) - ERREXIT(cinfo, JERR_CANT_SUSPEND); - } -} - - -LOCAL(void) -emit_marker (j_compress_ptr cinfo, JPEG_MARKER mark) -/* Emit a marker code */ -{ - emit_byte(cinfo, 0xFF); - emit_byte(cinfo, (int) mark); -} - - -LOCAL(void) -emit_2bytes (j_compress_ptr cinfo, int value) -/* Emit a 2-byte integer; these are always MSB first in JPEG files */ -{ - emit_byte(cinfo, (value >> 8) & 0xFF); - emit_byte(cinfo, value & 0xFF); -} - - -/* - * Routines to write specific marker types. - */ - -LOCAL(int) -emit_dqt (j_compress_ptr cinfo, int index) -/* Emit a DQT marker */ -/* Returns the precision used (0 = 8bits, 1 = 16bits) for baseline checking */ -{ - JQUANT_TBL * qtbl = cinfo->quant_tbl_ptrs[index]; - int prec; - int i; - - if (qtbl == NULL) - ERREXIT1(cinfo, JERR_NO_QUANT_TABLE, index); - - prec = 0; - for (i = 0; i <= cinfo->lim_Se; i++) { - if (qtbl->quantval[cinfo->natural_order[i]] > 255) - prec = 1; - } - - if (! qtbl->sent_table) { - emit_marker(cinfo, M_DQT); - - emit_2bytes(cinfo, - prec ? cinfo->lim_Se * 2 + 2 + 1 + 2 : cinfo->lim_Se + 1 + 1 + 2); - - emit_byte(cinfo, index + (prec<<4)); - - for (i = 0; i <= cinfo->lim_Se; i++) { - /* The table entries must be emitted in zigzag order. */ - unsigned int qval = qtbl->quantval[cinfo->natural_order[i]]; - if (prec) - emit_byte(cinfo, (int) (qval >> 8)); - emit_byte(cinfo, (int) (qval & 0xFF)); - } - - qtbl->sent_table = TRUE; - } - - return prec; -} - - -LOCAL(void) -emit_dht (j_compress_ptr cinfo, int index, boolean is_ac) -/* Emit a DHT marker */ -{ - JHUFF_TBL * htbl; - int length, i; - - if (is_ac) { - htbl = cinfo->ac_huff_tbl_ptrs[index]; - index += 0x10; /* output index has AC bit set */ - } else { - htbl = cinfo->dc_huff_tbl_ptrs[index]; - } - - if (htbl == NULL) - ERREXIT1(cinfo, JERR_NO_HUFF_TABLE, index); - - if (! htbl->sent_table) { - emit_marker(cinfo, M_DHT); - - length = 0; - for (i = 1; i <= 16; i++) - length += htbl->bits[i]; - - emit_2bytes(cinfo, length + 2 + 1 + 16); - emit_byte(cinfo, index); - - for (i = 1; i <= 16; i++) - emit_byte(cinfo, htbl->bits[i]); - - for (i = 0; i < length; i++) - emit_byte(cinfo, htbl->huffval[i]); - - htbl->sent_table = TRUE; - } -} - - -LOCAL(void) -emit_dac (j_compress_ptr cinfo) -/* Emit a DAC marker */ -/* Since the useful info is so small, we want to emit all the tables in */ -/* one DAC marker. Therefore this routine does its own scan of the table. */ -{ -#ifdef C_ARITH_CODING_SUPPORTED - char dc_in_use[NUM_ARITH_TBLS]; - char ac_in_use[NUM_ARITH_TBLS]; - int length, i; - jpeg_component_info *compptr; - - for (i = 0; i < NUM_ARITH_TBLS; i++) - dc_in_use[i] = ac_in_use[i] = 0; - - for (i = 0; i < cinfo->comps_in_scan; i++) { - compptr = cinfo->cur_comp_info[i]; - /* DC needs no table for refinement scan */ - if (cinfo->Ss == 0 && cinfo->Ah == 0) - dc_in_use[compptr->dc_tbl_no] = 1; - /* AC needs no table when not present */ - if (cinfo->Se) - ac_in_use[compptr->ac_tbl_no] = 1; - } - - length = 0; - for (i = 0; i < NUM_ARITH_TBLS; i++) - length += dc_in_use[i] + ac_in_use[i]; - - if (length) { - emit_marker(cinfo, M_DAC); - - emit_2bytes(cinfo, length*2 + 2); - - for (i = 0; i < NUM_ARITH_TBLS; i++) { - if (dc_in_use[i]) { - emit_byte(cinfo, i); - emit_byte(cinfo, cinfo->arith_dc_L[i] + (cinfo->arith_dc_U[i]<<4)); - } - if (ac_in_use[i]) { - emit_byte(cinfo, i + 0x10); - emit_byte(cinfo, cinfo->arith_ac_K[i]); - } - } - } -#endif /* C_ARITH_CODING_SUPPORTED */ -} - - -LOCAL(void) -emit_dri (j_compress_ptr cinfo) -/* Emit a DRI marker */ -{ - emit_marker(cinfo, M_DRI); - - emit_2bytes(cinfo, 4); /* fixed length */ - - emit_2bytes(cinfo, (int) cinfo->restart_interval); -} - - -LOCAL(void) -emit_lse_ict (j_compress_ptr cinfo) -/* Emit an LSE inverse color transform specification marker */ -{ - /* Support only 1 transform */ - if (cinfo->color_transform != JCT_SUBTRACT_GREEN || - cinfo->num_components < 3) - ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); - - emit_marker(cinfo, M_JPG8); - - emit_2bytes(cinfo, 24); /* fixed length */ - - emit_byte(cinfo, 0x0D); /* ID inverse transform specification */ - emit_2bytes(cinfo, MAXJSAMPLE); /* MAXTRANS */ - emit_byte(cinfo, 3); /* Nt=3 */ - emit_byte(cinfo, cinfo->comp_info[1].component_id); - emit_byte(cinfo, cinfo->comp_info[0].component_id); - emit_byte(cinfo, cinfo->comp_info[2].component_id); - emit_byte(cinfo, 0x80); /* F1: CENTER1=1, NORM1=0 */ - emit_2bytes(cinfo, 0); /* A(1,1)=0 */ - emit_2bytes(cinfo, 0); /* A(1,2)=0 */ - emit_byte(cinfo, 0); /* F2: CENTER2=0, NORM2=0 */ - emit_2bytes(cinfo, 1); /* A(2,1)=1 */ - emit_2bytes(cinfo, 0); /* A(2,2)=0 */ - emit_byte(cinfo, 0); /* F3: CENTER3=0, NORM3=0 */ - emit_2bytes(cinfo, 1); /* A(3,1)=1 */ - emit_2bytes(cinfo, 0); /* A(3,2)=0 */ -} - - -LOCAL(void) -emit_sof (j_compress_ptr cinfo, JPEG_MARKER code) -/* Emit a SOF marker */ -{ - int ci; - jpeg_component_info *compptr; - - emit_marker(cinfo, code); - - emit_2bytes(cinfo, 3 * cinfo->num_components + 2 + 5 + 1); /* length */ - - /* Make sure image isn't bigger than SOF field can handle */ - if ((long) cinfo->jpeg_height > 65535L || - (long) cinfo->jpeg_width > 65535L) - ERREXIT1(cinfo, JERR_IMAGE_TOO_BIG, (unsigned int) 65535); - - emit_byte(cinfo, cinfo->data_precision); - emit_2bytes(cinfo, (int) cinfo->jpeg_height); - emit_2bytes(cinfo, (int) cinfo->jpeg_width); - - emit_byte(cinfo, cinfo->num_components); - - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - emit_byte(cinfo, compptr->component_id); - emit_byte(cinfo, (compptr->h_samp_factor << 4) + compptr->v_samp_factor); - emit_byte(cinfo, compptr->quant_tbl_no); - } -} - - -LOCAL(void) -emit_sos (j_compress_ptr cinfo) -/* Emit a SOS marker */ -{ - int i, td, ta; - jpeg_component_info *compptr; - - emit_marker(cinfo, M_SOS); - - emit_2bytes(cinfo, 2 * cinfo->comps_in_scan + 2 + 1 + 3); /* length */ - - emit_byte(cinfo, cinfo->comps_in_scan); - - for (i = 0; i < cinfo->comps_in_scan; i++) { - compptr = cinfo->cur_comp_info[i]; - emit_byte(cinfo, compptr->component_id); - - /* We emit 0 for unused field(s); this is recommended by the P&M text - * but does not seem to be specified in the standard. - */ - - /* DC needs no table for refinement scan */ - td = cinfo->Ss == 0 && cinfo->Ah == 0 ? compptr->dc_tbl_no : 0; - /* AC needs no table when not present */ - ta = cinfo->Se ? compptr->ac_tbl_no : 0; - - emit_byte(cinfo, (td << 4) + ta); - } - - emit_byte(cinfo, cinfo->Ss); - emit_byte(cinfo, cinfo->Se); - emit_byte(cinfo, (cinfo->Ah << 4) + cinfo->Al); -} - - -LOCAL(void) -emit_pseudo_sos (j_compress_ptr cinfo) -/* Emit a pseudo SOS marker */ -{ - emit_marker(cinfo, M_SOS); - - emit_2bytes(cinfo, 2 + 1 + 3); /* length */ - - emit_byte(cinfo, 0); /* Ns */ - - emit_byte(cinfo, 0); /* Ss */ - emit_byte(cinfo, cinfo->block_size * cinfo->block_size - 1); /* Se */ - emit_byte(cinfo, 0); /* Ah/Al */ -} - - -LOCAL(void) -emit_jfif_app0 (j_compress_ptr cinfo) -/* Emit a JFIF-compliant APP0 marker */ -{ - /* - * Length of APP0 block (2 bytes) - * Block ID (4 bytes - ASCII "JFIF") - * Zero byte (1 byte to terminate the ID string) - * Version Major, Minor (2 bytes - major first) - * Units (1 byte - 0x00 = none, 0x01 = inch, 0x02 = cm) - * Xdpu (2 bytes - dots per unit horizontal) - * Ydpu (2 bytes - dots per unit vertical) - * Thumbnail X size (1 byte) - * Thumbnail Y size (1 byte) - */ - - emit_marker(cinfo, M_APP0); - - emit_2bytes(cinfo, 2 + 4 + 1 + 2 + 1 + 2 + 2 + 1 + 1); /* length */ - - emit_byte(cinfo, 0x4A); /* Identifier: ASCII "JFIF" */ - emit_byte(cinfo, 0x46); - emit_byte(cinfo, 0x49); - emit_byte(cinfo, 0x46); - emit_byte(cinfo, 0); - emit_byte(cinfo, cinfo->JFIF_major_version); /* Version fields */ - emit_byte(cinfo, cinfo->JFIF_minor_version); - emit_byte(cinfo, cinfo->density_unit); /* Pixel size information */ - emit_2bytes(cinfo, (int) cinfo->X_density); - emit_2bytes(cinfo, (int) cinfo->Y_density); - emit_byte(cinfo, 0); /* No thumbnail image */ - emit_byte(cinfo, 0); -} - - -LOCAL(void) -emit_adobe_app14 (j_compress_ptr cinfo) -/* Emit an Adobe APP14 marker */ -{ - /* - * Length of APP14 block (2 bytes) - * Block ID (5 bytes - ASCII "Adobe") - * Version Number (2 bytes - currently 100) - * Flags0 (2 bytes - currently 0) - * Flags1 (2 bytes - currently 0) - * Color transform (1 byte) - * - * Although Adobe TN 5116 mentions Version = 101, all the Adobe files - * now in circulation seem to use Version = 100, so that's what we write. - * - * We write the color transform byte as 1 if the JPEG color space is - * YCbCr, 2 if it's YCCK, 0 otherwise. Adobe's definition has to do with - * whether the encoder performed a transformation, which is pretty useless. - */ - - emit_marker(cinfo, M_APP14); - - emit_2bytes(cinfo, 2 + 5 + 2 + 2 + 2 + 1); /* length */ - - emit_byte(cinfo, 0x41); /* Identifier: ASCII "Adobe" */ - emit_byte(cinfo, 0x64); - emit_byte(cinfo, 0x6F); - emit_byte(cinfo, 0x62); - emit_byte(cinfo, 0x65); - emit_2bytes(cinfo, 100); /* Version */ - emit_2bytes(cinfo, 0); /* Flags0 */ - emit_2bytes(cinfo, 0); /* Flags1 */ - switch (cinfo->jpeg_color_space) { - case JCS_YCbCr: - emit_byte(cinfo, 1); /* Color transform = 1 */ - break; - case JCS_YCCK: - emit_byte(cinfo, 2); /* Color transform = 2 */ - break; - default: - emit_byte(cinfo, 0); /* Color transform = 0 */ - } -} - - -/* - * These routines allow writing an arbitrary marker with parameters. - * The only intended use is to emit COM or APPn markers after calling - * write_file_header and before calling write_frame_header. - * Other uses are not guaranteed to produce desirable results. - * Counting the parameter bytes properly is the caller's responsibility. - */ - -METHODDEF(void) -write_marker_header (j_compress_ptr cinfo, int marker, unsigned int datalen) -/* Emit an arbitrary marker header */ -{ - if (datalen > (unsigned int) 65533) /* safety check */ - ERREXIT(cinfo, JERR_BAD_LENGTH); - - emit_marker(cinfo, (JPEG_MARKER) marker); - - emit_2bytes(cinfo, (int) (datalen + 2)); /* total length */ -} - -METHODDEF(void) -write_marker_byte (j_compress_ptr cinfo, int val) -/* Emit one byte of marker parameters following write_marker_header */ -{ - emit_byte(cinfo, val); -} - - -/* - * Write datastream header. - * This consists of an SOI and optional APPn markers. - * We recommend use of the JFIF marker, but not the Adobe marker, - * when using YCbCr or grayscale data. The JFIF marker is also used - * for other standard JPEG colorspaces. The Adobe marker is helpful - * to distinguish RGB, CMYK, and YCCK colorspaces. - * Note that an application can write additional header markers after - * jpeg_start_compress returns. - */ - -METHODDEF(void) -write_file_header (j_compress_ptr cinfo) -{ - my_marker_ptr marker = (my_marker_ptr) cinfo->marker; - - emit_marker(cinfo, M_SOI); /* first the SOI */ - - /* SOI is defined to reset restart interval to 0 */ - marker->last_restart_interval = 0; - - if (cinfo->write_JFIF_header) /* next an optional JFIF APP0 */ - emit_jfif_app0(cinfo); - if (cinfo->write_Adobe_marker) /* next an optional Adobe APP14 */ - emit_adobe_app14(cinfo); -} - - -/* - * Write frame header. - * This consists of DQT and SOFn markers, - * a conditional LSE marker and a conditional pseudo SOS marker. - * Note that we do not emit the SOF until we have emitted the DQT(s). - * This avoids compatibility problems with incorrect implementations that - * try to error-check the quant table numbers as soon as they see the SOF. - */ - -METHODDEF(void) -write_frame_header (j_compress_ptr cinfo) -{ - int ci, prec; - boolean is_baseline; - jpeg_component_info *compptr; - - /* Emit DQT for each quantization table. - * Note that emit_dqt() suppresses any duplicate tables. - */ - prec = 0; - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - prec += emit_dqt(cinfo, compptr->quant_tbl_no); - } - /* now prec is nonzero iff there are any 16-bit quant tables. */ - - /* Check for a non-baseline specification. - * Note we assume that Huffman table numbers won't be changed later. - */ - if (cinfo->arith_code || cinfo->progressive_mode || - cinfo->data_precision != 8 || cinfo->block_size != DCTSIZE) { - is_baseline = FALSE; - } else { - is_baseline = TRUE; - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - if (compptr->dc_tbl_no > 1 || compptr->ac_tbl_no > 1) - is_baseline = FALSE; - } - if (prec && is_baseline) { - is_baseline = FALSE; - /* If it's baseline except for quantizer size, warn the user */ - TRACEMS(cinfo, 0, JTRC_16BIT_TABLES); - } - } - - /* Emit the proper SOF marker */ - if (cinfo->arith_code) { - if (cinfo->progressive_mode) - emit_sof(cinfo, M_SOF10); /* SOF code for progressive arithmetic */ - else - emit_sof(cinfo, M_SOF9); /* SOF code for sequential arithmetic */ - } else { - if (cinfo->progressive_mode) - emit_sof(cinfo, M_SOF2); /* SOF code for progressive Huffman */ - else if (is_baseline) - emit_sof(cinfo, M_SOF0); /* SOF code for baseline implementation */ - else - emit_sof(cinfo, M_SOF1); /* SOF code for non-baseline Huffman file */ - } - - /* Check to emit LSE inverse color transform specification marker */ - if (cinfo->color_transform) - emit_lse_ict(cinfo); - - /* Check to emit pseudo SOS marker */ - if (cinfo->progressive_mode && cinfo->block_size != DCTSIZE) - emit_pseudo_sos(cinfo); -} - - -/* - * Write scan header. - * This consists of DHT or DAC markers, optional DRI, and SOS. - * Compressed data will be written following the SOS. - */ - -METHODDEF(void) -write_scan_header (j_compress_ptr cinfo) -{ - my_marker_ptr marker = (my_marker_ptr) cinfo->marker; - int i; - jpeg_component_info *compptr; - - if (cinfo->arith_code) { - /* Emit arith conditioning info. We may have some duplication - * if the file has multiple scans, but it's so small it's hardly - * worth worrying about. - */ - emit_dac(cinfo); - } else { - /* Emit Huffman tables. - * Note that emit_dht() suppresses any duplicate tables. - */ - for (i = 0; i < cinfo->comps_in_scan; i++) { - compptr = cinfo->cur_comp_info[i]; - /* DC needs no table for refinement scan */ - if (cinfo->Ss == 0 && cinfo->Ah == 0) - emit_dht(cinfo, compptr->dc_tbl_no, FALSE); - /* AC needs no table when not present */ - if (cinfo->Se) - emit_dht(cinfo, compptr->ac_tbl_no, TRUE); - } - } - - /* Emit DRI if required --- note that DRI value could change for each scan. - * We avoid wasting space with unnecessary DRIs, however. - */ - if (cinfo->restart_interval != marker->last_restart_interval) { - emit_dri(cinfo); - marker->last_restart_interval = cinfo->restart_interval; - } - - emit_sos(cinfo); -} - - -/* - * Write datastream trailer. - */ - -METHODDEF(void) -write_file_trailer (j_compress_ptr cinfo) -{ - emit_marker(cinfo, M_EOI); -} - - -/* - * Write an abbreviated table-specification datastream. - * This consists of SOI, DQT and DHT tables, and EOI. - * Any table that is defined and not marked sent_table = TRUE will be - * emitted. Note that all tables will be marked sent_table = TRUE at exit. - */ - -METHODDEF(void) -write_tables_only (j_compress_ptr cinfo) -{ - int i; - - emit_marker(cinfo, M_SOI); - - for (i = 0; i < NUM_QUANT_TBLS; i++) { - if (cinfo->quant_tbl_ptrs[i] != NULL) - (void) emit_dqt(cinfo, i); - } - - if (! cinfo->arith_code) { - for (i = 0; i < NUM_HUFF_TBLS; i++) { - if (cinfo->dc_huff_tbl_ptrs[i] != NULL) - emit_dht(cinfo, i, FALSE); - if (cinfo->ac_huff_tbl_ptrs[i] != NULL) - emit_dht(cinfo, i, TRUE); - } - } - - emit_marker(cinfo, M_EOI); -} - - -/* - * Initialize the marker writer module. - */ - -GLOBAL(void) -jinit_marker_writer (j_compress_ptr cinfo) -{ - my_marker_ptr marker; - - /* Create the subobject */ - marker = (my_marker_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_marker_writer)); - cinfo->marker = &marker->pub; - /* Initialize method pointers */ - marker->pub.write_file_header = write_file_header; - marker->pub.write_frame_header = write_frame_header; - marker->pub.write_scan_header = write_scan_header; - marker->pub.write_file_trailer = write_file_trailer; - marker->pub.write_tables_only = write_tables_only; - marker->pub.write_marker_header = write_marker_header; - marker->pub.write_marker_byte = write_marker_byte; - /* Initialize private state */ - marker->last_restart_interval = 0; -} diff --git a/cmake/externals/libjpeg/jcmaster.c b/cmake/externals/libjpeg/jcmaster.c deleted file mode 100644 index a70af0c02..000000000 --- a/cmake/externals/libjpeg/jcmaster.c +++ /dev/null @@ -1,675 +0,0 @@ -/* - * jcmaster.c - * - * Copyright (C) 1991-1997, Thomas G. Lane. - * Modified 2003-2020 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains master control logic for the JPEG compressor. - * These routines are concerned with parameter validation, initial setup, - * and inter-pass control (determining the number of passes and the work - * to be done in each pass). - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* Private state */ - -typedef enum { - main_pass, /* input data, also do first output step */ - huff_opt_pass, /* Huffman code optimization pass */ - output_pass /* data output pass */ -} c_pass_type; - -typedef struct { - struct jpeg_comp_master pub; /* public fields */ - - c_pass_type pass_type; /* the type of the current pass */ - - int pass_number; /* # of passes completed */ - int total_passes; /* total # of passes needed */ - - int scan_number; /* current index in scan_info[] */ -} my_comp_master; - -typedef my_comp_master * my_master_ptr; - - -/* - * Support routines that do various essential calculations. - */ - -LOCAL(void) -initial_setup (j_compress_ptr cinfo) -/* Do computations that are needed before master selection phase */ -{ - int ci, ssize; - jpeg_component_info *compptr; - - /* Sanity check on block_size */ - if (cinfo->block_size < 1 || cinfo->block_size > 16) - ERREXIT2(cinfo, JERR_BAD_DCTSIZE, cinfo->block_size, cinfo->block_size); - - /* Derive natural_order from block_size */ - switch (cinfo->block_size) { - case 2: cinfo->natural_order = jpeg_natural_order2; break; - case 3: cinfo->natural_order = jpeg_natural_order3; break; - case 4: cinfo->natural_order = jpeg_natural_order4; break; - case 5: cinfo->natural_order = jpeg_natural_order5; break; - case 6: cinfo->natural_order = jpeg_natural_order6; break; - case 7: cinfo->natural_order = jpeg_natural_order7; break; - default: cinfo->natural_order = jpeg_natural_order; - } - - /* Derive lim_Se from block_size */ - cinfo->lim_Se = cinfo->block_size < DCTSIZE ? - cinfo->block_size * cinfo->block_size - 1 : DCTSIZE2-1; - - /* Sanity check on image dimensions */ - if (cinfo->jpeg_height <= 0 || cinfo->jpeg_width <= 0 || - cinfo->num_components <= 0) - ERREXIT(cinfo, JERR_EMPTY_IMAGE); - - /* Make sure image isn't bigger than I can handle */ - if ((long) cinfo->jpeg_height > (long) JPEG_MAX_DIMENSION || - (long) cinfo->jpeg_width > (long) JPEG_MAX_DIMENSION) - ERREXIT1(cinfo, JERR_IMAGE_TOO_BIG, (unsigned int) JPEG_MAX_DIMENSION); - - /* Only 8 to 12 bits data precision are supported for DCT based JPEG */ - if (cinfo->data_precision < 8 || cinfo->data_precision > 12) - ERREXIT1(cinfo, JERR_BAD_PRECISION, cinfo->data_precision); - - /* Check that number of components won't exceed internal array sizes */ - if (cinfo->num_components > MAX_COMPONENTS) - ERREXIT2(cinfo, JERR_COMPONENT_COUNT, cinfo->num_components, - MAX_COMPONENTS); - - /* Compute maximum sampling factors; check factor validity */ - cinfo->max_h_samp_factor = 1; - cinfo->max_v_samp_factor = 1; - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - if (compptr->h_samp_factor<=0 || compptr->h_samp_factor>MAX_SAMP_FACTOR || - compptr->v_samp_factor<=0 || compptr->v_samp_factor>MAX_SAMP_FACTOR) - ERREXIT(cinfo, JERR_BAD_SAMPLING); - cinfo->max_h_samp_factor = MAX(cinfo->max_h_samp_factor, - compptr->h_samp_factor); - cinfo->max_v_samp_factor = MAX(cinfo->max_v_samp_factor, - compptr->v_samp_factor); - } - - /* Compute dimensions of components */ - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - /* Fill in the correct component_index value; don't rely on application */ - compptr->component_index = ci; - /* In selecting the actual DCT scaling for each component, we try to - * scale down the chroma components via DCT scaling rather than downsampling. - * This saves time if the downsampler gets to use 1:1 scaling. - * Note this code adapts subsampling ratios which are powers of 2. - */ - ssize = 1; -#ifdef DCT_SCALING_SUPPORTED - if (! cinfo->raw_data_in) - while (cinfo->min_DCT_h_scaled_size * ssize <= - (cinfo->do_fancy_downsampling ? DCTSIZE : DCTSIZE / 2) && - (cinfo->max_h_samp_factor % (compptr->h_samp_factor * ssize * 2)) == - 0) { - ssize = ssize * 2; - } -#endif - compptr->DCT_h_scaled_size = cinfo->min_DCT_h_scaled_size * ssize; - ssize = 1; -#ifdef DCT_SCALING_SUPPORTED - if (! cinfo->raw_data_in) - while (cinfo->min_DCT_v_scaled_size * ssize <= - (cinfo->do_fancy_downsampling ? DCTSIZE : DCTSIZE / 2) && - (cinfo->max_v_samp_factor % (compptr->v_samp_factor * ssize * 2)) == - 0) { - ssize = ssize * 2; - } -#endif - compptr->DCT_v_scaled_size = cinfo->min_DCT_v_scaled_size * ssize; - - /* We don't support DCT ratios larger than 2. */ - if (compptr->DCT_h_scaled_size > compptr->DCT_v_scaled_size * 2) - compptr->DCT_h_scaled_size = compptr->DCT_v_scaled_size * 2; - else if (compptr->DCT_v_scaled_size > compptr->DCT_h_scaled_size * 2) - compptr->DCT_v_scaled_size = compptr->DCT_h_scaled_size * 2; - - /* Size in DCT blocks */ - compptr->width_in_blocks = (JDIMENSION) - jdiv_round_up((long) cinfo->jpeg_width * (long) compptr->h_samp_factor, - (long) (cinfo->max_h_samp_factor * cinfo->block_size)); - compptr->height_in_blocks = (JDIMENSION) - jdiv_round_up((long) cinfo->jpeg_height * (long) compptr->v_samp_factor, - (long) (cinfo->max_v_samp_factor * cinfo->block_size)); - /* Size in samples */ - compptr->downsampled_width = (JDIMENSION) - jdiv_round_up((long) cinfo->jpeg_width * - (long) (compptr->h_samp_factor * compptr->DCT_h_scaled_size), - (long) (cinfo->max_h_samp_factor * cinfo->block_size)); - compptr->downsampled_height = (JDIMENSION) - jdiv_round_up((long) cinfo->jpeg_height * - (long) (compptr->v_samp_factor * compptr->DCT_v_scaled_size), - (long) (cinfo->max_v_samp_factor * cinfo->block_size)); - /* Don't need quantization scale after DCT, - * until color conversion says otherwise. - */ - compptr->component_needed = FALSE; - } - - /* Compute number of fully interleaved MCU rows (number of times that - * main controller will call coefficient controller). - */ - cinfo->total_iMCU_rows = (JDIMENSION) - jdiv_round_up((long) cinfo->jpeg_height, - (long) (cinfo->max_v_samp_factor * cinfo->block_size)); -} - - -#ifdef C_MULTISCAN_FILES_SUPPORTED - -LOCAL(void) -validate_script (j_compress_ptr cinfo) -/* Verify that the scan script in cinfo->scan_info[] is valid; also - * determine whether it uses progressive JPEG, and set cinfo->progressive_mode. - */ -{ - const jpeg_scan_info * scanptr; - int scanno, ncomps, ci, coefi, thisi; - int Ss, Se, Ah, Al; - boolean component_sent[MAX_COMPONENTS]; -#ifdef C_PROGRESSIVE_SUPPORTED - int * last_bitpos_ptr; - int last_bitpos[MAX_COMPONENTS][DCTSIZE2]; - /* -1 until that coefficient has been seen; then last Al for it */ -#endif - - if (cinfo->num_scans <= 0) - ERREXIT1(cinfo, JERR_BAD_SCAN_SCRIPT, 0); - - /* For sequential JPEG, all scans must have Ss=0, Se=DCTSIZE2-1; - * for progressive JPEG, no scan can have this. - */ - scanptr = cinfo->scan_info; - if (scanptr->Ss != 0 || scanptr->Se != DCTSIZE2-1) { -#ifdef C_PROGRESSIVE_SUPPORTED - cinfo->progressive_mode = TRUE; - last_bitpos_ptr = & last_bitpos[0][0]; - for (ci = 0; ci < cinfo->num_components; ci++) - for (coefi = 0; coefi < DCTSIZE2; coefi++) - *last_bitpos_ptr++ = -1; -#else - ERREXIT(cinfo, JERR_NOT_COMPILED); -#endif - } else { - cinfo->progressive_mode = FALSE; - for (ci = 0; ci < cinfo->num_components; ci++) - component_sent[ci] = FALSE; - } - - for (scanno = 1; scanno <= cinfo->num_scans; scanptr++, scanno++) { - /* Validate component indexes */ - ncomps = scanptr->comps_in_scan; - if (ncomps <= 0 || ncomps > MAX_COMPS_IN_SCAN) - ERREXIT2(cinfo, JERR_COMPONENT_COUNT, ncomps, MAX_COMPS_IN_SCAN); - for (ci = 0; ci < ncomps; ci++) { - thisi = scanptr->component_index[ci]; - if (thisi < 0 || thisi >= cinfo->num_components) - ERREXIT1(cinfo, JERR_BAD_SCAN_SCRIPT, scanno); - /* Components must appear in SOF order within each scan */ - if (ci > 0 && thisi <= scanptr->component_index[ci-1]) - ERREXIT1(cinfo, JERR_BAD_SCAN_SCRIPT, scanno); - } - /* Validate progression parameters */ - Ss = scanptr->Ss; - Se = scanptr->Se; - Ah = scanptr->Ah; - Al = scanptr->Al; - if (cinfo->progressive_mode) { -#ifdef C_PROGRESSIVE_SUPPORTED - /* The JPEG spec simply gives the ranges 0..13 for Ah and Al, but that - * seems wrong: the upper bound ought to depend on data precision. - * Perhaps they really meant 0..N+1 for N-bit precision. - * Here we allow 0..10 for 8-bit data; Al larger than 10 results in - * out-of-range reconstructed DC values during the first DC scan, - * which might cause problems for some decoders. - */ - if (Ss < 0 || Ss >= DCTSIZE2 || Se < Ss || Se >= DCTSIZE2 || - Ah < 0 || Ah > (cinfo->data_precision > 8 ? 13 : 10) || - Al < 0 || Al > (cinfo->data_precision > 8 ? 13 : 10)) - ERREXIT1(cinfo, JERR_BAD_PROG_SCRIPT, scanno); - if (Ss == 0) { - if (Se != 0) /* DC and AC together not OK */ - ERREXIT1(cinfo, JERR_BAD_PROG_SCRIPT, scanno); - } else { - if (ncomps != 1) /* AC scans must be for only one component */ - ERREXIT1(cinfo, JERR_BAD_PROG_SCRIPT, scanno); - } - for (ci = 0; ci < ncomps; ci++) { - last_bitpos_ptr = & last_bitpos[scanptr->component_index[ci]][0]; - if (Ss != 0 && last_bitpos_ptr[0] < 0) /* AC without prior DC scan */ - ERREXIT1(cinfo, JERR_BAD_PROG_SCRIPT, scanno); - for (coefi = Ss; coefi <= Se; coefi++) { - if (last_bitpos_ptr[coefi] < 0) { - /* first scan of this coefficient */ - if (Ah != 0) - ERREXIT1(cinfo, JERR_BAD_PROG_SCRIPT, scanno); - } else { - /* not first scan */ - if (Ah != last_bitpos_ptr[coefi] || Al != Ah-1) - ERREXIT1(cinfo, JERR_BAD_PROG_SCRIPT, scanno); - } - last_bitpos_ptr[coefi] = Al; - } - } -#endif - } else { - /* For sequential JPEG, all progression parameters must be these: */ - if (Ss != 0 || Se != DCTSIZE2-1 || Ah != 0 || Al != 0) - ERREXIT1(cinfo, JERR_BAD_PROG_SCRIPT, scanno); - /* Make sure components are not sent twice */ - for (ci = 0; ci < ncomps; ci++) { - thisi = scanptr->component_index[ci]; - if (component_sent[thisi]) - ERREXIT1(cinfo, JERR_BAD_SCAN_SCRIPT, scanno); - component_sent[thisi] = TRUE; - } - } - } - - /* Now verify that everything got sent. */ - if (cinfo->progressive_mode) { -#ifdef C_PROGRESSIVE_SUPPORTED - /* For progressive mode, we only check that at least some DC data - * got sent for each component; the spec does not require that all bits - * of all coefficients be transmitted. Would it be wiser to enforce - * transmission of all coefficient bits?? - */ - for (ci = 0; ci < cinfo->num_components; ci++) { - if (last_bitpos[ci][0] < 0) - ERREXIT(cinfo, JERR_MISSING_DATA); - } -#endif - } else { - for (ci = 0; ci < cinfo->num_components; ci++) { - if (! component_sent[ci]) - ERREXIT(cinfo, JERR_MISSING_DATA); - } - } -} - - -LOCAL(void) -reduce_script (j_compress_ptr cinfo) -/* Adapt scan script for use with reduced block size; - * assume that script has been validated before. - */ -{ - jpeg_scan_info * scanptr; - int idxout, idxin; - - /* Circumvent const declaration for this function */ - scanptr = (jpeg_scan_info *) cinfo->scan_info; - idxout = 0; - - for (idxin = 0; idxin < cinfo->num_scans; idxin++) { - /* After skipping, idxout becomes smaller than idxin */ - if (idxin != idxout) - /* Copy rest of data; - * note we stay in given chunk of allocated memory. - */ - scanptr[idxout] = scanptr[idxin]; - if (scanptr[idxout].Ss > cinfo->lim_Se) - /* Entire scan out of range - skip this entry */ - continue; - if (scanptr[idxout].Se > cinfo->lim_Se) - /* Limit scan to end of block */ - scanptr[idxout].Se = cinfo->lim_Se; - idxout++; - } - - cinfo->num_scans = idxout; -} - -#endif /* C_MULTISCAN_FILES_SUPPORTED */ - - -LOCAL(void) -select_scan_parameters (j_compress_ptr cinfo) -/* Set up the scan parameters for the current scan */ -{ - int ci; - -#ifdef C_MULTISCAN_FILES_SUPPORTED - if (cinfo->scan_info != NULL) { - /* Prepare for current scan --- the script is already validated */ - my_master_ptr master = (my_master_ptr) cinfo->master; - const jpeg_scan_info * scanptr = cinfo->scan_info + master->scan_number; - - cinfo->comps_in_scan = scanptr->comps_in_scan; - for (ci = 0; ci < scanptr->comps_in_scan; ci++) { - cinfo->cur_comp_info[ci] = - &cinfo->comp_info[scanptr->component_index[ci]]; - } - if (cinfo->progressive_mode) { - cinfo->Ss = scanptr->Ss; - cinfo->Se = scanptr->Se; - cinfo->Ah = scanptr->Ah; - cinfo->Al = scanptr->Al; - return; - } - } - else -#endif - { - /* Prepare for single sequential-JPEG scan containing all components */ - if (cinfo->num_components > MAX_COMPS_IN_SCAN) - ERREXIT2(cinfo, JERR_COMPONENT_COUNT, cinfo->num_components, - MAX_COMPS_IN_SCAN); - cinfo->comps_in_scan = cinfo->num_components; - for (ci = 0; ci < cinfo->num_components; ci++) { - cinfo->cur_comp_info[ci] = &cinfo->comp_info[ci]; - } - } - cinfo->Ss = 0; - cinfo->Se = cinfo->block_size * cinfo->block_size - 1; - cinfo->Ah = 0; - cinfo->Al = 0; -} - - -LOCAL(void) -per_scan_setup (j_compress_ptr cinfo) -/* Do computations that are needed before processing a JPEG scan */ -/* cinfo->comps_in_scan and cinfo->cur_comp_info[] are already set */ -{ - int ci, mcublks, tmp; - jpeg_component_info *compptr; - - if (cinfo->comps_in_scan == 1) { - - /* Noninterleaved (single-component) scan */ - compptr = cinfo->cur_comp_info[0]; - - /* Overall image size in MCUs */ - cinfo->MCUs_per_row = compptr->width_in_blocks; - cinfo->MCU_rows_in_scan = compptr->height_in_blocks; - - /* For noninterleaved scan, always one block per MCU */ - compptr->MCU_width = 1; - compptr->MCU_height = 1; - compptr->MCU_blocks = 1; - compptr->MCU_sample_width = compptr->DCT_h_scaled_size; - compptr->last_col_width = 1; - /* For noninterleaved scans, it is convenient to define last_row_height - * as the number of block rows present in the last iMCU row. - */ - tmp = (int) (compptr->height_in_blocks % compptr->v_samp_factor); - if (tmp == 0) tmp = compptr->v_samp_factor; - compptr->last_row_height = tmp; - - /* Prepare array describing MCU composition */ - cinfo->blocks_in_MCU = 1; - cinfo->MCU_membership[0] = 0; - - } else { - - /* Interleaved (multi-component) scan */ - if (cinfo->comps_in_scan <= 0 || cinfo->comps_in_scan > MAX_COMPS_IN_SCAN) - ERREXIT2(cinfo, JERR_COMPONENT_COUNT, cinfo->comps_in_scan, - MAX_COMPS_IN_SCAN); - - /* Overall image size in MCUs */ - cinfo->MCUs_per_row = (JDIMENSION) - jdiv_round_up((long) cinfo->jpeg_width, - (long) (cinfo->max_h_samp_factor * cinfo->block_size)); - cinfo->MCU_rows_in_scan = cinfo->total_iMCU_rows; - - cinfo->blocks_in_MCU = 0; - - for (ci = 0; ci < cinfo->comps_in_scan; ci++) { - compptr = cinfo->cur_comp_info[ci]; - /* Sampling factors give # of blocks of component in each MCU */ - compptr->MCU_width = compptr->h_samp_factor; - compptr->MCU_height = compptr->v_samp_factor; - compptr->MCU_blocks = compptr->MCU_width * compptr->MCU_height; - compptr->MCU_sample_width = compptr->MCU_width * compptr->DCT_h_scaled_size; - /* Figure number of non-dummy blocks in last MCU column & row */ - tmp = (int) (compptr->width_in_blocks % compptr->MCU_width); - if (tmp == 0) tmp = compptr->MCU_width; - compptr->last_col_width = tmp; - tmp = (int) (compptr->height_in_blocks % compptr->MCU_height); - if (tmp == 0) tmp = compptr->MCU_height; - compptr->last_row_height = tmp; - /* Prepare array describing MCU composition */ - mcublks = compptr->MCU_blocks; - if (cinfo->blocks_in_MCU + mcublks > C_MAX_BLOCKS_IN_MCU) - ERREXIT(cinfo, JERR_BAD_MCU_SIZE); - while (mcublks-- > 0) { - cinfo->MCU_membership[cinfo->blocks_in_MCU++] = ci; - } - } - - } - - /* Convert restart specified in rows to actual MCU count. */ - /* Note that count must fit in 16 bits, so we provide limiting. */ - if (cinfo->restart_in_rows > 0) { - long nominal = (long) cinfo->restart_in_rows * (long) cinfo->MCUs_per_row; - cinfo->restart_interval = (unsigned int) MIN(nominal, 65535L); - } -} - - -/* - * Per-pass setup. - * This is called at the beginning of each pass. We determine which modules - * will be active during this pass and give them appropriate start_pass calls. - * We also set is_last_pass to indicate whether any more passes will be - * required. - */ - -METHODDEF(void) -prepare_for_pass (j_compress_ptr cinfo) -{ - my_master_ptr master = (my_master_ptr) cinfo->master; - - switch (master->pass_type) { - case main_pass: - /* Initial pass: will collect input data, and do either Huffman - * optimization or data output for the first scan. - */ - select_scan_parameters(cinfo); - per_scan_setup(cinfo); - if (! cinfo->raw_data_in) { - (*cinfo->cconvert->start_pass) (cinfo); - (*cinfo->downsample->start_pass) (cinfo); - (*cinfo->prep->start_pass) (cinfo, JBUF_PASS_THRU); - } - (*cinfo->fdct->start_pass) (cinfo); - (*cinfo->entropy->start_pass) (cinfo, cinfo->optimize_coding); - (*cinfo->coef->start_pass) (cinfo, - (master->total_passes > 1 ? - JBUF_SAVE_AND_PASS : JBUF_PASS_THRU)); - (*cinfo->main->start_pass) (cinfo, JBUF_PASS_THRU); - if (cinfo->optimize_coding) { - /* No immediate data output; postpone writing frame/scan headers */ - master->pub.call_pass_startup = FALSE; - } else { - /* Will write frame/scan headers at first jpeg_write_scanlines call */ - master->pub.call_pass_startup = TRUE; - } - break; -#ifdef ENTROPY_OPT_SUPPORTED - case huff_opt_pass: - /* Do Huffman optimization for a scan after the first one. */ - select_scan_parameters(cinfo); - per_scan_setup(cinfo); - if (cinfo->Ss != 0 || cinfo->Ah == 0) { - (*cinfo->entropy->start_pass) (cinfo, TRUE); - (*cinfo->coef->start_pass) (cinfo, JBUF_CRANK_DEST); - master->pub.call_pass_startup = FALSE; - break; - } - /* Special case: Huffman DC refinement scans need no Huffman table - * and therefore we can skip the optimization pass for them. - */ - master->pass_type = output_pass; - master->pass_number++; - /*FALLTHROUGH*/ -#endif - case output_pass: - /* Do a data-output pass. */ - /* We need not repeat per-scan setup if prior optimization pass did it. */ - if (! cinfo->optimize_coding) { - select_scan_parameters(cinfo); - per_scan_setup(cinfo); - } - (*cinfo->entropy->start_pass) (cinfo, FALSE); - (*cinfo->coef->start_pass) (cinfo, JBUF_CRANK_DEST); - /* We emit frame/scan headers now */ - if (master->scan_number == 0) - (*cinfo->marker->write_frame_header) (cinfo); - (*cinfo->marker->write_scan_header) (cinfo); - master->pub.call_pass_startup = FALSE; - break; - default: - ERREXIT(cinfo, JERR_NOT_COMPILED); - } - - master->pub.is_last_pass = (master->pass_number == master->total_passes-1); - - /* Set up progress monitor's pass info if present */ - if (cinfo->progress != NULL) { - cinfo->progress->completed_passes = master->pass_number; - cinfo->progress->total_passes = master->total_passes; - } -} - - -/* - * Special start-of-pass hook. - * This is called by jpeg_write_scanlines if call_pass_startup is TRUE. - * In single-pass processing, we need this hook because we don't want to - * write frame/scan headers during jpeg_start_compress; we want to let the - * application write COM markers etc. between jpeg_start_compress and the - * jpeg_write_scanlines loop. - * In multi-pass processing, this routine is not used. - */ - -METHODDEF(void) -pass_startup (j_compress_ptr cinfo) -{ - cinfo->master->call_pass_startup = FALSE; /* reset flag so call only once */ - - (*cinfo->marker->write_frame_header) (cinfo); - (*cinfo->marker->write_scan_header) (cinfo); -} - - -/* - * Finish up at end of pass. - */ - -METHODDEF(void) -finish_pass_master (j_compress_ptr cinfo) -{ - my_master_ptr master = (my_master_ptr) cinfo->master; - - /* The entropy coder always needs an end-of-pass call, - * either to analyze statistics or to flush its output buffer. - */ - (*cinfo->entropy->finish_pass) (cinfo); - - /* Update state for next pass */ - switch (master->pass_type) { - case main_pass: - /* next pass is either output of scan 0 (after optimization) - * or output of scan 1 (if no optimization). - */ - master->pass_type = output_pass; - if (! cinfo->optimize_coding) - master->scan_number++; - break; - case huff_opt_pass: - /* next pass is always output of current scan */ - master->pass_type = output_pass; - break; - case output_pass: - /* next pass is either optimization or output of next scan */ - if (cinfo->optimize_coding) - master->pass_type = huff_opt_pass; - master->scan_number++; - break; - } - - master->pass_number++; -} - - -/* - * Initialize master compression control. - */ - -GLOBAL(void) -jinit_c_master_control (j_compress_ptr cinfo, boolean transcode_only) -{ - my_master_ptr master; - - master = (my_master_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_comp_master)); - cinfo->master = &master->pub; - master->pub.prepare_for_pass = prepare_for_pass; - master->pub.pass_startup = pass_startup; - master->pub.finish_pass = finish_pass_master; - master->pub.is_last_pass = FALSE; - - /* Validate parameters, determine derived values */ - initial_setup(cinfo); - - if (cinfo->scan_info != NULL) { -#ifdef C_MULTISCAN_FILES_SUPPORTED - validate_script(cinfo); - if (cinfo->block_size < DCTSIZE) - reduce_script(cinfo); -#else - ERREXIT(cinfo, JERR_NOT_COMPILED); -#endif - } else { - cinfo->progressive_mode = FALSE; - cinfo->num_scans = 1; - } - - if (cinfo->optimize_coding) - cinfo->arith_code = FALSE; /* disable arithmetic coding */ - else if (! cinfo->arith_code && - (cinfo->progressive_mode || - (cinfo->block_size > 1 && cinfo->block_size < DCTSIZE))) - /* TEMPORARY HACK ??? */ - /* assume default tables no good for progressive or reduced AC mode */ - cinfo->optimize_coding = TRUE; /* force Huffman optimization */ - - /* Initialize my private state */ - if (transcode_only) { - /* no main pass in transcoding */ - if (cinfo->optimize_coding) - master->pass_type = huff_opt_pass; - else - master->pass_type = output_pass; - } else { - /* for normal compression, first pass is always this type: */ - master->pass_type = main_pass; - } - master->scan_number = 0; - master->pass_number = 0; - if (cinfo->optimize_coding) - master->total_passes = cinfo->num_scans * 2; - else - master->total_passes = cinfo->num_scans; -} diff --git a/cmake/externals/libjpeg/jcomapi.c b/cmake/externals/libjpeg/jcomapi.c deleted file mode 100644 index 678c5d12d..000000000 --- a/cmake/externals/libjpeg/jcomapi.c +++ /dev/null @@ -1,244 +0,0 @@ -/* - * jcomapi.c - * - * Copyright (C) 1994-1997, Thomas G. Lane. - * Modified 2019 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains application interface routines that are used for both - * compression and decompression. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* - * Abort processing of a JPEG compression or decompression operation, - * but don't destroy the object itself. - * - * For this, we merely clean up all the nonpermanent memory pools. - * Note that temp files (virtual arrays) are not allowed to belong to - * the permanent pool, so we will be able to close all temp files here. - * Closing a data source or destination, if necessary, is the application's - * responsibility. - */ - -GLOBAL(void) -jpeg_abort (j_common_ptr cinfo) -{ - int pool; - - /* Do nothing if called on a not-initialized or destroyed JPEG object. */ - if (cinfo->mem == NULL) - return; - - /* Releasing pools in reverse order might help avoid fragmentation - * with some (brain-damaged) malloc libraries. - */ - for (pool = JPOOL_NUMPOOLS-1; pool > JPOOL_PERMANENT; pool--) { - (*cinfo->mem->free_pool) (cinfo, pool); - } - - /* Reset overall state for possible reuse of object */ - if (cinfo->is_decompressor) { - cinfo->global_state = DSTATE_START; - /* Try to keep application from accessing now-deleted marker list. - * A bit kludgy to do it here, but this is the most central place. - */ - ((j_decompress_ptr) cinfo)->marker_list = NULL; - } else { - cinfo->global_state = CSTATE_START; - } -} - - -/* - * Destruction of a JPEG object. - * - * Everything gets deallocated except the master jpeg_compress_struct itself - * and the error manager struct. Both of these are supplied by the application - * and must be freed, if necessary, by the application. (Often they are on - * the stack and so don't need to be freed anyway.) - * Closing a data source or destination, if necessary, is the application's - * responsibility. - */ - -GLOBAL(void) -jpeg_destroy (j_common_ptr cinfo) -{ - /* We need only tell the memory manager to release everything. */ - /* NB: mem pointer is NULL if memory mgr failed to initialize. */ - if (cinfo->mem != NULL) - (*cinfo->mem->self_destruct) (cinfo); - cinfo->mem = NULL; /* be safe if jpeg_destroy is called twice */ - cinfo->global_state = 0; /* mark it destroyed */ -} - - -/* - * Convenience routines for allocating quantization and Huffman tables. - * (Would jutils.c be a more reasonable place to put these?) - */ - -GLOBAL(JQUANT_TBL *) -jpeg_alloc_quant_table (j_common_ptr cinfo) -{ - JQUANT_TBL *tbl; - - tbl = (JQUANT_TBL *) - (*cinfo->mem->alloc_small) (cinfo, JPOOL_PERMANENT, SIZEOF(JQUANT_TBL)); - tbl->sent_table = FALSE; /* make sure this is false in any new table */ - return tbl; -} - - -GLOBAL(JHUFF_TBL *) -jpeg_alloc_huff_table (j_common_ptr cinfo) -{ - JHUFF_TBL *tbl; - - tbl = (JHUFF_TBL *) - (*cinfo->mem->alloc_small) (cinfo, JPOOL_PERMANENT, SIZEOF(JHUFF_TBL)); - tbl->sent_table = FALSE; /* make sure this is false in any new table */ - return tbl; -} - - -/* - * Set up the standard Huffman tables (cf. JPEG standard section K.3). - * IMPORTANT: these are only valid for 8-bit data precision! - * (Would jutils.c be a more reasonable place to put this?) - */ - -GLOBAL(JHUFF_TBL *) -jpeg_std_huff_table (j_common_ptr cinfo, boolean isDC, int tblno) -{ - JHUFF_TBL **htblptr, *htbl; - const UINT8 *bits, *val; - int nsymbols, len; - - static const UINT8 bits_dc_luminance[17] = - { /* 0-base */ 0, 0, 1, 5, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0 }; - static const UINT8 val_dc_luminance[] = - { 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 }; - - static const UINT8 bits_dc_chrominance[17] = - { /* 0-base */ 0, 0, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0 }; - static const UINT8 val_dc_chrominance[] = - { 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11 }; - - static const UINT8 bits_ac_luminance[17] = - { /* 0-base */ 0, 0, 2, 1, 3, 3, 2, 4, 3, 5, 5, 4, 4, 0, 0, 1, 0x7d }; - static const UINT8 val_ac_luminance[] = - { 0x01, 0x02, 0x03, 0x00, 0x04, 0x11, 0x05, 0x12, - 0x21, 0x31, 0x41, 0x06, 0x13, 0x51, 0x61, 0x07, - 0x22, 0x71, 0x14, 0x32, 0x81, 0x91, 0xa1, 0x08, - 0x23, 0x42, 0xb1, 0xc1, 0x15, 0x52, 0xd1, 0xf0, - 0x24, 0x33, 0x62, 0x72, 0x82, 0x09, 0x0a, 0x16, - 0x17, 0x18, 0x19, 0x1a, 0x25, 0x26, 0x27, 0x28, - 0x29, 0x2a, 0x34, 0x35, 0x36, 0x37, 0x38, 0x39, - 0x3a, 0x43, 0x44, 0x45, 0x46, 0x47, 0x48, 0x49, - 0x4a, 0x53, 0x54, 0x55, 0x56, 0x57, 0x58, 0x59, - 0x5a, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, 0x69, - 0x6a, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, 0x79, - 0x7a, 0x83, 0x84, 0x85, 0x86, 0x87, 0x88, 0x89, - 0x8a, 0x92, 0x93, 0x94, 0x95, 0x96, 0x97, 0x98, - 0x99, 0x9a, 0xa2, 0xa3, 0xa4, 0xa5, 0xa6, 0xa7, - 0xa8, 0xa9, 0xaa, 0xb2, 0xb3, 0xb4, 0xb5, 0xb6, - 0xb7, 0xb8, 0xb9, 0xba, 0xc2, 0xc3, 0xc4, 0xc5, - 0xc6, 0xc7, 0xc8, 0xc9, 0xca, 0xd2, 0xd3, 0xd4, - 0xd5, 0xd6, 0xd7, 0xd8, 0xd9, 0xda, 0xe1, 0xe2, - 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, 0xe8, 0xe9, 0xea, - 0xf1, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, 0xf8, - 0xf9, 0xfa }; - - static const UINT8 bits_ac_chrominance[17] = - { /* 0-base */ 0, 0, 2, 1, 2, 4, 4, 3, 4, 7, 5, 4, 4, 0, 1, 2, 0x77 }; - static const UINT8 val_ac_chrominance[] = - { 0x00, 0x01, 0x02, 0x03, 0x11, 0x04, 0x05, 0x21, - 0x31, 0x06, 0x12, 0x41, 0x51, 0x07, 0x61, 0x71, - 0x13, 0x22, 0x32, 0x81, 0x08, 0x14, 0x42, 0x91, - 0xa1, 0xb1, 0xc1, 0x09, 0x23, 0x33, 0x52, 0xf0, - 0x15, 0x62, 0x72, 0xd1, 0x0a, 0x16, 0x24, 0x34, - 0xe1, 0x25, 0xf1, 0x17, 0x18, 0x19, 0x1a, 0x26, - 0x27, 0x28, 0x29, 0x2a, 0x35, 0x36, 0x37, 0x38, - 0x39, 0x3a, 0x43, 0x44, 0x45, 0x46, 0x47, 0x48, - 0x49, 0x4a, 0x53, 0x54, 0x55, 0x56, 0x57, 0x58, - 0x59, 0x5a, 0x63, 0x64, 0x65, 0x66, 0x67, 0x68, - 0x69, 0x6a, 0x73, 0x74, 0x75, 0x76, 0x77, 0x78, - 0x79, 0x7a, 0x82, 0x83, 0x84, 0x85, 0x86, 0x87, - 0x88, 0x89, 0x8a, 0x92, 0x93, 0x94, 0x95, 0x96, - 0x97, 0x98, 0x99, 0x9a, 0xa2, 0xa3, 0xa4, 0xa5, - 0xa6, 0xa7, 0xa8, 0xa9, 0xaa, 0xb2, 0xb3, 0xb4, - 0xb5, 0xb6, 0xb7, 0xb8, 0xb9, 0xba, 0xc2, 0xc3, - 0xc4, 0xc5, 0xc6, 0xc7, 0xc8, 0xc9, 0xca, 0xd2, - 0xd3, 0xd4, 0xd5, 0xd6, 0xd7, 0xd8, 0xd9, 0xda, - 0xe2, 0xe3, 0xe4, 0xe5, 0xe6, 0xe7, 0xe8, 0xe9, - 0xea, 0xf2, 0xf3, 0xf4, 0xf5, 0xf6, 0xf7, 0xf8, - 0xf9, 0xfa }; - - if (cinfo->is_decompressor) { - if (isDC) - htblptr = ((j_decompress_ptr) cinfo)->dc_huff_tbl_ptrs; - else - htblptr = ((j_decompress_ptr) cinfo)->ac_huff_tbl_ptrs; - } else { - if (isDC) - htblptr = ((j_compress_ptr) cinfo)->dc_huff_tbl_ptrs; - else - htblptr = ((j_compress_ptr) cinfo)->ac_huff_tbl_ptrs; - } - - switch (tblno) { - case 0: - if (isDC) { - bits = bits_dc_luminance; - val = val_dc_luminance; - } else { - bits = bits_ac_luminance; - val = val_ac_luminance; - } - break; - case 1: - if (isDC) { - bits = bits_dc_chrominance; - val = val_dc_chrominance; - } else { - bits = bits_ac_chrominance; - val = val_ac_chrominance; - } - break; - default: - ERREXIT1(cinfo, JERR_NO_HUFF_TABLE, tblno); - return NULL; /* avoid compiler warnings for uninitialized variables */ - } - - if (htblptr[tblno] == NULL) - htblptr[tblno] = jpeg_alloc_huff_table(cinfo); - - htbl = htblptr[tblno]; - - /* Copy the number-of-symbols-of-each-code-length counts */ - MEMCOPY(htbl->bits, bits, SIZEOF(htbl->bits)); - - /* Validate the counts. We do this here mainly so we can copy the right - * number of symbols from the val[] array, without risking marching off - * the end of memory. jxhuff.c will do a more thorough test later. - */ - nsymbols = 0; - for (len = 1; len <= 16; len++) - nsymbols += bits[len]; - if (nsymbols > 256) - ERREXIT(cinfo, JERR_BAD_HUFF_TABLE); - - if (nsymbols > 0) - MEMCOPY(htbl->huffval, val, nsymbols * SIZEOF(UINT8)); - - /* Initialize sent_table FALSE so table will be written to JPEG file. */ - htbl->sent_table = FALSE; - - return htbl; -} diff --git a/cmake/externals/libjpeg/jconfig.h b/cmake/externals/libjpeg/jconfig.h deleted file mode 100644 index 8d6c8f5b3..000000000 --- a/cmake/externals/libjpeg/jconfig.h +++ /dev/null @@ -1,64 +0,0 @@ -/* jconfig.vc --- jconfig.h for Microsoft Visual C++ on Windows 9x or NT. - * This file also works for Borland/Embarcadero C++ for Win32 or Win64 - * (CLI: bcc32, bcc32c, bcc32x, bcc64; GUI IDE: C++Builder/RAD Studio). - * See jconfig.txt for explanations. - */ - -#define HAVE_PROTOTYPES -#define HAVE_UNSIGNED_CHAR -#define HAVE_UNSIGNED_SHORT -/* #define void char */ -/* #define const */ -#undef CHAR_IS_UNSIGNED -#define HAVE_STDDEF_H -#define HAVE_STDLIB_H -#undef NEED_BSD_STRINGS -#undef NEED_SYS_TYPES_H -#undef NEED_FAR_POINTERS /* we presume a 32-bit flat memory model */ -#undef NEED_SHORT_EXTERNAL_NAMES -#undef INCOMPLETE_TYPES_BROKEN - -/* Define "boolean" as unsigned char, not enum, per Windows custom */ -#ifndef __RPCNDR_H__ /* don't conflict if rpcndr.h already read */ -typedef unsigned char boolean; -#endif -#ifndef FALSE /* in case these macros already exist */ -#define FALSE 0 /* values of boolean */ -#endif -#ifndef TRUE -#define TRUE 1 -#endif -#define HAVE_BOOLEAN /* prevent jmorecfg.h from redefining it */ - -/* Define custom RGB color order, prevent jmorecfg.h from redefinition */ -#undef JPEG_HAVE_RGB_CUSTOM -/* Use Windows custom BGR color order defined in jmorecfg.h */ -#undef JPEG_USE_RGB_CUSTOM - -/* Define custom file I/O functions, prevent jinclude.h from redefinition */ -#undef JPEG_HAVE_FILE_IO_CUSTOM -/* Use Delphi custom file I/O functions defined in jinclude.h */ -#undef JPEG_USE_FILE_IO_CUSTOM - - -#ifdef JPEG_INTERNALS - -#undef RIGHT_SHIFT_IS_UNSIGNED - -#endif /* JPEG_INTERNALS */ - -#ifdef JPEG_CJPEG_DJPEG - -#define BMP_SUPPORTED /* BMP image file format */ -#define GIF_SUPPORTED /* GIF image file format */ -#define PPM_SUPPORTED /* PBMPLUS PPM/PGM image file format */ -#undef RLE_SUPPORTED /* Utah RLE image file format */ -#define TARGA_SUPPORTED /* Targa image file format */ - -#define TWO_FILE_COMMANDLINE /* optional */ -#define USE_SETMODE /* Microsoft/Borland/Embarcadero have setmode() */ -#undef NEED_SIGNAL_CATCHER -#undef DONT_USE_B_MODE -#undef PROGRESS_REPORT /* optional */ - -#endif /* JPEG_CJPEG_DJPEG */ diff --git a/cmake/externals/libjpeg/jcparam.c b/cmake/externals/libjpeg/jcparam.c deleted file mode 100644 index 261ae86ca..000000000 --- a/cmake/externals/libjpeg/jcparam.c +++ /dev/null @@ -1,591 +0,0 @@ -/* - * jcparam.c - * - * Copyright (C) 1991-1998, Thomas G. Lane. - * Modified 2003-2022 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains optional default-setting code for the JPEG compressor. - * Applications do not have to use this file, but those that don't use it - * must know a lot more about the innards of the JPEG code. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* - * Quantization table setup routines - */ - -GLOBAL(void) -jpeg_add_quant_table (j_compress_ptr cinfo, int which_tbl, - const unsigned int *basic_table, - int scale_factor, boolean force_baseline) -/* Define a quantization table equal to the basic_table times - * a scale factor (given as a percentage). - * If force_baseline is TRUE, the computed quantization table entries - * are limited to 1..255 for JPEG baseline compatibility. - */ -{ - JQUANT_TBL ** qtblptr; - int i; - long temp; - - /* Safety check to ensure start_compress not called yet. */ - if (cinfo->global_state != CSTATE_START) - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - - if (which_tbl < 0 || which_tbl >= NUM_QUANT_TBLS) - ERREXIT1(cinfo, JERR_DQT_INDEX, which_tbl); - - qtblptr = & cinfo->quant_tbl_ptrs[which_tbl]; - - if (*qtblptr == NULL) - *qtblptr = jpeg_alloc_quant_table((j_common_ptr) cinfo); - - for (i = 0; i < DCTSIZE2; i++) { - temp = ((long) basic_table[i] * scale_factor + 50L) / 100L; - /* limit the values to the valid range */ - if (temp <= 0L) temp = 1L; - if (temp > 32767L) temp = 32767L; /* max quantizer needed for 12 bits */ - if (force_baseline && temp > 255L) - temp = 255L; /* limit to baseline range if requested */ - (*qtblptr)->quantval[i] = (UINT16) temp; - } - - /* Initialize sent_table FALSE so table will be written to JPEG file. */ - (*qtblptr)->sent_table = FALSE; -} - - -/* These are the sample quantization tables given in JPEG spec section K.1. - * NOTE: chrominance DC value is changed from 17 to 16 for lossless support. - * The spec says that the values given produce "good" quality, - * and when divided by 2, "very good" quality. - */ -static const unsigned int std_luminance_quant_tbl[DCTSIZE2] = { - 16, 11, 10, 16, 24, 40, 51, 61, - 12, 12, 14, 19, 26, 58, 60, 55, - 14, 13, 16, 24, 40, 57, 69, 56, - 14, 17, 22, 29, 51, 87, 80, 62, - 18, 22, 37, 56, 68, 109, 103, 77, - 24, 35, 55, 64, 81, 104, 113, 92, - 49, 64, 78, 87, 103, 121, 120, 101, - 72, 92, 95, 98, 112, 100, 103, 99 -}; -static const unsigned int std_chrominance_quant_tbl[DCTSIZE2] = { - 16, 18, 24, 47, 99, 99, 99, 99, - 18, 21, 26, 66, 99, 99, 99, 99, - 24, 26, 56, 99, 99, 99, 99, 99, - 47, 66, 99, 99, 99, 99, 99, 99, - 99, 99, 99, 99, 99, 99, 99, 99, - 99, 99, 99, 99, 99, 99, 99, 99, - 99, 99, 99, 99, 99, 99, 99, 99, - 99, 99, 99, 99, 99, 99, 99, 99 -}; - - -GLOBAL(void) -jpeg_default_qtables (j_compress_ptr cinfo, boolean force_baseline) -/* Set or change the 'quality' (quantization) setting, using default tables - * and straight percentage-scaling quality scales. - * This entry point allows different scalings for luminance and chrominance. - */ -{ - /* Set up two quantization tables using the specified scaling */ - jpeg_add_quant_table(cinfo, 0, std_luminance_quant_tbl, - cinfo->q_scale_factor[0], force_baseline); - jpeg_add_quant_table(cinfo, 1, std_chrominance_quant_tbl, - cinfo->q_scale_factor[1], force_baseline); -} - - -GLOBAL(void) -jpeg_set_linear_quality (j_compress_ptr cinfo, int scale_factor, - boolean force_baseline) -/* Set or change the 'quality' (quantization) setting, using default tables - * and a straight percentage-scaling quality scale. In most cases it's better - * to use jpeg_set_quality (below); this entry point is provided for - * applications that insist on a linear percentage scaling. - */ -{ - /* Set up two quantization tables using the specified scaling */ - jpeg_add_quant_table(cinfo, 0, std_luminance_quant_tbl, - scale_factor, force_baseline); - jpeg_add_quant_table(cinfo, 1, std_chrominance_quant_tbl, - scale_factor, force_baseline); -} - - -GLOBAL(int) -jpeg_quality_scaling (int quality) -/* Convert a user-specified quality rating to a percentage scaling factor - * for an underlying quantization table, using our recommended scaling curve. - * The input 'quality' factor should be 0 (terrible) to 100 (very good). - */ -{ - /* Safety limit on quality factor. Convert 0 to 1 to avoid zero divide. */ - if (quality <= 0) quality = 1; - if (quality > 100) quality = 100; - - /* The basic table is used as-is (scaling 100) for a quality of 50. - * Qualities 50..100 are converted to scaling percentage 200 - 2*Q; - * note that at Q=100 the scaling is 0, which will cause jpeg_add_quant_table - * to make all the table entries 1 (hence, minimum quantization loss). - * Qualities 1..50 are converted to scaling percentage 5000/Q. - */ - if (quality < 50) - quality = 5000 / quality; - else - quality = 200 - quality*2; - - return quality; -} - - -GLOBAL(void) -jpeg_set_quality (j_compress_ptr cinfo, int quality, boolean force_baseline) -/* Set or change the 'quality' (quantization) setting, using default tables. - * This is the standard quality-adjusting entry point for typical user - * interfaces; only those who want detailed control over quantization tables - * would use the preceding routines directly. - */ -{ - /* Convert user 0-100 rating to percentage scaling */ - quality = jpeg_quality_scaling(quality); - - /* Set up standard quality tables */ - jpeg_set_linear_quality(cinfo, quality, force_baseline); -} - - -/* - * Reset standard Huffman tables - */ - -LOCAL(void) -std_huff_tables (j_compress_ptr cinfo) -{ - if (cinfo->dc_huff_tbl_ptrs[0] != NULL) - (void) jpeg_std_huff_table((j_common_ptr) cinfo, TRUE, 0); - - if (cinfo->ac_huff_tbl_ptrs[0] != NULL) - (void) jpeg_std_huff_table((j_common_ptr) cinfo, FALSE, 0); - - if (cinfo->dc_huff_tbl_ptrs[1] != NULL) - (void) jpeg_std_huff_table((j_common_ptr) cinfo, TRUE, 1); - - if (cinfo->ac_huff_tbl_ptrs[1] != NULL) - (void) jpeg_std_huff_table((j_common_ptr) cinfo, FALSE, 1); -} - - -/* - * Default parameter setup for compression. - * - * Applications that don't choose to use this routine must do their - * own setup of all these parameters. Alternately, you can call this - * to establish defaults and then alter parameters selectively. This - * is the recommended approach since, if we add any new parameters, - * your code will still work (they'll be set to reasonable defaults). - */ - -GLOBAL(void) -jpeg_set_defaults (j_compress_ptr cinfo) -{ - int i; - - /* Safety check to ensure start_compress not called yet. */ - if (cinfo->global_state != CSTATE_START) - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - - /* Allocate comp_info array large enough for maximum component count. - * Array is made permanent in case application wants to compress - * multiple images at same param settings. - */ - if (cinfo->comp_info == NULL) - cinfo->comp_info = (jpeg_component_info *) - (*cinfo->mem->alloc_small) ((j_common_ptr) cinfo, JPOOL_PERMANENT, - MAX_COMPONENTS * SIZEOF(jpeg_component_info)); - - /* Initialize everything not dependent on the color space */ - - cinfo->scale_num = 1; /* 1:1 scaling */ - cinfo->scale_denom = 1; - cinfo->data_precision = BITS_IN_JSAMPLE; - /* Set up two quantization tables using default quality of 75 */ - jpeg_set_quality(cinfo, 75, TRUE); - /* Reset standard Huffman tables */ - std_huff_tables(cinfo); - - /* Initialize default arithmetic coding conditioning */ - for (i = 0; i < NUM_ARITH_TBLS; i++) { - cinfo->arith_dc_L[i] = 0; - cinfo->arith_dc_U[i] = 1; - cinfo->arith_ac_K[i] = 5; - } - - /* Default is no multiple-scan output */ - cinfo->scan_info = NULL; - cinfo->num_scans = 0; - - /* Expect normal source image, not raw downsampled data */ - cinfo->raw_data_in = FALSE; - - /* The standard Huffman tables are only valid for 8-bit data precision. - * If the precision is higher, use arithmetic coding. - * (Alternatively, using Huffman coding would be possible with forcing - * optimization on so that usable tables will be computed, or by - * supplying default tables that are valid for the desired precision.) - * Otherwise, use Huffman coding by default. - */ - cinfo->arith_code = cinfo->data_precision > 8 ? TRUE : FALSE; - - /* By default, don't do extra passes to optimize entropy coding */ - cinfo->optimize_coding = FALSE; - - /* By default, use the simpler non-cosited sampling alignment */ - cinfo->CCIR601_sampling = FALSE; - - /* By default, apply fancy downsampling */ - cinfo->do_fancy_downsampling = TRUE; - - /* No input smoothing */ - cinfo->smoothing_factor = 0; - - /* DCT algorithm preference */ - cinfo->dct_method = JDCT_DEFAULT; - - /* No restart markers */ - cinfo->restart_interval = 0; - cinfo->restart_in_rows = 0; - - /* Fill in default JFIF marker parameters. Note that whether the marker - * will actually be written is determined by jpeg_set_colorspace. - * - * By default, the library emits JFIF version code 1.01. - * An application that wants to emit JFIF 1.02 extension markers should set - * JFIF_minor_version to 2. We could probably get away with just defaulting - * to 1.02, but there may still be some decoders in use that will complain - * about that; saying 1.01 should minimize compatibility problems. - * - * For wide gamut colorspaces (BG_RGB and BG_YCC), the major version will be - * overridden by jpeg_set_colorspace and set to 2. - */ - cinfo->JFIF_major_version = 1; /* Default JFIF version = 1.01 */ - cinfo->JFIF_minor_version = 1; - cinfo->density_unit = 0; /* Pixel size is unknown by default */ - cinfo->X_density = 1; /* Pixel aspect ratio is square by default */ - cinfo->Y_density = 1; - - /* No color transform */ - cinfo->color_transform = JCT_NONE; - - /* Choose JPEG colorspace based on input space, set defaults accordingly */ - - jpeg_default_colorspace(cinfo); -} - - -/* - * Select an appropriate JPEG colorspace for in_color_space. - */ - -GLOBAL(void) -jpeg_default_colorspace (j_compress_ptr cinfo) -{ - switch (cinfo->in_color_space) { - case JCS_UNKNOWN: - jpeg_set_colorspace(cinfo, JCS_UNKNOWN); - break; - case JCS_GRAYSCALE: - jpeg_set_colorspace(cinfo, JCS_GRAYSCALE); - break; - case JCS_RGB: - jpeg_set_colorspace(cinfo, JCS_YCbCr); - break; - case JCS_YCbCr: - jpeg_set_colorspace(cinfo, JCS_YCbCr); - break; - case JCS_CMYK: - jpeg_set_colorspace(cinfo, JCS_CMYK); /* By default, no translation */ - break; - case JCS_YCCK: - jpeg_set_colorspace(cinfo, JCS_YCCK); - break; - case JCS_BG_RGB: - /* No translation for now -- conversion to BG_YCC not yet supportet */ - jpeg_set_colorspace(cinfo, JCS_BG_RGB); - break; - case JCS_BG_YCC: - jpeg_set_colorspace(cinfo, JCS_BG_YCC); - break; - default: - ERREXIT(cinfo, JERR_BAD_IN_COLORSPACE); - } -} - - -/* - * Set the JPEG colorspace, and choose colorspace-dependent default values. - */ - -GLOBAL(void) -jpeg_set_colorspace (j_compress_ptr cinfo, J_COLOR_SPACE colorspace) -{ - jpeg_component_info * compptr; - int ci; - -#define SET_COMP(index,id,hsamp,vsamp,quant,dctbl,actbl) \ - (compptr = &cinfo->comp_info[index], \ - compptr->component_id = (id), \ - compptr->h_samp_factor = (hsamp), \ - compptr->v_samp_factor = (vsamp), \ - compptr->quant_tbl_no = (quant), \ - compptr->dc_tbl_no = (dctbl), \ - compptr->ac_tbl_no = (actbl) ) - - /* Safety check to ensure start_compress not called yet. */ - if (cinfo->global_state != CSTATE_START) - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - - /* For all colorspaces, we use Q and Huff tables 0 for luminance components, - * tables 1 for chrominance components. - */ - - cinfo->jpeg_color_space = colorspace; - - cinfo->write_JFIF_header = FALSE; /* No marker for non-JFIF colorspaces */ - cinfo->write_Adobe_marker = FALSE; /* write no Adobe marker by default */ - - switch (colorspace) { - case JCS_UNKNOWN: - cinfo->num_components = cinfo->input_components; - if (cinfo->num_components < 1 || cinfo->num_components > MAX_COMPONENTS) - ERREXIT2(cinfo, JERR_COMPONENT_COUNT, cinfo->num_components, - MAX_COMPONENTS); - for (ci = 0; ci < cinfo->num_components; ci++) { - SET_COMP(ci, ci, 1,1, 0, 0,0); - } - break; - case JCS_GRAYSCALE: - cinfo->write_JFIF_header = TRUE; /* Write a JFIF marker */ - cinfo->num_components = 1; - /* JFIF specifies component ID 1 */ - SET_COMP(0, 0x01, 1,1, 0, 0,0); - break; - case JCS_RGB: - cinfo->write_Adobe_marker = TRUE; /* write Adobe marker to flag RGB */ - cinfo->num_components = 3; - SET_COMP(0, 0x52 /* 'R' */, 1,1, - cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0, - cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0, - cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0); - SET_COMP(1, 0x47 /* 'G' */, 1,1, 0, 0,0); - SET_COMP(2, 0x42 /* 'B' */, 1,1, - cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0, - cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0, - cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0); - break; - case JCS_YCbCr: - cinfo->write_JFIF_header = TRUE; /* Write a JFIF marker */ - cinfo->num_components = 3; - /* JFIF specifies component IDs 1,2,3 */ - /* We default to 2x2 subsamples of chrominance */ - SET_COMP(0, 0x01, 2,2, 0, 0,0); - SET_COMP(1, 0x02, 1,1, 1, 1,1); - SET_COMP(2, 0x03, 1,1, 1, 1,1); - break; - case JCS_CMYK: - cinfo->write_Adobe_marker = TRUE; /* write Adobe marker to flag CMYK */ - cinfo->num_components = 4; - SET_COMP(0, 0x43 /* 'C' */, 1,1, 0, 0,0); - SET_COMP(1, 0x4D /* 'M' */, 1,1, 0, 0,0); - SET_COMP(2, 0x59 /* 'Y' */, 1,1, 0, 0,0); - SET_COMP(3, 0x4B /* 'K' */, 1,1, 0, 0,0); - break; - case JCS_YCCK: - cinfo->write_Adobe_marker = TRUE; /* write Adobe marker to flag YCCK */ - cinfo->num_components = 4; - SET_COMP(0, 0x01, 2,2, 0, 0,0); - SET_COMP(1, 0x02, 1,1, 1, 1,1); - SET_COMP(2, 0x03, 1,1, 1, 1,1); - SET_COMP(3, 0x04, 2,2, 0, 0,0); - break; - case JCS_BG_RGB: - cinfo->write_JFIF_header = TRUE; /* Write a JFIF marker */ - cinfo->JFIF_major_version = 2; /* Set JFIF major version = 2 */ - cinfo->num_components = 3; - /* Add offset 0x20 to the normal R/G/B component IDs */ - SET_COMP(0, 0x72 /* 'r' */, 1,1, - cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0, - cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0, - cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0); - SET_COMP(1, 0x67 /* 'g' */, 1,1, 0, 0,0); - SET_COMP(2, 0x62 /* 'b' */, 1,1, - cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0, - cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0, - cinfo->color_transform == JCT_SUBTRACT_GREEN ? 1 : 0); - break; - case JCS_BG_YCC: - cinfo->write_JFIF_header = TRUE; /* Write a JFIF marker */ - cinfo->JFIF_major_version = 2; /* Set JFIF major version = 2 */ - cinfo->num_components = 3; - /* Add offset 0x20 to the normal Cb/Cr component IDs */ - /* We default to 2x2 subsamples of chrominance */ - SET_COMP(0, 0x01, 2,2, 0, 0,0); - SET_COMP(1, 0x22, 1,1, 1, 1,1); - SET_COMP(2, 0x23, 1,1, 1, 1,1); - break; - default: - ERREXIT(cinfo, JERR_BAD_J_COLORSPACE); - } -} - - -#ifdef C_PROGRESSIVE_SUPPORTED - -LOCAL(jpeg_scan_info *) -fill_a_scan (jpeg_scan_info * scanptr, int ci, - int Ss, int Se, int Ah, int Al) -/* Support routine: generate one scan for specified component */ -{ - scanptr->comps_in_scan = 1; - scanptr->component_index[0] = ci; - scanptr->Ss = Ss; - scanptr->Se = Se; - scanptr->Ah = Ah; - scanptr->Al = Al; - scanptr++; - return scanptr; -} - -LOCAL(jpeg_scan_info *) -fill_scans (jpeg_scan_info * scanptr, int ncomps, - int Ss, int Se, int Ah, int Al) -/* Support routine: generate one scan for each component */ -{ - int ci; - - for (ci = 0; ci < ncomps; ci++) { - scanptr->comps_in_scan = 1; - scanptr->component_index[0] = ci; - scanptr->Ss = Ss; - scanptr->Se = Se; - scanptr->Ah = Ah; - scanptr->Al = Al; - scanptr++; - } - return scanptr; -} - -LOCAL(jpeg_scan_info *) -fill_dc_scans (jpeg_scan_info * scanptr, int ncomps, int Ah, int Al) -/* Support routine: generate interleaved DC scan if possible, else N scans */ -{ - int ci; - - if (ncomps <= MAX_COMPS_IN_SCAN) { - /* Single interleaved DC scan */ - scanptr->comps_in_scan = ncomps; - for (ci = 0; ci < ncomps; ci++) - scanptr->component_index[ci] = ci; - scanptr->Ss = scanptr->Se = 0; - scanptr->Ah = Ah; - scanptr->Al = Al; - scanptr++; - } else { - /* Noninterleaved DC scan for each component */ - scanptr = fill_scans(scanptr, ncomps, 0, 0, Ah, Al); - } - return scanptr; -} - - -/* - * Create a recommended progressive-JPEG script. - * cinfo->num_components and cinfo->jpeg_color_space must be correct. - */ - -GLOBAL(void) -jpeg_simple_progression (j_compress_ptr cinfo) -{ - int ncomps = cinfo->num_components; - int nscans; - jpeg_scan_info * scanptr; - - /* Safety check to ensure start_compress not called yet. */ - if (cinfo->global_state != CSTATE_START) - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - - /* Figure space needed for script. Calculation must match code below! */ - if (ncomps == 3 && - (cinfo->jpeg_color_space == JCS_YCbCr || - cinfo->jpeg_color_space == JCS_BG_YCC)) { - /* Custom script for YCC color images. */ - nscans = 10; - } else { - /* All-purpose script for other color spaces. */ - if (ncomps > MAX_COMPS_IN_SCAN) - nscans = 6 * ncomps; /* 2 DC + 4 AC scans per component */ - else - nscans = 2 + 4 * ncomps; /* 2 DC scans; 4 AC scans per component */ - } - - /* Allocate space for script. - * We need to put it in the permanent pool in case the application performs - * multiple compressions without changing the settings. To avoid a memory - * leak if jpeg_simple_progression is called repeatedly for the same JPEG - * object, we try to re-use previously allocated space, and we allocate - * enough space to handle YCC even if initially asked for grayscale. - */ - if (cinfo->script_space == NULL || cinfo->script_space_size < nscans) { - cinfo->script_space_size = MAX(nscans, 10); - cinfo->script_space = (jpeg_scan_info *) - (*cinfo->mem->alloc_small) ((j_common_ptr) cinfo, JPOOL_PERMANENT, - cinfo->script_space_size * SIZEOF(jpeg_scan_info)); - } - scanptr = cinfo->script_space; - cinfo->scan_info = scanptr; - cinfo->num_scans = nscans; - - if (ncomps == 3 && - (cinfo->jpeg_color_space == JCS_YCbCr || - cinfo->jpeg_color_space == JCS_BG_YCC)) { - /* Custom script for YCC color images. */ - /* Initial DC scan */ - scanptr = fill_dc_scans(scanptr, ncomps, 0, 1); - /* Initial AC scan: get some luma data out in a hurry */ - scanptr = fill_a_scan(scanptr, 0, 1, 5, 0, 2); - /* Chroma data is too small to be worth expending many scans on */ - scanptr = fill_a_scan(scanptr, 2, 1, 63, 0, 1); - scanptr = fill_a_scan(scanptr, 1, 1, 63, 0, 1); - /* Complete spectral selection for luma AC */ - scanptr = fill_a_scan(scanptr, 0, 6, 63, 0, 2); - /* Refine next bit of luma AC */ - scanptr = fill_a_scan(scanptr, 0, 1, 63, 2, 1); - /* Finish DC successive approximation */ - scanptr = fill_dc_scans(scanptr, ncomps, 1, 0); - /* Finish AC successive approximation */ - scanptr = fill_a_scan(scanptr, 2, 1, 63, 1, 0); - scanptr = fill_a_scan(scanptr, 1, 1, 63, 1, 0); - /* Luma bottom bit comes last since it's usually largest scan */ - scanptr = fill_a_scan(scanptr, 0, 1, 63, 1, 0); - } else { - /* All-purpose script for other color spaces. */ - /* Successive approximation first pass */ - scanptr = fill_dc_scans(scanptr, ncomps, 0, 1); - scanptr = fill_scans(scanptr, ncomps, 1, 5, 0, 2); - scanptr = fill_scans(scanptr, ncomps, 6, 63, 0, 2); - /* Successive approximation second pass */ - scanptr = fill_scans(scanptr, ncomps, 1, 63, 2, 1); - /* Successive approximation final pass */ - scanptr = fill_dc_scans(scanptr, ncomps, 1, 0); - scanptr = fill_scans(scanptr, ncomps, 1, 63, 1, 0); - } -} - -#endif /* C_PROGRESSIVE_SUPPORTED */ diff --git a/cmake/externals/libjpeg/jcprepct.c b/cmake/externals/libjpeg/jcprepct.c deleted file mode 100644 index 586964bd4..000000000 --- a/cmake/externals/libjpeg/jcprepct.c +++ /dev/null @@ -1,358 +0,0 @@ -/* - * jcprepct.c - * - * Copyright (C) 1994-1996, Thomas G. Lane. - * Modified 2003-2020 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains the compression preprocessing controller. - * This controller manages the color conversion, downsampling, - * and edge expansion steps. - * - * Most of the complexity here is associated with buffering input rows - * as required by the downsampler. See the comments at the head of - * jcsample.c for the downsampler's needs. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* At present, jcsample.c can request context rows only for smoothing. - * In the future, we might also need context rows for CCIR601 sampling - * or other more-complex downsampling procedures. The code to support - * context rows should be compiled only if needed. - */ -#ifdef INPUT_SMOOTHING_SUPPORTED -#define CONTEXT_ROWS_SUPPORTED -#endif - - -/* - * For the simple (no-context-row) case, we just need to buffer one - * row group's worth of pixels for the downsampling step. At the bottom of - * the image, we pad to a full row group by replicating the last pixel row. - * The downsampler's last output row is then replicated if needed to pad - * out to a full iMCU row. - * - * When providing context rows, we must buffer three row groups' worth of - * pixels. Three row groups are physically allocated, but the row pointer - * arrays are made five row groups high, with the extra pointers above and - * below "wrapping around" to point to the last and first real row groups. - * This allows the downsampler to access the proper context rows. - * At the top and bottom of the image, we create dummy context rows by - * copying the first or last real pixel row. This copying could be avoided - * by pointer hacking as is done in jdmainct.c, but it doesn't seem worth the - * trouble on the compression side. - */ - - -/* Private buffer controller object */ - -typedef struct { - struct jpeg_c_prep_controller pub; /* public fields */ - - /* Downsampling input buffer. This buffer holds color-converted data - * until we have enough to do a downsample step. - */ - JSAMPARRAY color_buf[MAX_COMPONENTS]; - - JDIMENSION rows_to_go; /* counts rows remaining in source image */ - int next_buf_row; /* index of next row to store in color_buf */ - -#ifdef CONTEXT_ROWS_SUPPORTED /* only needed for context case */ - int this_row_group; /* starting row index of group to process */ - int next_buf_stop; /* downsample when we reach this index */ -#endif -} my_prep_controller; - -typedef my_prep_controller * my_prep_ptr; - - -/* - * Initialize for a processing pass. - */ - -METHODDEF(void) -start_pass_prep (j_compress_ptr cinfo, J_BUF_MODE pass_mode) -{ - my_prep_ptr prep = (my_prep_ptr) cinfo->prep; - - if (pass_mode != JBUF_PASS_THRU) - ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); - - /* Initialize total-height counter for detecting bottom of image */ - prep->rows_to_go = cinfo->image_height; - /* Mark the conversion buffer empty */ - prep->next_buf_row = 0; -#ifdef CONTEXT_ROWS_SUPPORTED - /* Preset additional state variables for context mode. - * These aren't used in non-context mode, so we needn't test which mode. - */ - prep->this_row_group = 0; - /* Set next_buf_stop to stop after two row groups have been read in. */ - prep->next_buf_stop = 2 * cinfo->max_v_samp_factor; -#endif -} - - -/* - * Expand an image vertically from height input_rows to height output_rows, - * by duplicating the bottom row. - */ - -LOCAL(void) -expand_bottom_edge (JSAMPARRAY image_data, JDIMENSION num_cols, - int input_rows, int output_rows) -{ - register int row; - - for (row = input_rows; row < output_rows; row++) { - jcopy_sample_rows(image_data + input_rows - 1, - image_data + row, - 1, num_cols); - } -} - - -/* - * Process some data in the simple no-context case. - * - * Preprocessor output data is counted in "row groups". A row group - * is defined to be v_samp_factor sample rows of each component. - * Downsampling will produce this much data from each max_v_samp_factor - * input rows. - */ - -METHODDEF(void) -pre_process_data (j_compress_ptr cinfo, - JSAMPARRAY input_buf, JDIMENSION *in_row_ctr, - JDIMENSION in_rows_avail, - JSAMPIMAGE output_buf, JDIMENSION *out_row_group_ctr, - JDIMENSION out_row_groups_avail) -{ - my_prep_ptr prep = (my_prep_ptr) cinfo->prep; - int numrows, ci; - JDIMENSION inrows; - jpeg_component_info * compptr; - - while (*in_row_ctr < in_rows_avail && - *out_row_group_ctr < out_row_groups_avail) { - /* Do color conversion to fill the conversion buffer. */ - inrows = in_rows_avail - *in_row_ctr; - numrows = cinfo->max_v_samp_factor - prep->next_buf_row; - numrows = (int) MIN((JDIMENSION) numrows, inrows); - (*cinfo->cconvert->color_convert) (cinfo, input_buf + *in_row_ctr, - prep->color_buf, - (JDIMENSION) prep->next_buf_row, - numrows); - *in_row_ctr += numrows; - prep->next_buf_row += numrows; - prep->rows_to_go -= numrows; - /* If at bottom of image, pad to fill the conversion buffer. */ - if (prep->rows_to_go == 0 && - prep->next_buf_row < cinfo->max_v_samp_factor) { - for (ci = 0; ci < cinfo->num_components; ci++) { - expand_bottom_edge(prep->color_buf[ci], cinfo->image_width, - prep->next_buf_row, cinfo->max_v_samp_factor); - } - prep->next_buf_row = cinfo->max_v_samp_factor; - } - /* If we've filled the conversion buffer, empty it. */ - if (prep->next_buf_row == cinfo->max_v_samp_factor) { - (*cinfo->downsample->downsample) (cinfo, - prep->color_buf, (JDIMENSION) 0, - output_buf, *out_row_group_ctr); - prep->next_buf_row = 0; - (*out_row_group_ctr)++; - } - /* If at bottom of image, pad the output to a full iMCU height. - * Note we assume the caller is providing a one-iMCU-height output buffer! - */ - if (prep->rows_to_go == 0 && - *out_row_group_ctr < out_row_groups_avail) { - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - numrows = (compptr->v_samp_factor * compptr->DCT_v_scaled_size) / - cinfo->min_DCT_v_scaled_size; - expand_bottom_edge(output_buf[ci], - compptr->width_in_blocks * compptr->DCT_h_scaled_size, - (int) (*out_row_group_ctr * numrows), - (int) (out_row_groups_avail * numrows)); - } - *out_row_group_ctr = out_row_groups_avail; - break; /* can exit outer loop without test */ - } - } -} - - -#ifdef CONTEXT_ROWS_SUPPORTED - -/* - * Process some data in the context case. - */ - -METHODDEF(void) -pre_process_context (j_compress_ptr cinfo, - JSAMPARRAY input_buf, JDIMENSION *in_row_ctr, - JDIMENSION in_rows_avail, - JSAMPIMAGE output_buf, JDIMENSION *out_row_group_ctr, - JDIMENSION out_row_groups_avail) -{ - my_prep_ptr prep = (my_prep_ptr) cinfo->prep; - int numrows, ci; - int buf_height = cinfo->max_v_samp_factor * 3; - JDIMENSION inrows; - - while (*out_row_group_ctr < out_row_groups_avail) { - if (*in_row_ctr < in_rows_avail) { - /* Do color conversion to fill the conversion buffer. */ - inrows = in_rows_avail - *in_row_ctr; - numrows = prep->next_buf_stop - prep->next_buf_row; - numrows = (int) MIN((JDIMENSION) numrows, inrows); - (*cinfo->cconvert->color_convert) (cinfo, input_buf + *in_row_ctr, - prep->color_buf, - (JDIMENSION) prep->next_buf_row, - numrows); - /* Pad at top of image, if first time through */ - if (prep->rows_to_go == cinfo->image_height) { - for (ci = 0; ci < cinfo->num_components; ci++) { - int row; - for (row = 1; row <= cinfo->max_v_samp_factor; row++) { - jcopy_sample_rows(prep->color_buf[ci], - prep->color_buf[ci] - row, - 1, cinfo->image_width); - } - } - } - *in_row_ctr += numrows; - prep->next_buf_row += numrows; - prep->rows_to_go -= numrows; - } else { - /* Return for more data, unless we are at the bottom of the image. */ - if (prep->rows_to_go != 0) - break; - /* When at bottom of image, pad to fill the conversion buffer. */ - if (prep->next_buf_row < prep->next_buf_stop) { - for (ci = 0; ci < cinfo->num_components; ci++) { - expand_bottom_edge(prep->color_buf[ci], cinfo->image_width, - prep->next_buf_row, prep->next_buf_stop); - } - prep->next_buf_row = prep->next_buf_stop; - } - } - /* If we've gotten enough data, downsample a row group. */ - if (prep->next_buf_row == prep->next_buf_stop) { - (*cinfo->downsample->downsample) (cinfo, - prep->color_buf, - (JDIMENSION) prep->this_row_group, - output_buf, *out_row_group_ctr); - (*out_row_group_ctr)++; - /* Advance pointers with wraparound as necessary. */ - prep->this_row_group += cinfo->max_v_samp_factor; - if (prep->this_row_group >= buf_height) - prep->this_row_group = 0; - if (prep->next_buf_row >= buf_height) - prep->next_buf_row = 0; - prep->next_buf_stop = prep->next_buf_row + cinfo->max_v_samp_factor; - } - } -} - - -/* - * Create the wrapped-around downsampling input buffer needed for context mode. - */ - -LOCAL(void) -create_context_buffer (j_compress_ptr cinfo) -{ - my_prep_ptr prep = (my_prep_ptr) cinfo->prep; - int rgroup_height = cinfo->max_v_samp_factor; - int ci, i; - jpeg_component_info * compptr; - JSAMPARRAY true_buffer, fake_buffer; - - /* Grab enough space for fake row pointers for all the components; - * we need five row groups' worth of pointers for each component. - */ - fake_buffer = (JSAMPARRAY) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, - (cinfo->num_components * 5 * rgroup_height) * SIZEOF(JSAMPROW)); - - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - /* Allocate the actual buffer space (3 row groups) for this component. - * We make the buffer wide enough to allow the downsampler to edge-expand - * horizontally within the buffer, if it so chooses. - */ - true_buffer = (*cinfo->mem->alloc_sarray) - ((j_common_ptr) cinfo, JPOOL_IMAGE, - (JDIMENSION) (((long) compptr->width_in_blocks * - cinfo->min_DCT_h_scaled_size * - cinfo->max_h_samp_factor) / compptr->h_samp_factor), - (JDIMENSION) (3 * rgroup_height)); - /* Copy true buffer row pointers into the middle of the fake row array */ - MEMCOPY(fake_buffer + rgroup_height, true_buffer, - 3 * rgroup_height * SIZEOF(JSAMPROW)); - /* Fill in the above and below wraparound pointers */ - for (i = 0; i < rgroup_height; i++) { - fake_buffer[i] = true_buffer[2 * rgroup_height + i]; - fake_buffer[4 * rgroup_height + i] = true_buffer[i]; - } - prep->color_buf[ci] = fake_buffer + rgroup_height; - fake_buffer += 5 * rgroup_height; /* point to space for next component */ - } -} - -#endif /* CONTEXT_ROWS_SUPPORTED */ - - -/* - * Initialize preprocessing controller. - */ - -GLOBAL(void) -jinit_c_prep_controller (j_compress_ptr cinfo, boolean need_full_buffer) -{ - my_prep_ptr prep; - int ci; - jpeg_component_info * compptr; - - if (need_full_buffer) /* safety check */ - ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); - - prep = (my_prep_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_prep_controller)); - cinfo->prep = &prep->pub; - prep->pub.start_pass = start_pass_prep; - - /* Allocate the color conversion buffer. - * We make the buffer wide enough to allow the downsampler to edge-expand - * horizontally within the buffer, if it so chooses. - */ - if (cinfo->downsample->need_context_rows) { - /* Set up to provide context rows */ -#ifdef CONTEXT_ROWS_SUPPORTED - prep->pub.pre_process_data = pre_process_context; - create_context_buffer(cinfo); -#else - ERREXIT(cinfo, JERR_NOT_COMPILED); -#endif - } else { - /* No context, just make it tall enough for one row group */ - prep->pub.pre_process_data = pre_process_data; - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - prep->color_buf[ci] = (*cinfo->mem->alloc_sarray) - ((j_common_ptr) cinfo, JPOOL_IMAGE, - (JDIMENSION) (((long) compptr->width_in_blocks * - cinfo->min_DCT_h_scaled_size * - cinfo->max_h_samp_factor) / compptr->h_samp_factor), - (JDIMENSION) cinfo->max_v_samp_factor); - } - } -} diff --git a/cmake/externals/libjpeg/jcsample.c b/cmake/externals/libjpeg/jcsample.c deleted file mode 100644 index 2372c4173..000000000 --- a/cmake/externals/libjpeg/jcsample.c +++ /dev/null @@ -1,545 +0,0 @@ -/* - * jcsample.c - * - * Copyright (C) 1991-1996, Thomas G. Lane. - * Modified 2003-2020 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains downsampling routines. - * - * Downsampling input data is counted in "row groups". A row group - * is defined to be max_v_samp_factor pixel rows of each component, - * from which the downsampler produces v_samp_factor sample rows. - * A single row group is processed in each call to the downsampler module. - * - * The downsampler is responsible for edge-expansion of its output data - * to fill an integral number of DCT blocks horizontally. The source buffer - * may be modified if it is helpful for this purpose (the source buffer is - * allocated wide enough to correspond to the desired output width). - * The caller (the prep controller) is responsible for vertical padding. - * - * The downsampler may request "context rows" by setting need_context_rows - * during startup. In this case, the input arrays will contain at least - * one row group's worth of pixels above and below the passed-in data; - * the caller will create dummy rows at image top and bottom by replicating - * the first or last real pixel row. - * - * An excellent reference for image resampling is - * Digital Image Warping, George Wolberg, 1990. - * Pub. by IEEE Computer Society Press, Los Alamitos, CA. ISBN 0-8186-8944-7. - * - * The downsampling algorithm used here is a simple average of the source - * pixels covered by the output pixel. The hi-falutin sampling literature - * refers to this as a "box filter". In general the characteristics of a box - * filter are not very good, but for the specific cases we normally use (1:1 - * and 2:1 ratios) the box is equivalent to a "triangle filter" which is not - * nearly so bad. If you intend to use other sampling ratios, you'd be well - * advised to improve this code. - * - * A simple input-smoothing capability is provided. This is mainly intended - * for cleaning up color-dithered GIF input files (if you find it inadequate, - * we suggest using an external filtering program such as pnmconvol). When - * enabled, each input pixel P is replaced by a weighted sum of itself and its - * eight neighbors. P's weight is 1-8*SF and each neighbor's weight is SF, - * where SF = (smoothing_factor / 1024). - * Currently, smoothing is only supported for 2h2v sampling factors. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* Pointer to routine to downsample a single component */ -typedef JMETHOD(void, downsample1_ptr, - (j_compress_ptr cinfo, jpeg_component_info * compptr, - JSAMPARRAY input_data, JSAMPARRAY output_data)); - -/* Private subobject */ - -typedef struct { - struct jpeg_downsampler pub; /* public fields */ - - /* Downsampling method pointers, one per component */ - downsample1_ptr methods[MAX_COMPONENTS]; - - /* Height of an output row group for each component. */ - int rowgroup_height[MAX_COMPONENTS]; - - /* These arrays save pixel expansion factors so that int_downsample need not - * recompute them each time. They are unused for other downsampling methods. - */ - UINT8 h_expand[MAX_COMPONENTS]; - UINT8 v_expand[MAX_COMPONENTS]; -} my_downsampler; - -typedef my_downsampler * my_downsample_ptr; - - -/* - * Initialize for a downsampling pass. - */ - -METHODDEF(void) -start_pass_downsample (j_compress_ptr cinfo) -{ - /* no work for now */ -} - - -/* - * Expand a component horizontally from width input_cols to width output_cols, - * by duplicating the rightmost samples. - */ - -LOCAL(void) -expand_right_edge (JSAMPARRAY image_data, int num_rows, - JDIMENSION input_cols, JDIMENSION output_cols) -{ - register JSAMPROW ptr; - register JSAMPLE pixval; - register int count; - int row; - int numcols = (int) (output_cols - input_cols); - - if (numcols > 0) { - for (row = 0; row < num_rows; row++) { - ptr = image_data[row] + input_cols; - pixval = ptr[-1]; /* don't need GETJSAMPLE() here */ - for (count = numcols; count > 0; count--) - *ptr++ = pixval; - } - } -} - - -/* - * Do downsampling for a whole row group (all components). - * - * In this version we simply downsample each component independently. - */ - -METHODDEF(void) -sep_downsample (j_compress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION in_row_index, - JSAMPIMAGE output_buf, JDIMENSION out_row_group_index) -{ - my_downsample_ptr downsample = (my_downsample_ptr) cinfo->downsample; - int ci; - jpeg_component_info * compptr; - JSAMPARRAY in_ptr, out_ptr; - - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - in_ptr = input_buf[ci] + in_row_index; - out_ptr = output_buf[ci] + - (out_row_group_index * downsample->rowgroup_height[ci]); - (*downsample->methods[ci]) (cinfo, compptr, in_ptr, out_ptr); - } -} - - -/* - * Downsample pixel values of a single component. - * One row group is processed per call. - * This version handles arbitrary integral sampling ratios, without smoothing. - * Note that this version is not actually used for customary sampling ratios. - */ - -METHODDEF(void) -int_downsample (j_compress_ptr cinfo, jpeg_component_info * compptr, - JSAMPARRAY input_data, JSAMPARRAY output_data) -{ - my_downsample_ptr downsample = (my_downsample_ptr) cinfo->downsample; - int inrow, outrow, h_expand, v_expand, numpix, numpix2, h, v; - JDIMENSION outcol, outcol_h; /* outcol_h == outcol*h_expand */ - JDIMENSION output_cols = compptr->width_in_blocks * compptr->DCT_h_scaled_size; - JSAMPROW inptr, outptr; - INT32 outvalue; - - h_expand = downsample->h_expand[compptr->component_index]; - v_expand = downsample->v_expand[compptr->component_index]; - numpix = h_expand * v_expand; - numpix2 = numpix/2; - - /* Expand input data enough to let all the output samples be generated - * by the standard loop. Special-casing padded output would be more - * efficient. - */ - expand_right_edge(input_data, cinfo->max_v_samp_factor, - cinfo->image_width, output_cols * h_expand); - - inrow = outrow = 0; - while (inrow < cinfo->max_v_samp_factor) { - outptr = output_data[outrow]; - for (outcol = 0, outcol_h = 0; outcol < output_cols; - outcol++, outcol_h += h_expand) { - outvalue = 0; - for (v = 0; v < v_expand; v++) { - inptr = input_data[inrow+v] + outcol_h; - for (h = 0; h < h_expand; h++) { - outvalue += (INT32) GETJSAMPLE(*inptr++); - } - } - *outptr++ = (JSAMPLE) ((outvalue + numpix2) / numpix); - } - inrow += v_expand; - outrow++; - } -} - - -/* - * Downsample pixel values of a single component. - * This version handles the special case of a full-size component, - * without smoothing. - */ - -METHODDEF(void) -fullsize_downsample (j_compress_ptr cinfo, jpeg_component_info * compptr, - JSAMPARRAY input_data, JSAMPARRAY output_data) -{ - /* Copy the data */ - jcopy_sample_rows(input_data, output_data, - cinfo->max_v_samp_factor, cinfo->image_width); - /* Edge-expand */ - expand_right_edge(output_data, cinfo->max_v_samp_factor, cinfo->image_width, - compptr->width_in_blocks * compptr->DCT_h_scaled_size); -} - - -/* - * Downsample pixel values of a single component. - * This version handles the common case of 2:1 horizontal and 1:1 vertical, - * without smoothing. - * - * A note about the "bias" calculations: when rounding fractional values to - * integer, we do not want to always round 0.5 up to the next integer. - * If we did that, we'd introduce a noticeable bias towards larger values. - * Instead, this code is arranged so that 0.5 will be rounded up or down at - * alternate pixel locations (a simple ordered dither pattern). - */ - -METHODDEF(void) -h2v1_downsample (j_compress_ptr cinfo, jpeg_component_info * compptr, - JSAMPARRAY input_data, JSAMPARRAY output_data) -{ - int inrow; - JDIMENSION outcol; - JDIMENSION output_cols = compptr->width_in_blocks * compptr->DCT_h_scaled_size; - register JSAMPROW inptr, outptr; - register int bias; - - /* Expand input data enough to let all the output samples be generated - * by the standard loop. Special-casing padded output would be more - * efficient. - */ - expand_right_edge(input_data, cinfo->max_v_samp_factor, - cinfo->image_width, output_cols * 2); - - for (inrow = 0; inrow < cinfo->max_v_samp_factor; inrow++) { - outptr = output_data[inrow]; - inptr = input_data[inrow]; - bias = 0; /* bias = 0,1,0,1,... for successive samples */ - for (outcol = 0; outcol < output_cols; outcol++) { - *outptr++ = (JSAMPLE) ((GETJSAMPLE(*inptr) + GETJSAMPLE(inptr[1]) - + bias) >> 1); - bias ^= 1; /* 0=>1, 1=>0 */ - inptr += 2; - } - } -} - - -/* - * Downsample pixel values of a single component. - * This version handles the standard case of 2:1 horizontal and 2:1 vertical, - * without smoothing. - */ - -METHODDEF(void) -h2v2_downsample (j_compress_ptr cinfo, jpeg_component_info * compptr, - JSAMPARRAY input_data, JSAMPARRAY output_data) -{ - int inrow, outrow; - JDIMENSION outcol; - JDIMENSION output_cols = compptr->width_in_blocks * compptr->DCT_h_scaled_size; - register JSAMPROW inptr0, inptr1, outptr; - register int bias; - - /* Expand input data enough to let all the output samples be generated - * by the standard loop. Special-casing padded output would be more - * efficient. - */ - expand_right_edge(input_data, cinfo->max_v_samp_factor, - cinfo->image_width, output_cols * 2); - - inrow = outrow = 0; - while (inrow < cinfo->max_v_samp_factor) { - outptr = output_data[outrow]; - inptr0 = input_data[inrow]; - inptr1 = input_data[inrow+1]; - bias = 1; /* bias = 1,2,1,2,... for successive samples */ - for (outcol = 0; outcol < output_cols; outcol++) { - *outptr++ = (JSAMPLE) ((GETJSAMPLE(*inptr0) + GETJSAMPLE(inptr0[1]) + - GETJSAMPLE(*inptr1) + GETJSAMPLE(inptr1[1]) - + bias) >> 2); - bias ^= 3; /* 1=>2, 2=>1 */ - inptr0 += 2; inptr1 += 2; - } - inrow += 2; - outrow++; - } -} - - -#ifdef INPUT_SMOOTHING_SUPPORTED - -/* - * Downsample pixel values of a single component. - * This version handles the standard case of 2:1 horizontal and 2:1 vertical, - * with smoothing. One row of context is required. - */ - -METHODDEF(void) -h2v2_smooth_downsample (j_compress_ptr cinfo, jpeg_component_info * compptr, - JSAMPARRAY input_data, JSAMPARRAY output_data) -{ - int inrow, outrow; - JDIMENSION colctr; - JDIMENSION output_cols = compptr->width_in_blocks * compptr->DCT_h_scaled_size; - register JSAMPROW inptr0, inptr1, above_ptr, below_ptr, outptr; - INT32 membersum, neighsum, memberscale, neighscale; - - /* Expand input data enough to let all the output samples be generated - * by the standard loop. Special-casing padded output would be more - * efficient. - */ - expand_right_edge(input_data - 1, cinfo->max_v_samp_factor + 2, - cinfo->image_width, output_cols * 2); - - /* We don't bother to form the individual "smoothed" input pixel values; - * we can directly compute the output which is the average of the four - * smoothed values. Each of the four member pixels contributes a fraction - * (1-8*SF) to its own smoothed image and a fraction SF to each of the three - * other smoothed pixels, therefore a total fraction (1-5*SF)/4 to the final - * output. The four corner-adjacent neighbor pixels contribute a fraction - * SF to just one smoothed pixel, or SF/4 to the final output; while the - * eight edge-adjacent neighbors contribute SF to each of two smoothed - * pixels, or SF/2 overall. In order to use integer arithmetic, these - * factors are scaled by 2^16 = 65536. - * Also recall that SF = smoothing_factor / 1024. - */ - - memberscale = 16384 - cinfo->smoothing_factor * 80; /* scaled (1-5*SF)/4 */ - neighscale = cinfo->smoothing_factor * 16; /* scaled SF/4 */ - - inrow = outrow = 0; - while (inrow < cinfo->max_v_samp_factor) { - outptr = output_data[outrow]; - inptr0 = input_data[inrow]; - inptr1 = input_data[inrow+1]; - above_ptr = input_data[inrow-1]; - below_ptr = input_data[inrow+2]; - - /* Special case for first column: pretend column -1 is same as column 0 */ - membersum = GETJSAMPLE(*inptr0) + GETJSAMPLE(inptr0[1]) + - GETJSAMPLE(*inptr1) + GETJSAMPLE(inptr1[1]); - neighsum = GETJSAMPLE(*above_ptr) + GETJSAMPLE(above_ptr[1]) + - GETJSAMPLE(*below_ptr) + GETJSAMPLE(below_ptr[1]) + - GETJSAMPLE(*inptr0) + GETJSAMPLE(inptr0[2]) + - GETJSAMPLE(*inptr1) + GETJSAMPLE(inptr1[2]); - neighsum += neighsum; - neighsum += GETJSAMPLE(*above_ptr) + GETJSAMPLE(above_ptr[2]) + - GETJSAMPLE(*below_ptr) + GETJSAMPLE(below_ptr[2]); - membersum = membersum * memberscale + neighsum * neighscale; - *outptr++ = (JSAMPLE) ((membersum + 32768) >> 16); - inptr0 += 2; inptr1 += 2; above_ptr += 2; below_ptr += 2; - - for (colctr = output_cols - 2; colctr > 0; colctr--) { - /* sum of pixels directly mapped to this output element */ - membersum = GETJSAMPLE(*inptr0) + GETJSAMPLE(inptr0[1]) + - GETJSAMPLE(*inptr1) + GETJSAMPLE(inptr1[1]); - /* sum of edge-neighbor pixels */ - neighsum = GETJSAMPLE(*above_ptr) + GETJSAMPLE(above_ptr[1]) + - GETJSAMPLE(*below_ptr) + GETJSAMPLE(below_ptr[1]) + - GETJSAMPLE(inptr0[-1]) + GETJSAMPLE(inptr0[2]) + - GETJSAMPLE(inptr1[-1]) + GETJSAMPLE(inptr1[2]); - /* The edge-neighbors count twice as much as corner-neighbors */ - neighsum += neighsum; - /* Add in the corner-neighbors */ - neighsum += GETJSAMPLE(above_ptr[-1]) + GETJSAMPLE(above_ptr[2]) + - GETJSAMPLE(below_ptr[-1]) + GETJSAMPLE(below_ptr[2]); - /* form final output scaled up by 2^16 */ - membersum = membersum * memberscale + neighsum * neighscale; - /* round, descale and output it */ - *outptr++ = (JSAMPLE) ((membersum + 32768) >> 16); - inptr0 += 2; inptr1 += 2; above_ptr += 2; below_ptr += 2; - } - - /* Special case for last column */ - membersum = GETJSAMPLE(*inptr0) + GETJSAMPLE(inptr0[1]) + - GETJSAMPLE(*inptr1) + GETJSAMPLE(inptr1[1]); - neighsum = GETJSAMPLE(*above_ptr) + GETJSAMPLE(above_ptr[1]) + - GETJSAMPLE(*below_ptr) + GETJSAMPLE(below_ptr[1]) + - GETJSAMPLE(inptr0[-1]) + GETJSAMPLE(inptr0[1]) + - GETJSAMPLE(inptr1[-1]) + GETJSAMPLE(inptr1[1]); - neighsum += neighsum; - neighsum += GETJSAMPLE(above_ptr[-1]) + GETJSAMPLE(above_ptr[1]) + - GETJSAMPLE(below_ptr[-1]) + GETJSAMPLE(below_ptr[1]); - membersum = membersum * memberscale + neighsum * neighscale; - *outptr = (JSAMPLE) ((membersum + 32768) >> 16); - - inrow += 2; - outrow++; - } -} - - -/* - * Downsample pixel values of a single component. - * This version handles the special case of a full-size component, - * with smoothing. One row of context is required. - */ - -METHODDEF(void) -fullsize_smooth_downsample (j_compress_ptr cinfo, jpeg_component_info *compptr, - JSAMPARRAY input_data, JSAMPARRAY output_data) -{ - int inrow; - JDIMENSION colctr; - JDIMENSION output_cols = compptr->width_in_blocks * compptr->DCT_h_scaled_size; - register JSAMPROW inptr, above_ptr, below_ptr, outptr; - INT32 membersum, neighsum, memberscale, neighscale; - int colsum, lastcolsum, nextcolsum; - - /* Expand input data enough to let all the output samples be generated - * by the standard loop. Special-casing padded output would be more - * efficient. - */ - expand_right_edge(input_data - 1, cinfo->max_v_samp_factor + 2, - cinfo->image_width, output_cols); - - /* Each of the eight neighbor pixels contributes a fraction SF to the - * smoothed pixel, while the main pixel contributes (1-8*SF). In order - * to use integer arithmetic, these factors are multiplied by 2^16 = 65536. - * Also recall that SF = smoothing_factor / 1024. - */ - - memberscale = 65536L - cinfo->smoothing_factor * 512L; /* scaled 1-8*SF */ - neighscale = cinfo->smoothing_factor * 64; /* scaled SF */ - - for (inrow = 0; inrow < cinfo->max_v_samp_factor; inrow++) { - outptr = output_data[inrow]; - inptr = input_data[inrow]; - above_ptr = input_data[inrow-1]; - below_ptr = input_data[inrow+1]; - - /* Special case for first column */ - colsum = GETJSAMPLE(*above_ptr++) + GETJSAMPLE(*below_ptr++) + - GETJSAMPLE(*inptr); - membersum = GETJSAMPLE(*inptr++); - nextcolsum = GETJSAMPLE(*above_ptr) + GETJSAMPLE(*below_ptr) + - GETJSAMPLE(*inptr); - neighsum = colsum + (colsum - membersum) + nextcolsum; - membersum = membersum * memberscale + neighsum * neighscale; - *outptr++ = (JSAMPLE) ((membersum + 32768) >> 16); - lastcolsum = colsum; colsum = nextcolsum; - - for (colctr = output_cols - 2; colctr > 0; colctr--) { - membersum = GETJSAMPLE(*inptr++); - above_ptr++; below_ptr++; - nextcolsum = GETJSAMPLE(*above_ptr) + GETJSAMPLE(*below_ptr) + - GETJSAMPLE(*inptr); - neighsum = lastcolsum + (colsum - membersum) + nextcolsum; - membersum = membersum * memberscale + neighsum * neighscale; - *outptr++ = (JSAMPLE) ((membersum + 32768) >> 16); - lastcolsum = colsum; colsum = nextcolsum; - } - - /* Special case for last column */ - membersum = GETJSAMPLE(*inptr); - neighsum = lastcolsum + (colsum - membersum) + colsum; - membersum = membersum * memberscale + neighsum * neighscale; - *outptr = (JSAMPLE) ((membersum + 32768) >> 16); - - } -} - -#endif /* INPUT_SMOOTHING_SUPPORTED */ - - -/* - * Module initialization routine for downsampling. - * Note that we must select a routine for each component. - */ - -GLOBAL(void) -jinit_downsampler (j_compress_ptr cinfo) -{ - my_downsample_ptr downsample; - int ci; - jpeg_component_info * compptr; - boolean smoothok = TRUE; - int h_in_group, v_in_group, h_out_group, v_out_group; - - downsample = (my_downsample_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_downsampler)); - cinfo->downsample = &downsample->pub; - downsample->pub.start_pass = start_pass_downsample; - downsample->pub.downsample = sep_downsample; - downsample->pub.need_context_rows = FALSE; - - if (cinfo->CCIR601_sampling) - ERREXIT(cinfo, JERR_CCIR601_NOTIMPL); - - /* Verify we can handle the sampling factors, and set up method pointers */ - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - /* Compute size of an "output group" for DCT scaling. This many samples - * are to be converted from max_h_samp_factor * max_v_samp_factor pixels. - */ - h_out_group = (compptr->h_samp_factor * compptr->DCT_h_scaled_size) / - cinfo->min_DCT_h_scaled_size; - v_out_group = (compptr->v_samp_factor * compptr->DCT_v_scaled_size) / - cinfo->min_DCT_v_scaled_size; - h_in_group = cinfo->max_h_samp_factor; - v_in_group = cinfo->max_v_samp_factor; - downsample->rowgroup_height[ci] = v_out_group; /* save for use later */ - if (h_in_group == h_out_group && v_in_group == v_out_group) { -#ifdef INPUT_SMOOTHING_SUPPORTED - if (cinfo->smoothing_factor) { - downsample->methods[ci] = fullsize_smooth_downsample; - downsample->pub.need_context_rows = TRUE; - } else -#endif - downsample->methods[ci] = fullsize_downsample; - } else if (h_in_group == h_out_group * 2 && - v_in_group == v_out_group) { - smoothok = FALSE; - downsample->methods[ci] = h2v1_downsample; - } else if (h_in_group == h_out_group * 2 && - v_in_group == v_out_group * 2) { -#ifdef INPUT_SMOOTHING_SUPPORTED - if (cinfo->smoothing_factor) { - downsample->methods[ci] = h2v2_smooth_downsample; - downsample->pub.need_context_rows = TRUE; - } else -#endif - downsample->methods[ci] = h2v2_downsample; - } else if ((h_in_group % h_out_group) == 0 && - (v_in_group % v_out_group) == 0) { - smoothok = FALSE; - downsample->methods[ci] = int_downsample; - downsample->h_expand[ci] = (UINT8) (h_in_group / h_out_group); - downsample->v_expand[ci] = (UINT8) (v_in_group / v_out_group); - } else - ERREXIT(cinfo, JERR_FRACT_SAMPLE_NOTIMPL); - } - -#ifdef INPUT_SMOOTHING_SUPPORTED - if (cinfo->smoothing_factor && !smoothok) - TRACEMS(cinfo, 0, JTRC_SMOOTH_NOTIMPL); -#endif -} diff --git a/cmake/externals/libjpeg/jctrans.c b/cmake/externals/libjpeg/jctrans.c deleted file mode 100644 index 261dd2996..000000000 --- a/cmake/externals/libjpeg/jctrans.c +++ /dev/null @@ -1,399 +0,0 @@ -/* - * jctrans.c - * - * Copyright (C) 1995-1998, Thomas G. Lane. - * Modified 2000-2020 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains library routines for transcoding compression, - * that is, writing raw DCT coefficient arrays to an output JPEG file. - * The routines in jcapimin.c will also be needed by a transcoder. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* Forward declarations */ -LOCAL(void) transencode_master_selection - JPP((j_compress_ptr cinfo, jvirt_barray_ptr * coef_arrays)); -LOCAL(void) transencode_coef_controller - JPP((j_compress_ptr cinfo, jvirt_barray_ptr * coef_arrays)); - - -/* - * Compression initialization for writing raw-coefficient data. - * Before calling this, all parameters and a data destination must be set up. - * Call jpeg_finish_compress() to actually write the data. - * - * The number of passed virtual arrays must match cinfo->num_components. - * Note that the virtual arrays need not be filled or even realized at - * the time write_coefficients is called; indeed, if the virtual arrays - * were requested from this compression object's memory manager, they - * typically will be realized during this routine and filled afterwards. - */ - -GLOBAL(void) -jpeg_write_coefficients (j_compress_ptr cinfo, jvirt_barray_ptr * coef_arrays) -{ - if (cinfo->global_state != CSTATE_START) - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - /* Mark all tables to be written */ - jpeg_suppress_tables(cinfo, FALSE); - /* (Re)initialize error mgr and destination modules */ - (*cinfo->err->reset_error_mgr) ((j_common_ptr) cinfo); - (*cinfo->dest->init_destination) (cinfo); - /* Perform master selection of active modules */ - transencode_master_selection(cinfo, coef_arrays); - /* Wait for jpeg_finish_compress() call */ - cinfo->next_scanline = 0; /* so jpeg_write_marker works */ - cinfo->global_state = CSTATE_WRCOEFS; -} - - -/* - * Initialize the compression object with default parameters, - * then copy from the source object all parameters needed for lossless - * transcoding. Parameters that can be varied without loss (such as - * scan script and Huffman optimization) are left in their default states. - */ - -GLOBAL(void) -jpeg_copy_critical_parameters (j_decompress_ptr srcinfo, - j_compress_ptr dstinfo) -{ - JQUANT_TBL ** qtblptr; - jpeg_component_info *incomp, *outcomp; - JQUANT_TBL *c_quant, *slot_quant; - int tblno, ci, coefi; - - /* Safety check to ensure start_compress not called yet. */ - if (dstinfo->global_state != CSTATE_START) - ERREXIT1(dstinfo, JERR_BAD_STATE, dstinfo->global_state); - /* Copy fundamental image dimensions */ - dstinfo->image_width = srcinfo->image_width; - dstinfo->image_height = srcinfo->image_height; - dstinfo->input_components = srcinfo->num_components; - dstinfo->in_color_space = srcinfo->jpeg_color_space; - dstinfo->jpeg_width = srcinfo->output_width; - dstinfo->jpeg_height = srcinfo->output_height; - dstinfo->min_DCT_h_scaled_size = srcinfo->min_DCT_h_scaled_size; - dstinfo->min_DCT_v_scaled_size = srcinfo->min_DCT_v_scaled_size; - /* Initialize all parameters to default values */ - jpeg_set_defaults(dstinfo); - /* jpeg_set_defaults may choose wrong colorspace, eg YCbCr if input is RGB. - * Fix it to get the right header markers for the image colorspace. - * Note: Entropy table assignment in jpeg_set_colorspace - * depends on color_transform. - * Adaption is also required for setting the appropriate - * entropy coding mode dependent on image data precision. - */ - dstinfo->color_transform = srcinfo->color_transform; - jpeg_set_colorspace(dstinfo, srcinfo->jpeg_color_space); - dstinfo->data_precision = srcinfo->data_precision; - dstinfo->arith_code = srcinfo->data_precision > 8 ? TRUE : FALSE; - dstinfo->CCIR601_sampling = srcinfo->CCIR601_sampling; - /* Copy the source's quantization tables. */ - for (tblno = 0; tblno < NUM_QUANT_TBLS; tblno++) { - if (srcinfo->quant_tbl_ptrs[tblno] != NULL) { - qtblptr = & dstinfo->quant_tbl_ptrs[tblno]; - if (*qtblptr == NULL) - *qtblptr = jpeg_alloc_quant_table((j_common_ptr) dstinfo); - MEMCOPY((*qtblptr)->quantval, - srcinfo->quant_tbl_ptrs[tblno]->quantval, - SIZEOF((*qtblptr)->quantval)); - (*qtblptr)->sent_table = FALSE; - } - } - /* Copy the source's per-component info. - * Note we assume jpeg_set_defaults has allocated the dest comp_info array. - */ - dstinfo->num_components = srcinfo->num_components; - if (dstinfo->num_components < 1 || dstinfo->num_components > MAX_COMPONENTS) - ERREXIT2(dstinfo, JERR_COMPONENT_COUNT, dstinfo->num_components, - MAX_COMPONENTS); - for (ci = 0, incomp = srcinfo->comp_info, outcomp = dstinfo->comp_info; - ci < dstinfo->num_components; ci++, incomp++, outcomp++) { - outcomp->component_id = incomp->component_id; - outcomp->h_samp_factor = incomp->h_samp_factor; - outcomp->v_samp_factor = incomp->v_samp_factor; - outcomp->quant_tbl_no = incomp->quant_tbl_no; - /* Make sure saved quantization table for component matches the qtable - * slot. If not, the input file re-used this qtable slot. - * IJG encoder currently cannot duplicate this. - */ - tblno = outcomp->quant_tbl_no; - if (tblno < 0 || tblno >= NUM_QUANT_TBLS || - srcinfo->quant_tbl_ptrs[tblno] == NULL) - ERREXIT1(dstinfo, JERR_NO_QUANT_TABLE, tblno); - slot_quant = srcinfo->quant_tbl_ptrs[tblno]; - c_quant = incomp->quant_table; - if (c_quant != NULL) { - for (coefi = 0; coefi < DCTSIZE2; coefi++) { - if (c_quant->quantval[coefi] != slot_quant->quantval[coefi]) - ERREXIT1(dstinfo, JERR_MISMATCHED_QUANT_TABLE, tblno); - } - } - /* Note: we do not copy the source's entropy table assignments; - * instead we rely on jpeg_set_colorspace to have made a suitable choice. - */ - } - /* Also copy JFIF version and resolution information, if available. - * Strictly speaking this isn't "critical" info, but it's nearly - * always appropriate to copy it if available. In particular, - * if the application chooses to copy JFIF 1.02 extension markers from - * the source file, we need to copy the version to make sure we don't - * emit a file that has 1.02 extensions but a claimed version of 1.01. - */ - if (srcinfo->saw_JFIF_marker) { - if (srcinfo->JFIF_major_version == 1 || - srcinfo->JFIF_major_version == 2) { - dstinfo->JFIF_major_version = srcinfo->JFIF_major_version; - dstinfo->JFIF_minor_version = srcinfo->JFIF_minor_version; - } - dstinfo->density_unit = srcinfo->density_unit; - dstinfo->X_density = srcinfo->X_density; - dstinfo->Y_density = srcinfo->Y_density; - } -} - - -LOCAL(void) -jpeg_calc_trans_dimensions (j_compress_ptr cinfo) -/* Do computations that are needed before master selection phase */ -{ - if (cinfo->min_DCT_h_scaled_size != cinfo->min_DCT_v_scaled_size) - ERREXIT2(cinfo, JERR_BAD_DCTSIZE, - cinfo->min_DCT_h_scaled_size, cinfo->min_DCT_v_scaled_size); - - cinfo->block_size = cinfo->min_DCT_h_scaled_size; -} - - -/* - * Master selection of compression modules for transcoding. - * This substitutes for jcinit.c's initialization of the full compressor. - */ - -LOCAL(void) -transencode_master_selection (j_compress_ptr cinfo, - jvirt_barray_ptr * coef_arrays) -{ - /* Do computations that are needed before master selection phase */ - jpeg_calc_trans_dimensions(cinfo); - - /* Initialize master control (includes parameter checking/processing) */ - jinit_c_master_control(cinfo, TRUE /* transcode only */); - - /* Entropy encoding: either Huffman or arithmetic coding. */ - if (cinfo->arith_code) - jinit_arith_encoder(cinfo); - else { - jinit_huff_encoder(cinfo); - } - - /* We need a special coefficient buffer controller. */ - transencode_coef_controller(cinfo, coef_arrays); - - jinit_marker_writer(cinfo); - - /* We can now tell the memory manager to allocate virtual arrays. */ - (*cinfo->mem->realize_virt_arrays) ((j_common_ptr) cinfo); - - /* Write the datastream header (SOI, JFIF) immediately. - * Frame and scan headers are postponed till later. - * This lets application insert special markers after the SOI. - */ - (*cinfo->marker->write_file_header) (cinfo); -} - - -/* - * The rest of this file is a special implementation of the coefficient - * buffer controller. This is similar to jccoefct.c, but it handles only - * output from presupplied virtual arrays. Furthermore, we generate any - * dummy padding blocks on-the-fly rather than expecting them to be present - * in the arrays. - */ - -/* Private buffer controller object */ - -typedef struct { - struct jpeg_c_coef_controller pub; /* public fields */ - - JDIMENSION iMCU_row_num; /* iMCU row # within image */ - JDIMENSION MCU_ctr; /* counts MCUs processed in current row */ - int MCU_vert_offset; /* counts MCU rows within iMCU row */ - int MCU_rows_per_iMCU_row; /* number of such rows needed */ - - /* Virtual block array for each component. */ - jvirt_barray_ptr * whole_image; - - /* Workspace for constructing dummy blocks at right/bottom edges. */ - JBLOCK dummy_buffer[C_MAX_BLOCKS_IN_MCU]; -} my_coef_controller; - -typedef my_coef_controller * my_coef_ptr; - - -LOCAL(void) -start_iMCU_row (j_compress_ptr cinfo) -/* Reset within-iMCU-row counters for a new row */ -{ - my_coef_ptr coef = (my_coef_ptr) cinfo->coef; - - /* In an interleaved scan, an MCU row is the same as an iMCU row. - * In a noninterleaved scan, an iMCU row has v_samp_factor MCU rows. - * But at the bottom of the image, process only what's left. - */ - if (cinfo->comps_in_scan > 1) { - coef->MCU_rows_per_iMCU_row = 1; - } else { - if (coef->iMCU_row_num < (cinfo->total_iMCU_rows-1)) - coef->MCU_rows_per_iMCU_row = cinfo->cur_comp_info[0]->v_samp_factor; - else - coef->MCU_rows_per_iMCU_row = cinfo->cur_comp_info[0]->last_row_height; - } - - coef->MCU_ctr = 0; - coef->MCU_vert_offset = 0; -} - - -/* - * Initialize for a processing pass. - */ - -METHODDEF(void) -start_pass_coef (j_compress_ptr cinfo, J_BUF_MODE pass_mode) -{ - my_coef_ptr coef = (my_coef_ptr) cinfo->coef; - - if (pass_mode != JBUF_CRANK_DEST) - ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); - - coef->iMCU_row_num = 0; - start_iMCU_row(cinfo); -} - - -/* - * Process some data. - * We process the equivalent of one fully interleaved MCU row ("iMCU" row) - * per call, ie, v_samp_factor block rows for each component in the scan. - * The data is obtained from the virtual arrays and fed to the entropy coder. - * Returns TRUE if the iMCU row is completed, FALSE if suspended. - * - * NB: input_buf is ignored; it is likely to be a NULL pointer. - */ - -METHODDEF(boolean) -compress_output (j_compress_ptr cinfo, JSAMPIMAGE input_buf) -{ - my_coef_ptr coef = (my_coef_ptr) cinfo->coef; - JDIMENSION MCU_col_num; /* index of current MCU within row */ - JDIMENSION last_MCU_col = cinfo->MCUs_per_row - 1; - JDIMENSION last_iMCU_row = cinfo->total_iMCU_rows - 1; - int blkn, ci, xindex, yindex, yoffset, blockcnt; - JDIMENSION start_col; - JBLOCKARRAY buffer[MAX_COMPS_IN_SCAN]; - JBLOCKROW MCU_buffer[C_MAX_BLOCKS_IN_MCU]; - JBLOCKROW buffer_ptr; - jpeg_component_info *compptr; - - /* Align the virtual buffers for the components used in this scan. */ - for (ci = 0; ci < cinfo->comps_in_scan; ci++) { - compptr = cinfo->cur_comp_info[ci]; - buffer[ci] = (*cinfo->mem->access_virt_barray) - ((j_common_ptr) cinfo, coef->whole_image[compptr->component_index], - coef->iMCU_row_num * compptr->v_samp_factor, - (JDIMENSION) compptr->v_samp_factor, FALSE); - } - - /* Loop to process one whole iMCU row */ - for (yoffset = coef->MCU_vert_offset; yoffset < coef->MCU_rows_per_iMCU_row; - yoffset++) { - for (MCU_col_num = coef->MCU_ctr; MCU_col_num <= last_MCU_col; - MCU_col_num++) { - /* Construct list of pointers to DCT blocks belonging to this MCU */ - blkn = 0; /* index of current DCT block within MCU */ - for (ci = 0; ci < cinfo->comps_in_scan; ci++) { - compptr = cinfo->cur_comp_info[ci]; - blockcnt = (MCU_col_num < last_MCU_col) ? compptr->MCU_width - : compptr->last_col_width; - start_col = MCU_col_num * compptr->MCU_width; - for (yindex = 0; yindex < compptr->MCU_height; yindex++) { - if (coef->iMCU_row_num < last_iMCU_row || - yoffset + yindex < compptr->last_row_height) { - /* Fill in pointers to real blocks in this row */ - buffer_ptr = buffer[ci][yoffset + yindex] + start_col; - xindex = blockcnt; - do { - MCU_buffer[blkn++] = buffer_ptr++; - } while (--xindex); - /* Dummy blocks at right edge */ - if ((xindex = compptr->MCU_width - blockcnt) == 0) - continue; - } else { - /* At bottom of image, need a whole row of dummy blocks */ - xindex = compptr->MCU_width; - } - /* Fill in any dummy blocks needed in this row. - * Dummy blocks are filled in the same way as in jccoefct.c: - * all zeroes in the AC entries, DC entries equal to previous - * block's DC value. The init routine has already zeroed the - * AC entries, so we need only set the DC entries correctly. - */ - buffer_ptr = coef->dummy_buffer + blkn; - do { - buffer_ptr[0][0] = MCU_buffer[blkn-1][0][0]; - MCU_buffer[blkn++] = buffer_ptr++; - } while (--xindex); - } - } - /* Try to write the MCU. */ - if (! (*cinfo->entropy->encode_mcu) (cinfo, MCU_buffer)) { - /* Suspension forced; update state counters and exit */ - coef->MCU_vert_offset = yoffset; - coef->MCU_ctr = MCU_col_num; - return FALSE; - } - } - /* Completed an MCU row, but perhaps not an iMCU row */ - coef->MCU_ctr = 0; - } - /* Completed the iMCU row, advance counters for next one */ - coef->iMCU_row_num++; - start_iMCU_row(cinfo); - return TRUE; -} - - -/* - * Initialize coefficient buffer controller. - * - * Each passed coefficient array must be the right size for that - * coefficient: width_in_blocks wide and height_in_blocks high, - * with unitheight at least v_samp_factor. - */ - -LOCAL(void) -transencode_coef_controller (j_compress_ptr cinfo, - jvirt_barray_ptr * coef_arrays) -{ - my_coef_ptr coef; - - coef = (my_coef_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_coef_controller)); - cinfo->coef = &coef->pub; - coef->pub.start_pass = start_pass_coef; - coef->pub.compress_data = compress_output; - - /* Save pointer to virtual arrays */ - coef->whole_image = coef_arrays; - - /* Pre-zero space for dummy DCT blocks */ - MEMZERO(coef->dummy_buffer, SIZEOF(coef->dummy_buffer)); -} diff --git a/cmake/externals/libjpeg/jdapimin.c b/cmake/externals/libjpeg/jdapimin.c deleted file mode 100644 index 785e52722..000000000 --- a/cmake/externals/libjpeg/jdapimin.c +++ /dev/null @@ -1,412 +0,0 @@ -/* - * jdapimin.c - * - * Copyright (C) 1994-1998, Thomas G. Lane. - * Modified 2009-2020 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains application interface code for the decompression half - * of the JPEG library. These are the "minimum" API routines that may be - * needed in either the normal full-decompression case or the - * transcoding-only case. - * - * Most of the routines intended to be called directly by an application - * are in this file or in jdapistd.c. But also see jcomapi.c for routines - * shared by compression and decompression, and jdtrans.c for the transcoding - * case. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* - * Initialization of a JPEG decompression object. - * The error manager must already be set up (in case memory manager fails). - */ - -GLOBAL(void) -jpeg_CreateDecompress (j_decompress_ptr cinfo, int version, size_t structsize) -{ - int i; - - /* Guard against version mismatches between library and caller. */ - cinfo->mem = NULL; /* so jpeg_destroy knows mem mgr not called */ - if (version != JPEG_LIB_VERSION) - ERREXIT2(cinfo, JERR_BAD_LIB_VERSION, JPEG_LIB_VERSION, version); - if (structsize != SIZEOF(struct jpeg_decompress_struct)) - ERREXIT2(cinfo, JERR_BAD_STRUCT_SIZE, - (int) SIZEOF(struct jpeg_decompress_struct), (int) structsize); - - /* For debugging purposes, we zero the whole master structure. - * But the application has already set the err pointer, and may have set - * client_data, so we have to save and restore those fields. - * Note: if application hasn't set client_data, tools like Purify may - * complain here. - */ - { - struct jpeg_error_mgr * err = cinfo->err; - void * client_data = cinfo->client_data; /* ignore Purify complaint here */ - MEMZERO(cinfo, SIZEOF(struct jpeg_decompress_struct)); - cinfo->err = err; - cinfo->client_data = client_data; - } - cinfo->is_decompressor = TRUE; - - /* Initialize a memory manager instance for this object */ - jinit_memory_mgr((j_common_ptr) cinfo); - - /* Zero out pointers to permanent structures. */ - cinfo->progress = NULL; - cinfo->src = NULL; - - for (i = 0; i < NUM_QUANT_TBLS; i++) - cinfo->quant_tbl_ptrs[i] = NULL; - - for (i = 0; i < NUM_HUFF_TBLS; i++) { - cinfo->dc_huff_tbl_ptrs[i] = NULL; - cinfo->ac_huff_tbl_ptrs[i] = NULL; - } - - /* Initialize marker processor so application can override methods - * for COM, APPn markers before calling jpeg_read_header. - */ - cinfo->marker_list = NULL; - jinit_marker_reader(cinfo); - - /* And initialize the overall input controller. */ - jinit_input_controller(cinfo); - - /* OK, I'm ready */ - cinfo->global_state = DSTATE_START; -} - - -/* - * Destruction of a JPEG decompression object - */ - -GLOBAL(void) -jpeg_destroy_decompress (j_decompress_ptr cinfo) -{ - jpeg_destroy((j_common_ptr) cinfo); /* use common routine */ -} - - -/* - * Abort processing of a JPEG decompression operation, - * but don't destroy the object itself. - */ - -GLOBAL(void) -jpeg_abort_decompress (j_decompress_ptr cinfo) -{ - jpeg_abort((j_common_ptr) cinfo); /* use common routine */ -} - - -/* - * Set default decompression parameters. - */ - -LOCAL(void) -default_decompress_parms (j_decompress_ptr cinfo) -{ - int cid0, cid1, cid2, cid3; - - /* Guess the input colorspace, and set output colorspace accordingly. */ - /* Note application may override our guesses. */ - switch (cinfo->num_components) { - case 1: - cinfo->jpeg_color_space = JCS_GRAYSCALE; - cinfo->out_color_space = JCS_GRAYSCALE; - break; - - case 3: - cid0 = cinfo->comp_info[0].component_id; - cid1 = cinfo->comp_info[1].component_id; - cid2 = cinfo->comp_info[2].component_id; - - /* For robust detection of standard colorspaces - * regardless of the presence of special markers, - * check component IDs from SOF marker first. - */ - if (cid0 == 0x01 && cid1 == 0x02 && cid2 == 0x03) - cinfo->jpeg_color_space = JCS_YCbCr; - else if (cid0 == 0x01 && cid1 == 0x22 && cid2 == 0x23) - cinfo->jpeg_color_space = JCS_BG_YCC; - else if (cid0 == 0x52 && cid1 == 0x47 && cid2 == 0x42) - cinfo->jpeg_color_space = JCS_RGB; /* ASCII 'R', 'G', 'B' */ - else if (cid0 == 0x72 && cid1 == 0x67 && cid2 == 0x62) - cinfo->jpeg_color_space = JCS_BG_RGB; /* ASCII 'r', 'g', 'b' */ - else if (cinfo->saw_JFIF_marker) - cinfo->jpeg_color_space = JCS_YCbCr; /* assume it's YCbCr */ - else if (cinfo->saw_Adobe_marker) { - switch (cinfo->Adobe_transform) { - case 0: - cinfo->jpeg_color_space = JCS_RGB; - break; - case 1: - cinfo->jpeg_color_space = JCS_YCbCr; - break; - default: - WARNMS1(cinfo, JWRN_ADOBE_XFORM, cinfo->Adobe_transform); - cinfo->jpeg_color_space = JCS_YCbCr; /* assume it's YCbCr */ - } - } else { - TRACEMS3(cinfo, 1, JTRC_UNKNOWN_IDS, cid0, cid1, cid2); - cinfo->jpeg_color_space = JCS_YCbCr; /* assume it's YCbCr */ - } - /* Always guess RGB is proper output colorspace. */ - cinfo->out_color_space = JCS_RGB; - break; - - case 4: - cid0 = cinfo->comp_info[0].component_id; - cid1 = cinfo->comp_info[1].component_id; - cid2 = cinfo->comp_info[2].component_id; - cid3 = cinfo->comp_info[3].component_id; - - /* For robust detection of standard colorspaces - * regardless of the presence of special markers, - * check component IDs from SOF marker first. - */ - if (cid0 == 0x01 && cid1 == 0x02 && cid2 == 0x03 && cid3 == 0x04) - cinfo->jpeg_color_space = JCS_YCCK; - else if (cid0 == 0x43 && cid1 == 0x4D && cid2 == 0x59 && cid3 == 0x4B) - cinfo->jpeg_color_space = JCS_CMYK; /* ASCII 'C', 'M', 'Y', 'K' */ - else if (cinfo->saw_Adobe_marker) { - switch (cinfo->Adobe_transform) { - case 0: - cinfo->jpeg_color_space = JCS_CMYK; - break; - case 2: - cinfo->jpeg_color_space = JCS_YCCK; - break; - default: - WARNMS1(cinfo, JWRN_ADOBE_XFORM, cinfo->Adobe_transform); - cinfo->jpeg_color_space = JCS_YCCK; /* assume it's YCCK */ - } - } else { - /* Unknown IDs and no special markers, assume straight CMYK. */ - cinfo->jpeg_color_space = JCS_CMYK; - } - cinfo->out_color_space = JCS_CMYK; - break; - - default: - cinfo->jpeg_color_space = JCS_UNKNOWN; - cinfo->out_color_space = JCS_UNKNOWN; - } - - /* Set defaults for other decompression parameters. */ - cinfo->scale_num = cinfo->block_size; /* 1:1 scaling */ - cinfo->scale_denom = cinfo->block_size; - cinfo->output_gamma = 1.0; - cinfo->buffered_image = FALSE; - cinfo->raw_data_out = FALSE; - cinfo->dct_method = JDCT_DEFAULT; - cinfo->do_fancy_upsampling = TRUE; - cinfo->do_block_smoothing = TRUE; - cinfo->quantize_colors = FALSE; - /* We set these in case application only sets quantize_colors. */ - cinfo->dither_mode = JDITHER_FS; -#ifdef QUANT_2PASS_SUPPORTED - cinfo->two_pass_quantize = TRUE; -#else - cinfo->two_pass_quantize = FALSE; -#endif - cinfo->desired_number_of_colors = 256; - cinfo->colormap = NULL; - /* Initialize for no mode change in buffered-image mode. */ - cinfo->enable_1pass_quant = FALSE; - cinfo->enable_external_quant = FALSE; - cinfo->enable_2pass_quant = FALSE; -} - - -/* - * Decompression startup: read start of JPEG datastream to see what's there. - * Need only initialize JPEG object and supply a data source before calling. - * - * This routine will read as far as the first SOS marker (ie, actual start of - * compressed data), and will save all tables and parameters in the JPEG - * object. It will also initialize the decompression parameters to default - * values, and finally return JPEG_HEADER_OK. On return, the application may - * adjust the decompression parameters and then call jpeg_start_decompress. - * (Or, if the application only wanted to determine the image parameters, - * the data need not be decompressed. In that case, call jpeg_abort or - * jpeg_destroy to release any temporary space.) - * If an abbreviated (tables only) datastream is presented, the routine will - * return JPEG_HEADER_TABLES_ONLY upon reaching EOI. The application may then - * re-use the JPEG object to read the abbreviated image datastream(s). - * It is unnecessary (but OK) to call jpeg_abort in this case. - * The JPEG_SUSPENDED return code only occurs if the data source module - * requests suspension of the decompressor. In this case the application - * should load more source data and then re-call jpeg_read_header to resume - * processing. - * If a non-suspending data source is used and require_image is TRUE, then the - * return code need not be inspected since only JPEG_HEADER_OK is possible. - * - * This routine is now just a front end to jpeg_consume_input, with some - * extra error checking. - */ - -GLOBAL(int) -jpeg_read_header (j_decompress_ptr cinfo, boolean require_image) -{ - int retcode; - - if (cinfo->global_state != DSTATE_START && - cinfo->global_state != DSTATE_INHEADER) - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - - retcode = jpeg_consume_input(cinfo); - - switch (retcode) { - case JPEG_REACHED_SOS: - retcode = JPEG_HEADER_OK; - break; - case JPEG_REACHED_EOI: - if (require_image) /* Complain if application wanted an image */ - ERREXIT(cinfo, JERR_NO_IMAGE); - /* Reset to start state; it would be safer to require the application to - * call jpeg_abort, but we can't change it now for compatibility reasons. - * A side effect is to free any temporary memory (there shouldn't be any). - */ - jpeg_abort((j_common_ptr) cinfo); /* sets state = DSTATE_START */ - retcode = JPEG_HEADER_TABLES_ONLY; - break; - case JPEG_SUSPENDED: - /* no work */ - break; - } - - return retcode; -} - - -/* - * Consume data in advance of what the decompressor requires. - * This can be called at any time once the decompressor object has - * been created and a data source has been set up. - * - * This routine is essentially a state machine that handles a couple - * of critical state-transition actions, namely initial setup and - * transition from header scanning to ready-for-start_decompress. - * All the actual input is done via the input controller's consume_input - * method. - */ - -GLOBAL(int) -jpeg_consume_input (j_decompress_ptr cinfo) -{ - int retcode = JPEG_SUSPENDED; - - /* NB: every possible DSTATE value should be listed in this switch */ - switch (cinfo->global_state) { - case DSTATE_START: - /* Start-of-datastream actions: reset appropriate modules */ - (*cinfo->inputctl->reset_input_controller) (cinfo); - /* Initialize application's data source module */ - (*cinfo->src->init_source) (cinfo); - cinfo->global_state = DSTATE_INHEADER; - /*FALLTHROUGH*/ - case DSTATE_INHEADER: - retcode = (*cinfo->inputctl->consume_input) (cinfo); - if (retcode == JPEG_REACHED_SOS) { /* Found SOS, prepare to decompress */ - /* Set up default parameters based on header data */ - default_decompress_parms(cinfo); - /* Set global state: ready for start_decompress */ - cinfo->global_state = DSTATE_READY; - } - break; - case DSTATE_READY: - /* Can't advance past first SOS until start_decompress is called */ - retcode = JPEG_REACHED_SOS; - break; - case DSTATE_PRELOAD: - case DSTATE_PRESCAN: - case DSTATE_SCANNING: - case DSTATE_RAW_OK: - case DSTATE_BUFIMAGE: - case DSTATE_BUFPOST: - case DSTATE_STOPPING: - retcode = (*cinfo->inputctl->consume_input) (cinfo); - break; - default: - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - } - return retcode; -} - - -/* - * Have we finished reading the input file? - */ - -GLOBAL(boolean) -jpeg_input_complete (j_decompress_ptr cinfo) -{ - /* Check for valid jpeg object */ - if (cinfo->global_state < DSTATE_START || - cinfo->global_state > DSTATE_STOPPING) - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - return cinfo->inputctl->eoi_reached; -} - - -/* - * Is there more than one scan? - */ - -GLOBAL(boolean) -jpeg_has_multiple_scans (j_decompress_ptr cinfo) -{ - /* Only valid after jpeg_read_header completes */ - if (cinfo->global_state < DSTATE_READY || - cinfo->global_state > DSTATE_STOPPING) - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - return cinfo->inputctl->has_multiple_scans; -} - - -/* - * Finish JPEG decompression. - * - * This will normally just verify the file trailer and release temp storage. - * - * Returns FALSE if suspended. The return value need be inspected only if - * a suspending data source is used. - */ - -GLOBAL(boolean) -jpeg_finish_decompress (j_decompress_ptr cinfo) -{ - if ((cinfo->global_state == DSTATE_SCANNING || - cinfo->global_state == DSTATE_RAW_OK) && ! cinfo->buffered_image) { - /* Terminate final pass of non-buffered mode */ - if (cinfo->output_scanline < cinfo->output_height) - ERREXIT(cinfo, JERR_TOO_LITTLE_DATA); - (*cinfo->master->finish_output_pass) (cinfo); - cinfo->global_state = DSTATE_STOPPING; - } else if (cinfo->global_state == DSTATE_BUFIMAGE) { - /* Finishing after a buffered-image operation */ - cinfo->global_state = DSTATE_STOPPING; - } else if (cinfo->global_state != DSTATE_STOPPING) { - /* STOPPING = repeat call after a suspension, anything else is error */ - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - } - /* Read until EOI */ - while (! cinfo->inputctl->eoi_reached) { - if ((*cinfo->inputctl->consume_input) (cinfo) == JPEG_SUSPENDED) - return FALSE; /* Suspend, come back later */ - } - /* Do final cleanup */ - (*cinfo->src->term_source) (cinfo); - /* We can use jpeg_abort to release memory and reset global_state */ - jpeg_abort((j_common_ptr) cinfo); - return TRUE; -} diff --git a/cmake/externals/libjpeg/jdapistd.c b/cmake/externals/libjpeg/jdapistd.c deleted file mode 100644 index 7f3a78b25..000000000 --- a/cmake/externals/libjpeg/jdapistd.c +++ /dev/null @@ -1,276 +0,0 @@ -/* - * jdapistd.c - * - * Copyright (C) 1994-1996, Thomas G. Lane. - * Modified 2002-2013 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains application interface code for the decompression half - * of the JPEG library. These are the "standard" API routines that are - * used in the normal full-decompression case. They are not used by a - * transcoding-only application. Note that if an application links in - * jpeg_start_decompress, it will end up linking in the entire decompressor. - * We thus must separate this file from jdapimin.c to avoid linking the - * whole decompression library into a transcoder. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* Forward declarations */ -LOCAL(boolean) output_pass_setup JPP((j_decompress_ptr cinfo)); - - -/* - * Decompression initialization. - * jpeg_read_header must be completed before calling this. - * - * If a multipass operating mode was selected, this will do all but the - * last pass, and thus may take a great deal of time. - * - * Returns FALSE if suspended. The return value need be inspected only if - * a suspending data source is used. - */ - -GLOBAL(boolean) -jpeg_start_decompress (j_decompress_ptr cinfo) -{ - if (cinfo->global_state == DSTATE_READY) { - /* First call: initialize master control, select active modules */ - jinit_master_decompress(cinfo); - if (cinfo->buffered_image) { - /* No more work here; expecting jpeg_start_output next */ - cinfo->global_state = DSTATE_BUFIMAGE; - return TRUE; - } - cinfo->global_state = DSTATE_PRELOAD; - } - if (cinfo->global_state == DSTATE_PRELOAD) { - /* If file has multiple scans, absorb them all into the coef buffer */ - if (cinfo->inputctl->has_multiple_scans) { -#ifdef D_MULTISCAN_FILES_SUPPORTED - for (;;) { - int retcode; - /* Call progress monitor hook if present */ - if (cinfo->progress != NULL) - (*cinfo->progress->progress_monitor) ((j_common_ptr) cinfo); - /* Absorb some more input */ - retcode = (*cinfo->inputctl->consume_input) (cinfo); - if (retcode == JPEG_SUSPENDED) - return FALSE; - if (retcode == JPEG_REACHED_EOI) - break; - /* Advance progress counter if appropriate */ - if (cinfo->progress != NULL && - (retcode == JPEG_ROW_COMPLETED || retcode == JPEG_REACHED_SOS)) { - if (++cinfo->progress->pass_counter >= cinfo->progress->pass_limit) { - /* jdmaster underestimated number of scans; ratchet up one scan */ - cinfo->progress->pass_limit += (long) cinfo->total_iMCU_rows; - } - } - } -#else - ERREXIT(cinfo, JERR_NOT_COMPILED); -#endif /* D_MULTISCAN_FILES_SUPPORTED */ - } - cinfo->output_scan_number = cinfo->input_scan_number; - } else if (cinfo->global_state != DSTATE_PRESCAN) - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - /* Perform any dummy output passes, and set up for the final pass */ - return output_pass_setup(cinfo); -} - - -/* - * Set up for an output pass, and perform any dummy pass(es) needed. - * Common subroutine for jpeg_start_decompress and jpeg_start_output. - * Entry: global_state = DSTATE_PRESCAN only if previously suspended. - * Exit: If done, returns TRUE and sets global_state for proper output mode. - * If suspended, returns FALSE and sets global_state = DSTATE_PRESCAN. - */ - -LOCAL(boolean) -output_pass_setup (j_decompress_ptr cinfo) -{ - if (cinfo->global_state != DSTATE_PRESCAN) { - /* First call: do pass setup */ - (*cinfo->master->prepare_for_output_pass) (cinfo); - cinfo->output_scanline = 0; - cinfo->global_state = DSTATE_PRESCAN; - } - /* Loop over any required dummy passes */ - while (cinfo->master->is_dummy_pass) { -#ifdef QUANT_2PASS_SUPPORTED - /* Crank through the dummy pass */ - while (cinfo->output_scanline < cinfo->output_height) { - JDIMENSION last_scanline; - /* Call progress monitor hook if present */ - if (cinfo->progress != NULL) { - cinfo->progress->pass_counter = (long) cinfo->output_scanline; - cinfo->progress->pass_limit = (long) cinfo->output_height; - (*cinfo->progress->progress_monitor) ((j_common_ptr) cinfo); - } - /* Process some data */ - last_scanline = cinfo->output_scanline; - (*cinfo->main->process_data) (cinfo, (JSAMPARRAY) NULL, - &cinfo->output_scanline, (JDIMENSION) 0); - if (cinfo->output_scanline == last_scanline) - return FALSE; /* No progress made, must suspend */ - } - /* Finish up dummy pass, and set up for another one */ - (*cinfo->master->finish_output_pass) (cinfo); - (*cinfo->master->prepare_for_output_pass) (cinfo); - cinfo->output_scanline = 0; -#else - ERREXIT(cinfo, JERR_NOT_COMPILED); -#endif /* QUANT_2PASS_SUPPORTED */ - } - /* Ready for application to drive output pass through - * jpeg_read_scanlines or jpeg_read_raw_data. - */ - cinfo->global_state = cinfo->raw_data_out ? DSTATE_RAW_OK : DSTATE_SCANNING; - return TRUE; -} - - -/* - * Read some scanlines of data from the JPEG decompressor. - * - * The return value will be the number of lines actually read. - * This may be less than the number requested in several cases, - * including bottom of image, data source suspension, and operating - * modes that emit multiple scanlines at a time. - * - * Note: we warn about excess calls to jpeg_read_scanlines() since - * this likely signals an application programmer error. However, - * an oversize buffer (max_lines > scanlines remaining) is not an error. - */ - -GLOBAL(JDIMENSION) -jpeg_read_scanlines (j_decompress_ptr cinfo, JSAMPARRAY scanlines, - JDIMENSION max_lines) -{ - JDIMENSION row_ctr; - - if (cinfo->global_state != DSTATE_SCANNING) - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - if (cinfo->output_scanline >= cinfo->output_height) { - WARNMS(cinfo, JWRN_TOO_MUCH_DATA); - return 0; - } - - /* Call progress monitor hook if present */ - if (cinfo->progress != NULL) { - cinfo->progress->pass_counter = (long) cinfo->output_scanline; - cinfo->progress->pass_limit = (long) cinfo->output_height; - (*cinfo->progress->progress_monitor) ((j_common_ptr) cinfo); - } - - /* Process some data */ - row_ctr = 0; - (*cinfo->main->process_data) (cinfo, scanlines, &row_ctr, max_lines); - cinfo->output_scanline += row_ctr; - return row_ctr; -} - - -/* - * Alternate entry point to read raw data. - * Processes exactly one iMCU row per call, unless suspended. - */ - -GLOBAL(JDIMENSION) -jpeg_read_raw_data (j_decompress_ptr cinfo, JSAMPIMAGE data, - JDIMENSION max_lines) -{ - JDIMENSION lines_per_iMCU_row; - - if (cinfo->global_state != DSTATE_RAW_OK) - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - if (cinfo->output_scanline >= cinfo->output_height) { - WARNMS(cinfo, JWRN_TOO_MUCH_DATA); - return 0; - } - - /* Call progress monitor hook if present */ - if (cinfo->progress != NULL) { - cinfo->progress->pass_counter = (long) cinfo->output_scanline; - cinfo->progress->pass_limit = (long) cinfo->output_height; - (*cinfo->progress->progress_monitor) ((j_common_ptr) cinfo); - } - - /* Verify that at least one iMCU row can be returned. */ - lines_per_iMCU_row = cinfo->max_v_samp_factor * cinfo->min_DCT_v_scaled_size; - if (max_lines < lines_per_iMCU_row) - ERREXIT(cinfo, JERR_BUFFER_SIZE); - - /* Decompress directly into user's buffer. */ - if (! (*cinfo->coef->decompress_data) (cinfo, data)) - return 0; /* suspension forced, can do nothing more */ - - /* OK, we processed one iMCU row. */ - cinfo->output_scanline += lines_per_iMCU_row; - return lines_per_iMCU_row; -} - - -/* Additional entry points for buffered-image mode. */ - -#ifdef D_MULTISCAN_FILES_SUPPORTED - -/* - * Initialize for an output pass in buffered-image mode. - */ - -GLOBAL(boolean) -jpeg_start_output (j_decompress_ptr cinfo, int scan_number) -{ - if (cinfo->global_state != DSTATE_BUFIMAGE && - cinfo->global_state != DSTATE_PRESCAN) - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - /* Limit scan number to valid range */ - if (scan_number <= 0) - scan_number = 1; - if (cinfo->inputctl->eoi_reached && - scan_number > cinfo->input_scan_number) - scan_number = cinfo->input_scan_number; - cinfo->output_scan_number = scan_number; - /* Perform any dummy output passes, and set up for the real pass */ - return output_pass_setup(cinfo); -} - - -/* - * Finish up after an output pass in buffered-image mode. - * - * Returns FALSE if suspended. The return value need be inspected only if - * a suspending data source is used. - */ - -GLOBAL(boolean) -jpeg_finish_output (j_decompress_ptr cinfo) -{ - if ((cinfo->global_state == DSTATE_SCANNING || - cinfo->global_state == DSTATE_RAW_OK) && cinfo->buffered_image) { - /* Terminate this pass. */ - /* We do not require the whole pass to have been completed. */ - (*cinfo->master->finish_output_pass) (cinfo); - cinfo->global_state = DSTATE_BUFPOST; - } else if (cinfo->global_state != DSTATE_BUFPOST) { - /* BUFPOST = repeat call after a suspension, anything else is error */ - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - } - /* Read markers looking for SOS or EOI */ - while (cinfo->input_scan_number <= cinfo->output_scan_number && - ! cinfo->inputctl->eoi_reached) { - if ((*cinfo->inputctl->consume_input) (cinfo) == JPEG_SUSPENDED) - return FALSE; /* Suspend, come back later */ - } - cinfo->global_state = DSTATE_BUFIMAGE; - return TRUE; -} - -#endif /* D_MULTISCAN_FILES_SUPPORTED */ diff --git a/cmake/externals/libjpeg/jdarith.c b/cmake/externals/libjpeg/jdarith.c deleted file mode 100644 index 2c9abe23f..000000000 --- a/cmake/externals/libjpeg/jdarith.c +++ /dev/null @@ -1,796 +0,0 @@ -/* - * jdarith.c - * - * Developed 1997-2020 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains portable arithmetic entropy decoding routines for JPEG - * (implementing the ISO/IEC IS 10918-1 and CCITT Recommendation ITU-T T.81). - * - * Both sequential and progressive modes are supported in this single module. - * - * Suspension is not currently supported in this module. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* Expanded entropy decoder object for arithmetic decoding. */ - -typedef struct { - struct jpeg_entropy_decoder pub; /* public fields */ - - INT32 c; /* C register, base of coding interval + input bit buffer */ - INT32 a; /* A register, normalized size of coding interval */ - int ct; /* bit shift counter, # of bits left in bit buffer part of C */ - /* init: ct = -16 */ - /* run: ct = 0..7 */ - /* error: ct = -1 */ - int last_dc_val[MAX_COMPS_IN_SCAN]; /* last DC coef for each component */ - int dc_context[MAX_COMPS_IN_SCAN]; /* context index for DC conditioning */ - - unsigned int restarts_to_go; /* MCUs left in this restart interval */ - - /* Pointers to statistics areas (these workspaces have image lifespan) */ - unsigned char * dc_stats[NUM_ARITH_TBLS]; - unsigned char * ac_stats[NUM_ARITH_TBLS]; - - /* Statistics bin for coding with fixed probability 0.5 */ - unsigned char fixed_bin[4]; -} arith_entropy_decoder; - -typedef arith_entropy_decoder * arith_entropy_ptr; - -/* The following two definitions specify the allocation chunk size - * for the statistics area. - * According to sections F.1.4.4.1.3 and F.1.4.4.2, we need at least - * 49 statistics bins for DC, and 245 statistics bins for AC coding. - * - * We use a compact representation with 1 byte per statistics bin, - * thus the numbers directly represent byte sizes. - * This 1 byte per statistics bin contains the meaning of the MPS - * (more probable symbol) in the highest bit (mask 0x80), and the - * index into the probability estimation state machine table - * in the lower bits (mask 0x7F). - */ - -#define DC_STAT_BINS 64 -#define AC_STAT_BINS 256 - - -LOCAL(int) -get_byte (j_decompress_ptr cinfo) -/* Read next input byte; we do not support suspension in this module. */ -{ - struct jpeg_source_mgr * src = cinfo->src; - - if (src->bytes_in_buffer == 0) - if (! (*src->fill_input_buffer) (cinfo)) - ERREXIT(cinfo, JERR_CANT_SUSPEND); - src->bytes_in_buffer--; - return GETJOCTET(*src->next_input_byte++); -} - - -/* - * The core arithmetic decoding routine (common in JPEG and JBIG). - * This needs to go as fast as possible. - * Machine-dependent optimization facilities - * are not utilized in this portable implementation. - * However, this code should be fairly efficient and - * may be a good base for further optimizations anyway. - * - * Return value is 0 or 1 (binary decision). - * - * Note: I've changed the handling of the code base & bit - * buffer register C compared to other implementations - * based on the standards layout & procedures. - * While it also contains both the actual base of the - * coding interval (16 bits) and the next-bits buffer, - * the cut-point between these two parts is floating - * (instead of fixed) with the bit shift counter CT. - * Thus, we also need only one (variable instead of - * fixed size) shift for the LPS/MPS decision, and - * we can do away with any renormalization update - * of C (except for new data insertion, of course). - * - * I've also introduced a new scheme for accessing - * the probability estimation state machine table, - * derived from Markus Kuhn's JBIG implementation. - */ - -LOCAL(int) -arith_decode (j_decompress_ptr cinfo, unsigned char *st) -{ - register arith_entropy_ptr e = (arith_entropy_ptr) cinfo->entropy; - register unsigned char nl, nm; - register INT32 qe, temp; - register int sv, data; - - /* Renormalization & data input per section D.2.6 */ - while (e->a < 0x8000L) { - if (--e->ct < 0) { - /* Need to fetch next data byte */ - if (cinfo->unread_marker) - data = 0; /* stuff zero data */ - else { - data = get_byte(cinfo); /* read next input byte */ - if (data == 0xFF) { /* zero stuff or marker code */ - do data = get_byte(cinfo); - while (data == 0xFF); /* swallow extra 0xFF bytes */ - if (data == 0) - data = 0xFF; /* discard stuffed zero byte */ - else { - /* Note: Different from the Huffman decoder, hitting - * a marker while processing the compressed data - * segment is legal in arithmetic coding. - * The convention is to supply zero data - * then until decoding is complete. - */ - cinfo->unread_marker = data; - data = 0; - } - } - } - e->c = (e->c << 8) | data; /* insert data into C register */ - if ((e->ct += 8) < 0) /* update bit shift counter */ - /* Need more initial bytes */ - if (++e->ct == 0) - /* Got 2 initial bytes -> re-init A and exit loop */ - e->a = 0x8000L; /* => e->a = 0x10000L after loop exit */ - } - e->a <<= 1; - } - - /* Fetch values from our compact representation of Table D.3(D.2): - * Qe values and probability estimation state machine - */ - sv = *st; - qe = jpeg_aritab[sv & 0x7F]; /* => Qe_Value */ - nl = qe & 0xFF; qe >>= 8; /* Next_Index_LPS + Switch_MPS */ - nm = qe & 0xFF; qe >>= 8; /* Next_Index_MPS */ - - /* Decode & estimation procedures per sections D.2.4 & D.2.5 */ - temp = e->a - qe; - e->a = temp; - temp <<= e->ct; - if (e->c >= temp) { - e->c -= temp; - /* Conditional LPS (less probable symbol) exchange */ - if (e->a < qe) { - e->a = qe; - *st = (sv & 0x80) ^ nm; /* Estimate_after_MPS */ - } else { - e->a = qe; - *st = (sv & 0x80) ^ nl; /* Estimate_after_LPS */ - sv ^= 0x80; /* Exchange LPS/MPS */ - } - } else if (e->a < 0x8000L) { - /* Conditional MPS (more probable symbol) exchange */ - if (e->a < qe) { - *st = (sv & 0x80) ^ nl; /* Estimate_after_LPS */ - sv ^= 0x80; /* Exchange LPS/MPS */ - } else { - *st = (sv & 0x80) ^ nm; /* Estimate_after_MPS */ - } - } - - return sv >> 7; -} - - -/* - * Check for a restart marker & resynchronize decoder. - */ - -LOCAL(void) -process_restart (j_decompress_ptr cinfo) -{ - arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; - int ci; - jpeg_component_info * compptr; - - /* Advance past the RSTn marker */ - if (! (*cinfo->marker->read_restart_marker) (cinfo)) - ERREXIT(cinfo, JERR_CANT_SUSPEND); - - /* Re-initialize statistics areas */ - for (ci = 0; ci < cinfo->comps_in_scan; ci++) { - compptr = cinfo->cur_comp_info[ci]; - if (! cinfo->progressive_mode || (cinfo->Ss == 0 && cinfo->Ah == 0)) { - MEMZERO(entropy->dc_stats[compptr->dc_tbl_no], DC_STAT_BINS); - /* Reset DC predictions to 0 */ - entropy->last_dc_val[ci] = 0; - entropy->dc_context[ci] = 0; - } - if ((! cinfo->progressive_mode && cinfo->lim_Se) || - (cinfo->progressive_mode && cinfo->Ss)) { - MEMZERO(entropy->ac_stats[compptr->ac_tbl_no], AC_STAT_BINS); - } - } - - /* Reset arithmetic decoding variables */ - entropy->c = 0; - entropy->a = 0; - entropy->ct = -16; /* force reading 2 initial bytes to fill C */ - - /* Reset restart counter */ - entropy->restarts_to_go = cinfo->restart_interval; -} - - -/* - * Arithmetic MCU decoding. - * Each of these routines decodes and returns one MCU's worth of - * arithmetic-compressed coefficients. - * The coefficients are reordered from zigzag order into natural array order, - * but are not dequantized. - * - * The i'th block of the MCU is stored into the block pointed to by - * MCU_data[i]. WE ASSUME THIS AREA IS INITIALLY ZEROED BY THE CALLER. - */ - -/* - * MCU decoding for DC initial scan (either spectral selection, - * or first pass of successive approximation). - */ - -METHODDEF(boolean) -decode_mcu_DC_first (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data) -{ - arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; - JBLOCKROW block; - unsigned char *st; - int blkn, ci, tbl, sign; - int v, m; - - /* Process restart marker if needed */ - if (cinfo->restart_interval) { - if (entropy->restarts_to_go == 0) - process_restart(cinfo); - entropy->restarts_to_go--; - } - - if (entropy->ct == -1) return TRUE; /* if error do nothing */ - - /* Outer loop handles each block in the MCU */ - - for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { - block = MCU_data[blkn]; - ci = cinfo->MCU_membership[blkn]; - tbl = cinfo->cur_comp_info[ci]->dc_tbl_no; - - /* Sections F.2.4.1 & F.1.4.4.1: Decoding of DC coefficients */ - - /* Table F.4: Point to statistics bin S0 for DC coefficient coding */ - st = entropy->dc_stats[tbl] + entropy->dc_context[ci]; - - /* Figure F.19: Decode_DC_DIFF */ - if (arith_decode(cinfo, st) == 0) - entropy->dc_context[ci] = 0; - else { - /* Figure F.21: Decoding nonzero value v */ - /* Figure F.22: Decoding the sign of v */ - sign = arith_decode(cinfo, st + 1); - st += 2; st += sign; - /* Figure F.23: Decoding the magnitude category of v */ - if ((m = arith_decode(cinfo, st)) != 0) { - st = entropy->dc_stats[tbl] + 20; /* Table F.4: X1 = 20 */ - while (arith_decode(cinfo, st)) { - if ((m <<= 1) == (int) 0x8000U) { - WARNMS(cinfo, JWRN_ARITH_BAD_CODE); - entropy->ct = -1; /* magnitude overflow */ - return TRUE; - } - st += 1; - } - } - /* Section F.1.4.4.1.2: Establish dc_context conditioning category */ - if (m < (int) ((1L << cinfo->arith_dc_L[tbl]) >> 1)) - entropy->dc_context[ci] = 0; /* zero diff category */ - else if (m > (int) ((1L << cinfo->arith_dc_U[tbl]) >> 1)) - entropy->dc_context[ci] = 12 + (sign * 4); /* large diff category */ - else - entropy->dc_context[ci] = 4 + (sign * 4); /* small diff category */ - v = m; - /* Figure F.24: Decoding the magnitude bit pattern of v */ - st += 14; - while (m >>= 1) - if (arith_decode(cinfo, st)) v |= m; - v += 1; if (sign) v = -v; - entropy->last_dc_val[ci] += v; - } - - /* Scale and output the DC coefficient (assumes jpeg_natural_order[0]=0) */ - (*block)[0] = (JCOEF) (entropy->last_dc_val[ci] << cinfo->Al); - } - - return TRUE; -} - - -/* - * MCU decoding for AC initial scan (either spectral selection, - * or first pass of successive approximation). - */ - -METHODDEF(boolean) -decode_mcu_AC_first (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data) -{ - arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; - JBLOCKROW block; - unsigned char *st; - int tbl, sign, k; - int v, m; - const int * natural_order; - - /* Process restart marker if needed */ - if (cinfo->restart_interval) { - if (entropy->restarts_to_go == 0) - process_restart(cinfo); - entropy->restarts_to_go--; - } - - if (entropy->ct == -1) return TRUE; /* if error do nothing */ - - natural_order = cinfo->natural_order; - - /* There is always only one block per MCU */ - block = MCU_data[0]; - tbl = cinfo->cur_comp_info[0]->ac_tbl_no; - - /* Sections F.2.4.2 & F.1.4.4.2: Decoding of AC coefficients */ - - /* Figure F.20: Decode_AC_coefficients */ - k = cinfo->Ss - 1; - do { - st = entropy->ac_stats[tbl] + 3 * k; - if (arith_decode(cinfo, st)) break; /* EOB flag */ - for (;;) { - k++; - if (arith_decode(cinfo, st + 1)) break; - st += 3; - if (k >= cinfo->Se) { - WARNMS(cinfo, JWRN_ARITH_BAD_CODE); - entropy->ct = -1; /* spectral overflow */ - return TRUE; - } - } - /* Figure F.21: Decoding nonzero value v */ - /* Figure F.22: Decoding the sign of v */ - sign = arith_decode(cinfo, entropy->fixed_bin); - st += 2; - /* Figure F.23: Decoding the magnitude category of v */ - if ((m = arith_decode(cinfo, st)) != 0) { - if (arith_decode(cinfo, st)) { - m <<= 1; - st = entropy->ac_stats[tbl] + - (k <= cinfo->arith_ac_K[tbl] ? 189 : 217); - while (arith_decode(cinfo, st)) { - if ((m <<= 1) == (int) 0x8000U) { - WARNMS(cinfo, JWRN_ARITH_BAD_CODE); - entropy->ct = -1; /* magnitude overflow */ - return TRUE; - } - st += 1; - } - } - } - v = m; - /* Figure F.24: Decoding the magnitude bit pattern of v */ - st += 14; - while (m >>= 1) - if (arith_decode(cinfo, st)) v |= m; - v += 1; if (sign) v = -v; - /* Scale and output coefficient in natural (dezigzagged) order */ - (*block)[natural_order[k]] = (JCOEF) (v << cinfo->Al); - } while (k < cinfo->Se); - - return TRUE; -} - - -/* - * MCU decoding for DC successive approximation refinement scan. - * Note: we assume such scans can be multi-component, - * although the spec is not very clear on the point. - */ - -METHODDEF(boolean) -decode_mcu_DC_refine (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data) -{ - arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; - unsigned char *st; - JCOEF p1; - int blkn; - - /* Process restart marker if needed */ - if (cinfo->restart_interval) { - if (entropy->restarts_to_go == 0) - process_restart(cinfo); - entropy->restarts_to_go--; - } - - st = entropy->fixed_bin; /* use fixed probability estimation */ - p1 = 1 << cinfo->Al; /* 1 in the bit position being coded */ - - /* Outer loop handles each block in the MCU */ - - for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { - /* Encoded data is simply the next bit of the two's-complement DC value */ - if (arith_decode(cinfo, st)) - MCU_data[blkn][0][0] |= p1; - } - - return TRUE; -} - - -/* - * MCU decoding for AC successive approximation refinement scan. - */ - -METHODDEF(boolean) -decode_mcu_AC_refine (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data) -{ - arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; - JBLOCKROW block; - JCOEFPTR thiscoef; - unsigned char *st; - int tbl, k, kex; - JCOEF p1, m1; - const int * natural_order; - - /* Process restart marker if needed */ - if (cinfo->restart_interval) { - if (entropy->restarts_to_go == 0) - process_restart(cinfo); - entropy->restarts_to_go--; - } - - if (entropy->ct == -1) return TRUE; /* if error do nothing */ - - natural_order = cinfo->natural_order; - - /* There is always only one block per MCU */ - block = MCU_data[0]; - tbl = cinfo->cur_comp_info[0]->ac_tbl_no; - - p1 = 1 << cinfo->Al; /* 1 in the bit position being coded */ - m1 = -p1; /* -1 in the bit position being coded */ - - /* Establish EOBx (previous stage end-of-block) index */ - kex = cinfo->Se; - do { - if ((*block)[natural_order[kex]]) break; - } while (--kex); - - k = cinfo->Ss - 1; - do { - st = entropy->ac_stats[tbl] + 3 * k; - if (k >= kex) - if (arith_decode(cinfo, st)) break; /* EOB flag */ - for (;;) { - thiscoef = *block + natural_order[++k]; - if (*thiscoef) { /* previously nonzero coef */ - if (arith_decode(cinfo, st + 2)) { - if (*thiscoef < 0) - *thiscoef += m1; - else - *thiscoef += p1; - } - break; - } - if (arith_decode(cinfo, st + 1)) { /* newly nonzero coef */ - if (arith_decode(cinfo, entropy->fixed_bin)) - *thiscoef = m1; - else - *thiscoef = p1; - break; - } - st += 3; - if (k >= cinfo->Se) { - WARNMS(cinfo, JWRN_ARITH_BAD_CODE); - entropy->ct = -1; /* spectral overflow */ - return TRUE; - } - } - } while (k < cinfo->Se); - - return TRUE; -} - - -/* - * Decode one MCU's worth of arithmetic-compressed coefficients. - */ - -METHODDEF(boolean) -decode_mcu (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data) -{ - arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; - jpeg_component_info * compptr; - JBLOCKROW block; - unsigned char *st; - int blkn, ci, tbl, sign, k; - int v, m; - const int * natural_order; - - /* Process restart marker if needed */ - if (cinfo->restart_interval) { - if (entropy->restarts_to_go == 0) - process_restart(cinfo); - entropy->restarts_to_go--; - } - - if (entropy->ct == -1) return TRUE; /* if error do nothing */ - - natural_order = cinfo->natural_order; - - /* Outer loop handles each block in the MCU */ - - for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { - block = MCU_data[blkn]; - ci = cinfo->MCU_membership[blkn]; - compptr = cinfo->cur_comp_info[ci]; - - /* Sections F.2.4.1 & F.1.4.4.1: Decoding of DC coefficients */ - - tbl = compptr->dc_tbl_no; - - /* Table F.4: Point to statistics bin S0 for DC coefficient coding */ - st = entropy->dc_stats[tbl] + entropy->dc_context[ci]; - - /* Figure F.19: Decode_DC_DIFF */ - if (arith_decode(cinfo, st) == 0) - entropy->dc_context[ci] = 0; - else { - /* Figure F.21: Decoding nonzero value v */ - /* Figure F.22: Decoding the sign of v */ - sign = arith_decode(cinfo, st + 1); - st += 2; st += sign; - /* Figure F.23: Decoding the magnitude category of v */ - if ((m = arith_decode(cinfo, st)) != 0) { - st = entropy->dc_stats[tbl] + 20; /* Table F.4: X1 = 20 */ - while (arith_decode(cinfo, st)) { - if ((m <<= 1) == (int) 0x8000U) { - WARNMS(cinfo, JWRN_ARITH_BAD_CODE); - entropy->ct = -1; /* magnitude overflow */ - return TRUE; - } - st += 1; - } - } - /* Section F.1.4.4.1.2: Establish dc_context conditioning category */ - if (m < (int) ((1L << cinfo->arith_dc_L[tbl]) >> 1)) - entropy->dc_context[ci] = 0; /* zero diff category */ - else if (m > (int) ((1L << cinfo->arith_dc_U[tbl]) >> 1)) - entropy->dc_context[ci] = 12 + (sign * 4); /* large diff category */ - else - entropy->dc_context[ci] = 4 + (sign * 4); /* small diff category */ - v = m; - /* Figure F.24: Decoding the magnitude bit pattern of v */ - st += 14; - while (m >>= 1) - if (arith_decode(cinfo, st)) v |= m; - v += 1; if (sign) v = -v; - entropy->last_dc_val[ci] += v; - } - - (*block)[0] = (JCOEF) entropy->last_dc_val[ci]; - - /* Sections F.2.4.2 & F.1.4.4.2: Decoding of AC coefficients */ - - if (cinfo->lim_Se == 0) continue; - tbl = compptr->ac_tbl_no; - k = 0; - - /* Figure F.20: Decode_AC_coefficients */ - do { - st = entropy->ac_stats[tbl] + 3 * k; - if (arith_decode(cinfo, st)) break; /* EOB flag */ - for (;;) { - k++; - if (arith_decode(cinfo, st + 1)) break; - st += 3; - if (k >= cinfo->lim_Se) { - WARNMS(cinfo, JWRN_ARITH_BAD_CODE); - entropy->ct = -1; /* spectral overflow */ - return TRUE; - } - } - /* Figure F.21: Decoding nonzero value v */ - /* Figure F.22: Decoding the sign of v */ - sign = arith_decode(cinfo, entropy->fixed_bin); - st += 2; - /* Figure F.23: Decoding the magnitude category of v */ - if ((m = arith_decode(cinfo, st)) != 0) { - if (arith_decode(cinfo, st)) { - m <<= 1; - st = entropy->ac_stats[tbl] + - (k <= cinfo->arith_ac_K[tbl] ? 189 : 217); - while (arith_decode(cinfo, st)) { - if ((m <<= 1) == (int) 0x8000U) { - WARNMS(cinfo, JWRN_ARITH_BAD_CODE); - entropy->ct = -1; /* magnitude overflow */ - return TRUE; - } - st += 1; - } - } - } - v = m; - /* Figure F.24: Decoding the magnitude bit pattern of v */ - st += 14; - while (m >>= 1) - if (arith_decode(cinfo, st)) v |= m; - v += 1; if (sign) v = -v; - (*block)[natural_order[k]] = (JCOEF) v; - } while (k < cinfo->lim_Se); - } - - return TRUE; -} - - -/* - * Initialize for an arithmetic-compressed scan. - */ - -METHODDEF(void) -start_pass (j_decompress_ptr cinfo) -{ - arith_entropy_ptr entropy = (arith_entropy_ptr) cinfo->entropy; - int ci, tbl; - jpeg_component_info * compptr; - - if (cinfo->progressive_mode) { - /* Validate progressive scan parameters */ - if (cinfo->Ss == 0) { - if (cinfo->Se != 0) - goto bad; - } else { - /* need not check Ss/Se < 0 since they came from unsigned bytes */ - if (cinfo->Se < cinfo->Ss || cinfo->Se > cinfo->lim_Se) - goto bad; - /* AC scans may have only one component */ - if (cinfo->comps_in_scan != 1) - goto bad; - } - if (cinfo->Ah != 0) { - /* Successive approximation refinement scan: must have Al = Ah-1. */ - if (cinfo->Ah-1 != cinfo->Al) - goto bad; - } - if (cinfo->Al > 13) { /* need not check for < 0 */ - bad: - ERREXIT4(cinfo, JERR_BAD_PROGRESSION, - cinfo->Ss, cinfo->Se, cinfo->Ah, cinfo->Al); - } - /* Update progression status, and verify that scan order is legal. - * Note that inter-scan inconsistencies are treated as warnings - * not fatal errors ... not clear if this is right way to behave. - */ - for (ci = 0; ci < cinfo->comps_in_scan; ci++) { - int coefi, cindex = cinfo->cur_comp_info[ci]->component_index; - int *coef_bit_ptr = & cinfo->coef_bits[cindex][0]; - if (cinfo->Ss && coef_bit_ptr[0] < 0) /* AC without prior DC scan */ - WARNMS2(cinfo, JWRN_BOGUS_PROGRESSION, cindex, 0); - for (coefi = cinfo->Ss; coefi <= cinfo->Se; coefi++) { - int expected = (coef_bit_ptr[coefi] < 0) ? 0 : coef_bit_ptr[coefi]; - if (cinfo->Ah != expected) - WARNMS2(cinfo, JWRN_BOGUS_PROGRESSION, cindex, coefi); - coef_bit_ptr[coefi] = cinfo->Al; - } - } - /* Select MCU decoding routine */ - if (cinfo->Ah == 0) { - if (cinfo->Ss == 0) - entropy->pub.decode_mcu = decode_mcu_DC_first; - else - entropy->pub.decode_mcu = decode_mcu_AC_first; - } else { - if (cinfo->Ss == 0) - entropy->pub.decode_mcu = decode_mcu_DC_refine; - else - entropy->pub.decode_mcu = decode_mcu_AC_refine; - } - } else { - /* Check that the scan parameters Ss, Se, Ah/Al are OK for sequential JPEG. - * This ought to be an error condition, but we make it a warning. - */ - if (cinfo->Ss != 0 || cinfo->Ah != 0 || cinfo->Al != 0 || - (cinfo->Se < DCTSIZE2 && cinfo->Se != cinfo->lim_Se)) - WARNMS(cinfo, JWRN_NOT_SEQUENTIAL); - /* Select MCU decoding routine */ - entropy->pub.decode_mcu = decode_mcu; - } - - /* Allocate & initialize requested statistics areas */ - for (ci = 0; ci < cinfo->comps_in_scan; ci++) { - compptr = cinfo->cur_comp_info[ci]; - if (! cinfo->progressive_mode || (cinfo->Ss == 0 && cinfo->Ah == 0)) { - tbl = compptr->dc_tbl_no; - if (tbl < 0 || tbl >= NUM_ARITH_TBLS) - ERREXIT1(cinfo, JERR_NO_ARITH_TABLE, tbl); - if (entropy->dc_stats[tbl] == NULL) - entropy->dc_stats[tbl] = (unsigned char *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, DC_STAT_BINS); - MEMZERO(entropy->dc_stats[tbl], DC_STAT_BINS); - /* Initialize DC predictions to 0 */ - entropy->last_dc_val[ci] = 0; - entropy->dc_context[ci] = 0; - } - if ((! cinfo->progressive_mode && cinfo->lim_Se) || - (cinfo->progressive_mode && cinfo->Ss)) { - tbl = compptr->ac_tbl_no; - if (tbl < 0 || tbl >= NUM_ARITH_TBLS) - ERREXIT1(cinfo, JERR_NO_ARITH_TABLE, tbl); - if (entropy->ac_stats[tbl] == NULL) - entropy->ac_stats[tbl] = (unsigned char *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, AC_STAT_BINS); - MEMZERO(entropy->ac_stats[tbl], AC_STAT_BINS); - } - } - - /* Initialize arithmetic decoding variables */ - entropy->c = 0; - entropy->a = 0; - entropy->ct = -16; /* force reading 2 initial bytes to fill C */ - - /* Initialize restart counter */ - entropy->restarts_to_go = cinfo->restart_interval; -} - - -/* - * Finish up at the end of an arithmetic-compressed scan. - */ - -METHODDEF(void) -finish_pass (j_decompress_ptr cinfo) -{ - /* no work necessary here */ -} - - -/* - * Module initialization routine for arithmetic entropy decoding. - */ - -GLOBAL(void) -jinit_arith_decoder (j_decompress_ptr cinfo) -{ - arith_entropy_ptr entropy; - int i; - - entropy = (arith_entropy_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(arith_entropy_decoder)); - cinfo->entropy = &entropy->pub; - entropy->pub.start_pass = start_pass; - entropy->pub.finish_pass = finish_pass; - - /* Mark tables unallocated */ - for (i = 0; i < NUM_ARITH_TBLS; i++) { - entropy->dc_stats[i] = NULL; - entropy->ac_stats[i] = NULL; - } - - /* Initialize index for fixed probability estimation */ - entropy->fixed_bin[0] = 113; - - if (cinfo->progressive_mode) { - /* Create progression status table */ - int *coef_bit_ptr, ci; - cinfo->coef_bits = (int (*)[DCTSIZE2]) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, - cinfo->num_components * DCTSIZE2 * SIZEOF(int)); - coef_bit_ptr = & cinfo->coef_bits[0][0]; - for (ci = 0; ci < cinfo->num_components; ci++) - for (i = 0; i < DCTSIZE2; i++) - *coef_bit_ptr++ = -1; - } -} diff --git a/cmake/externals/libjpeg/jdatadst.c b/cmake/externals/libjpeg/jdatadst.c deleted file mode 100644 index b3b4798ea..000000000 --- a/cmake/externals/libjpeg/jdatadst.c +++ /dev/null @@ -1,263 +0,0 @@ -/* - * jdatadst.c - * - * Copyright (C) 1994-1996, Thomas G. Lane. - * Modified 2009-2022 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains compression data destination routines for the case of - * emitting JPEG data to memory or to a file (or any stdio stream). - * While these routines are sufficient for most applications, - * some will want to use a different destination manager. - * IMPORTANT: we assume that fwrite() will correctly transcribe an array of - * JOCTETs into 8-bit-wide elements on external storage. If char is wider - * than 8 bits on your machine, you may need to do some tweaking. - */ - -/* this is not a core library module, so it doesn't define JPEG_INTERNALS */ -#include "jinclude.h" -#include "jpeglib.h" -#include "jerror.h" - -#ifndef HAVE_STDLIB_H /* should declare malloc(),free() */ -extern void * malloc JPP((size_t size)); -extern void free JPP((void *ptr)); -#endif - - -/* Expanded data destination object for stdio output */ - -#define OUTPUT_BUF_SIZE 4096 /* choose an efficiently fwrite'able size */ - -typedef struct { - struct jpeg_destination_mgr pub; /* public fields */ - - FILE * outfile; /* target stream */ - JOCTET buffer[OUTPUT_BUF_SIZE]; /* output buffer */ -} my_destination_mgr; - -typedef my_destination_mgr * my_dest_ptr; - - -/* Expanded data destination object for memory output */ - -typedef struct { - struct jpeg_destination_mgr pub; /* public fields */ - - unsigned char ** outbuffer; /* target buffer */ - size_t * outsize; - unsigned char * newbuffer; /* newly allocated buffer */ - JOCTET * buffer; /* start of buffer */ - size_t bufsize; -} my_mem_destination_mgr; - -typedef my_mem_destination_mgr * my_mem_dest_ptr; - - -/* - * Initialize destination --- called by jpeg_start_compress - * before any data is actually written. - */ - -METHODDEF(void) -init_destination (j_compress_ptr cinfo) -{ - my_dest_ptr dest = (my_dest_ptr) cinfo->dest; - - dest->pub.next_output_byte = dest->buffer; - dest->pub.free_in_buffer = OUTPUT_BUF_SIZE; -} - -METHODDEF(void) -init_mem_destination (j_compress_ptr cinfo) -{ - /* no work necessary here */ -} - - -/* - * Empty the output buffer --- called whenever buffer fills up. - * - * In typical applications, this should write the entire output buffer - * (ignoring the current state of next_output_byte & free_in_buffer), - * reset the pointer & count to the start of the buffer, and return TRUE - * indicating that the buffer has been dumped. - * - * In applications that need to be able to suspend compression due to output - * overrun, a FALSE return indicates that the buffer cannot be emptied now. - * In this situation, the compressor will return to its caller (possibly with - * an indication that it has not accepted all the supplied scanlines). The - * application should resume compression after it has made more room in the - * output buffer. Note that there are substantial restrictions on the use of - * suspension --- see the documentation. - * - * When suspending, the compressor will back up to a convenient restart point - * (typically the start of the current MCU). next_output_byte & free_in_buffer - * indicate where the restart point will be if the current call returns FALSE. - * Data beyond this point will be regenerated after resumption, so do not - * write it out when emptying the buffer externally. - */ - -METHODDEF(boolean) -empty_output_buffer (j_compress_ptr cinfo) -{ - my_dest_ptr dest = (my_dest_ptr) cinfo->dest; - - if (JFWRITE(dest->outfile, dest->buffer, OUTPUT_BUF_SIZE) != - (size_t) OUTPUT_BUF_SIZE) - ERREXIT(cinfo, JERR_FILE_WRITE); - - dest->pub.next_output_byte = dest->buffer; - dest->pub.free_in_buffer = OUTPUT_BUF_SIZE; - - return TRUE; -} - -METHODDEF(boolean) -empty_mem_output_buffer (j_compress_ptr cinfo) -{ - size_t nextsize; - JOCTET * nextbuffer; - my_mem_dest_ptr dest = (my_mem_dest_ptr) cinfo->dest; - - /* Try to allocate new buffer with double size */ - nextsize = dest->bufsize * 2; - nextbuffer = (JOCTET *) malloc(nextsize); - - if (nextbuffer == NULL) - ERREXIT1(cinfo, JERR_OUT_OF_MEMORY, 11); - - MEMCOPY(nextbuffer, dest->buffer, dest->bufsize); - - if (dest->newbuffer != NULL) - free(dest->newbuffer); - - dest->newbuffer = nextbuffer; - - dest->pub.next_output_byte = nextbuffer + dest->bufsize; - dest->pub.free_in_buffer = dest->bufsize; - - dest->buffer = nextbuffer; - dest->bufsize = nextsize; - - return TRUE; -} - - -/* - * Terminate destination --- called by jpeg_finish_compress - * after all data has been written. Usually needs to flush buffer. - * - * NB: *not* called by jpeg_abort or jpeg_destroy; surrounding - * application must deal with any cleanup that should happen even - * for error exit. - */ - -METHODDEF(void) -term_destination (j_compress_ptr cinfo) -{ - my_dest_ptr dest = (my_dest_ptr) cinfo->dest; - size_t datacount = OUTPUT_BUF_SIZE - dest->pub.free_in_buffer; - - /* Write any data remaining in the buffer */ - if (datacount > 0) { - if (JFWRITE(dest->outfile, dest->buffer, datacount) != datacount) - ERREXIT(cinfo, JERR_FILE_WRITE); - } - JFFLUSH(dest->outfile); - /* Make sure we wrote the output file OK */ - if (JFERROR(dest->outfile)) - ERREXIT(cinfo, JERR_FILE_WRITE); -} - -METHODDEF(void) -term_mem_destination (j_compress_ptr cinfo) -{ - my_mem_dest_ptr dest = (my_mem_dest_ptr) cinfo->dest; - - *dest->outbuffer = dest->buffer; - *dest->outsize = dest->bufsize - dest->pub.free_in_buffer; -} - - -/* - * Prepare for output to a stdio stream. - * The caller must have already opened the stream, - * and is responsible for closing it after finishing compression. - */ - -GLOBAL(void) -jpeg_stdio_dest (j_compress_ptr cinfo, FILE * outfile) -{ - my_dest_ptr dest; - - /* The destination object is made permanent so that multiple JPEG images - * can be written to the same file without re-executing jpeg_stdio_dest. - * This makes it dangerous to use this manager and a different destination - * manager serially with the same JPEG object, because their private object - * sizes may be different. Caveat programmer. - */ - if (cinfo->dest == NULL) { /* first time for this JPEG object? */ - cinfo->dest = (struct jpeg_destination_mgr *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_PERMANENT, SIZEOF(my_destination_mgr)); - } - - dest = (my_dest_ptr) cinfo->dest; - dest->pub.init_destination = init_destination; - dest->pub.empty_output_buffer = empty_output_buffer; - dest->pub.term_destination = term_destination; - dest->outfile = outfile; -} - - -/* - * Prepare for output to a memory buffer. - * The caller may supply an own initial buffer with appropriate size. - * Otherwise, or when the actual data output exceeds the given size, - * the library adapts the buffer size as necessary. - * The standard library functions malloc/free are used for allocating - * larger memory, so the buffer is available to the application after - * finishing compression, and then the application is responsible for - * freeing the requested memory. - * Note: An initial buffer supplied by the caller is expected to be - * managed by the application. The library does not free such buffer - * when allocating a larger buffer. - */ - -GLOBAL(void) -jpeg_mem_dest (j_compress_ptr cinfo, - unsigned char ** outbuffer, size_t * outsize) -{ - my_mem_dest_ptr dest; - - if (outbuffer == NULL || outsize == NULL) /* sanity check */ - ERREXIT(cinfo, JERR_BUFFER_SIZE); - - /* The destination object is made permanent so that multiple JPEG images - * can be written to the same buffer without re-executing jpeg_mem_dest. - */ - if (cinfo->dest == NULL) { /* first time for this JPEG object? */ - cinfo->dest = (struct jpeg_destination_mgr *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_PERMANENT, SIZEOF(my_mem_destination_mgr)); - } - - dest = (my_mem_dest_ptr) cinfo->dest; - dest->pub.init_destination = init_mem_destination; - dest->pub.empty_output_buffer = empty_mem_output_buffer; - dest->pub.term_destination = term_mem_destination; - dest->outbuffer = outbuffer; - dest->outsize = outsize; - dest->newbuffer = NULL; - - if (*outbuffer == NULL || *outsize == 0) { - /* Allocate initial buffer */ - dest->newbuffer = *outbuffer = (unsigned char *) malloc(OUTPUT_BUF_SIZE); - if (dest->newbuffer == NULL) - ERREXIT1(cinfo, JERR_OUT_OF_MEMORY, 10); - *outsize = OUTPUT_BUF_SIZE; - } - - dest->pub.next_output_byte = dest->buffer = *outbuffer; - dest->pub.free_in_buffer = dest->bufsize = *outsize; -} diff --git a/cmake/externals/libjpeg/jdatasrc.c b/cmake/externals/libjpeg/jdatasrc.c deleted file mode 100644 index fd7a1a594..000000000 --- a/cmake/externals/libjpeg/jdatasrc.c +++ /dev/null @@ -1,271 +0,0 @@ -/* - * jdatasrc.c - * - * Copyright (C) 1994-1996, Thomas G. Lane. - * Modified 2009-2022 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains decompression data source routines for the case of - * reading JPEG data from memory or from a file (or any stdio stream). - * While these routines are sufficient for most applications, - * some will want to use a different source manager. - * IMPORTANT: we assume that fread() will correctly transcribe an array of - * JOCTETs from 8-bit-wide elements on external storage. If char is wider - * than 8 bits on your machine, you may need to do some tweaking. - */ - -/* this is not a core library module, so it doesn't define JPEG_INTERNALS */ -#include "jinclude.h" -#include "jpeglib.h" -#include "jerror.h" - - -/* Expanded data source object for stdio input */ - -#define INPUT_BUF_SIZE 4096 /* choose an efficiently fread'able size */ - -typedef struct { - struct jpeg_source_mgr pub; /* public fields */ - - FILE * infile; /* source stream */ - JOCTET buffer[INPUT_BUF_SIZE]; /* input buffer */ - boolean start_of_file; /* have we gotten any data yet? */ -} my_source_mgr; - -typedef my_source_mgr * my_src_ptr; - - -/* - * Initialize source --- called by jpeg_read_header - * before any data is actually read. - */ - -METHODDEF(void) -init_source (j_decompress_ptr cinfo) -{ - my_src_ptr src = (my_src_ptr) cinfo->src; - - /* We reset the empty-input-file flag for each image, - * but we don't clear the input buffer. - * This is correct behavior for reading a series of images from one source. - */ - src->start_of_file = TRUE; -} - -METHODDEF(void) -init_mem_source (j_decompress_ptr cinfo) -{ - /* no work necessary here */ -} - - -/* - * Fill the input buffer --- called whenever buffer is emptied. - * - * In typical applications, this should read fresh data into the buffer - * (ignoring the current state of next_input_byte & bytes_in_buffer), - * reset the pointer & count to the start of the buffer, and return TRUE - * indicating that the buffer has been reloaded. It is not necessary to - * fill the buffer entirely, only to obtain at least one more byte. - * - * There is no such thing as an EOF return. If the end of the file has been - * reached, the routine has a choice of ERREXIT() or inserting fake data into - * the buffer. In most cases, generating a warning message and inserting a - * fake EOI marker is the best course of action --- this will allow the - * decompressor to output however much of the image is there. However, - * the resulting error message is misleading if the real problem is an empty - * input file, so we handle that case specially. - * - * In applications that need to be able to suspend compression due to input - * not being available yet, a FALSE return indicates that no more data can be - * obtained right now, but more may be forthcoming later. In this situation, - * the decompressor will return to its caller (with an indication of the - * number of scanlines it has read, if any). The application should resume - * decompression after it has loaded more data into the input buffer. Note - * that there are substantial restrictions on the use of suspension --- see - * the documentation. - * - * When suspending, the decompressor will back up to a convenient restart point - * (typically the start of the current MCU). next_input_byte & bytes_in_buffer - * indicate where the restart point will be if the current call returns FALSE. - * Data beyond this point must be rescanned after resumption, so move it to - * the front of the buffer rather than discarding it. - */ - -METHODDEF(boolean) -fill_input_buffer (j_decompress_ptr cinfo) -{ - my_src_ptr src = (my_src_ptr) cinfo->src; - size_t nbytes; - - nbytes = JFREAD(src->infile, src->buffer, INPUT_BUF_SIZE); - - if (nbytes <= 0) { - if (src->start_of_file) /* Treat empty input file as fatal error */ - ERREXIT(cinfo, JERR_INPUT_EMPTY); - WARNMS(cinfo, JWRN_JPEG_EOF); - /* Insert a fake EOI marker */ - src->buffer[0] = (JOCTET) 0xFF; - src->buffer[1] = (JOCTET) JPEG_EOI; - nbytes = 2; - } - - src->pub.next_input_byte = src->buffer; - src->pub.bytes_in_buffer = nbytes; - src->start_of_file = FALSE; - - return TRUE; -} - -METHODDEF(boolean) -fill_mem_input_buffer (j_decompress_ptr cinfo) -{ - static const JOCTET mybuffer[4] = { - (JOCTET) 0xFF, (JOCTET) JPEG_EOI, 0, 0 - }; - - /* The whole JPEG data is expected to reside in the supplied memory - * buffer, so any request for more data beyond the given buffer size - * is treated as an error. - */ - WARNMS(cinfo, JWRN_JPEG_EOF); - - /* Insert a fake EOI marker */ - - cinfo->src->next_input_byte = mybuffer; - cinfo->src->bytes_in_buffer = 2; - - return TRUE; -} - - -/* - * Skip data --- used to skip over a potentially large amount of - * uninteresting data (such as an APPn marker). - * - * Writers of suspendable-input applications must note that skip_input_data - * is not granted the right to give a suspension return. If the skip extends - * beyond the data currently in the buffer, the buffer can be marked empty so - * that the next read will cause a fill_input_buffer call that can suspend. - * Arranging for additional bytes to be discarded before reloading the input - * buffer is the application writer's problem. - */ - -METHODDEF(void) -skip_input_data (j_decompress_ptr cinfo, long num_bytes) -{ - struct jpeg_source_mgr * src = cinfo->src; - size_t nbytes; - - /* Just a dumb implementation for now. Could use fseek() except - * it doesn't work on pipes. Not clear that being smart is worth - * any trouble anyway --- large skips are infrequent. - */ - if (num_bytes > 0) { - nbytes = (size_t) num_bytes; - while (nbytes > src->bytes_in_buffer) { - nbytes -= src->bytes_in_buffer; - (void) (*src->fill_input_buffer) (cinfo); - /* note we assume that fill_input_buffer will never return FALSE, - * so suspension need not be handled. - */ - } - src->next_input_byte += nbytes; - src->bytes_in_buffer -= nbytes; - } -} - - -/* - * An additional method that can be provided by data source modules is the - * resync_to_restart method for error recovery in the presence of RST markers. - * For the moment, this source module just uses the default resync method - * provided by the JPEG library. That method assumes that no backtracking - * is possible. - */ - - -/* - * Terminate source --- called by jpeg_finish_decompress - * after all data has been read. Often a no-op. - * - * NB: *not* called by jpeg_abort or jpeg_destroy; surrounding - * application must deal with any cleanup that should happen even - * for error exit. - */ - -METHODDEF(void) -term_source (j_decompress_ptr cinfo) -{ - /* no work necessary here */ -} - - -/* - * Prepare for input from a stdio stream. - * The caller must have already opened the stream, - * and is responsible for closing it after finishing decompression. - */ - -GLOBAL(void) -jpeg_stdio_src (j_decompress_ptr cinfo, FILE * infile) -{ - my_src_ptr src; - - /* The source object including the input buffer is made permanent so that - * a series of JPEG images can be read from the same file by calling - * jpeg_stdio_src only before the first one. (If we discarded the buffer - * at the end of one image, we'd likely lose the start of the next one.) - * This makes it unsafe to use this manager and a different source - * manager serially with the same JPEG object. Caveat programmer. - */ - if (cinfo->src == NULL) { /* first time for this JPEG object? */ - cinfo->src = (struct jpeg_source_mgr *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_PERMANENT, SIZEOF(my_source_mgr)); - } - - src = (my_src_ptr) cinfo->src; - src->pub.init_source = init_source; - src->pub.fill_input_buffer = fill_input_buffer; - src->pub.skip_input_data = skip_input_data; - src->pub.resync_to_restart = jpeg_resync_to_restart; /* use default method */ - src->pub.term_source = term_source; - src->infile = infile; - src->pub.bytes_in_buffer = 0; /* forces fill_input_buffer on first read */ - src->pub.next_input_byte = NULL; /* until buffer loaded */ -} - - -/* - * Prepare for input from a supplied memory buffer. - * The buffer must contain the whole JPEG data. - */ - -GLOBAL(void) -jpeg_mem_src (j_decompress_ptr cinfo, - const unsigned char * inbuffer, size_t insize) -{ - struct jpeg_source_mgr * src; - - if (inbuffer == NULL || insize == 0) /* Treat empty input as fatal error */ - ERREXIT(cinfo, JERR_INPUT_EMPTY); - - /* The source object is made permanent so that a series of JPEG images - * can be read from the same buffer by calling jpeg_mem_src only before - * the first one. - */ - if (cinfo->src == NULL) { /* first time for this JPEG object? */ - cinfo->src = (struct jpeg_source_mgr *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_PERMANENT, SIZEOF(struct jpeg_source_mgr)); - } - - src = cinfo->src; - src->init_source = init_mem_source; - src->fill_input_buffer = fill_mem_input_buffer; - src->skip_input_data = skip_input_data; - src->resync_to_restart = jpeg_resync_to_restart; /* use default method */ - src->term_source = term_source; - src->bytes_in_buffer = insize; - src->next_input_byte = (const JOCTET *) inbuffer; -} diff --git a/cmake/externals/libjpeg/jdcoefct.c b/cmake/externals/libjpeg/jdcoefct.c deleted file mode 100644 index 79ba42014..000000000 --- a/cmake/externals/libjpeg/jdcoefct.c +++ /dev/null @@ -1,744 +0,0 @@ -/* - * jdcoefct.c - * - * Copyright (C) 1994-1997, Thomas G. Lane. - * Modified 2002-2020 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains the coefficient buffer controller for decompression. - * This controller is the top level of the JPEG decompressor proper. - * The coefficient buffer lies between entropy decoding and inverse-DCT steps. - * - * In buffered-image mode, this controller is the interface between - * input-oriented processing and output-oriented processing. - * Also, the input side (only) is used when reading a file for transcoding. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* Block smoothing is only applicable for progressive JPEG, so: */ -#ifndef D_PROGRESSIVE_SUPPORTED -#undef BLOCK_SMOOTHING_SUPPORTED -#endif - - -/* Private buffer controller object */ - -typedef struct { - struct jpeg_d_coef_controller pub; /* public fields */ - - /* These variables keep track of the current location of the input side. */ - /* cinfo->input_iMCU_row is also used for this. */ - JDIMENSION MCU_ctr; /* counts MCUs processed in current row */ - int MCU_vert_offset; /* counts MCU rows within iMCU row */ - int MCU_rows_per_iMCU_row; /* number of such rows needed */ - - /* The output side's location is represented by cinfo->output_iMCU_row. */ - - /* In single-pass modes, it's sufficient to buffer just one MCU. - * We append a workspace of D_MAX_BLOCKS_IN_MCU coefficient blocks, - * and let the entropy decoder write into that workspace each time. - * In multi-pass modes, this array points to the current MCU's blocks - * within the virtual arrays; it is used only by the input side. - */ - JBLOCKROW MCU_buffer[D_MAX_BLOCKS_IN_MCU]; - -#ifdef D_MULTISCAN_FILES_SUPPORTED - /* In multi-pass modes, we need a virtual block array for each component. */ - jvirt_barray_ptr whole_image[MAX_COMPONENTS]; -#endif - -#ifdef BLOCK_SMOOTHING_SUPPORTED - /* When doing block smoothing, we latch coefficient Al values here */ - int * coef_bits_latch; -#define SAVED_COEFS 6 /* we save coef_bits[0..5] */ -#endif - - /* Workspace for single-pass modes (omitted otherwise). */ - JBLOCK blk_buffer[D_MAX_BLOCKS_IN_MCU]; -} my_coef_controller; - -typedef my_coef_controller * my_coef_ptr; - - -/* Forward declarations */ -METHODDEF(int) decompress_onepass - JPP((j_decompress_ptr cinfo, JSAMPIMAGE output_buf)); -#ifdef D_MULTISCAN_FILES_SUPPORTED -METHODDEF(int) decompress_data - JPP((j_decompress_ptr cinfo, JSAMPIMAGE output_buf)); -#endif -#ifdef BLOCK_SMOOTHING_SUPPORTED -LOCAL(boolean) smoothing_ok JPP((j_decompress_ptr cinfo)); -METHODDEF(int) decompress_smooth_data - JPP((j_decompress_ptr cinfo, JSAMPIMAGE output_buf)); -#endif - - -LOCAL(void) -start_iMCU_row (j_decompress_ptr cinfo) -/* Reset within-iMCU-row counters for a new row (input side) */ -{ - my_coef_ptr coef = (my_coef_ptr) cinfo->coef; - - /* In an interleaved scan, an MCU row is the same as an iMCU row. - * In a noninterleaved scan, an iMCU row has v_samp_factor MCU rows. - * But at the bottom of the image, process only what's left. - */ - if (cinfo->comps_in_scan > 1) { - coef->MCU_rows_per_iMCU_row = 1; - } else { - if (cinfo->input_iMCU_row < (cinfo->total_iMCU_rows-1)) - coef->MCU_rows_per_iMCU_row = cinfo->cur_comp_info[0]->v_samp_factor; - else - coef->MCU_rows_per_iMCU_row = cinfo->cur_comp_info[0]->last_row_height; - } - - coef->MCU_ctr = 0; - coef->MCU_vert_offset = 0; -} - - -/* - * Initialize for an input processing pass. - */ - -METHODDEF(void) -start_input_pass (j_decompress_ptr cinfo) -{ - cinfo->input_iMCU_row = 0; - start_iMCU_row(cinfo); -} - - -/* - * Initialize for an output processing pass. - */ - -METHODDEF(void) -start_output_pass (j_decompress_ptr cinfo) -{ -#ifdef BLOCK_SMOOTHING_SUPPORTED - my_coef_ptr coef = (my_coef_ptr) cinfo->coef; - - /* If multipass, check to see whether to use block smoothing on this pass */ - if (coef->pub.coef_arrays != NULL) { - if (cinfo->do_block_smoothing && smoothing_ok(cinfo)) - coef->pub.decompress_data = decompress_smooth_data; - else - coef->pub.decompress_data = decompress_data; - } -#endif - cinfo->output_iMCU_row = 0; -} - - -/* - * Decompress and return some data in the single-pass case. - * Always attempts to emit one fully interleaved MCU row ("iMCU" row). - * Input and output must run in lockstep since we have only a one-MCU buffer. - * Return value is JPEG_ROW_COMPLETED, JPEG_SCAN_COMPLETED, or JPEG_SUSPENDED. - * - * NB: output_buf contains a plane for each component in image, - * which we index according to the component's SOF position. - */ - -METHODDEF(int) -decompress_onepass (j_decompress_ptr cinfo, JSAMPIMAGE output_buf) -{ - my_coef_ptr coef = (my_coef_ptr) cinfo->coef; - JDIMENSION MCU_col_num; /* index of current MCU within row */ - JDIMENSION last_MCU_col = cinfo->MCUs_per_row - 1; - JDIMENSION last_iMCU_row = cinfo->total_iMCU_rows - 1; - int ci, xindex, yindex, yoffset, useful_width; - JBLOCKROW blkp; - JSAMPARRAY output_ptr; - JDIMENSION start_col, output_col; - jpeg_component_info *compptr; - inverse_DCT_method_ptr inverse_DCT; - - /* Loop to process as much as one whole iMCU row */ - for (yoffset = coef->MCU_vert_offset; yoffset < coef->MCU_rows_per_iMCU_row; - yoffset++) { - for (MCU_col_num = coef->MCU_ctr; MCU_col_num <= last_MCU_col; - MCU_col_num++) { - blkp = coef->blk_buffer; /* pointer to current DCT block within MCU */ - /* Try to fetch an MCU. Entropy decoder expects buffer to be zeroed. */ - if (cinfo->lim_Se) /* can bypass in DC only case */ - MEMZERO(blkp, cinfo->blocks_in_MCU * SIZEOF(JBLOCK)); - if (! (*cinfo->entropy->decode_mcu) (cinfo, coef->MCU_buffer)) { - /* Suspension forced; update state counters and exit */ - coef->MCU_vert_offset = yoffset; - coef->MCU_ctr = MCU_col_num; - return JPEG_SUSPENDED; - } - /* Determine where data should go in output_buf and do the IDCT thing. - * We skip dummy blocks at the right and bottom edges (but blkp gets - * incremented past them!). - */ - for (ci = 0; ci < cinfo->comps_in_scan; ci++) { - compptr = cinfo->cur_comp_info[ci]; - /* Don't bother to IDCT an uninteresting component. */ - if (! compptr->component_needed) { - blkp += compptr->MCU_blocks; - continue; - } - inverse_DCT = cinfo->idct->inverse_DCT[compptr->component_index]; - output_ptr = output_buf[compptr->component_index] + - yoffset * compptr->DCT_v_scaled_size; - useful_width = (MCU_col_num < last_MCU_col) ? compptr->MCU_width - : compptr->last_col_width; - start_col = MCU_col_num * compptr->MCU_sample_width; - for (yindex = 0; yindex < compptr->MCU_height; yindex++) { - if (cinfo->input_iMCU_row < last_iMCU_row || - yoffset + yindex < compptr->last_row_height) { - output_col = start_col; - for (xindex = 0; xindex < useful_width; xindex++) { - (*inverse_DCT) (cinfo, compptr, (JCOEFPTR) (blkp + xindex), - output_ptr, output_col); - output_col += compptr->DCT_h_scaled_size; - } - output_ptr += compptr->DCT_v_scaled_size; - } - blkp += compptr->MCU_width; - } - } - } - /* Completed an MCU row, but perhaps not an iMCU row */ - coef->MCU_ctr = 0; - } - /* Completed the iMCU row, advance counters for next one */ - cinfo->output_iMCU_row++; - if (++(cinfo->input_iMCU_row) <= last_iMCU_row) { - start_iMCU_row(cinfo); - return JPEG_ROW_COMPLETED; - } - /* Completed the scan */ - (*cinfo->inputctl->finish_input_pass) (cinfo); - return JPEG_SCAN_COMPLETED; -} - - -/* - * Dummy consume-input routine for single-pass operation. - */ - -METHODDEF(int) -dummy_consume_data (j_decompress_ptr cinfo) -{ - return JPEG_SUSPENDED; /* Always indicate nothing was done */ -} - - -#ifdef D_MULTISCAN_FILES_SUPPORTED - -/* - * Consume input data and store it in the full-image coefficient buffer. - * We read as much as one fully interleaved MCU row ("iMCU" row) per call, - * ie, v_samp_factor block rows for each component in the scan. - * Return value is JPEG_ROW_COMPLETED, JPEG_SCAN_COMPLETED, or JPEG_SUSPENDED. - */ - -METHODDEF(int) -consume_data (j_decompress_ptr cinfo) -{ - my_coef_ptr coef = (my_coef_ptr) cinfo->coef; - JDIMENSION MCU_col_num; /* index of current MCU within row */ - int ci, xindex, yindex, yoffset; - JDIMENSION start_col; - JBLOCKARRAY blkp; - JBLOCKARRAY buffer[MAX_COMPS_IN_SCAN]; - JBLOCKROW buffer_ptr; - jpeg_component_info *compptr; - - /* Align the virtual buffers for the components used in this scan. */ - for (ci = 0; ci < cinfo->comps_in_scan; ci++) { - compptr = cinfo->cur_comp_info[ci]; - buffer[ci] = (*cinfo->mem->access_virt_barray) - ((j_common_ptr) cinfo, coef->whole_image[compptr->component_index], - cinfo->input_iMCU_row * compptr->v_samp_factor, - (JDIMENSION) compptr->v_samp_factor, TRUE); - /* Note: entropy decoder expects buffer to be zeroed, - * but this is handled automatically by the memory manager - * because we requested a pre-zeroed array. - */ - } - - /* Loop to process one whole iMCU row */ - for (yoffset = coef->MCU_vert_offset; yoffset < coef->MCU_rows_per_iMCU_row; - yoffset++) { - for (MCU_col_num = coef->MCU_ctr; MCU_col_num < cinfo->MCUs_per_row; - MCU_col_num++) { - /* Construct list of pointers to DCT blocks belonging to this MCU */ - blkp = coef->MCU_buffer; /* pointer to current DCT block within MCU */ - for (ci = 0; ci < cinfo->comps_in_scan; ci++) { - compptr = cinfo->cur_comp_info[ci]; - start_col = MCU_col_num * compptr->MCU_width; - for (yindex = 0; yindex < compptr->MCU_height; yindex++) { - buffer_ptr = buffer[ci][yoffset + yindex] + start_col; - xindex = compptr->MCU_width; - do { - *blkp++ = buffer_ptr++; - } while (--xindex); - } - } - /* Try to fetch the MCU. */ - if (! (*cinfo->entropy->decode_mcu) (cinfo, coef->MCU_buffer)) { - /* Suspension forced; update state counters and exit */ - coef->MCU_vert_offset = yoffset; - coef->MCU_ctr = MCU_col_num; - return JPEG_SUSPENDED; - } - } - /* Completed an MCU row, but perhaps not an iMCU row */ - coef->MCU_ctr = 0; - } - /* Completed the iMCU row, advance counters for next one */ - if (++(cinfo->input_iMCU_row) < cinfo->total_iMCU_rows) { - start_iMCU_row(cinfo); - return JPEG_ROW_COMPLETED; - } - /* Completed the scan */ - (*cinfo->inputctl->finish_input_pass) (cinfo); - return JPEG_SCAN_COMPLETED; -} - - -/* - * Decompress and return some data in the multi-pass case. - * Always attempts to emit one fully interleaved MCU row ("iMCU" row). - * Return value is JPEG_ROW_COMPLETED, JPEG_SCAN_COMPLETED, or JPEG_SUSPENDED. - * - * NB: output_buf contains a plane for each component in image. - */ - -METHODDEF(int) -decompress_data (j_decompress_ptr cinfo, JSAMPIMAGE output_buf) -{ - my_coef_ptr coef = (my_coef_ptr) cinfo->coef; - JDIMENSION last_iMCU_row = cinfo->total_iMCU_rows - 1; - JDIMENSION block_num; - int ci, block_row, block_rows; - JBLOCKARRAY buffer; - JBLOCKROW buffer_ptr; - JSAMPARRAY output_ptr; - JDIMENSION output_col; - jpeg_component_info *compptr; - inverse_DCT_method_ptr inverse_DCT; - - /* Force some input to be done if we are getting ahead of the input. */ - while (cinfo->input_scan_number < cinfo->output_scan_number || - (cinfo->input_scan_number == cinfo->output_scan_number && - cinfo->input_iMCU_row <= cinfo->output_iMCU_row)) { - if ((*cinfo->inputctl->consume_input)(cinfo) == JPEG_SUSPENDED) - return JPEG_SUSPENDED; - } - - /* OK, output from the virtual arrays. */ - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - /* Don't bother to IDCT an uninteresting component. */ - if (! compptr->component_needed) - continue; - /* Align the virtual buffer for this component. */ - buffer = (*cinfo->mem->access_virt_barray) - ((j_common_ptr) cinfo, coef->whole_image[ci], - cinfo->output_iMCU_row * compptr->v_samp_factor, - (JDIMENSION) compptr->v_samp_factor, FALSE); - /* Count non-dummy DCT block rows in this iMCU row. */ - if (cinfo->output_iMCU_row < last_iMCU_row) - block_rows = compptr->v_samp_factor; - else { - /* NB: can't use last_row_height here; it is input-side-dependent! */ - block_rows = (int) (compptr->height_in_blocks % compptr->v_samp_factor); - if (block_rows == 0) block_rows = compptr->v_samp_factor; - } - inverse_DCT = cinfo->idct->inverse_DCT[ci]; - output_ptr = output_buf[ci]; - /* Loop over all DCT blocks to be processed. */ - for (block_row = 0; block_row < block_rows; block_row++) { - buffer_ptr = buffer[block_row]; - output_col = 0; - for (block_num = 0; block_num < compptr->width_in_blocks; block_num++) { - (*inverse_DCT) (cinfo, compptr, (JCOEFPTR) buffer_ptr, - output_ptr, output_col); - buffer_ptr++; - output_col += compptr->DCT_h_scaled_size; - } - output_ptr += compptr->DCT_v_scaled_size; - } - } - - if (++(cinfo->output_iMCU_row) <= last_iMCU_row) - return JPEG_ROW_COMPLETED; - return JPEG_SCAN_COMPLETED; -} - -#endif /* D_MULTISCAN_FILES_SUPPORTED */ - - -#ifdef BLOCK_SMOOTHING_SUPPORTED - -/* - * This code applies interblock smoothing as described by section K.8 - * of the JPEG standard: the first 5 AC coefficients are estimated from - * the DC values of a DCT block and its 8 neighboring blocks. - * We apply smoothing only for progressive JPEG decoding, and only if - * the coefficients it can estimate are not yet known to full precision. - */ - -/* Natural-order array positions of the first 5 zigzag-order coefficients */ -#define Q01_POS 1 -#define Q10_POS 8 -#define Q20_POS 16 -#define Q11_POS 9 -#define Q02_POS 2 - -/* - * Determine whether block smoothing is applicable and safe. - * We also latch the current states of the coef_bits[] entries for the - * AC coefficients; otherwise, if the input side of the decompressor - * advances into a new scan, we might think the coefficients are known - * more accurately than they really are. - */ - -LOCAL(boolean) -smoothing_ok (j_decompress_ptr cinfo) -{ - my_coef_ptr coef = (my_coef_ptr) cinfo->coef; - boolean smoothing_useful = FALSE; - int ci, coefi; - jpeg_component_info *compptr; - JQUANT_TBL * qtable; - int * coef_bits; - int * coef_bits_latch; - - if (! cinfo->progressive_mode || cinfo->coef_bits == NULL) - return FALSE; - - /* Allocate latch area if not already done */ - if (coef->coef_bits_latch == NULL) - coef->coef_bits_latch = (int *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, - cinfo->num_components * (SAVED_COEFS * SIZEOF(int))); - coef_bits_latch = coef->coef_bits_latch; - - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - /* All components' quantization values must already be latched. */ - if ((qtable = compptr->quant_table) == NULL) - return FALSE; - /* Verify DC & first 5 AC quantizers are nonzero to avoid zero-divide. */ - if (qtable->quantval[0] == 0 || - qtable->quantval[Q01_POS] == 0 || - qtable->quantval[Q10_POS] == 0 || - qtable->quantval[Q20_POS] == 0 || - qtable->quantval[Q11_POS] == 0 || - qtable->quantval[Q02_POS] == 0) - return FALSE; - /* DC values must be at least partly known for all components. */ - coef_bits = cinfo->coef_bits[ci]; - if (coef_bits[0] < 0) - return FALSE; - /* Block smoothing is helpful if some AC coefficients remain inaccurate. */ - for (coefi = 1; coefi <= 5; coefi++) { - coef_bits_latch[coefi] = coef_bits[coefi]; - if (coef_bits[coefi] != 0) - smoothing_useful = TRUE; - } - coef_bits_latch += SAVED_COEFS; - } - - return smoothing_useful; -} - - -/* - * Variant of decompress_data for use when doing block smoothing. - */ - -METHODDEF(int) -decompress_smooth_data (j_decompress_ptr cinfo, JSAMPIMAGE output_buf) -{ - my_coef_ptr coef = (my_coef_ptr) cinfo->coef; - JDIMENSION last_iMCU_row = cinfo->total_iMCU_rows - 1; - JDIMENSION block_num, last_block_column; - int ci, block_row, block_rows, access_rows; - JBLOCKARRAY buffer; - JBLOCKROW buffer_ptr, prev_block_row, next_block_row; - JSAMPARRAY output_ptr; - JDIMENSION output_col; - jpeg_component_info *compptr; - inverse_DCT_method_ptr inverse_DCT; - boolean first_row, last_row; - JBLOCK workspace; - int *coef_bits; - JQUANT_TBL *quanttbl; - INT32 Q00,Q01,Q02,Q10,Q11,Q20, num; - int DC1,DC2,DC3,DC4,DC5,DC6,DC7,DC8,DC9; - int Al, pred; - - /* Force some input to be done if we are getting ahead of the input. */ - while (cinfo->input_scan_number <= cinfo->output_scan_number && - ! cinfo->inputctl->eoi_reached) { - if (cinfo->input_scan_number == cinfo->output_scan_number) { - /* If input is working on current scan, we ordinarily want it to - * have completed the current row. But if input scan is DC, - * we want it to keep one row ahead so that next block row's DC - * values are up to date. - */ - JDIMENSION delta = (cinfo->Ss == 0) ? 1 : 0; - if (cinfo->input_iMCU_row > cinfo->output_iMCU_row+delta) - break; - } - if ((*cinfo->inputctl->consume_input)(cinfo) == JPEG_SUSPENDED) - return JPEG_SUSPENDED; - } - - /* OK, output from the virtual arrays. */ - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - /* Don't bother to IDCT an uninteresting component. */ - if (! compptr->component_needed) - continue; - /* Count non-dummy DCT block rows in this iMCU row. */ - if (cinfo->output_iMCU_row < last_iMCU_row) { - block_rows = compptr->v_samp_factor; - access_rows = block_rows * 2; /* this and next iMCU row */ - last_row = FALSE; - } else { - /* NB: can't use last_row_height here; it is input-side-dependent! */ - block_rows = (int) (compptr->height_in_blocks % compptr->v_samp_factor); - if (block_rows == 0) block_rows = compptr->v_samp_factor; - access_rows = block_rows; /* this iMCU row only */ - last_row = TRUE; - } - /* Align the virtual buffer for this component. */ - if (cinfo->output_iMCU_row > 0) { - access_rows += compptr->v_samp_factor; /* prior iMCU row too */ - buffer = (*cinfo->mem->access_virt_barray) - ((j_common_ptr) cinfo, coef->whole_image[ci], - (cinfo->output_iMCU_row - 1) * compptr->v_samp_factor, - (JDIMENSION) access_rows, FALSE); - buffer += compptr->v_samp_factor; /* point to current iMCU row */ - first_row = FALSE; - } else { - buffer = (*cinfo->mem->access_virt_barray) - ((j_common_ptr) cinfo, coef->whole_image[ci], - (JDIMENSION) 0, (JDIMENSION) access_rows, FALSE); - first_row = TRUE; - } - /* Fetch component-dependent info */ - coef_bits = coef->coef_bits_latch + (ci * SAVED_COEFS); - quanttbl = compptr->quant_table; - Q00 = quanttbl->quantval[0]; - Q01 = quanttbl->quantval[Q01_POS]; - Q10 = quanttbl->quantval[Q10_POS]; - Q20 = quanttbl->quantval[Q20_POS]; - Q11 = quanttbl->quantval[Q11_POS]; - Q02 = quanttbl->quantval[Q02_POS]; - inverse_DCT = cinfo->idct->inverse_DCT[ci]; - output_ptr = output_buf[ci]; - /* Loop over all DCT blocks to be processed. */ - for (block_row = 0; block_row < block_rows; block_row++) { - buffer_ptr = buffer[block_row]; - if (first_row && block_row == 0) - prev_block_row = buffer_ptr; - else - prev_block_row = buffer[block_row-1]; - if (last_row && block_row == block_rows-1) - next_block_row = buffer_ptr; - else - next_block_row = buffer[block_row+1]; - /* We fetch the surrounding DC values using a sliding-register approach. - * Initialize all nine here so as to do the right thing on narrow pics. - */ - DC1 = DC2 = DC3 = (int) prev_block_row[0][0]; - DC4 = DC5 = DC6 = (int) buffer_ptr[0][0]; - DC7 = DC8 = DC9 = (int) next_block_row[0][0]; - output_col = 0; - last_block_column = compptr->width_in_blocks - 1; - for (block_num = 0; block_num <= last_block_column; block_num++) { - /* Fetch current DCT block into workspace so we can modify it. */ - jcopy_block_row(buffer_ptr, (JBLOCKROW) workspace, (JDIMENSION) 1); - /* Update DC values */ - if (block_num < last_block_column) { - DC3 = (int) prev_block_row[1][0]; - DC6 = (int) buffer_ptr[1][0]; - DC9 = (int) next_block_row[1][0]; - } - /* Compute coefficient estimates per K.8. - * An estimate is applied only if coefficient is still zero, - * and is not known to be fully accurate. - */ - /* AC01 */ - if ((Al=coef_bits[1]) != 0 && workspace[1] == 0) { - num = 36 * Q00 * (DC4 - DC6); - if (num >= 0) { - pred = (int) (((Q01<<7) + num) / (Q01<<8)); - if (Al > 0 && pred >= (1< 0 && pred >= (1<= 0) { - pred = (int) (((Q10<<7) + num) / (Q10<<8)); - if (Al > 0 && pred >= (1< 0 && pred >= (1<= 0) { - pred = (int) (((Q20<<7) + num) / (Q20<<8)); - if (Al > 0 && pred >= (1< 0 && pred >= (1<= 0) { - pred = (int) (((Q11<<7) + num) / (Q11<<8)); - if (Al > 0 && pred >= (1< 0 && pred >= (1<= 0) { - pred = (int) (((Q02<<7) + num) / (Q02<<8)); - if (Al > 0 && pred >= (1< 0 && pred >= (1<DCT_h_scaled_size; - } - output_ptr += compptr->DCT_v_scaled_size; - } - } - - if (++(cinfo->output_iMCU_row) <= last_iMCU_row) - return JPEG_ROW_COMPLETED; - return JPEG_SCAN_COMPLETED; -} - -#endif /* BLOCK_SMOOTHING_SUPPORTED */ - - -/* - * Initialize coefficient buffer controller. - */ - -GLOBAL(void) -jinit_d_coef_controller (j_decompress_ptr cinfo, boolean need_full_buffer) -{ - my_coef_ptr coef; - - if (need_full_buffer) { -#ifdef D_MULTISCAN_FILES_SUPPORTED - /* Allocate a full-image virtual array for each component, */ - /* padded to a multiple of samp_factor DCT blocks in each direction. */ - /* Note we ask for a pre-zeroed array. */ - int ci, access_rows; - jpeg_component_info *compptr; - - coef = (my_coef_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, - SIZEOF(my_coef_controller) - SIZEOF(coef->blk_buffer)); - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - access_rows = compptr->v_samp_factor; -#ifdef BLOCK_SMOOTHING_SUPPORTED - /* If block smoothing could be used, need a bigger window */ - if (cinfo->progressive_mode) - access_rows *= 3; -#endif - coef->whole_image[ci] = (*cinfo->mem->request_virt_barray) - ((j_common_ptr) cinfo, JPOOL_IMAGE, TRUE, - (JDIMENSION) jround_up((long) compptr->width_in_blocks, - (long) compptr->h_samp_factor), - (JDIMENSION) jround_up((long) compptr->height_in_blocks, - (long) compptr->v_samp_factor), - (JDIMENSION) access_rows); - } - coef->pub.consume_data = consume_data; - coef->pub.decompress_data = decompress_data; - coef->pub.coef_arrays = coef->whole_image; /* link to virtual arrays */ -#else - ERREXIT(cinfo, JERR_NOT_COMPILED); -#endif - } else { - /* We only need a single-MCU buffer. */ - JBLOCKARRAY blkp; - JBLOCKROW buffer_ptr; - int bi; - - coef = (my_coef_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_coef_controller)); - buffer_ptr = coef->blk_buffer; - if (cinfo->lim_Se == 0) /* DC only case: want to bypass later */ - MEMZERO(buffer_ptr, SIZEOF(coef->blk_buffer)); - blkp = coef->MCU_buffer; - bi = D_MAX_BLOCKS_IN_MCU; - do { - *blkp++ = buffer_ptr++; - } while (--bi); - coef->pub.consume_data = dummy_consume_data; - coef->pub.decompress_data = decompress_onepass; - coef->pub.coef_arrays = NULL; /* flag for no virtual arrays */ - } - - coef->pub.start_input_pass = start_input_pass; - coef->pub.start_output_pass = start_output_pass; -#ifdef BLOCK_SMOOTHING_SUPPORTED - coef->coef_bits_latch = NULL; -#endif - cinfo->coef = &coef->pub; -} diff --git a/cmake/externals/libjpeg/jdcolor.c b/cmake/externals/libjpeg/jdcolor.c deleted file mode 100644 index 6b40fb534..000000000 --- a/cmake/externals/libjpeg/jdcolor.c +++ /dev/null @@ -1,769 +0,0 @@ -/* - * jdcolor.c - * - * Copyright (C) 1991-1997, Thomas G. Lane. - * Modified 2011-2023 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains output colorspace conversion routines. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -#if RANGE_BITS < 2 - /* Deliberate syntax err */ - Sorry, this code requires 2 or more range extension bits. -#endif - - -/* Private subobject */ - -typedef struct { - struct jpeg_color_deconverter pub; /* public fields */ - - /* Private state for YCbCr->RGB and BG_YCC->RGB conversion */ - int * Cr_r_tab; /* => table for Cr to R conversion */ - int * Cb_b_tab; /* => table for Cb to B conversion */ - INT32 * Cr_g_tab; /* => table for Cr to G conversion */ - INT32 * Cb_g_tab; /* => table for Cb to G conversion */ - - /* Private state for RGB->Y conversion */ - INT32 * R_y_tab; /* => table for R to Y conversion */ - INT32 * G_y_tab; /* => table for G to Y conversion */ - INT32 * B_y_tab; /* => table for B to Y conversion */ -} my_color_deconverter; - -typedef my_color_deconverter * my_cconvert_ptr; - - -/*************** YCbCr -> RGB conversion: most common case **************/ -/*************** BG_YCC -> RGB conversion: less common case **************/ -/*************** RGB -> Y conversion: less common case **************/ - -/* - * YCbCr is defined per Recommendation ITU-R BT.601-7 (03/2011), - * previously known as Recommendation CCIR 601-1, except that Cb and Cr - * are normalized to the range 0..MAXJSAMPLE rather than -0.5 .. 0.5. - * sRGB (standard RGB color space) is defined per IEC 61966-2-1:1999. - * sYCC (standard luma-chroma-chroma color space with extended gamut) - * is defined per IEC 61966-2-1:1999 Amendment A1:2003 Annex F. - * bg-sRGB and bg-sYCC (big gamut standard color spaces) - * are defined per IEC 61966-2-1:1999 Amendment A1:2003 Annex G. - * Note that the derived conversion coefficients given in some of these - * documents are imprecise. The general conversion equations are - * - * R = Y + K * (1 - Kr) * Cr - * G = Y - K * (Kb * (1 - Kb) * Cb + Kr * (1 - Kr) * Cr) / (1 - Kr - Kb) - * B = Y + K * (1 - Kb) * Cb - * - * Y = Kr * R + (1 - Kr - Kb) * G + Kb * B - * - * With Kr = 0.299 and Kb = 0.114 (derived according to SMPTE RP 177-1993 - * from the 1953 FCC NTSC primaries and CIE Illuminant C), K = 2 for sYCC, - * the conversion equations to be implemented are therefore - * - * R = Y + 1.402 * Cr - * G = Y - 0.344136286 * Cb - 0.714136286 * Cr - * B = Y + 1.772 * Cb - * - * Y = 0.299 * R + 0.587 * G + 0.114 * B - * - * where Cb and Cr represent the incoming values less CENTERJSAMPLE. - * For bg-sYCC, with K = 4, the equations are - * - * R = Y + 2.804 * Cr - * G = Y - 0.688272572 * Cb - 1.428272572 * Cr - * B = Y + 3.544 * Cb - * - * To avoid floating-point arithmetic, we represent the fractional constants - * as integers scaled up by 2^16 (about 4 digits precision); we have to divide - * the products by 2^16, with appropriate rounding, to get the correct answer. - * Notice that Y, being an integral input, does not contribute any fraction - * so it need not participate in the rounding. - * - * For even more speed, we avoid doing any multiplications in the inner loop - * by precalculating the constants times Cb and Cr for all possible values. - * For 8-bit JSAMPLEs this is very reasonable (only 256 entries per table); - * for 9-bit to 12-bit samples it is still acceptable. It's not very - * reasonable for 16-bit samples, but if you want lossless storage - * you shouldn't be changing colorspace anyway. - * The Cr=>R and Cb=>B values can be rounded to integers in advance; - * the values for the G calculation are left scaled up, - * since we must add them together before rounding. - */ - -#define SCALEBITS 16 /* speediest right-shift on some machines */ -#define ONE_HALF ((INT32) 1 << (SCALEBITS-1)) -#define FIX(x) ((INT32) ((x) * (1L<RGB and BG_YCC->RGB colorspace conversion. - */ - -LOCAL(void) -build_ycc_rgb_table (j_decompress_ptr cinfo) -/* Normal case, sYCC */ -{ - my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; - int i; - INT32 x; - SHIFT_TEMPS - - cconvert->Cr_r_tab = (int *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(int)); - cconvert->Cb_b_tab = (int *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(int)); - cconvert->Cr_g_tab = (INT32 *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(INT32)); - cconvert->Cb_g_tab = (INT32 *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(INT32)); - - for (i = 0, x = -CENTERJSAMPLE; i <= MAXJSAMPLE; i++, x++) { - /* i is the actual input pixel value, in the range 0..MAXJSAMPLE */ - /* The Cb or Cr value we are thinking of is x = i - CENTERJSAMPLE */ - /* Cr=>R value is nearest int to 1.402 * x */ - cconvert->Cr_r_tab[i] = (int) DESCALE(FIX(1.402) * x, SCALEBITS); - /* Cb=>B value is nearest int to 1.772 * x */ - cconvert->Cb_b_tab[i] = (int) DESCALE(FIX(1.772) * x, SCALEBITS); - /* Cr=>G value is scaled-up -0.714136286 * x */ - cconvert->Cr_g_tab[i] = (- FIX(0.714136286)) * x; - /* Cb=>G value is scaled-up -0.344136286 * x */ - /* We also add in ONE_HALF so that need not do it in inner loop */ - cconvert->Cb_g_tab[i] = (- FIX(0.344136286)) * x + ONE_HALF; - } -} - - -LOCAL(void) -build_bg_ycc_rgb_table (j_decompress_ptr cinfo) -/* Wide gamut case, bg-sYCC */ -{ - my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; - int i; - INT32 x; - SHIFT_TEMPS - - cconvert->Cr_r_tab = (int *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(int)); - cconvert->Cb_b_tab = (int *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(int)); - cconvert->Cr_g_tab = (INT32 *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(INT32)); - cconvert->Cb_g_tab = (INT32 *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(INT32)); - - for (i = 0, x = -CENTERJSAMPLE; i <= MAXJSAMPLE; i++, x++) { - /* i is the actual input pixel value, in the range 0..MAXJSAMPLE */ - /* The Cb or Cr value we are thinking of is x = i - CENTERJSAMPLE */ - /* Cr=>R value is nearest int to 2.804 * x */ - cconvert->Cr_r_tab[i] = (int) DESCALE(FIX(2.804) * x, SCALEBITS); - /* Cb=>B value is nearest int to 3.544 * x */ - cconvert->Cb_b_tab[i] = (int) DESCALE(FIX(3.544) * x, SCALEBITS); - /* Cr=>G value is scaled-up -1.428272572 * x */ - cconvert->Cr_g_tab[i] = (- FIX(1.428272572)) * x; - /* Cb=>G value is scaled-up -0.688272572 * x */ - /* We also add in ONE_HALF so that need not do it in inner loop */ - cconvert->Cb_g_tab[i] = (- FIX(0.688272572)) * x + ONE_HALF; - } -} - - -/* - * Convert some rows of samples to the output colorspace. - * - * Note that we change from noninterleaved, one-plane-per-component format - * to interleaved-pixel format. The output buffer is therefore three times - * as wide as the input buffer. - * - * A starting row offset is provided only for the input buffer. The caller - * can easily adjust the passed output_buf value to accommodate any row - * offset required on that side. - */ - -METHODDEF(void) -ycc_rgb_convert (j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION input_row, - JSAMPARRAY output_buf, int num_rows) -{ - my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; - register int y, cb, cr; - register JSAMPROW outptr; - register JSAMPROW inptr0, inptr1, inptr2; - register JDIMENSION col; - JDIMENSION num_cols = cinfo->output_width; - /* copy these pointers into registers if possible */ - register JSAMPLE * range_limit = cinfo->sample_range_limit; - register int * Crrtab = cconvert->Cr_r_tab; - register int * Cbbtab = cconvert->Cb_b_tab; - register INT32 * Crgtab = cconvert->Cr_g_tab; - register INT32 * Cbgtab = cconvert->Cb_g_tab; - SHIFT_TEMPS - - while (--num_rows >= 0) { - inptr0 = input_buf[0][input_row]; - inptr1 = input_buf[1][input_row]; - inptr2 = input_buf[2][input_row]; - input_row++; - outptr = *output_buf++; - for (col = 0; col < num_cols; col++) { - y = GETJSAMPLE(inptr0[col]); - cb = GETJSAMPLE(inptr1[col]); - cr = GETJSAMPLE(inptr2[col]); - /* Range-limiting is essential due to noise introduced by DCT losses, - * for extended gamut (sYCC) and wide gamut (bg-sYCC) encodings. - */ - outptr[RGB_RED] = range_limit[y + Crrtab[cr]]; - outptr[RGB_GREEN] = range_limit[y + - ((int) RIGHT_SHIFT(Cbgtab[cb] + Crgtab[cr], - SCALEBITS))]; - outptr[RGB_BLUE] = range_limit[y + Cbbtab[cb]]; - outptr += RGB_PIXELSIZE; - } - } -} - - -/**************** Cases other than YCC -> RGB ****************/ - - -/* - * Initialize for RGB->grayscale colorspace conversion. - */ - -LOCAL(void) -build_rgb_y_table (j_decompress_ptr cinfo) -{ - my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; - INT32 i; - - cconvert->R_y_tab = (INT32 *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(INT32)); - cconvert->G_y_tab = (INT32 *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(INT32)); - cconvert->B_y_tab = (INT32 *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(INT32)); - - for (i = 0; i <= MAXJSAMPLE; i++) { - cconvert->R_y_tab[i] = FIX(0.299) * i; - cconvert->G_y_tab[i] = FIX(0.587) * i; - cconvert->B_y_tab[i] = FIX(0.114) * i + ONE_HALF; - } -} - - -/* - * Convert RGB to grayscale. - */ - -METHODDEF(void) -rgb_gray_convert (j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION input_row, - JSAMPARRAY output_buf, int num_rows) -{ - my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; - register INT32 y; - register INT32 * Rytab = cconvert->R_y_tab; - register INT32 * Gytab = cconvert->G_y_tab; - register INT32 * Bytab = cconvert->B_y_tab; - register JSAMPROW outptr; - register JSAMPROW inptr0, inptr1, inptr2; - register JDIMENSION col; - JDIMENSION num_cols = cinfo->output_width; - - while (--num_rows >= 0) { - inptr0 = input_buf[0][input_row]; - inptr1 = input_buf[1][input_row]; - inptr2 = input_buf[2][input_row]; - input_row++; - outptr = *output_buf++; - for (col = 0; col < num_cols; col++) { - y = Rytab[GETJSAMPLE(inptr0[col])]; - y += Gytab[GETJSAMPLE(inptr1[col])]; - y += Bytab[GETJSAMPLE(inptr2[col])]; - outptr[col] = (JSAMPLE) (y >> SCALEBITS); - } - } -} - - -/* - * Convert some rows of samples to the output colorspace. - * [R-G,G,B-G] to [R,G,B] conversion with modulo calculation - * (inverse color transform). - * This can be seen as an adaption of the general YCbCr->RGB - * conversion equation with Kr = Kb = 0, while replacing the - * normalization by modulo calculation. - */ - -METHODDEF(void) -rgb1_rgb_convert (j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION input_row, - JSAMPARRAY output_buf, int num_rows) -{ - register int r, g, b; - register JSAMPROW outptr; - register JSAMPROW inptr0, inptr1, inptr2; - register JDIMENSION col; - JDIMENSION num_cols = cinfo->output_width; - - while (--num_rows >= 0) { - inptr0 = input_buf[0][input_row]; - inptr1 = input_buf[1][input_row]; - inptr2 = input_buf[2][input_row]; - input_row++; - outptr = *output_buf++; - for (col = 0; col < num_cols; col++) { - r = GETJSAMPLE(inptr0[col]); - g = GETJSAMPLE(inptr1[col]); - b = GETJSAMPLE(inptr2[col]); - /* Assume that MAXJSAMPLE+1 is a power of 2, so that the MOD - * (modulo) operator is equivalent to the bitmask operator AND. - */ - outptr[RGB_RED] = (JSAMPLE) ((r + g - CENTERJSAMPLE) & MAXJSAMPLE); - outptr[RGB_GREEN] = (JSAMPLE) g; - outptr[RGB_BLUE] = (JSAMPLE) ((b + g - CENTERJSAMPLE) & MAXJSAMPLE); - outptr += RGB_PIXELSIZE; - } - } -} - - -/* - * [R-G,G,B-G] to grayscale conversion with modulo calculation - * (inverse color transform). - */ - -METHODDEF(void) -rgb1_gray_convert (j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION input_row, - JSAMPARRAY output_buf, int num_rows) -{ - my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; - register int r, g, b; - register INT32 y; - register INT32 * Rytab = cconvert->R_y_tab; - register INT32 * Gytab = cconvert->G_y_tab; - register INT32 * Bytab = cconvert->B_y_tab; - register JSAMPROW outptr; - register JSAMPROW inptr0, inptr1, inptr2; - register JDIMENSION col; - JDIMENSION num_cols = cinfo->output_width; - - while (--num_rows >= 0) { - inptr0 = input_buf[0][input_row]; - inptr1 = input_buf[1][input_row]; - inptr2 = input_buf[2][input_row]; - input_row++; - outptr = *output_buf++; - for (col = 0; col < num_cols; col++) { - r = GETJSAMPLE(inptr0[col]); - g = GETJSAMPLE(inptr1[col]); - b = GETJSAMPLE(inptr2[col]); - /* Assume that MAXJSAMPLE+1 is a power of 2, so that the MOD - * (modulo) operator is equivalent to the bitmask operator AND. - */ - y = Rytab[(r + g - CENTERJSAMPLE) & MAXJSAMPLE]; - y += Gytab[g]; - y += Bytab[(b + g - CENTERJSAMPLE) & MAXJSAMPLE]; - outptr[col] = (JSAMPLE) (y >> SCALEBITS); - } - } -} - - -/* - * Convert some rows of samples to the output colorspace. - * No colorspace change, but conversion from separate-planes - * to interleaved representation. - */ - -METHODDEF(void) -rgb_convert (j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION input_row, - JSAMPARRAY output_buf, int num_rows) -{ - register JSAMPROW outptr; - register JSAMPROW inptr0, inptr1, inptr2; - register JDIMENSION col; - JDIMENSION num_cols = cinfo->output_width; - - while (--num_rows >= 0) { - inptr0 = input_buf[0][input_row]; - inptr1 = input_buf[1][input_row]; - inptr2 = input_buf[2][input_row]; - input_row++; - outptr = *output_buf++; - for (col = 0; col < num_cols; col++) { - /* We can dispense with GETJSAMPLE() here */ - outptr[RGB_RED] = inptr0[col]; - outptr[RGB_GREEN] = inptr1[col]; - outptr[RGB_BLUE] = inptr2[col]; - outptr += RGB_PIXELSIZE; - } - } -} - - -/* - * Color conversion for no colorspace change: just copy the data, - * converting from separate-planes to interleaved representation. - * Note: Omit uninteresting components in output buffer. - */ - -METHODDEF(void) -null_convert (j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION input_row, - JSAMPARRAY output_buf, int num_rows) -{ - register JSAMPROW outptr; - register JSAMPROW inptr; - register JDIMENSION count; - register int out_comps = cinfo->out_color_components; - JDIMENSION num_cols = cinfo->output_width; - JSAMPROW startptr; - int ci; - jpeg_component_info *compptr; - - while (--num_rows >= 0) { - /* It seems fastest to make a separate pass for each component. */ - startptr = *output_buf++; - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - if (! compptr->component_needed) - continue; /* skip uninteresting component */ - inptr = input_buf[ci][input_row]; - outptr = startptr++; - for (count = num_cols; count > 0; count--) { - *outptr = *inptr++; /* don't need GETJSAMPLE() here */ - outptr += out_comps; - } - } - input_row++; - } -} - - -/* - * Color conversion for grayscale: just copy the data. - * This also works for YCC -> grayscale conversion, in which - * we just copy the Y (luminance) component and ignore chrominance. - */ - -METHODDEF(void) -grayscale_convert (j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION input_row, - JSAMPARRAY output_buf, int num_rows) -{ - jcopy_sample_rows(input_buf[0] + input_row, output_buf, - num_rows, cinfo->output_width); -} - - -/* - * Convert grayscale to RGB: just duplicate the graylevel three times. - * This is provided to support applications that don't want to cope - * with grayscale as a separate case. - */ - -METHODDEF(void) -gray_rgb_convert (j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION input_row, - JSAMPARRAY output_buf, int num_rows) -{ - register JSAMPROW outptr; - register JSAMPROW inptr; - register JDIMENSION col; - JDIMENSION num_cols = cinfo->output_width; - - while (--num_rows >= 0) { - inptr = input_buf[0][input_row++]; - outptr = *output_buf++; - for (col = 0; col < num_cols; col++) { - /* We can dispense with GETJSAMPLE() here */ - outptr[RGB_RED] = outptr[RGB_GREEN] = outptr[RGB_BLUE] = inptr[col]; - outptr += RGB_PIXELSIZE; - } - } -} - - -/* - * Convert some rows of samples to the output colorspace. - * This version handles Adobe-style YCCK->CMYK conversion, - * where we convert YCbCr to R=1-C, G=1-M, and B=1-Y using the - * same conversion as above, while passing K (black) unchanged. - * We assume build_ycc_rgb_table has been called. - */ - -METHODDEF(void) -ycck_cmyk_convert (j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION input_row, - JSAMPARRAY output_buf, int num_rows) -{ - my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; - register int y, cb, cr; - register JSAMPROW outptr; - register JSAMPROW inptr0, inptr1, inptr2, inptr3; - register JDIMENSION col; - JDIMENSION num_cols = cinfo->output_width; - /* copy these pointers into registers if possible */ - register JSAMPLE * range_limit = cinfo->sample_range_limit; - register int * Crrtab = cconvert->Cr_r_tab; - register int * Cbbtab = cconvert->Cb_b_tab; - register INT32 * Crgtab = cconvert->Cr_g_tab; - register INT32 * Cbgtab = cconvert->Cb_g_tab; - SHIFT_TEMPS - - while (--num_rows >= 0) { - inptr0 = input_buf[0][input_row]; - inptr1 = input_buf[1][input_row]; - inptr2 = input_buf[2][input_row]; - inptr3 = input_buf[3][input_row]; - input_row++; - outptr = *output_buf++; - for (col = 0; col < num_cols; col++) { - y = GETJSAMPLE(inptr0[col]); - cb = GETJSAMPLE(inptr1[col]); - cr = GETJSAMPLE(inptr2[col]); - /* Range-limiting is essential due to noise introduced by DCT losses, - * and for extended gamut encodings (sYCC). - */ - outptr[0] = range_limit[MAXJSAMPLE - (y + Crrtab[cr])]; /* red */ - outptr[1] = range_limit[MAXJSAMPLE - (y + /* green */ - ((int) RIGHT_SHIFT(Cbgtab[cb] + Crgtab[cr], - SCALEBITS)))]; - outptr[2] = range_limit[MAXJSAMPLE - (y + Cbbtab[cb])]; /* blue */ - /* K passes through unchanged */ - outptr[3] = inptr3[col]; /* don't need GETJSAMPLE here */ - outptr += 4; - } - } -} - - -/* - * Convert CMYK to YK part of YCCK for colorless output. - * We assume build_rgb_y_table has been called. - */ - -METHODDEF(void) -cmyk_yk_convert (j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION input_row, - JSAMPARRAY output_buf, int num_rows) -{ - my_cconvert_ptr cconvert = (my_cconvert_ptr) cinfo->cconvert; - register INT32 y; - register INT32 * Rytab = cconvert->R_y_tab; - register INT32 * Gytab = cconvert->G_y_tab; - register INT32 * Bytab = cconvert->B_y_tab; - register JSAMPROW outptr; - register JSAMPROW inptr0, inptr1, inptr2, inptr3; - register JDIMENSION col; - JDIMENSION num_cols = cinfo->output_width; - - while (--num_rows >= 0) { - inptr0 = input_buf[0][input_row]; - inptr1 = input_buf[1][input_row]; - inptr2 = input_buf[2][input_row]; - inptr3 = input_buf[3][input_row]; - input_row++; - outptr = *output_buf++; - for (col = 0; col < num_cols; col++) { - y = Rytab[MAXJSAMPLE - GETJSAMPLE(inptr0[col])]; - y += Gytab[MAXJSAMPLE - GETJSAMPLE(inptr1[col])]; - y += Bytab[MAXJSAMPLE - GETJSAMPLE(inptr2[col])]; - outptr[0] = (JSAMPLE) (y >> SCALEBITS); - /* K passes through unchanged */ - outptr[1] = inptr3[col]; /* don't need GETJSAMPLE here */ - outptr += 2; - } - } -} - - -/* - * Empty method for start_pass. - */ - -METHODDEF(void) -start_pass_dcolor (j_decompress_ptr cinfo) -{ - /* no work needed */ -} - - -/* - * Module initialization routine for output colorspace conversion. - */ - -GLOBAL(void) -jinit_color_deconverter (j_decompress_ptr cinfo) -{ - my_cconvert_ptr cconvert; - int ci, i; - - cconvert = (my_cconvert_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_color_deconverter)); - cinfo->cconvert = &cconvert->pub; - cconvert->pub.start_pass = start_pass_dcolor; - - /* Make sure num_components agrees with jpeg_color_space */ - switch (cinfo->jpeg_color_space) { - case JCS_GRAYSCALE: - if (cinfo->num_components != 1) - ERREXIT(cinfo, JERR_BAD_J_COLORSPACE); - break; - - case JCS_RGB: - case JCS_YCbCr: - case JCS_BG_RGB: - case JCS_BG_YCC: - if (cinfo->num_components != 3) - ERREXIT(cinfo, JERR_BAD_J_COLORSPACE); - break; - - case JCS_CMYK: - case JCS_YCCK: - if (cinfo->num_components != 4) - ERREXIT(cinfo, JERR_BAD_J_COLORSPACE); - break; - - default: /* JCS_UNKNOWN can be anything */ - if (cinfo->num_components < 1) - ERREXIT(cinfo, JERR_BAD_J_COLORSPACE); - } - - /* Support color transform only for RGB colorspaces */ - if (cinfo->color_transform && - cinfo->jpeg_color_space != JCS_RGB && - cinfo->jpeg_color_space != JCS_BG_RGB) - ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); - - /* Set out_color_components and conversion method based on requested space. - * Also adjust the component_needed flags for any unused components, - * so that earlier pipeline stages can avoid useless computation. - */ - - switch (cinfo->out_color_space) { - case JCS_GRAYSCALE: - cinfo->out_color_components = 1; - switch (cinfo->jpeg_color_space) { - case JCS_GRAYSCALE: - case JCS_YCbCr: - case JCS_BG_YCC: - cconvert->pub.color_convert = grayscale_convert; - /* For color->grayscale conversion, only the Y (0) component is needed */ - for (ci = 1; ci < cinfo->num_components; ci++) - cinfo->comp_info[ci].component_needed = FALSE; - break; - case JCS_RGB: - switch (cinfo->color_transform) { - case JCT_NONE: - cconvert->pub.color_convert = rgb_gray_convert; - break; - case JCT_SUBTRACT_GREEN: - cconvert->pub.color_convert = rgb1_gray_convert; - break; - default: - ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); - } - build_rgb_y_table(cinfo); - break; - default: - ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); - } - break; - - case JCS_RGB: - cinfo->out_color_components = RGB_PIXELSIZE; - switch (cinfo->jpeg_color_space) { - case JCS_GRAYSCALE: - cconvert->pub.color_convert = gray_rgb_convert; - break; - case JCS_YCbCr: - cconvert->pub.color_convert = ycc_rgb_convert; - build_ycc_rgb_table(cinfo); - break; - case JCS_BG_YCC: - cconvert->pub.color_convert = ycc_rgb_convert; - build_bg_ycc_rgb_table(cinfo); - break; - case JCS_RGB: - switch (cinfo->color_transform) { - case JCT_NONE: - cconvert->pub.color_convert = rgb_convert; - break; - case JCT_SUBTRACT_GREEN: - cconvert->pub.color_convert = rgb1_rgb_convert; - break; - default: - ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); - } - break; - default: - ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); - } - break; - - case JCS_BG_RGB: - if (cinfo->jpeg_color_space != JCS_BG_RGB) - ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); - cinfo->out_color_components = RGB_PIXELSIZE; - switch (cinfo->color_transform) { - case JCT_NONE: - cconvert->pub.color_convert = rgb_convert; - break; - case JCT_SUBTRACT_GREEN: - cconvert->pub.color_convert = rgb1_rgb_convert; - break; - default: - ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); - } - break; - - case JCS_CMYK: - if (cinfo->jpeg_color_space != JCS_YCCK) - goto def_label; - cinfo->out_color_components = 4; - cconvert->pub.color_convert = ycck_cmyk_convert; - build_ycc_rgb_table(cinfo); - break; - - case JCS_YCCK: - if (cinfo->jpeg_color_space != JCS_CMYK || - /* Support only YK part of YCCK for colorless output */ - ! cinfo->comp_info[0].component_needed || - cinfo->comp_info[1].component_needed || - cinfo->comp_info[2].component_needed || - ! cinfo->comp_info[3].component_needed) - goto def_label; - cinfo->out_color_components = 2; - /* Need all components on input side */ - cinfo->comp_info[1].component_needed = TRUE; - cinfo->comp_info[2].component_needed = TRUE; - cconvert->pub.color_convert = cmyk_yk_convert; - build_rgb_y_table(cinfo); - break; - - default: def_label: /* permit null conversion to same output space */ - if (cinfo->out_color_space != cinfo->jpeg_color_space) - /* unsupported non-null conversion */ - ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); - i = 0; - for (ci = 0; ci < cinfo->num_components; ci++) - if (cinfo->comp_info[ci].component_needed) - i++; /* count output color components */ - cinfo->out_color_components = i; - cconvert->pub.color_convert = null_convert; - } - - if (cinfo->quantize_colors) - cinfo->output_components = 1; /* single colormapped output component */ - else - cinfo->output_components = cinfo->out_color_components; -} diff --git a/cmake/externals/libjpeg/jdct.h b/cmake/externals/libjpeg/jdct.h deleted file mode 100644 index 0f251590c..000000000 --- a/cmake/externals/libjpeg/jdct.h +++ /dev/null @@ -1,409 +0,0 @@ -/* - * jdct.h - * - * Copyright (C) 1994-1996, Thomas G. Lane. - * Modified 2002-2023 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This include file contains common declarations for the forward and - * inverse DCT modules. These declarations are private to the DCT managers - * (jcdctmgr.c, jddctmgr.c) and the individual DCT algorithms. - * The individual DCT algorithms are kept in separate files to ease - * machine-dependent tuning (e.g., assembly coding). - */ - - -/* - * A forward DCT routine is given a pointer to an input sample array and - * a pointer to a work area of type DCTELEM[]; the DCT is to be performed - * in-place in that buffer. Type DCTELEM is int for 8-bit samples, INT32 - * for 12-bit samples. (NOTE: Floating-point DCT implementations use an - * array of type FAST_FLOAT, instead.) - * The input data is to be fetched from the sample array starting at a - * specified column. (Any row offset needed will be applied to the array - * pointer before it is passed to the FDCT code.) - * Note that the number of samples fetched by the FDCT routine is - * DCT_h_scaled_size * DCT_v_scaled_size. - * The DCT outputs are returned scaled up by a factor of 8; they therefore - * have a range of +-8K for 8-bit data, +-128K for 12-bit data. This - * convention improves accuracy in integer implementations and saves some - * work in floating-point ones. - * Quantization of the output coefficients is done by jcdctmgr.c. - */ - -#if BITS_IN_JSAMPLE == 8 -typedef int DCTELEM; /* 16 or 32 bits is fine */ -#else -typedef INT32 DCTELEM; /* must have 32 bits */ -#endif - -typedef JMETHOD(void, forward_DCT_method_ptr, (DCTELEM * data, - JSAMPARRAY sample_data, - JDIMENSION start_col)); -typedef JMETHOD(void, float_DCT_method_ptr, (FAST_FLOAT * data, - JSAMPARRAY sample_data, - JDIMENSION start_col)); - - -/* - * An inverse DCT routine is given a pointer to the input JBLOCK and a pointer - * to an output sample array. The routine must dequantize the input data as - * well as perform the IDCT; for dequantization, it uses the multiplier table - * pointed to by compptr->dct_table. The output data is to be placed into the - * sample array starting at a specified column. (Any row offset needed will - * be applied to the array pointer before it is passed to the IDCT code.) - * Note that the number of samples emitted by the IDCT routine is - * DCT_h_scaled_size * DCT_v_scaled_size. - */ - -/* typedef inverse_DCT_method_ptr is declared in jpegint.h */ - -/* - * Each IDCT routine has its own ideas about the best dct_table element type. - */ - -typedef MULTIPLIER ISLOW_MULT_TYPE; /* short or int, whichever is faster */ -#if BITS_IN_JSAMPLE == 8 -typedef MULTIPLIER IFAST_MULT_TYPE; /* 16 bits is OK, use short if faster */ -#define IFAST_SCALE_BITS 2 /* fractional bits in scale factors */ -#else -typedef INT32 IFAST_MULT_TYPE; /* need 32 bits for scaled quantizers */ -#define IFAST_SCALE_BITS 13 /* fractional bits in scale factors */ -#endif -typedef FAST_FLOAT FLOAT_MULT_TYPE; /* preferred floating type */ - - -/* - * Each IDCT routine is responsible for range-limiting its results and - * converting them to unsigned form (0..MAXJSAMPLE). The raw outputs could - * be quite far out of range if the input data is corrupt, so a bulletproof - * range-limiting step is required. We use a mask-and-table-lookup method - * to do the combined operations quickly, assuming that RANGE_CENTER - * (defined in jpegint.h) is a power of 2. See the comments with - * prepare_range_limit_table (in jdmaster.c) for more info. - */ - -#define RANGE_MASK (RANGE_CENTER * 2 - 1) -#define RANGE_SUBSET (RANGE_CENTER - CENTERJSAMPLE) - -#define IDCT_range_limit(cinfo) ((cinfo)->sample_range_limit - RANGE_SUBSET) - - -/* Short forms of external names for systems with brain-damaged linkers. */ - -#ifdef NEED_SHORT_EXTERNAL_NAMES -#define jpeg_fdct_islow jFDislow -#define jpeg_fdct_ifast jFDifast -#define jpeg_fdct_float jFDfloat -#define jpeg_fdct_7x7 jFD7x7 -#define jpeg_fdct_6x6 jFD6x6 -#define jpeg_fdct_5x5 jFD5x5 -#define jpeg_fdct_4x4 jFD4x4 -#define jpeg_fdct_3x3 jFD3x3 -#define jpeg_fdct_2x2 jFD2x2 -#define jpeg_fdct_1x1 jFD1x1 -#define jpeg_fdct_9x9 jFD9x9 -#define jpeg_fdct_10x10 jFD10x10 -#define jpeg_fdct_11x11 jFD11x11 -#define jpeg_fdct_12x12 jFD12x12 -#define jpeg_fdct_13x13 jFD13x13 -#define jpeg_fdct_14x14 jFD14x14 -#define jpeg_fdct_15x15 jFD15x15 -#define jpeg_fdct_16x16 jFD16x16 -#define jpeg_fdct_16x8 jFD16x8 -#define jpeg_fdct_14x7 jFD14x7 -#define jpeg_fdct_12x6 jFD12x6 -#define jpeg_fdct_10x5 jFD10x5 -#define jpeg_fdct_8x4 jFD8x4 -#define jpeg_fdct_6x3 jFD6x3 -#define jpeg_fdct_4x2 jFD4x2 -#define jpeg_fdct_2x1 jFD2x1 -#define jpeg_fdct_8x16 jFD8x16 -#define jpeg_fdct_7x14 jFD7x14 -#define jpeg_fdct_6x12 jFD6x12 -#define jpeg_fdct_5x10 jFD5x10 -#define jpeg_fdct_4x8 jFD4x8 -#define jpeg_fdct_3x6 jFD3x6 -#define jpeg_fdct_2x4 jFD2x4 -#define jpeg_fdct_1x2 jFD1x2 -#define jpeg_idct_islow jRDislow -#define jpeg_idct_ifast jRDifast -#define jpeg_idct_float jRDfloat -#define jpeg_idct_7x7 jRD7x7 -#define jpeg_idct_6x6 jRD6x6 -#define jpeg_idct_5x5 jRD5x5 -#define jpeg_idct_4x4 jRD4x4 -#define jpeg_idct_3x3 jRD3x3 -#define jpeg_idct_2x2 jRD2x2 -#define jpeg_idct_1x1 jRD1x1 -#define jpeg_idct_9x9 jRD9x9 -#define jpeg_idct_10x10 jRD10x10 -#define jpeg_idct_11x11 jRD11x11 -#define jpeg_idct_12x12 jRD12x12 -#define jpeg_idct_13x13 jRD13x13 -#define jpeg_idct_14x14 jRD14x14 -#define jpeg_idct_15x15 jRD15x15 -#define jpeg_idct_16x16 jRD16x16 -#define jpeg_idct_16x8 jRD16x8 -#define jpeg_idct_14x7 jRD14x7 -#define jpeg_idct_12x6 jRD12x6 -#define jpeg_idct_10x5 jRD10x5 -#define jpeg_idct_8x4 jRD8x4 -#define jpeg_idct_6x3 jRD6x3 -#define jpeg_idct_4x2 jRD4x2 -#define jpeg_idct_2x1 jRD2x1 -#define jpeg_idct_8x16 jRD8x16 -#define jpeg_idct_7x14 jRD7x14 -#define jpeg_idct_6x12 jRD6x12 -#define jpeg_idct_5x10 jRD5x10 -#define jpeg_idct_4x8 jRD4x8 -#define jpeg_idct_3x6 jRD3x6 -#define jpeg_idct_2x4 jRD2x4 -#define jpeg_idct_1x2 jRD1x2 -#endif /* NEED_SHORT_EXTERNAL_NAMES */ - -/* Extern declarations for the forward and inverse DCT routines. */ - -EXTERN(void) jpeg_fdct_islow - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_ifast - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_float - JPP((FAST_FLOAT * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_7x7 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_6x6 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_5x5 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_4x4 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_3x3 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_2x2 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_1x1 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_9x9 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_10x10 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_11x11 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_12x12 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_13x13 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_14x14 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_15x15 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_16x16 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_16x8 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_14x7 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_12x6 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_10x5 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_8x4 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_6x3 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_4x2 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_2x1 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_8x16 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_7x14 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_6x12 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_5x10 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_4x8 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_3x6 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_2x4 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); -EXTERN(void) jpeg_fdct_1x2 - JPP((DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col)); - -EXTERN(void) jpeg_idct_islow - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_ifast - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_float - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_7x7 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_6x6 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_5x5 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_4x4 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_3x3 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_2x2 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_1x1 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_9x9 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_10x10 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_11x11 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_12x12 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_13x13 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_14x14 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_15x15 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_16x16 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_16x8 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_14x7 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_12x6 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_10x5 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_8x4 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_6x3 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_4x2 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_2x1 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_8x16 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_7x14 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_6x12 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_5x10 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_4x8 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_3x6 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_2x4 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); -EXTERN(void) jpeg_idct_1x2 - JPP((j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, JSAMPARRAY output_buf, JDIMENSION output_col)); - - -/* - * Macros for handling fixed-point arithmetic; these are used by many - * but not all of the DCT/IDCT modules. - * - * All values are expected to be of type INT32. - * Fractional constants are scaled left by CONST_BITS bits. - * CONST_BITS is defined within each module using these macros, - * and may differ from one module to the next. - */ - -#define ONE ((INT32) 1) -#define CONST_SCALE (ONE << CONST_BITS) - -/* Convert a positive real constant to an integer scaled by CONST_SCALE. - * Caution: some C compilers fail to reduce "FIX(constant)" at compile time, - * thus causing a lot of useless floating-point operations at run time. - */ - -#define FIX(x) ((INT32) ((x) * CONST_SCALE + 0.5)) - -/* Multiply an INT32 variable by an INT32 constant to yield an INT32 result. - * This macro is used only when the two inputs will actually be no more than - * 16 bits wide, so that a 16x16->32 bit multiply can be used instead of a - * full 32x32 multiply. This provides a useful speedup on many machines. - * Unfortunately there is no way to specify a 16x16->32 multiply portably - * in C, but some C compilers will do the right thing if you provide the - * correct combination of casts. - */ - -#ifdef SHORTxSHORT_32 /* may work if 'int' is 32 bits */ -#define MULTIPLY16C16(var,const) (((INT16) (var)) * ((INT16) (const))) -#endif -#ifdef SHORTxLCONST_32 /* known to work with Microsoft C 6.0 */ -#define MULTIPLY16C16(var,const) (((INT16) (var)) * ((INT32) (const))) -#endif - -#ifndef MULTIPLY16C16 /* default definition */ -#define MULTIPLY16C16(var,const) ((var) * (const)) -#endif - -/* Same except both inputs are variables. */ - -#ifdef SHORTxSHORT_32 /* may work if 'int' is 32 bits */ -#define MULTIPLY16V16(var1,var2) (((INT16) (var1)) * ((INT16) (var2))) -#endif - -#ifndef MULTIPLY16V16 /* default definition */ -#define MULTIPLY16V16(var1,var2) ((var1) * (var2)) -#endif - -/* Like RIGHT_SHIFT, but applies to a DCTELEM. - * We assume that int right shift is unsigned if INT32 right shift is. - */ - -#ifdef RIGHT_SHIFT_IS_UNSIGNED -#define ISHIFT_TEMPS DCTELEM ishift_temp; -#if BITS_IN_JSAMPLE == 8 -#define DCTELEMBITS 16 /* DCTELEM may be 16 or 32 bits */ -#else -#define DCTELEMBITS 32 /* DCTELEM must be 32 bits */ -#endif -#define IRIGHT_SHIFT(x,shft) \ - ((ishift_temp = (x)) < 0 ? \ - (ishift_temp >> (shft)) | ((~((DCTELEM) 0)) << (DCTELEMBITS-(shft))) : \ - (ishift_temp >> (shft))) -#else -#define ISHIFT_TEMPS -#define IRIGHT_SHIFT(x,shft) ((x) >> (shft)) -#endif diff --git a/cmake/externals/libjpeg/jddctmgr.c b/cmake/externals/libjpeg/jddctmgr.c deleted file mode 100644 index 9ecfbb510..000000000 --- a/cmake/externals/libjpeg/jddctmgr.c +++ /dev/null @@ -1,384 +0,0 @@ -/* - * jddctmgr.c - * - * Copyright (C) 1994-1996, Thomas G. Lane. - * Modified 2002-2013 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains the inverse-DCT management logic. - * This code selects a particular IDCT implementation to be used, - * and it performs related housekeeping chores. No code in this file - * is executed per IDCT step, only during output pass setup. - * - * Note that the IDCT routines are responsible for performing coefficient - * dequantization as well as the IDCT proper. This module sets up the - * dequantization multiplier table needed by the IDCT routine. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" -#include "jdct.h" /* Private declarations for DCT subsystem */ - - -/* - * The decompressor input side (jdinput.c) saves away the appropriate - * quantization table for each component at the start of the first scan - * involving that component. (This is necessary in order to correctly - * decode files that reuse Q-table slots.) - * When we are ready to make an output pass, the saved Q-table is converted - * to a multiplier table that will actually be used by the IDCT routine. - * The multiplier table contents are IDCT-method-dependent. To support - * application changes in IDCT method between scans, we can remake the - * multiplier tables if necessary. - * In buffered-image mode, the first output pass may occur before any data - * has been seen for some components, and thus before their Q-tables have - * been saved away. To handle this case, multiplier tables are preset - * to zeroes; the result of the IDCT will be a neutral gray level. - */ - - -/* Private subobject for this module */ - -typedef struct { - struct jpeg_inverse_dct pub; /* public fields */ - - /* This array contains the IDCT method code that each multiplier table - * is currently set up for, or -1 if it's not yet set up. - * The actual multiplier tables are pointed to by dct_table in the - * per-component comp_info structures. - */ - int cur_method[MAX_COMPONENTS]; -} my_idct_controller; - -typedef my_idct_controller * my_idct_ptr; - - -/* Allocated multiplier tables: big enough for any supported variant */ - -typedef union { - ISLOW_MULT_TYPE islow_array[DCTSIZE2]; -#ifdef DCT_IFAST_SUPPORTED - IFAST_MULT_TYPE ifast_array[DCTSIZE2]; -#endif -#ifdef DCT_FLOAT_SUPPORTED - FLOAT_MULT_TYPE float_array[DCTSIZE2]; -#endif -} multiplier_table; - - -/* The current scaled-IDCT routines require ISLOW-style multiplier tables, - * so be sure to compile that code if either ISLOW or SCALING is requested. - */ -#ifdef DCT_ISLOW_SUPPORTED -#define PROVIDE_ISLOW_TABLES -#else -#ifdef IDCT_SCALING_SUPPORTED -#define PROVIDE_ISLOW_TABLES -#endif -#endif - - -/* - * Prepare for an output pass. - * Here we select the proper IDCT routine for each component and build - * a matching multiplier table. - */ - -METHODDEF(void) -start_pass (j_decompress_ptr cinfo) -{ - my_idct_ptr idct = (my_idct_ptr) cinfo->idct; - int ci, i; - jpeg_component_info *compptr; - int method = 0; - inverse_DCT_method_ptr method_ptr = NULL; - JQUANT_TBL * qtbl; - - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - /* Select the proper IDCT routine for this component's scaling */ - switch ((compptr->DCT_h_scaled_size << 8) + compptr->DCT_v_scaled_size) { -#ifdef IDCT_SCALING_SUPPORTED - case ((1 << 8) + 1): - method_ptr = jpeg_idct_1x1; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((2 << 8) + 2): - method_ptr = jpeg_idct_2x2; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((3 << 8) + 3): - method_ptr = jpeg_idct_3x3; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((4 << 8) + 4): - method_ptr = jpeg_idct_4x4; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((5 << 8) + 5): - method_ptr = jpeg_idct_5x5; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((6 << 8) + 6): - method_ptr = jpeg_idct_6x6; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((7 << 8) + 7): - method_ptr = jpeg_idct_7x7; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((9 << 8) + 9): - method_ptr = jpeg_idct_9x9; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((10 << 8) + 10): - method_ptr = jpeg_idct_10x10; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((11 << 8) + 11): - method_ptr = jpeg_idct_11x11; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((12 << 8) + 12): - method_ptr = jpeg_idct_12x12; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((13 << 8) + 13): - method_ptr = jpeg_idct_13x13; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((14 << 8) + 14): - method_ptr = jpeg_idct_14x14; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((15 << 8) + 15): - method_ptr = jpeg_idct_15x15; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((16 << 8) + 16): - method_ptr = jpeg_idct_16x16; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((16 << 8) + 8): - method_ptr = jpeg_idct_16x8; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((14 << 8) + 7): - method_ptr = jpeg_idct_14x7; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((12 << 8) + 6): - method_ptr = jpeg_idct_12x6; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((10 << 8) + 5): - method_ptr = jpeg_idct_10x5; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((8 << 8) + 4): - method_ptr = jpeg_idct_8x4; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((6 << 8) + 3): - method_ptr = jpeg_idct_6x3; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((4 << 8) + 2): - method_ptr = jpeg_idct_4x2; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((2 << 8) + 1): - method_ptr = jpeg_idct_2x1; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((8 << 8) + 16): - method_ptr = jpeg_idct_8x16; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((7 << 8) + 14): - method_ptr = jpeg_idct_7x14; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((6 << 8) + 12): - method_ptr = jpeg_idct_6x12; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((5 << 8) + 10): - method_ptr = jpeg_idct_5x10; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((4 << 8) + 8): - method_ptr = jpeg_idct_4x8; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((3 << 8) + 6): - method_ptr = jpeg_idct_3x6; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((2 << 8) + 4): - method_ptr = jpeg_idct_2x4; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; - case ((1 << 8) + 2): - method_ptr = jpeg_idct_1x2; - method = JDCT_ISLOW; /* jidctint uses islow-style table */ - break; -#endif - case ((DCTSIZE << 8) + DCTSIZE): - switch (cinfo->dct_method) { -#ifdef DCT_ISLOW_SUPPORTED - case JDCT_ISLOW: - method_ptr = jpeg_idct_islow; - method = JDCT_ISLOW; - break; -#endif -#ifdef DCT_IFAST_SUPPORTED - case JDCT_IFAST: - method_ptr = jpeg_idct_ifast; - method = JDCT_IFAST; - break; -#endif -#ifdef DCT_FLOAT_SUPPORTED - case JDCT_FLOAT: - method_ptr = jpeg_idct_float; - method = JDCT_FLOAT; - break; -#endif - default: - ERREXIT(cinfo, JERR_NOT_COMPILED); - break; - } - break; - default: - ERREXIT2(cinfo, JERR_BAD_DCTSIZE, - compptr->DCT_h_scaled_size, compptr->DCT_v_scaled_size); - break; - } - idct->pub.inverse_DCT[ci] = method_ptr; - /* Create multiplier table from quant table. - * However, we can skip this if the component is uninteresting - * or if we already built the table. Also, if no quant table - * has yet been saved for the component, we leave the - * multiplier table all-zero; we'll be reading zeroes from the - * coefficient controller's buffer anyway. - */ - if (! compptr->component_needed || idct->cur_method[ci] == method) - continue; - qtbl = compptr->quant_table; - if (qtbl == NULL) /* happens if no data yet for component */ - continue; - idct->cur_method[ci] = method; - switch (method) { -#ifdef PROVIDE_ISLOW_TABLES - case JDCT_ISLOW: - { - /* For LL&M IDCT method, multipliers are equal to raw quantization - * coefficients, but are stored as ints to ensure access efficiency. - */ - ISLOW_MULT_TYPE * ismtbl = (ISLOW_MULT_TYPE *) compptr->dct_table; - for (i = 0; i < DCTSIZE2; i++) { - ismtbl[i] = (ISLOW_MULT_TYPE) qtbl->quantval[i]; - } - } - break; -#endif -#ifdef DCT_IFAST_SUPPORTED - case JDCT_IFAST: - { - /* For AA&N IDCT method, multipliers are equal to quantization - * coefficients scaled by scalefactor[row]*scalefactor[col], where - * scalefactor[0] = 1 - * scalefactor[k] = cos(k*PI/16) * sqrt(2) for k=1..7 - * For integer operation, the multiplier table is to be scaled by - * IFAST_SCALE_BITS. - */ - IFAST_MULT_TYPE * ifmtbl = (IFAST_MULT_TYPE *) compptr->dct_table; -#define CONST_BITS 14 - static const INT16 aanscales[DCTSIZE2] = { - /* precomputed values scaled up by 14 bits */ - 16384, 22725, 21407, 19266, 16384, 12873, 8867, 4520, - 22725, 31521, 29692, 26722, 22725, 17855, 12299, 6270, - 21407, 29692, 27969, 25172, 21407, 16819, 11585, 5906, - 19266, 26722, 25172, 22654, 19266, 15137, 10426, 5315, - 16384, 22725, 21407, 19266, 16384, 12873, 8867, 4520, - 12873, 17855, 16819, 15137, 12873, 10114, 6967, 3552, - 8867, 12299, 11585, 10426, 8867, 6967, 4799, 2446, - 4520, 6270, 5906, 5315, 4520, 3552, 2446, 1247 - }; - SHIFT_TEMPS - - for (i = 0; i < DCTSIZE2; i++) { - ifmtbl[i] = (IFAST_MULT_TYPE) - DESCALE(MULTIPLY16V16((INT32) qtbl->quantval[i], - (INT32) aanscales[i]), - CONST_BITS-IFAST_SCALE_BITS); - } - } - break; -#endif -#ifdef DCT_FLOAT_SUPPORTED - case JDCT_FLOAT: - { - /* For float AA&N IDCT method, multipliers are equal to quantization - * coefficients scaled by scalefactor[row]*scalefactor[col], where - * scalefactor[0] = 1 - * scalefactor[k] = cos(k*PI/16) * sqrt(2) for k=1..7 - * We apply a further scale factor of 1/8. - */ - FLOAT_MULT_TYPE * fmtbl = (FLOAT_MULT_TYPE *) compptr->dct_table; - int row, col; - static const double aanscalefactor[DCTSIZE] = { - 1.0, 1.387039845, 1.306562965, 1.175875602, - 1.0, 0.785694958, 0.541196100, 0.275899379 - }; - - i = 0; - for (row = 0; row < DCTSIZE; row++) { - for (col = 0; col < DCTSIZE; col++) { - fmtbl[i] = (FLOAT_MULT_TYPE) - ((double) qtbl->quantval[i] * - aanscalefactor[row] * aanscalefactor[col] * 0.125); - i++; - } - } - } - break; -#endif - default: - ERREXIT(cinfo, JERR_NOT_COMPILED); - break; - } - } -} - - -/* - * Initialize IDCT manager. - */ - -GLOBAL(void) -jinit_inverse_dct (j_decompress_ptr cinfo) -{ - my_idct_ptr idct; - int ci; - jpeg_component_info *compptr; - - idct = (my_idct_ptr) - (*cinfo->mem->alloc_small) ((j_common_ptr) cinfo, JPOOL_IMAGE, - SIZEOF(my_idct_controller)); - cinfo->idct = &idct->pub; - idct->pub.start_pass = start_pass; - - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - /* Allocate and pre-zero a multiplier table for each component */ - compptr->dct_table = - (*cinfo->mem->alloc_small) ((j_common_ptr) cinfo, JPOOL_IMAGE, - SIZEOF(multiplier_table)); - MEMZERO(compptr->dct_table, SIZEOF(multiplier_table)); - /* Mark multiplier table not yet set up for any method */ - idct->cur_method[ci] = -1; - } -} diff --git a/cmake/externals/libjpeg/jdhuff.c b/cmake/externals/libjpeg/jdhuff.c deleted file mode 100644 index f175f0c32..000000000 --- a/cmake/externals/libjpeg/jdhuff.c +++ /dev/null @@ -1,1559 +0,0 @@ -/* - * jdhuff.c - * - * Copyright (C) 1991-1997, Thomas G. Lane. - * Modified 2006-2020 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains Huffman entropy decoding routines. - * Both sequential and progressive modes are supported in this single module. - * - * Much of the complexity here has to do with supporting input suspension. - * If the data source module demands suspension, we want to be able to back - * up to the start of the current MCU. To do this, we copy state variables - * into local working storage, and update them back to the permanent - * storage only upon successful completion of an MCU. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* Derived data constructed for each Huffman table */ - -#define HUFF_LOOKAHEAD 8 /* # of bits of lookahead */ - -typedef struct { - /* Basic tables: (element [0] of each array is unused) */ - INT32 maxcode[18]; /* largest code of length k (-1 if none) */ - /* (maxcode[17] is a sentinel to ensure jpeg_huff_decode terminates) */ - INT32 valoffset[17]; /* huffval[] offset for codes of length k */ - /* valoffset[k] = huffval[] index of 1st symbol of code length k, less - * the smallest code of length k; so given a code of length k, the - * corresponding symbol is huffval[code + valoffset[k]] - */ - - /* Link to public Huffman table (needed only in jpeg_huff_decode) */ - JHUFF_TBL *pub; - - /* Lookahead tables: indexed by the next HUFF_LOOKAHEAD bits of - * the input data stream. If the next Huffman code is no more - * than HUFF_LOOKAHEAD bits long, we can obtain its length and - * the corresponding symbol directly from these tables. - */ - int look_nbits[1< 32 bits on your machine, and shifting/masking longs is - * reasonably fast, making bit_buf_type be long and setting BIT_BUF_SIZE - * appropriately should be a win. Unfortunately we can't define the size - * with something like #define BIT_BUF_SIZE (sizeof(bit_buf_type)*8) - * because not all machines measure sizeof in 8-bit bytes. - */ - -typedef struct { /* Bitreading state saved across MCUs */ - bit_buf_type get_buffer; /* current bit-extraction buffer */ - int bits_left; /* # of unused bits in it */ -} bitread_perm_state; - -typedef struct { /* Bitreading working state within an MCU */ - /* Current data source location */ - /* We need a copy, rather than munging the original, in case of suspension */ - const JOCTET * next_input_byte; /* => next byte to read from source */ - size_t bytes_in_buffer; /* # of bytes remaining in source buffer */ - /* Bit input buffer --- note these values are kept in register variables, - * not in this struct, inside the inner loops. - */ - bit_buf_type get_buffer; /* current bit-extraction buffer */ - int bits_left; /* # of unused bits in it */ - /* Pointer needed by jpeg_fill_bit_buffer. */ - j_decompress_ptr cinfo; /* back link to decompress master record */ -} bitread_working_state; - -/* Macros to declare and load/save bitread local variables. */ -#define BITREAD_STATE_VARS \ - register bit_buf_type get_buffer; \ - register int bits_left; \ - bitread_working_state br_state - -#define BITREAD_LOAD_STATE(cinfop,permstate) \ - br_state.cinfo = cinfop; \ - br_state.next_input_byte = cinfop->src->next_input_byte; \ - br_state.bytes_in_buffer = cinfop->src->bytes_in_buffer; \ - get_buffer = permstate.get_buffer; \ - bits_left = permstate.bits_left; - -#define BITREAD_SAVE_STATE(cinfop,permstate) \ - cinfop->src->next_input_byte = br_state.next_input_byte; \ - cinfop->src->bytes_in_buffer = br_state.bytes_in_buffer; \ - permstate.get_buffer = get_buffer; \ - permstate.bits_left = bits_left - -/* - * These macros provide the in-line portion of bit fetching. - * Use CHECK_BIT_BUFFER to ensure there are N bits in get_buffer - * before using GET_BITS, PEEK_BITS, or DROP_BITS. - * The variables get_buffer and bits_left are assumed to be locals, - * but the state struct might not be (jpeg_huff_decode needs this). - * CHECK_BIT_BUFFER(state,n,action); - * Ensure there are N bits in get_buffer; if suspend, take action. - * val = GET_BITS(n); - * Fetch next N bits. - * val = PEEK_BITS(n); - * Fetch next N bits without removing them from the buffer. - * DROP_BITS(n); - * Discard next N bits. - * The value N should be a simple variable, not an expression, because it - * is evaluated multiple times. - */ - -#define CHECK_BIT_BUFFER(state,nbits,action) \ - { if (bits_left < (nbits)) { \ - if (! jpeg_fill_bit_buffer(&(state),get_buffer,bits_left,nbits)) \ - { action; } \ - get_buffer = (state).get_buffer; bits_left = (state).bits_left; } } - -#define GET_BITS(nbits) \ - (((int) (get_buffer >> (bits_left -= (nbits)))) & BIT_MASK(nbits)) - -#define PEEK_BITS(nbits) \ - (((int) (get_buffer >> (bits_left - (nbits)))) & BIT_MASK(nbits)) - -#define DROP_BITS(nbits) \ - (bits_left -= (nbits)) - - -/* - * Code for extracting next Huffman-coded symbol from input bit stream. - * Again, this is time-critical and we make the main paths be macros. - * - * We use a lookahead table to process codes of up to HUFF_LOOKAHEAD bits - * without looping. Usually, more than 95% of the Huffman codes will be 8 - * or fewer bits long. The few overlength codes are handled with a loop, - * which need not be inline code. - * - * Notes about the HUFF_DECODE macro: - * 1. Near the end of the data segment, we may fail to get enough bits - * for a lookahead. In that case, we do it the hard way. - * 2. If the lookahead table contains no entry, the next code must be - * more than HUFF_LOOKAHEAD bits long. - * 3. jpeg_huff_decode returns -1 if forced to suspend. - */ - -#define HUFF_DECODE(result,state,htbl,failaction,slowlabel) \ -{ register int nb, look; \ - if (bits_left < HUFF_LOOKAHEAD) { \ - if (! jpeg_fill_bit_buffer(&state,get_buffer,bits_left, 0)) {failaction;} \ - get_buffer = state.get_buffer; bits_left = state.bits_left; \ - if (bits_left < HUFF_LOOKAHEAD) { \ - nb = 1; goto slowlabel; \ - } \ - } \ - look = PEEK_BITS(HUFF_LOOKAHEAD); \ - if ((nb = htbl->look_nbits[look]) != 0) { \ - DROP_BITS(nb); \ - result = htbl->look_sym[look]; \ - } else { \ - nb = HUFF_LOOKAHEAD+1; \ -slowlabel: \ - if ((result=jpeg_huff_decode(&state,get_buffer,bits_left,htbl,nb)) < 0) \ - { failaction; } \ - get_buffer = state.get_buffer; bits_left = state.bits_left; \ - } \ -} - - -/* - * Expanded entropy decoder object for Huffman decoding. - * - * The savable_state subrecord contains fields that change within an MCU, - * but must not be updated permanently until we complete the MCU. - */ - -typedef struct { - unsigned int EOBRUN; /* remaining EOBs in EOBRUN */ - int last_dc_val[MAX_COMPS_IN_SCAN]; /* last DC coef for each component */ -} savable_state; - -/* This macro is to work around compilers with missing or broken - * structure assignment. You'll need to fix this code if you have - * such a compiler and you change MAX_COMPS_IN_SCAN. - */ - -#ifndef NO_STRUCT_ASSIGN -#define ASSIGN_STATE(dest,src) ((dest) = (src)) -#else -#if MAX_COMPS_IN_SCAN == 4 -#define ASSIGN_STATE(dest,src) \ - ((dest).EOBRUN = (src).EOBRUN, \ - (dest).last_dc_val[0] = (src).last_dc_val[0], \ - (dest).last_dc_val[1] = (src).last_dc_val[1], \ - (dest).last_dc_val[2] = (src).last_dc_val[2], \ - (dest).last_dc_val[3] = (src).last_dc_val[3]) -#endif -#endif - - -typedef struct { - struct jpeg_entropy_decoder pub; /* public fields */ - - /* These fields are loaded into local variables at start of each MCU. - * In case of suspension, we exit WITHOUT updating them. - */ - bitread_perm_state bitstate; /* Bit buffer at start of MCU */ - savable_state saved; /* Other state at start of MCU */ - - /* These fields are NOT loaded into local working state. */ - boolean insufficient_data; /* set TRUE after emitting warning */ - unsigned int restarts_to_go; /* MCUs left in this restart interval */ - - /* Following two fields used only in progressive mode */ - - /* Pointers to derived tables (these workspaces have image lifespan) */ - d_derived_tbl * derived_tbls[NUM_HUFF_TBLS]; - - d_derived_tbl * ac_derived_tbl; /* active table during an AC scan */ - - /* Following fields used only in sequential mode */ - - /* Pointers to derived tables (these workspaces have image lifespan) */ - d_derived_tbl * dc_derived_tbls[NUM_HUFF_TBLS]; - d_derived_tbl * ac_derived_tbls[NUM_HUFF_TBLS]; - - /* Precalculated info set up by start_pass for use in decode_mcu: */ - - /* Pointers to derived tables to be used for each block within an MCU */ - d_derived_tbl * dc_cur_tbls[D_MAX_BLOCKS_IN_MCU]; - d_derived_tbl * ac_cur_tbls[D_MAX_BLOCKS_IN_MCU]; - /* Whether we care about the DC and AC coefficient values for each block */ - int coef_limit[D_MAX_BLOCKS_IN_MCU]; -} huff_entropy_decoder; - -typedef huff_entropy_decoder * huff_entropy_ptr; - - -static const int jpeg_zigzag_order[8][8] = { - { 0, 1, 5, 6, 14, 15, 27, 28 }, - { 2, 4, 7, 13, 16, 26, 29, 42 }, - { 3, 8, 12, 17, 25, 30, 41, 43 }, - { 9, 11, 18, 24, 31, 40, 44, 53 }, - { 10, 19, 23, 32, 39, 45, 52, 54 }, - { 20, 22, 33, 38, 46, 51, 55, 60 }, - { 21, 34, 37, 47, 50, 56, 59, 61 }, - { 35, 36, 48, 49, 57, 58, 62, 63 } -}; - -static const int jpeg_zigzag_order7[7][7] = { - { 0, 1, 5, 6, 14, 15, 27 }, - { 2, 4, 7, 13, 16, 26, 28 }, - { 3, 8, 12, 17, 25, 29, 38 }, - { 9, 11, 18, 24, 30, 37, 39 }, - { 10, 19, 23, 31, 36, 40, 45 }, - { 20, 22, 32, 35, 41, 44, 46 }, - { 21, 33, 34, 42, 43, 47, 48 } -}; - -static const int jpeg_zigzag_order6[6][6] = { - { 0, 1, 5, 6, 14, 15 }, - { 2, 4, 7, 13, 16, 25 }, - { 3, 8, 12, 17, 24, 26 }, - { 9, 11, 18, 23, 27, 32 }, - { 10, 19, 22, 28, 31, 33 }, - { 20, 21, 29, 30, 34, 35 } -}; - -static const int jpeg_zigzag_order5[5][5] = { - { 0, 1, 5, 6, 14 }, - { 2, 4, 7, 13, 15 }, - { 3, 8, 12, 16, 21 }, - { 9, 11, 17, 20, 22 }, - { 10, 18, 19, 23, 24 } -}; - -static const int jpeg_zigzag_order4[4][4] = { - { 0, 1, 5, 6 }, - { 2, 4, 7, 12 }, - { 3, 8, 11, 13 }, - { 9, 10, 14, 15 } -}; - -static const int jpeg_zigzag_order3[3][3] = { - { 0, 1, 5 }, - { 2, 4, 6 }, - { 3, 7, 8 } -}; - -static const int jpeg_zigzag_order2[2][2] = { - { 0, 1 }, - { 2, 3 } -}; - - -/* - * Compute the derived values for a Huffman table. - * This routine also performs some validation checks on the table. - */ - -LOCAL(void) -jpeg_make_d_derived_tbl (j_decompress_ptr cinfo, boolean isDC, int tblno, - d_derived_tbl ** pdtbl) -{ - JHUFF_TBL *htbl; - d_derived_tbl *dtbl; - int p, i, l, si, numsymbols; - int lookbits, ctr; - char huffsize[257]; - unsigned int huffcode[257]; - unsigned int code; - - /* Note that huffsize[] and huffcode[] are filled in code-length order, - * paralleling the order of the symbols themselves in htbl->huffval[]. - */ - - /* Find the input Huffman table */ - if (tblno < 0 || tblno >= NUM_HUFF_TBLS) - ERREXIT1(cinfo, JERR_NO_HUFF_TABLE, tblno); - htbl = - isDC ? cinfo->dc_huff_tbl_ptrs[tblno] : cinfo->ac_huff_tbl_ptrs[tblno]; - if (htbl == NULL) - htbl = jpeg_std_huff_table((j_common_ptr) cinfo, isDC, tblno); - - /* Allocate a workspace if we haven't already done so. */ - if (*pdtbl == NULL) - *pdtbl = (d_derived_tbl *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(d_derived_tbl)); - dtbl = *pdtbl; - dtbl->pub = htbl; /* fill in back link */ - - /* Figure C.1: make table of Huffman code length for each symbol */ - - p = 0; - for (l = 1; l <= 16; l++) { - i = (int) htbl->bits[l]; - if (i < 0 || p + i > 256) /* protect against table overrun */ - ERREXIT(cinfo, JERR_BAD_HUFF_TABLE); - while (i--) - huffsize[p++] = (char) l; - } - huffsize[p] = 0; - numsymbols = p; - - /* Figure C.2: generate the codes themselves */ - /* We also validate that the counts represent a legal Huffman code tree. */ - - code = 0; - si = huffsize[0]; - p = 0; - while (huffsize[p]) { - while (((int) huffsize[p]) == si) { - huffcode[p++] = code; - code++; - } - /* code is now 1 more than the last code used for codelength si; but - * it must still fit in si bits, since no code is allowed to be all ones. - */ - if (((INT32) code) >= (((INT32) 1) << si)) - ERREXIT(cinfo, JERR_BAD_HUFF_TABLE); - code <<= 1; - si++; - } - - /* Figure F.15: generate decoding tables for bit-sequential decoding */ - - p = 0; - for (l = 1; l <= 16; l++) { - if (htbl->bits[l]) { - /* valoffset[l] = huffval[] index of 1st symbol of code length l, - * minus the minimum code of length l - */ - dtbl->valoffset[l] = (INT32) p - (INT32) huffcode[p]; - p += htbl->bits[l]; - dtbl->maxcode[l] = huffcode[p-1]; /* maximum code of length l */ - } else { - dtbl->maxcode[l] = -1; /* -1 if no codes of this length */ - } - } - dtbl->maxcode[17] = 0xFFFFFL; /* ensures jpeg_huff_decode terminates */ - - /* Compute lookahead tables to speed up decoding. - * First we set all the table entries to 0, indicating "too long"; - * then we iterate through the Huffman codes that are short enough and - * fill in all the entries that correspond to bit sequences starting - * with that code. - */ - - MEMZERO(dtbl->look_nbits, SIZEOF(dtbl->look_nbits)); - - p = 0; - for (l = 1; l <= HUFF_LOOKAHEAD; l++) { - for (i = 1; i <= (int) htbl->bits[l]; i++, p++) { - /* l = current code's length, p = its index in huffcode[] & huffval[]. */ - /* Generate left-justified code followed by all possible bit sequences */ - lookbits = huffcode[p] << (HUFF_LOOKAHEAD-l); - for (ctr = 1 << (HUFF_LOOKAHEAD-l); ctr > 0; ctr--) { - dtbl->look_nbits[lookbits] = l; - dtbl->look_sym[lookbits] = htbl->huffval[p]; - lookbits++; - } - } - } - - /* Validate symbols as being reasonable. - * For AC tables, we make no check, but accept all byte values 0..255. - * For DC tables, we require the symbols to be in range 0..15. - * (Tighter bounds could be applied depending on the data depth and mode, - * but this is sufficient to ensure safe decoding.) - */ - if (isDC) { - for (i = 0; i < numsymbols; i++) { - int sym = htbl->huffval[i]; - if (sym < 0 || sym > 15) - ERREXIT(cinfo, JERR_BAD_HUFF_TABLE); - } - } -} - - -/* - * Out-of-line code for bit fetching. - * Note: current values of get_buffer and bits_left are passed as parameters, - * but are returned in the corresponding fields of the state struct. - * - * On most machines MIN_GET_BITS should be 25 to allow the full 32-bit width - * of get_buffer to be used. (On machines with wider words, an even larger - * buffer could be used.) However, on some machines 32-bit shifts are - * quite slow and take time proportional to the number of places shifted. - * (This is true with most PC compilers, for instance.) In this case it may - * be a win to set MIN_GET_BITS to the minimum value of 15. This reduces the - * average shift distance at the cost of more calls to jpeg_fill_bit_buffer. - */ - -#ifdef SLOW_SHIFT_32 -#define MIN_GET_BITS 15 /* minimum allowable value */ -#else -#define MIN_GET_BITS (BIT_BUF_SIZE-7) -#endif - - -LOCAL(boolean) -jpeg_fill_bit_buffer (bitread_working_state * state, - register bit_buf_type get_buffer, register int bits_left, - int nbits) -/* Load up the bit buffer to a depth of at least nbits */ -{ - /* Copy heavily used state fields into locals (hopefully registers) */ - register const JOCTET * next_input_byte = state->next_input_byte; - register size_t bytes_in_buffer = state->bytes_in_buffer; - j_decompress_ptr cinfo = state->cinfo; - - /* Attempt to load at least MIN_GET_BITS bits into get_buffer. */ - /* (It is assumed that no request will be for more than that many bits.) */ - /* We fail to do so only if we hit a marker or are forced to suspend. */ - - if (cinfo->unread_marker == 0) { /* cannot advance past a marker */ - while (bits_left < MIN_GET_BITS) { - register int c; - - /* Attempt to read a byte */ - if (bytes_in_buffer == 0) { - if (! (*cinfo->src->fill_input_buffer) (cinfo)) - return FALSE; - next_input_byte = cinfo->src->next_input_byte; - bytes_in_buffer = cinfo->src->bytes_in_buffer; - } - bytes_in_buffer--; - c = GETJOCTET(*next_input_byte++); - - /* If it's 0xFF, check and discard stuffed zero byte */ - if (c == 0xFF) { - /* Loop here to discard any padding FF's on terminating marker, - * so that we can save a valid unread_marker value. NOTE: we will - * accept multiple FF's followed by a 0 as meaning a single FF data - * byte. This data pattern is not valid according to the standard. - */ - do { - if (bytes_in_buffer == 0) { - if (! (*cinfo->src->fill_input_buffer) (cinfo)) - return FALSE; - next_input_byte = cinfo->src->next_input_byte; - bytes_in_buffer = cinfo->src->bytes_in_buffer; - } - bytes_in_buffer--; - c = GETJOCTET(*next_input_byte++); - } while (c == 0xFF); - - if (c == 0) { - /* Found FF/00, which represents an FF data byte */ - c = 0xFF; - } else { - /* Oops, it's actually a marker indicating end of compressed data. - * Save the marker code for later use. - * Fine point: it might appear that we should save the marker into - * bitread working state, not straight into permanent state. But - * once we have hit a marker, we cannot need to suspend within the - * current MCU, because we will read no more bytes from the data - * source. So it is OK to update permanent state right away. - */ - cinfo->unread_marker = c; - /* See if we need to insert some fake zero bits. */ - goto no_more_bytes; - } - } - - /* OK, load c into get_buffer */ - get_buffer = (get_buffer << 8) | c; - bits_left += 8; - } /* end while */ - } else { - no_more_bytes: - /* We get here if we've read the marker that terminates the compressed - * data segment. There should be enough bits in the buffer register - * to satisfy the request; if so, no problem. - */ - if (nbits > bits_left) { - /* Uh-oh. Report corrupted data to user and stuff zeroes into - * the data stream, so that we can produce some kind of image. - * We use a nonvolatile flag to ensure that only one warning message - * appears per data segment. - */ - if (! ((huff_entropy_ptr) cinfo->entropy)->insufficient_data) { - WARNMS(cinfo, JWRN_HIT_MARKER); - ((huff_entropy_ptr) cinfo->entropy)->insufficient_data = TRUE; - } - /* Fill the buffer with zero bits */ - get_buffer <<= MIN_GET_BITS - bits_left; - bits_left = MIN_GET_BITS; - } - } - - /* Unload the local registers */ - state->next_input_byte = next_input_byte; - state->bytes_in_buffer = bytes_in_buffer; - state->get_buffer = get_buffer; - state->bits_left = bits_left; - - return TRUE; -} - - -/* - * Figure F.12: extend sign bit. - * On some machines, a shift and sub will be faster than a table lookup. - */ - -#ifdef AVOID_TABLES - -#define BIT_MASK(nbits) ((1<<(nbits))-1) -#define HUFF_EXTEND(x,s) ((x) < (1<<((s)-1)) ? (x) - ((1<<(s))-1) : (x)) - -#else - -#define BIT_MASK(nbits) bmask[nbits] -#define HUFF_EXTEND(x,s) ((x) <= bmask[(s) - 1] ? (x) - bmask[s] : (x)) - -static const int bmask[16] = /* bmask[n] is mask for n rightmost bits */ - { 0, 0x0001, 0x0003, 0x0007, 0x000F, 0x001F, 0x003F, 0x007F, 0x00FF, - 0x01FF, 0x03FF, 0x07FF, 0x0FFF, 0x1FFF, 0x3FFF, 0x7FFF }; - -#endif /* AVOID_TABLES */ - - -/* - * Out-of-line code for Huffman code decoding. - */ - -LOCAL(int) -jpeg_huff_decode (bitread_working_state * state, - register bit_buf_type get_buffer, register int bits_left, - d_derived_tbl * htbl, int min_bits) -{ - register int l = min_bits; - register INT32 code; - - /* HUFF_DECODE has determined that the code is at least min_bits */ - /* bits long, so fetch that many bits in one swoop. */ - - CHECK_BIT_BUFFER(*state, l, return -1); - code = GET_BITS(l); - - /* Collect the rest of the Huffman code one bit at a time. */ - /* This is per Figure F.16 in the JPEG spec. */ - - while (code > htbl->maxcode[l]) { - code <<= 1; - CHECK_BIT_BUFFER(*state, 1, return -1); - code |= GET_BITS(1); - l++; - } - - /* Unload the local registers */ - state->get_buffer = get_buffer; - state->bits_left = bits_left; - - /* With garbage input we may reach the sentinel value l = 17. */ - - if (l > 16) { - WARNMS(state->cinfo, JWRN_HUFF_BAD_CODE); - return 0; /* fake a zero as the safest result */ - } - - return htbl->pub->huffval[ (int) (code + htbl->valoffset[l]) ]; -} - - -/* - * Finish up at the end of a Huffman-compressed scan. - */ - -METHODDEF(void) -finish_pass_huff (j_decompress_ptr cinfo) -{ - huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; - - /* Throw away any unused bits remaining in bit buffer; */ - /* include any full bytes in next_marker's count of discarded bytes */ - cinfo->marker->discarded_bytes += entropy->bitstate.bits_left / 8; - entropy->bitstate.bits_left = 0; -} - - -/* - * Check for a restart marker & resynchronize decoder. - * Returns FALSE if must suspend. - */ - -LOCAL(boolean) -process_restart (j_decompress_ptr cinfo) -{ - huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; - int ci; - - finish_pass_huff(cinfo); - - /* Advance past the RSTn marker */ - if (! (*cinfo->marker->read_restart_marker) (cinfo)) - return FALSE; - - /* Re-initialize DC predictions to 0 */ - for (ci = 0; ci < cinfo->comps_in_scan; ci++) - entropy->saved.last_dc_val[ci] = 0; - /* Re-init EOB run count, too */ - entropy->saved.EOBRUN = 0; - - /* Reset restart counter */ - entropy->restarts_to_go = cinfo->restart_interval; - - /* Reset out-of-data flag, unless read_restart_marker left us smack up - * against a marker. In that case we will end up treating the next data - * segment as empty, and we can avoid producing bogus output pixels by - * leaving the flag set. - */ - if (cinfo->unread_marker == 0) - entropy->insufficient_data = FALSE; - - return TRUE; -} - - -/* - * Huffman MCU decoding. - * Each of these routines decodes and returns one MCU's worth of - * Huffman-compressed coefficients. - * The coefficients are reordered from zigzag order into natural array order, - * but are not dequantized. - * - * The i'th block of the MCU is stored into the block pointed to by - * MCU_data[i]. WE ASSUME THIS AREA IS INITIALLY ZEROED BY THE CALLER. - * (Wholesale zeroing is usually a little faster than retail...) - * - * We return FALSE if data source requested suspension. In that case no - * changes have been made to permanent state. (Exception: some output - * coefficients may already have been assigned. This is harmless for - * spectral selection, since we'll just re-assign them on the next call. - * Successive approximation AC refinement has to be more careful, however.) - */ - -/* - * MCU decoding for DC initial scan (either spectral selection, - * or first pass of successive approximation). - */ - -METHODDEF(boolean) -decode_mcu_DC_first (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data) -{ - huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; - int Al = cinfo->Al; - register int s, r; - int blkn, ci; - JBLOCKROW block; - BITREAD_STATE_VARS; - savable_state state; - d_derived_tbl * tbl; - jpeg_component_info * compptr; - - /* Process restart marker if needed; may have to suspend */ - if (cinfo->restart_interval) { - if (entropy->restarts_to_go == 0) - if (! process_restart(cinfo)) - return FALSE; - } - - /* If we've run out of data, just leave the MCU set to zeroes. - * This way, we return uniform gray for the remainder of the segment. - */ - if (! entropy->insufficient_data) { - - /* Load up working state */ - BITREAD_LOAD_STATE(cinfo, entropy->bitstate); - ASSIGN_STATE(state, entropy->saved); - - /* Outer loop handles each block in the MCU */ - - for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { - block = MCU_data[blkn]; - ci = cinfo->MCU_membership[blkn]; - compptr = cinfo->cur_comp_info[ci]; - tbl = entropy->derived_tbls[compptr->dc_tbl_no]; - - /* Decode a single block's worth of coefficients */ - - /* Section F.2.2.1: decode the DC coefficient difference */ - HUFF_DECODE(s, br_state, tbl, return FALSE, label1); - if (s) { - CHECK_BIT_BUFFER(br_state, s, return FALSE); - r = GET_BITS(s); - s = HUFF_EXTEND(r, s); - } - - /* Convert DC difference to actual value, update last_dc_val */ - s += state.last_dc_val[ci]; - state.last_dc_val[ci] = s; - /* Scale and output the coefficient (assumes jpeg_natural_order[0]=0) */ - (*block)[0] = (JCOEF) (s << Al); - } - - /* Completed MCU, so update state */ - BITREAD_SAVE_STATE(cinfo, entropy->bitstate); - ASSIGN_STATE(entropy->saved, state); - } - - /* Account for restart interval if using restarts */ - if (cinfo->restart_interval) - entropy->restarts_to_go--; - - return TRUE; -} - - -/* - * MCU decoding for AC initial scan (either spectral selection, - * or first pass of successive approximation). - */ - -METHODDEF(boolean) -decode_mcu_AC_first (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data) -{ - huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; - register int s, k, r; - unsigned int EOBRUN; - int Se, Al; - const int * natural_order; - JBLOCKROW block; - BITREAD_STATE_VARS; - d_derived_tbl * tbl; - - /* Process restart marker if needed; may have to suspend */ - if (cinfo->restart_interval) { - if (entropy->restarts_to_go == 0) - if (! process_restart(cinfo)) - return FALSE; - } - - /* If we've run out of data, just leave the MCU set to zeroes. - * This way, we return uniform gray for the remainder of the segment. - */ - if (! entropy->insufficient_data) { - - /* Load up working state. - * We can avoid loading/saving bitread state if in an EOB run. - */ - EOBRUN = entropy->saved.EOBRUN; /* only part of saved state we need */ - - /* There is always only one block per MCU */ - - if (EOBRUN) /* if it's a band of zeroes... */ - EOBRUN--; /* ...process it now (we do nothing) */ - else { - BITREAD_LOAD_STATE(cinfo, entropy->bitstate); - Se = cinfo->Se; - Al = cinfo->Al; - natural_order = cinfo->natural_order; - block = MCU_data[0]; - tbl = entropy->ac_derived_tbl; - - for (k = cinfo->Ss; k <= Se; k++) { - HUFF_DECODE(s, br_state, tbl, return FALSE, label2); - r = s >> 4; - s &= 15; - if (s) { - k += r; - CHECK_BIT_BUFFER(br_state, s, return FALSE); - r = GET_BITS(s); - s = HUFF_EXTEND(r, s); - /* Scale and output coefficient in natural (dezigzagged) order */ - (*block)[natural_order[k]] = (JCOEF) (s << Al); - } else { - if (r != 15) { /* EOBr, run length is 2^r + appended bits */ - if (r) { /* EOBr, r > 0 */ - EOBRUN = 1 << r; - CHECK_BIT_BUFFER(br_state, r, return FALSE); - r = GET_BITS(r); - EOBRUN += r; - EOBRUN--; /* this band is processed at this moment */ - } - break; /* force end-of-band */ - } - k += 15; /* ZRL: skip 15 zeroes in band */ - } - } - - BITREAD_SAVE_STATE(cinfo, entropy->bitstate); - } - - /* Completed MCU, so update state */ - entropy->saved.EOBRUN = EOBRUN; /* only part of saved state we need */ - } - - /* Account for restart interval if using restarts */ - if (cinfo->restart_interval) - entropy->restarts_to_go--; - - return TRUE; -} - - -/* - * MCU decoding for DC successive approximation refinement scan. - * Note: we assume such scans can be multi-component, - * although the spec is not very clear on the point. - */ - -METHODDEF(boolean) -decode_mcu_DC_refine (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data) -{ - huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; - JCOEF p1; - int blkn; - BITREAD_STATE_VARS; - - /* Process restart marker if needed; may have to suspend */ - if (cinfo->restart_interval) { - if (entropy->restarts_to_go == 0) - if (! process_restart(cinfo)) - return FALSE; - } - - /* Not worth the cycles to check insufficient_data here, - * since we will not change the data anyway if we read zeroes. - */ - - /* Load up working state */ - BITREAD_LOAD_STATE(cinfo, entropy->bitstate); - - p1 = 1 << cinfo->Al; /* 1 in the bit position being coded */ - - /* Outer loop handles each block in the MCU */ - - for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { - /* Encoded data is simply the next bit of the two's-complement DC value */ - CHECK_BIT_BUFFER(br_state, 1, return FALSE); - if (GET_BITS(1)) - MCU_data[blkn][0][0] |= p1; - /* Note: since we use |=, repeating the assignment later is safe */ - } - - /* Completed MCU, so update state */ - BITREAD_SAVE_STATE(cinfo, entropy->bitstate); - - /* Account for restart interval if using restarts */ - if (cinfo->restart_interval) - entropy->restarts_to_go--; - - return TRUE; -} - - -/* - * MCU decoding for AC successive approximation refinement scan. - */ - -METHODDEF(boolean) -decode_mcu_AC_refine (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data) -{ - huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; - register int s, k, r; - unsigned int EOBRUN; - int Se; - JCOEF p1, m1; - const int * natural_order; - JBLOCKROW block; - JCOEFPTR thiscoef; - BITREAD_STATE_VARS; - d_derived_tbl * tbl; - int num_newnz; - int newnz_pos[DCTSIZE2]; - - /* Process restart marker if needed; may have to suspend */ - if (cinfo->restart_interval) { - if (entropy->restarts_to_go == 0) - if (! process_restart(cinfo)) - return FALSE; - } - - /* If we've run out of data, don't modify the MCU. - */ - if (! entropy->insufficient_data) { - - Se = cinfo->Se; - p1 = 1 << cinfo->Al; /* 1 in the bit position being coded */ - m1 = -p1; /* -1 in the bit position being coded */ - natural_order = cinfo->natural_order; - - /* Load up working state */ - BITREAD_LOAD_STATE(cinfo, entropy->bitstate); - EOBRUN = entropy->saved.EOBRUN; /* only part of saved state we need */ - - /* There is always only one block per MCU */ - block = MCU_data[0]; - tbl = entropy->ac_derived_tbl; - - /* If we are forced to suspend, we must undo the assignments to any newly - * nonzero coefficients in the block, because otherwise we'd get confused - * next time about which coefficients were already nonzero. - * But we need not undo addition of bits to already-nonzero coefficients; - * instead, we can test the current bit to see if we already did it. - */ - num_newnz = 0; - - /* initialize coefficient loop counter to start of band */ - k = cinfo->Ss; - - if (EOBRUN == 0) { - do { - HUFF_DECODE(s, br_state, tbl, goto undoit, label3); - r = s >> 4; - s &= 15; - if (s) { - if (s != 1) /* size of new coef should always be 1 */ - WARNMS(cinfo, JWRN_HUFF_BAD_CODE); - CHECK_BIT_BUFFER(br_state, 1, goto undoit); - if (GET_BITS(1)) - s = p1; /* newly nonzero coef is positive */ - else - s = m1; /* newly nonzero coef is negative */ - } else { - if (r != 15) { - EOBRUN = 1 << r; /* EOBr, run length is 2^r + appended bits */ - if (r) { - CHECK_BIT_BUFFER(br_state, r, goto undoit); - r = GET_BITS(r); - EOBRUN += r; - } - break; /* rest of block is handled by EOB logic */ - } - /* note s = 0 for processing ZRL */ - } - /* Advance over already-nonzero coefs and r still-zero coefs, - * appending correction bits to the nonzeroes. A correction bit is 1 - * if the absolute value of the coefficient must be increased. - */ - do { - thiscoef = *block + natural_order[k]; - if (*thiscoef) { - CHECK_BIT_BUFFER(br_state, 1, goto undoit); - if (GET_BITS(1)) { - if ((*thiscoef & p1) == 0) { /* do nothing if already set it */ - if (*thiscoef >= 0) - *thiscoef += p1; - else - *thiscoef += m1; - } - } - } else { - if (--r < 0) - break; /* reached target zero coefficient */ - } - k++; - } while (k <= Se); - if (s) { - int pos = natural_order[k]; - /* Output newly nonzero coefficient */ - (*block)[pos] = (JCOEF) s; - /* Remember its position in case we have to suspend */ - newnz_pos[num_newnz++] = pos; - } - k++; - } while (k <= Se); - } - - if (EOBRUN) { - /* Scan any remaining coefficient positions after the end-of-band - * (the last newly nonzero coefficient, if any). Append a correction - * bit to each already-nonzero coefficient. A correction bit is 1 - * if the absolute value of the coefficient must be increased. - */ - do { - thiscoef = *block + natural_order[k]; - if (*thiscoef) { - CHECK_BIT_BUFFER(br_state, 1, goto undoit); - if (GET_BITS(1)) { - if ((*thiscoef & p1) == 0) { /* do nothing if already changed it */ - if (*thiscoef >= 0) - *thiscoef += p1; - else - *thiscoef += m1; - } - } - } - k++; - } while (k <= Se); - /* Count one block completed in EOB run */ - EOBRUN--; - } - - /* Completed MCU, so update state */ - BITREAD_SAVE_STATE(cinfo, entropy->bitstate); - entropy->saved.EOBRUN = EOBRUN; /* only part of saved state we need */ - } - - /* Account for restart interval if using restarts */ - if (cinfo->restart_interval) - entropy->restarts_to_go--; - - return TRUE; - -undoit: - /* Re-zero any output coefficients that we made newly nonzero */ - while (num_newnz) - (*block)[newnz_pos[--num_newnz]] = 0; - - return FALSE; -} - - -/* - * Decode one MCU's worth of Huffman-compressed coefficients, - * partial blocks. - */ - -METHODDEF(boolean) -decode_mcu_sub (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data) -{ - huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; - const int * natural_order; - int Se, blkn; - BITREAD_STATE_VARS; - savable_state state; - - /* Process restart marker if needed; may have to suspend */ - if (cinfo->restart_interval) { - if (entropy->restarts_to_go == 0) - if (! process_restart(cinfo)) - return FALSE; - } - - /* If we've run out of data, just leave the MCU set to zeroes. - * This way, we return uniform gray for the remainder of the segment. - */ - if (! entropy->insufficient_data) { - - natural_order = cinfo->natural_order; - Se = cinfo->lim_Se; - - /* Load up working state */ - BITREAD_LOAD_STATE(cinfo, entropy->bitstate); - ASSIGN_STATE(state, entropy->saved); - - /* Outer loop handles each block in the MCU */ - - for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { - JBLOCKROW block = MCU_data[blkn]; - d_derived_tbl * htbl; - register int s, k, r; - int coef_limit, ci; - - /* Decode a single block's worth of coefficients */ - - /* Section F.2.2.1: decode the DC coefficient difference */ - htbl = entropy->dc_cur_tbls[blkn]; - HUFF_DECODE(s, br_state, htbl, return FALSE, label1); - - htbl = entropy->ac_cur_tbls[blkn]; - k = 1; - coef_limit = entropy->coef_limit[blkn]; - if (coef_limit) { - /* Convert DC difference to actual value, update last_dc_val */ - if (s) { - CHECK_BIT_BUFFER(br_state, s, return FALSE); - r = GET_BITS(s); - s = HUFF_EXTEND(r, s); - } - ci = cinfo->MCU_membership[blkn]; - s += state.last_dc_val[ci]; - state.last_dc_val[ci] = s; - /* Output the DC coefficient */ - (*block)[0] = (JCOEF) s; - - /* Section F.2.2.2: decode the AC coefficients */ - /* Since zeroes are skipped, output area must be cleared beforehand */ - for (; k < coef_limit; k++) { - HUFF_DECODE(s, br_state, htbl, return FALSE, label2); - - r = s >> 4; - s &= 15; - - if (s) { - k += r; - CHECK_BIT_BUFFER(br_state, s, return FALSE); - r = GET_BITS(s); - s = HUFF_EXTEND(r, s); - /* Output coefficient in natural (dezigzagged) order. - * Note: the extra entries in natural_order[] will save us - * if k > Se, which could happen if the data is corrupted. - */ - (*block)[natural_order[k]] = (JCOEF) s; - } else { - if (r != 15) - goto EndOfBlock; - k += 15; - } - } - } else { - if (s) { - CHECK_BIT_BUFFER(br_state, s, return FALSE); - DROP_BITS(s); - } - } - - /* Section F.2.2.2: decode the AC coefficients */ - /* In this path we just discard the values */ - for (; k <= Se; k++) { - HUFF_DECODE(s, br_state, htbl, return FALSE, label3); - - r = s >> 4; - s &= 15; - - if (s) { - k += r; - CHECK_BIT_BUFFER(br_state, s, return FALSE); - DROP_BITS(s); - } else { - if (r != 15) - break; - k += 15; - } - } - - EndOfBlock: ; - } - - /* Completed MCU, so update state */ - BITREAD_SAVE_STATE(cinfo, entropy->bitstate); - ASSIGN_STATE(entropy->saved, state); - } - - /* Account for restart interval if using restarts */ - if (cinfo->restart_interval) - entropy->restarts_to_go--; - - return TRUE; -} - - -/* - * Decode one MCU's worth of Huffman-compressed coefficients, - * full-size blocks. - */ - -METHODDEF(boolean) -decode_mcu (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data) -{ - huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; - int blkn; - BITREAD_STATE_VARS; - savable_state state; - - /* Process restart marker if needed; may have to suspend */ - if (cinfo->restart_interval) { - if (entropy->restarts_to_go == 0) - if (! process_restart(cinfo)) - return FALSE; - } - - /* If we've run out of data, just leave the MCU set to zeroes. - * This way, we return uniform gray for the remainder of the segment. - */ - if (! entropy->insufficient_data) { - - /* Load up working state */ - BITREAD_LOAD_STATE(cinfo, entropy->bitstate); - ASSIGN_STATE(state, entropy->saved); - - /* Outer loop handles each block in the MCU */ - - for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { - JBLOCKROW block = MCU_data[blkn]; - d_derived_tbl * htbl; - register int s, k, r; - int coef_limit, ci; - - /* Decode a single block's worth of coefficients */ - - /* Section F.2.2.1: decode the DC coefficient difference */ - htbl = entropy->dc_cur_tbls[blkn]; - HUFF_DECODE(s, br_state, htbl, return FALSE, label1); - - htbl = entropy->ac_cur_tbls[blkn]; - k = 1; - coef_limit = entropy->coef_limit[blkn]; - if (coef_limit) { - /* Convert DC difference to actual value, update last_dc_val */ - if (s) { - CHECK_BIT_BUFFER(br_state, s, return FALSE); - r = GET_BITS(s); - s = HUFF_EXTEND(r, s); - } - ci = cinfo->MCU_membership[blkn]; - s += state.last_dc_val[ci]; - state.last_dc_val[ci] = s; - /* Output the DC coefficient */ - (*block)[0] = (JCOEF) s; - - /* Section F.2.2.2: decode the AC coefficients */ - /* Since zeroes are skipped, output area must be cleared beforehand */ - for (; k < coef_limit; k++) { - HUFF_DECODE(s, br_state, htbl, return FALSE, label2); - - r = s >> 4; - s &= 15; - - if (s) { - k += r; - CHECK_BIT_BUFFER(br_state, s, return FALSE); - r = GET_BITS(s); - s = HUFF_EXTEND(r, s); - /* Output coefficient in natural (dezigzagged) order. - * Note: the extra entries in jpeg_natural_order[] will save us - * if k >= DCTSIZE2, which could happen if the data is corrupted. - */ - (*block)[jpeg_natural_order[k]] = (JCOEF) s; - } else { - if (r != 15) - goto EndOfBlock; - k += 15; - } - } - } else { - if (s) { - CHECK_BIT_BUFFER(br_state, s, return FALSE); - DROP_BITS(s); - } - } - - /* Section F.2.2.2: decode the AC coefficients */ - /* In this path we just discard the values */ - for (; k < DCTSIZE2; k++) { - HUFF_DECODE(s, br_state, htbl, return FALSE, label3); - - r = s >> 4; - s &= 15; - - if (s) { - k += r; - CHECK_BIT_BUFFER(br_state, s, return FALSE); - DROP_BITS(s); - } else { - if (r != 15) - break; - k += 15; - } - } - - EndOfBlock: ; - } - - /* Completed MCU, so update state */ - BITREAD_SAVE_STATE(cinfo, entropy->bitstate); - ASSIGN_STATE(entropy->saved, state); - } - - /* Account for restart interval if using restarts */ - if (cinfo->restart_interval) - entropy->restarts_to_go--; - - return TRUE; -} - - -/* - * Initialize for a Huffman-compressed scan. - */ - -METHODDEF(void) -start_pass_huff_decoder (j_decompress_ptr cinfo) -{ - huff_entropy_ptr entropy = (huff_entropy_ptr) cinfo->entropy; - int ci, blkn, tbl, i; - jpeg_component_info * compptr; - - if (cinfo->progressive_mode) { - /* Validate progressive scan parameters */ - if (cinfo->Ss == 0) { - if (cinfo->Se != 0) - goto bad; - } else { - /* need not check Ss/Se < 0 since they came from unsigned bytes */ - if (cinfo->Se < cinfo->Ss || cinfo->Se > cinfo->lim_Se) - goto bad; - /* AC scans may have only one component */ - if (cinfo->comps_in_scan != 1) - goto bad; - } - if (cinfo->Ah != 0) { - /* Successive approximation refinement scan: must have Al = Ah-1. */ - if (cinfo->Ah-1 != cinfo->Al) - goto bad; - } - if (cinfo->Al > 13) { /* need not check for < 0 */ - /* Arguably the maximum Al value should be less than 13 for 8-bit - * precision, but the spec doesn't say so, and we try to be liberal - * about what we accept. Note: large Al values could result in - * out-of-range DC coefficients during early scans, leading to bizarre - * displays due to overflows in the IDCT math. But we won't crash. - */ - bad: - ERREXIT4(cinfo, JERR_BAD_PROGRESSION, - cinfo->Ss, cinfo->Se, cinfo->Ah, cinfo->Al); - } - /* Update progression status, and verify that scan order is legal. - * Note that inter-scan inconsistencies are treated as warnings - * not fatal errors ... not clear if this is right way to behave. - */ - for (ci = 0; ci < cinfo->comps_in_scan; ci++) { - int coefi, cindex = cinfo->cur_comp_info[ci]->component_index; - int *coef_bit_ptr = & cinfo->coef_bits[cindex][0]; - if (cinfo->Ss && coef_bit_ptr[0] < 0) /* AC without prior DC scan */ - WARNMS2(cinfo, JWRN_BOGUS_PROGRESSION, cindex, 0); - for (coefi = cinfo->Ss; coefi <= cinfo->Se; coefi++) { - int expected = (coef_bit_ptr[coefi] < 0) ? 0 : coef_bit_ptr[coefi]; - if (cinfo->Ah != expected) - WARNMS2(cinfo, JWRN_BOGUS_PROGRESSION, cindex, coefi); - coef_bit_ptr[coefi] = cinfo->Al; - } - } - - /* Select MCU decoding routine */ - if (cinfo->Ah == 0) { - if (cinfo->Ss == 0) - entropy->pub.decode_mcu = decode_mcu_DC_first; - else - entropy->pub.decode_mcu = decode_mcu_AC_first; - } else { - if (cinfo->Ss == 0) - entropy->pub.decode_mcu = decode_mcu_DC_refine; - else - entropy->pub.decode_mcu = decode_mcu_AC_refine; - } - - for (ci = 0; ci < cinfo->comps_in_scan; ci++) { - compptr = cinfo->cur_comp_info[ci]; - /* Make sure requested tables are present, and compute derived tables. - * We may build same derived table more than once, but it's not expensive. - */ - if (cinfo->Ss == 0) { - if (cinfo->Ah == 0) { /* DC refinement needs no table */ - tbl = compptr->dc_tbl_no; - jpeg_make_d_derived_tbl(cinfo, TRUE, tbl, - & entropy->derived_tbls[tbl]); - } - } else { - tbl = compptr->ac_tbl_no; - jpeg_make_d_derived_tbl(cinfo, FALSE, tbl, - & entropy->derived_tbls[tbl]); - /* remember the single active table */ - entropy->ac_derived_tbl = entropy->derived_tbls[tbl]; - } - /* Initialize DC predictions to 0 */ - entropy->saved.last_dc_val[ci] = 0; - } - - /* Initialize private state variables */ - entropy->saved.EOBRUN = 0; - } else { - /* Check that the scan parameters Ss, Se, Ah/Al are OK for sequential JPEG. - * This ought to be an error condition, but we make it a warning because - * there are some baseline files out there with all zeroes in these bytes. - */ - if (cinfo->Ss != 0 || cinfo->Ah != 0 || cinfo->Al != 0 || - ((cinfo->is_baseline || cinfo->Se < DCTSIZE2) && - cinfo->Se != cinfo->lim_Se)) - WARNMS(cinfo, JWRN_NOT_SEQUENTIAL); - - /* Select MCU decoding routine */ - /* We retain the hard-coded case for full-size blocks. - * This is not necessary, but it appears that this version is slightly - * more performant in the given implementation. - * With an improved implementation we would prefer a single optimized - * function. - */ - if (cinfo->lim_Se != DCTSIZE2-1) - entropy->pub.decode_mcu = decode_mcu_sub; - else - entropy->pub.decode_mcu = decode_mcu; - - for (ci = 0; ci < cinfo->comps_in_scan; ci++) { - compptr = cinfo->cur_comp_info[ci]; - /* Compute derived values for Huffman tables */ - /* We may do this more than once for a table, but it's not expensive */ - tbl = compptr->dc_tbl_no; - jpeg_make_d_derived_tbl(cinfo, TRUE, tbl, - & entropy->dc_derived_tbls[tbl]); - if (cinfo->lim_Se) { /* AC needs no table when not present */ - tbl = compptr->ac_tbl_no; - jpeg_make_d_derived_tbl(cinfo, FALSE, tbl, - & entropy->ac_derived_tbls[tbl]); - } - /* Initialize DC predictions to 0 */ - entropy->saved.last_dc_val[ci] = 0; - } - - /* Precalculate decoding info for each block in an MCU of this scan */ - for (blkn = 0; blkn < cinfo->blocks_in_MCU; blkn++) { - ci = cinfo->MCU_membership[blkn]; - compptr = cinfo->cur_comp_info[ci]; - /* Precalculate which table to use for each block */ - entropy->dc_cur_tbls[blkn] = entropy->dc_derived_tbls[compptr->dc_tbl_no]; - entropy->ac_cur_tbls[blkn] = /* AC needs no table when not present */ - cinfo->lim_Se ? entropy->ac_derived_tbls[compptr->ac_tbl_no] : NULL; - /* Decide whether we really care about the coefficient values */ - if (compptr->component_needed) { - ci = compptr->DCT_v_scaled_size; - i = compptr->DCT_h_scaled_size; - switch (cinfo->lim_Se) { - case (1*1-1): - entropy->coef_limit[blkn] = 1; - break; - case (2*2-1): - if (ci <= 0 || ci > 2) ci = 2; - if (i <= 0 || i > 2) i = 2; - entropy->coef_limit[blkn] = 1 + jpeg_zigzag_order2[ci - 1][i - 1]; - break; - case (3*3-1): - if (ci <= 0 || ci > 3) ci = 3; - if (i <= 0 || i > 3) i = 3; - entropy->coef_limit[blkn] = 1 + jpeg_zigzag_order3[ci - 1][i - 1]; - break; - case (4*4-1): - if (ci <= 0 || ci > 4) ci = 4; - if (i <= 0 || i > 4) i = 4; - entropy->coef_limit[blkn] = 1 + jpeg_zigzag_order4[ci - 1][i - 1]; - break; - case (5*5-1): - if (ci <= 0 || ci > 5) ci = 5; - if (i <= 0 || i > 5) i = 5; - entropy->coef_limit[blkn] = 1 + jpeg_zigzag_order5[ci - 1][i - 1]; - break; - case (6*6-1): - if (ci <= 0 || ci > 6) ci = 6; - if (i <= 0 || i > 6) i = 6; - entropy->coef_limit[blkn] = 1 + jpeg_zigzag_order6[ci - 1][i - 1]; - break; - case (7*7-1): - if (ci <= 0 || ci > 7) ci = 7; - if (i <= 0 || i > 7) i = 7; - entropy->coef_limit[blkn] = 1 + jpeg_zigzag_order7[ci - 1][i - 1]; - break; - default: - if (ci <= 0 || ci > 8) ci = 8; - if (i <= 0 || i > 8) i = 8; - entropy->coef_limit[blkn] = 1 + jpeg_zigzag_order[ci - 1][i - 1]; - } - } else { - entropy->coef_limit[blkn] = 0; - } - } - } - - /* Initialize bitread state variables */ - entropy->bitstate.bits_left = 0; - entropy->bitstate.get_buffer = 0; /* unnecessary, but keeps Purify quiet */ - entropy->insufficient_data = FALSE; - - /* Initialize restart counter */ - entropy->restarts_to_go = cinfo->restart_interval; -} - - -/* - * Module initialization routine for Huffman entropy decoding. - */ - -GLOBAL(void) -jinit_huff_decoder (j_decompress_ptr cinfo) -{ - huff_entropy_ptr entropy; - int i; - - entropy = (huff_entropy_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(huff_entropy_decoder)); - cinfo->entropy = &entropy->pub; - entropy->pub.start_pass = start_pass_huff_decoder; - entropy->pub.finish_pass = finish_pass_huff; - - if (cinfo->progressive_mode) { - /* Create progression status table */ - int *coef_bit_ptr, ci; - cinfo->coef_bits = (int (*)[DCTSIZE2]) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, - cinfo->num_components * DCTSIZE2 * SIZEOF(int)); - coef_bit_ptr = & cinfo->coef_bits[0][0]; - for (ci = 0; ci < cinfo->num_components; ci++) - for (i = 0; i < DCTSIZE2; i++) - *coef_bit_ptr++ = -1; - - /* Mark derived tables unallocated */ - for (i = 0; i < NUM_HUFF_TBLS; i++) { - entropy->derived_tbls[i] = NULL; - } - } else { - /* Mark derived tables unallocated */ - for (i = 0; i < NUM_HUFF_TBLS; i++) { - entropy->dc_derived_tbls[i] = entropy->ac_derived_tbls[i] = NULL; - } - } -} diff --git a/cmake/externals/libjpeg/jdinput.c b/cmake/externals/libjpeg/jdinput.c deleted file mode 100644 index 29fbef90b..000000000 --- a/cmake/externals/libjpeg/jdinput.c +++ /dev/null @@ -1,657 +0,0 @@ -/* - * jdinput.c - * - * Copyright (C) 1991-1997, Thomas G. Lane. - * Modified 2002-2020 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains input control logic for the JPEG decompressor. - * These routines are concerned with controlling the decompressor's input - * processing (marker reading and coefficient decoding). The actual input - * reading is done in jdmarker.c, jdhuff.c, and jdarith.c. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* Private state */ - -typedef struct { - struct jpeg_input_controller pub; /* public fields */ - - int inheaders; /* Nonzero until first SOS is reached */ -} my_input_controller; - -typedef my_input_controller * my_inputctl_ptr; - - -/* Forward declarations */ -METHODDEF(int) consume_markers JPP((j_decompress_ptr cinfo)); - - -/* - * Routines to calculate various quantities related to the size of the image. - */ - - -/* - * Compute output image dimensions and related values. - * NOTE: this is exported for possible use by application. - * Hence it mustn't do anything that can't be done twice. - */ - -GLOBAL(void) -jpeg_core_output_dimensions (j_decompress_ptr cinfo) -/* Do computations that are needed before master selection phase. - * This function is used for transcoding and full decompression. - */ -{ -#ifdef IDCT_SCALING_SUPPORTED - int ci; - jpeg_component_info *compptr; - - /* Compute actual output image dimensions and DCT scaling choices. */ - if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom) { - /* Provide 1/block_size scaling */ - cinfo->output_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width, (long) cinfo->block_size); - cinfo->output_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height, (long) cinfo->block_size); - cinfo->min_DCT_h_scaled_size = 1; - cinfo->min_DCT_v_scaled_size = 1; - } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 2) { - /* Provide 2/block_size scaling */ - cinfo->output_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * 2L, (long) cinfo->block_size); - cinfo->output_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * 2L, (long) cinfo->block_size); - cinfo->min_DCT_h_scaled_size = 2; - cinfo->min_DCT_v_scaled_size = 2; - } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 3) { - /* Provide 3/block_size scaling */ - cinfo->output_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * 3L, (long) cinfo->block_size); - cinfo->output_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * 3L, (long) cinfo->block_size); - cinfo->min_DCT_h_scaled_size = 3; - cinfo->min_DCT_v_scaled_size = 3; - } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 4) { - /* Provide 4/block_size scaling */ - cinfo->output_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * 4L, (long) cinfo->block_size); - cinfo->output_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * 4L, (long) cinfo->block_size); - cinfo->min_DCT_h_scaled_size = 4; - cinfo->min_DCT_v_scaled_size = 4; - } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 5) { - /* Provide 5/block_size scaling */ - cinfo->output_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * 5L, (long) cinfo->block_size); - cinfo->output_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * 5L, (long) cinfo->block_size); - cinfo->min_DCT_h_scaled_size = 5; - cinfo->min_DCT_v_scaled_size = 5; - } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 6) { - /* Provide 6/block_size scaling */ - cinfo->output_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * 6L, (long) cinfo->block_size); - cinfo->output_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * 6L, (long) cinfo->block_size); - cinfo->min_DCT_h_scaled_size = 6; - cinfo->min_DCT_v_scaled_size = 6; - } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 7) { - /* Provide 7/block_size scaling */ - cinfo->output_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * 7L, (long) cinfo->block_size); - cinfo->output_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * 7L, (long) cinfo->block_size); - cinfo->min_DCT_h_scaled_size = 7; - cinfo->min_DCT_v_scaled_size = 7; - } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 8) { - /* Provide 8/block_size scaling */ - cinfo->output_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * 8L, (long) cinfo->block_size); - cinfo->output_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * 8L, (long) cinfo->block_size); - cinfo->min_DCT_h_scaled_size = 8; - cinfo->min_DCT_v_scaled_size = 8; - } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 9) { - /* Provide 9/block_size scaling */ - cinfo->output_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * 9L, (long) cinfo->block_size); - cinfo->output_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * 9L, (long) cinfo->block_size); - cinfo->min_DCT_h_scaled_size = 9; - cinfo->min_DCT_v_scaled_size = 9; - } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 10) { - /* Provide 10/block_size scaling */ - cinfo->output_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * 10L, (long) cinfo->block_size); - cinfo->output_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * 10L, (long) cinfo->block_size); - cinfo->min_DCT_h_scaled_size = 10; - cinfo->min_DCT_v_scaled_size = 10; - } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 11) { - /* Provide 11/block_size scaling */ - cinfo->output_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * 11L, (long) cinfo->block_size); - cinfo->output_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * 11L, (long) cinfo->block_size); - cinfo->min_DCT_h_scaled_size = 11; - cinfo->min_DCT_v_scaled_size = 11; - } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 12) { - /* Provide 12/block_size scaling */ - cinfo->output_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * 12L, (long) cinfo->block_size); - cinfo->output_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * 12L, (long) cinfo->block_size); - cinfo->min_DCT_h_scaled_size = 12; - cinfo->min_DCT_v_scaled_size = 12; - } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 13) { - /* Provide 13/block_size scaling */ - cinfo->output_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * 13L, (long) cinfo->block_size); - cinfo->output_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * 13L, (long) cinfo->block_size); - cinfo->min_DCT_h_scaled_size = 13; - cinfo->min_DCT_v_scaled_size = 13; - } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 14) { - /* Provide 14/block_size scaling */ - cinfo->output_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * 14L, (long) cinfo->block_size); - cinfo->output_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * 14L, (long) cinfo->block_size); - cinfo->min_DCT_h_scaled_size = 14; - cinfo->min_DCT_v_scaled_size = 14; - } else if (cinfo->scale_num * cinfo->block_size <= cinfo->scale_denom * 15) { - /* Provide 15/block_size scaling */ - cinfo->output_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * 15L, (long) cinfo->block_size); - cinfo->output_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * 15L, (long) cinfo->block_size); - cinfo->min_DCT_h_scaled_size = 15; - cinfo->min_DCT_v_scaled_size = 15; - } else { - /* Provide 16/block_size scaling */ - cinfo->output_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * 16L, (long) cinfo->block_size); - cinfo->output_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * 16L, (long) cinfo->block_size); - cinfo->min_DCT_h_scaled_size = 16; - cinfo->min_DCT_v_scaled_size = 16; - } - - /* Recompute dimensions of components */ - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - compptr->DCT_h_scaled_size = cinfo->min_DCT_h_scaled_size; - compptr->DCT_v_scaled_size = cinfo->min_DCT_v_scaled_size; - } - -#else /* !IDCT_SCALING_SUPPORTED */ - - /* Hardwire it to "no scaling" */ - cinfo->output_width = cinfo->image_width; - cinfo->output_height = cinfo->image_height; - /* initial_setup has already initialized DCT_scaled_size, - * and has computed unscaled downsampled_width and downsampled_height. - */ - -#endif /* IDCT_SCALING_SUPPORTED */ -} - - -LOCAL(void) -initial_setup (j_decompress_ptr cinfo) -/* Called once, when first SOS marker is reached */ -{ - int ci; - jpeg_component_info *compptr; - - /* Make sure image isn't bigger than I can handle */ - if ((long) cinfo->image_height > (long) JPEG_MAX_DIMENSION || - (long) cinfo->image_width > (long) JPEG_MAX_DIMENSION) - ERREXIT1(cinfo, JERR_IMAGE_TOO_BIG, (unsigned int) JPEG_MAX_DIMENSION); - - /* Only 8 to 12 bits data precision are supported for DCT based JPEG */ - if (cinfo->data_precision < 8 || cinfo->data_precision > 12) - ERREXIT1(cinfo, JERR_BAD_PRECISION, cinfo->data_precision); - - /* Check that number of components won't exceed internal array sizes */ - if (cinfo->num_components > MAX_COMPONENTS) - ERREXIT2(cinfo, JERR_COMPONENT_COUNT, cinfo->num_components, - MAX_COMPONENTS); - - /* Compute maximum sampling factors; check factor validity */ - cinfo->max_h_samp_factor = 1; - cinfo->max_v_samp_factor = 1; - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - if (compptr->h_samp_factor<=0 || compptr->h_samp_factor>MAX_SAMP_FACTOR || - compptr->v_samp_factor<=0 || compptr->v_samp_factor>MAX_SAMP_FACTOR) - ERREXIT(cinfo, JERR_BAD_SAMPLING); - cinfo->max_h_samp_factor = MAX(cinfo->max_h_samp_factor, - compptr->h_samp_factor); - cinfo->max_v_samp_factor = MAX(cinfo->max_v_samp_factor, - compptr->v_samp_factor); - } - - /* Derive block_size, natural_order, and lim_Se */ - if (cinfo->is_baseline || (cinfo->progressive_mode && - cinfo->comps_in_scan)) { /* no pseudo SOS marker */ - cinfo->block_size = DCTSIZE; - cinfo->natural_order = jpeg_natural_order; - cinfo->lim_Se = DCTSIZE2-1; - } else - switch (cinfo->Se) { - case (1*1-1): - cinfo->block_size = 1; - cinfo->natural_order = jpeg_natural_order; /* not needed */ - cinfo->lim_Se = cinfo->Se; - break; - case (2*2-1): - cinfo->block_size = 2; - cinfo->natural_order = jpeg_natural_order2; - cinfo->lim_Se = cinfo->Se; - break; - case (3*3-1): - cinfo->block_size = 3; - cinfo->natural_order = jpeg_natural_order3; - cinfo->lim_Se = cinfo->Se; - break; - case (4*4-1): - cinfo->block_size = 4; - cinfo->natural_order = jpeg_natural_order4; - cinfo->lim_Se = cinfo->Se; - break; - case (5*5-1): - cinfo->block_size = 5; - cinfo->natural_order = jpeg_natural_order5; - cinfo->lim_Se = cinfo->Se; - break; - case (6*6-1): - cinfo->block_size = 6; - cinfo->natural_order = jpeg_natural_order6; - cinfo->lim_Se = cinfo->Se; - break; - case (7*7-1): - cinfo->block_size = 7; - cinfo->natural_order = jpeg_natural_order7; - cinfo->lim_Se = cinfo->Se; - break; - case (8*8-1): - cinfo->block_size = 8; - cinfo->natural_order = jpeg_natural_order; - cinfo->lim_Se = DCTSIZE2-1; - break; - case (9*9-1): - cinfo->block_size = 9; - cinfo->natural_order = jpeg_natural_order; - cinfo->lim_Se = DCTSIZE2-1; - break; - case (10*10-1): - cinfo->block_size = 10; - cinfo->natural_order = jpeg_natural_order; - cinfo->lim_Se = DCTSIZE2-1; - break; - case (11*11-1): - cinfo->block_size = 11; - cinfo->natural_order = jpeg_natural_order; - cinfo->lim_Se = DCTSIZE2-1; - break; - case (12*12-1): - cinfo->block_size = 12; - cinfo->natural_order = jpeg_natural_order; - cinfo->lim_Se = DCTSIZE2-1; - break; - case (13*13-1): - cinfo->block_size = 13; - cinfo->natural_order = jpeg_natural_order; - cinfo->lim_Se = DCTSIZE2-1; - break; - case (14*14-1): - cinfo->block_size = 14; - cinfo->natural_order = jpeg_natural_order; - cinfo->lim_Se = DCTSIZE2-1; - break; - case (15*15-1): - cinfo->block_size = 15; - cinfo->natural_order = jpeg_natural_order; - cinfo->lim_Se = DCTSIZE2-1; - break; - case (16*16-1): - cinfo->block_size = 16; - cinfo->natural_order = jpeg_natural_order; - cinfo->lim_Se = DCTSIZE2-1; - break; - default: - ERREXIT4(cinfo, JERR_BAD_PROGRESSION, - cinfo->Ss, cinfo->Se, cinfo->Ah, cinfo->Al); - } - - /* We initialize DCT_scaled_size and min_DCT_scaled_size to block_size. - * In the full decompressor, - * this will be overridden by jpeg_calc_output_dimensions in jdmaster.c; - * but in the transcoder, - * jpeg_calc_output_dimensions is not used, so we must do it here. - */ - cinfo->min_DCT_h_scaled_size = cinfo->block_size; - cinfo->min_DCT_v_scaled_size = cinfo->block_size; - - /* Compute dimensions of components */ - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - compptr->DCT_h_scaled_size = cinfo->block_size; - compptr->DCT_v_scaled_size = cinfo->block_size; - /* Size in DCT blocks */ - compptr->width_in_blocks = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * (long) compptr->h_samp_factor, - (long) (cinfo->max_h_samp_factor * cinfo->block_size)); - compptr->height_in_blocks = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * (long) compptr->v_samp_factor, - (long) (cinfo->max_v_samp_factor * cinfo->block_size)); - /* downsampled_width and downsampled_height will also be overridden by - * jdmaster.c if we are doing full decompression. The transcoder library - * doesn't use these values, but the calling application might. - */ - /* Size in samples */ - compptr->downsampled_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * (long) compptr->h_samp_factor, - (long) cinfo->max_h_samp_factor); - compptr->downsampled_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * (long) compptr->v_samp_factor, - (long) cinfo->max_v_samp_factor); - /* Mark component needed, until color conversion says otherwise */ - compptr->component_needed = TRUE; - /* Mark no quantization table yet saved for component */ - compptr->quant_table = NULL; - } - - /* Compute number of fully interleaved MCU rows. */ - cinfo->total_iMCU_rows = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height, - (long) (cinfo->max_v_samp_factor * cinfo->block_size)); - - /* Decide whether file contains multiple scans */ - if (cinfo->comps_in_scan < cinfo->num_components || cinfo->progressive_mode) - cinfo->inputctl->has_multiple_scans = TRUE; - else - cinfo->inputctl->has_multiple_scans = FALSE; -} - - -LOCAL(void) -per_scan_setup (j_decompress_ptr cinfo) -/* Do computations that are needed before processing a JPEG scan */ -/* cinfo->comps_in_scan and cinfo->cur_comp_info[] were set from SOS marker */ -{ - int ci, mcublks, tmp; - jpeg_component_info *compptr; - - if (cinfo->comps_in_scan == 1) { - - /* Noninterleaved (single-component) scan */ - compptr = cinfo->cur_comp_info[0]; - - /* Overall image size in MCUs */ - cinfo->MCUs_per_row = compptr->width_in_blocks; - cinfo->MCU_rows_in_scan = compptr->height_in_blocks; - - /* For noninterleaved scan, always one block per MCU */ - compptr->MCU_width = 1; - compptr->MCU_height = 1; - compptr->MCU_blocks = 1; - compptr->MCU_sample_width = compptr->DCT_h_scaled_size; - compptr->last_col_width = 1; - /* For noninterleaved scans, it is convenient to define last_row_height - * as the number of block rows present in the last iMCU row. - */ - tmp = (int) (compptr->height_in_blocks % compptr->v_samp_factor); - if (tmp == 0) tmp = compptr->v_samp_factor; - compptr->last_row_height = tmp; - - /* Prepare array describing MCU composition */ - cinfo->blocks_in_MCU = 1; - cinfo->MCU_membership[0] = 0; - - } else { - - /* Interleaved (multi-component) scan */ - if (cinfo->comps_in_scan <= 0 || cinfo->comps_in_scan > MAX_COMPS_IN_SCAN) - ERREXIT2(cinfo, JERR_COMPONENT_COUNT, cinfo->comps_in_scan, - MAX_COMPS_IN_SCAN); - - /* Overall image size in MCUs */ - cinfo->MCUs_per_row = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width, - (long) (cinfo->max_h_samp_factor * cinfo->block_size)); - cinfo->MCU_rows_in_scan = cinfo->total_iMCU_rows; - - cinfo->blocks_in_MCU = 0; - - for (ci = 0; ci < cinfo->comps_in_scan; ci++) { - compptr = cinfo->cur_comp_info[ci]; - /* Sampling factors give # of blocks of component in each MCU */ - compptr->MCU_width = compptr->h_samp_factor; - compptr->MCU_height = compptr->v_samp_factor; - compptr->MCU_blocks = compptr->MCU_width * compptr->MCU_height; - compptr->MCU_sample_width = compptr->MCU_width * compptr->DCT_h_scaled_size; - /* Figure number of non-dummy blocks in last MCU column & row */ - tmp = (int) (compptr->width_in_blocks % compptr->MCU_width); - if (tmp == 0) tmp = compptr->MCU_width; - compptr->last_col_width = tmp; - tmp = (int) (compptr->height_in_blocks % compptr->MCU_height); - if (tmp == 0) tmp = compptr->MCU_height; - compptr->last_row_height = tmp; - /* Prepare array describing MCU composition */ - mcublks = compptr->MCU_blocks; - if (cinfo->blocks_in_MCU + mcublks > D_MAX_BLOCKS_IN_MCU) - ERREXIT(cinfo, JERR_BAD_MCU_SIZE); - while (mcublks-- > 0) { - cinfo->MCU_membership[cinfo->blocks_in_MCU++] = ci; - } - } - - } -} - - -/* - * Save away a copy of the Q-table referenced by each component present - * in the current scan, unless already saved during a prior scan. - * - * In a multiple-scan JPEG file, the encoder could assign different components - * the same Q-table slot number, but change table definitions between scans - * so that each component uses a different Q-table. (The IJG encoder is not - * currently capable of doing this, but other encoders might.) Since we want - * to be able to dequantize all the components at the end of the file, this - * means that we have to save away the table actually used for each component. - * We do this by copying the table at the start of the first scan containing - * the component. - * The JPEG spec prohibits the encoder from changing the contents of a Q-table - * slot between scans of a component using that slot. If the encoder does so - * anyway, this decoder will simply use the Q-table values that were current - * at the start of the first scan for the component. - * - * The decompressor output side looks only at the saved quant tables, - * not at the current Q-table slots. - */ - -LOCAL(void) -latch_quant_tables (j_decompress_ptr cinfo) -{ - int ci, qtblno; - jpeg_component_info *compptr; - JQUANT_TBL * qtbl; - - for (ci = 0; ci < cinfo->comps_in_scan; ci++) { - compptr = cinfo->cur_comp_info[ci]; - /* No work if we already saved Q-table for this component */ - if (compptr->quant_table != NULL) - continue; - /* Make sure specified quantization table is present */ - qtblno = compptr->quant_tbl_no; - if (qtblno < 0 || qtblno >= NUM_QUANT_TBLS || - cinfo->quant_tbl_ptrs[qtblno] == NULL) - ERREXIT1(cinfo, JERR_NO_QUANT_TABLE, qtblno); - /* OK, save away the quantization table */ - qtbl = (JQUANT_TBL *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(JQUANT_TBL)); - MEMCOPY(qtbl, cinfo->quant_tbl_ptrs[qtblno], SIZEOF(JQUANT_TBL)); - compptr->quant_table = qtbl; - } -} - - -/* - * Initialize the input modules to read a scan of compressed data. - * The first call to this is done by jdmaster.c after initializing - * the entire decompressor (during jpeg_start_decompress). - * Subsequent calls come from consume_markers, below. - */ - -METHODDEF(void) -start_input_pass (j_decompress_ptr cinfo) -{ - per_scan_setup(cinfo); - latch_quant_tables(cinfo); - (*cinfo->entropy->start_pass) (cinfo); - (*cinfo->coef->start_input_pass) (cinfo); - cinfo->inputctl->consume_input = cinfo->coef->consume_data; -} - - -/* - * Finish up after inputting a compressed-data scan. - * This is called by the coefficient controller after it's read all - * the expected data of the scan. - */ - -METHODDEF(void) -finish_input_pass (j_decompress_ptr cinfo) -{ - (*cinfo->entropy->finish_pass) (cinfo); - cinfo->inputctl->consume_input = consume_markers; -} - - -/* - * Read JPEG markers before, between, or after compressed-data scans. - * Change state as necessary when a new scan is reached. - * Return value is JPEG_SUSPENDED, JPEG_REACHED_SOS, or JPEG_REACHED_EOI. - * - * The consume_input method pointer points either here or to the - * coefficient controller's consume_data routine, depending on whether - * we are reading a compressed data segment or inter-segment markers. - * - * Note: This function should NOT return a pseudo SOS marker (with zero - * component number) to the caller. A pseudo marker received by - * read_markers is processed and then skipped for other markers. - */ - -METHODDEF(int) -consume_markers (j_decompress_ptr cinfo) -{ - my_inputctl_ptr inputctl = (my_inputctl_ptr) cinfo->inputctl; - int val; - - if (inputctl->pub.eoi_reached) /* After hitting EOI, read no further */ - return JPEG_REACHED_EOI; - - for (;;) { /* Loop to pass pseudo SOS marker */ - val = (*cinfo->marker->read_markers) (cinfo); - - switch (val) { - case JPEG_REACHED_SOS: /* Found SOS */ - if (inputctl->inheaders) { /* 1st SOS */ - if (inputctl->inheaders == 1) - initial_setup(cinfo); - if (cinfo->comps_in_scan == 0) { /* pseudo SOS marker */ - inputctl->inheaders = 2; - break; - } - inputctl->inheaders = 0; - /* Note: start_input_pass must be called by jdmaster.c - * before any more input can be consumed. jdapimin.c is - * responsible for enforcing this sequencing. - */ - } else { /* 2nd or later SOS marker */ - if (! inputctl->pub.has_multiple_scans) - ERREXIT(cinfo, JERR_EOI_EXPECTED); /* Oops, I wasn't expecting this! */ - if (cinfo->comps_in_scan == 0) /* unexpected pseudo SOS marker */ - break; - start_input_pass(cinfo); - } - return val; - case JPEG_REACHED_EOI: /* Found EOI */ - inputctl->pub.eoi_reached = TRUE; - if (inputctl->inheaders) { /* Tables-only datastream, apparently */ - if (cinfo->marker->saw_SOF) - ERREXIT(cinfo, JERR_SOF_NO_SOS); - } else { - /* Prevent infinite loop in coef ctlr's decompress_data routine - * if user set output_scan_number larger than number of scans. - */ - if (cinfo->output_scan_number > cinfo->input_scan_number) - cinfo->output_scan_number = cinfo->input_scan_number; - } - return val; - case JPEG_SUSPENDED: - return val; - default: - return val; - } - } -} - - -/* - * Reset state to begin a fresh datastream. - */ - -METHODDEF(void) -reset_input_controller (j_decompress_ptr cinfo) -{ - my_inputctl_ptr inputctl = (my_inputctl_ptr) cinfo->inputctl; - - inputctl->pub.consume_input = consume_markers; - inputctl->pub.has_multiple_scans = FALSE; /* "unknown" would be better */ - inputctl->pub.eoi_reached = FALSE; - inputctl->inheaders = 1; - /* Reset other modules */ - (*cinfo->err->reset_error_mgr) ((j_common_ptr) cinfo); - (*cinfo->marker->reset_marker_reader) (cinfo); - /* Reset progression state -- would be cleaner if entropy decoder did this */ - cinfo->coef_bits = NULL; -} - - -/* - * Initialize the input controller module. - * This is called only once, when the decompression object is created. - */ - -GLOBAL(void) -jinit_input_controller (j_decompress_ptr cinfo) -{ - my_inputctl_ptr inputctl; - - /* Create subobject in permanent pool */ - inputctl = (my_inputctl_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_PERMANENT, SIZEOF(my_input_controller)); - cinfo->inputctl = &inputctl->pub; - /* Initialize method pointers */ - inputctl->pub.consume_input = consume_markers; - inputctl->pub.reset_input_controller = reset_input_controller; - inputctl->pub.start_input_pass = start_input_pass; - inputctl->pub.finish_input_pass = finish_input_pass; - /* Initialize state: can't use reset_input_controller since we don't - * want to try to reset other modules yet. - */ - inputctl->pub.has_multiple_scans = FALSE; /* "unknown" would be better */ - inputctl->pub.eoi_reached = FALSE; - inputctl->inheaders = 1; -} diff --git a/cmake/externals/libjpeg/jdmainct.c b/cmake/externals/libjpeg/jdmainct.c deleted file mode 100644 index 1cd66d853..000000000 --- a/cmake/externals/libjpeg/jdmainct.c +++ /dev/null @@ -1,511 +0,0 @@ -/* - * jdmainct.c - * - * Copyright (C) 1994-1996, Thomas G. Lane. - * Modified 2002-2020 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains the main buffer controller for decompression. - * The main buffer lies between the JPEG decompressor proper and the - * post-processor; it holds downsampled data in the JPEG colorspace. - * - * Note that this code is bypassed in raw-data mode, since the application - * supplies the equivalent of the main buffer in that case. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* - * In the current system design, the main buffer need never be a full-image - * buffer; any full-height buffers will be found inside the coefficient or - * postprocessing controllers. Nonetheless, the main controller is not - * trivial. Its responsibility is to provide context rows for upsampling/ - * rescaling, and doing this in an efficient fashion is a bit tricky. - * - * Postprocessor input data is counted in "row groups". A row group is - * defined to be (v_samp_factor * DCT_v_scaled_size / min_DCT_v_scaled_size) - * sample rows of each component. (We require DCT_scaled_size values to be - * chosen such that these numbers are integers. In practice DCT_scaled_size - * values will likely be powers of two, so we actually have the stronger - * condition that DCT_scaled_size / min_DCT_scaled_size is an integer.) - * Upsampling will typically produce max_v_samp_factor pixel rows from each - * row group (times any additional scale factor that the upsampler is - * applying). - * - * The coefficient controller will deliver data to us one iMCU row at a time; - * each iMCU row contains v_samp_factor * DCT_v_scaled_size sample rows, or - * exactly min_DCT_v_scaled_size row groups. (This amount of data corresponds - * to one row of MCUs when the image is fully interleaved.) Note that the - * number of sample rows varies across components, but the number of row - * groups does not. Some garbage sample rows may be included in the last iMCU - * row at the bottom of the image. - * - * Depending on the vertical scaling algorithm used, the upsampler may need - * access to the sample row(s) above and below its current input row group. - * The upsampler is required to set need_context_rows TRUE at global selection - * time if so. When need_context_rows is FALSE, this controller can simply - * obtain one iMCU row at a time from the coefficient controller and dole it - * out as row groups to the postprocessor. - * - * When need_context_rows is TRUE, this controller guarantees that the buffer - * passed to postprocessing contains at least one row group's worth of samples - * above and below the row group(s) being processed. Note that the context - * rows "above" the first passed row group appear at negative row offsets in - * the passed buffer. At the top and bottom of the image, the required - * context rows are manufactured by duplicating the first or last real sample - * row; this avoids having special cases in the upsampling inner loops. - * - * The amount of context is fixed at one row group just because that's a - * convenient number for this controller to work with. The existing - * upsamplers really only need one sample row of context. An upsampler - * supporting arbitrary output rescaling might wish for more than one row - * group of context when shrinking the image; tough, we don't handle that. - * (This is justified by the assumption that downsizing will be handled mostly - * by adjusting the DCT_scaled_size values, so that the actual scale factor at - * the upsample step needn't be much less than one.) - * - * To provide the desired context, we have to retain the last two row groups - * of one iMCU row while reading in the next iMCU row. (The last row group - * can't be processed until we have another row group for its below-context, - * and so we have to save the next-to-last group too for its above-context.) - * We could do this most simply by copying data around in our buffer, but - * that'd be very slow. We can avoid copying any data by creating a rather - * strange pointer structure. Here's how it works. We allocate a workspace - * consisting of M+2 row groups (where M = min_DCT_v_scaled_size is the number - * of row groups per iMCU row). We create two sets of redundant pointers to - * the workspace. Labeling the physical row groups 0 to M+1, the synthesized - * pointer lists look like this: - * M+1 M-1 - * master pointer --> 0 master pointer --> 0 - * 1 1 - * ... ... - * M-3 M-3 - * M-2 M - * M-1 M+1 - * M M-2 - * M+1 M-1 - * 0 0 - * We read alternate iMCU rows using each master pointer; thus the last two - * row groups of the previous iMCU row remain un-overwritten in the workspace. - * The pointer lists are set up so that the required context rows appear to - * be adjacent to the proper places when we pass the pointer lists to the - * upsampler. - * - * The above pictures describe the normal state of the pointer lists. - * At top and bottom of the image, we diddle the pointer lists to duplicate - * the first or last sample row as necessary (this is cheaper than copying - * sample rows around). - * - * This scheme breaks down if M < 2, ie, min_DCT_v_scaled_size is 1. In that - * situation each iMCU row provides only one row group so the buffering logic - * must be different (eg, we must read two iMCU rows before we can emit the - * first row group). For now, we simply do not support providing context - * rows when min_DCT_v_scaled_size is 1. That combination seems unlikely to - * be worth providing --- if someone wants a 1/8th-size preview, they probably - * want it quick and dirty, so a context-free upsampler is sufficient. - */ - - -/* Private buffer controller object */ - -typedef struct { - struct jpeg_d_main_controller pub; /* public fields */ - - /* Pointer to allocated workspace (M or M+2 row groups). */ - JSAMPARRAY buffer[MAX_COMPONENTS]; - - JDIMENSION rowgroup_ctr; /* counts row groups output to postprocessor */ - JDIMENSION rowgroups_avail; /* row groups available to postprocessor */ - - /* Remaining fields are only used in the context case. */ - - boolean buffer_full; /* Have we gotten an iMCU row from decoder? */ - - /* These are the master pointers to the funny-order pointer lists. */ - JSAMPIMAGE xbuffer[2]; /* pointers to weird pointer lists */ - - int whichptr; /* indicates which pointer set is now in use */ - int context_state; /* process_data state machine status */ - JDIMENSION iMCU_row_ctr; /* counts iMCU rows to detect image top/bot */ -} my_main_controller; - -typedef my_main_controller * my_main_ptr; - -/* context_state values: */ -#define CTX_PREPARE_FOR_IMCU 0 /* need to prepare for MCU row */ -#define CTX_PROCESS_IMCU 1 /* feeding iMCU to postprocessor */ -#define CTX_POSTPONED_ROW 2 /* feeding postponed row group */ - - -/* Forward declarations */ -METHODDEF(void) process_data_simple_main - JPP((j_decompress_ptr cinfo, JSAMPARRAY output_buf, - JDIMENSION *out_row_ctr, JDIMENSION out_rows_avail)); -METHODDEF(void) process_data_context_main - JPP((j_decompress_ptr cinfo, JSAMPARRAY output_buf, - JDIMENSION *out_row_ctr, JDIMENSION out_rows_avail)); -#ifdef QUANT_2PASS_SUPPORTED -METHODDEF(void) process_data_crank_post - JPP((j_decompress_ptr cinfo, JSAMPARRAY output_buf, - JDIMENSION *out_row_ctr, JDIMENSION out_rows_avail)); -#endif - - -LOCAL(void) -alloc_funny_pointers (j_decompress_ptr cinfo) -/* Allocate space for the funny pointer lists. - * This is done only once, not once per pass. - */ -{ - my_main_ptr mainp = (my_main_ptr) cinfo->main; - int ci, rgroup; - int M = cinfo->min_DCT_v_scaled_size; - jpeg_component_info *compptr; - JSAMPARRAY xbuf; - - /* Get top-level space for component array pointers. - * We alloc both arrays with one call to save a few cycles. - */ - mainp->xbuffer[0] = (JSAMPIMAGE) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, - cinfo->num_components * 2 * SIZEOF(JSAMPARRAY)); - mainp->xbuffer[1] = mainp->xbuffer[0] + cinfo->num_components; - - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - if (! compptr->component_needed) - continue; /* skip uninteresting component */ - rgroup = (compptr->v_samp_factor * compptr->DCT_v_scaled_size) / - cinfo->min_DCT_v_scaled_size; /* height of a row group of component */ - /* Get space for pointer lists --- M+4 row groups in each list. - * We alloc both pointer lists with one call to save a few cycles. - */ - xbuf = (JSAMPARRAY) (*cinfo->mem->alloc_small) ((j_common_ptr) cinfo, - JPOOL_IMAGE, 2 * (rgroup * (M + 4)) * SIZEOF(JSAMPROW)); - xbuf += rgroup; /* want one row group at negative offsets */ - mainp->xbuffer[0][ci] = xbuf; - xbuf += rgroup * (M + 4); - mainp->xbuffer[1][ci] = xbuf; - } -} - - -LOCAL(void) -make_funny_pointers (j_decompress_ptr cinfo) -/* Create the funny pointer lists discussed in the comments above. - * The actual workspace is already allocated (in mainp->buffer), - * and the space for the pointer lists is allocated too. - * This routine just fills in the curiously ordered lists. - * This will be repeated at the beginning of each pass. - */ -{ - my_main_ptr mainp = (my_main_ptr) cinfo->main; - int ci, i, rgroup; - int M = cinfo->min_DCT_v_scaled_size; - jpeg_component_info *compptr; - JSAMPARRAY buf, xbuf0, xbuf1; - - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - if (! compptr->component_needed) - continue; /* skip uninteresting component */ - rgroup = (compptr->v_samp_factor * compptr->DCT_v_scaled_size) / - cinfo->min_DCT_v_scaled_size; /* height of a row group of component */ - xbuf0 = mainp->xbuffer[0][ci]; - xbuf1 = mainp->xbuffer[1][ci]; - /* First copy the workspace pointers as-is */ - buf = mainp->buffer[ci]; - for (i = 0; i < rgroup * (M + 2); i++) { - xbuf0[i] = xbuf1[i] = buf[i]; - } - /* In the second list, put the last four row groups in swapped order */ - for (i = 0; i < rgroup * 2; i++) { - xbuf1[rgroup*(M-2) + i] = buf[rgroup*M + i]; - xbuf1[rgroup*M + i] = buf[rgroup*(M-2) + i]; - } - /* The wraparound pointers at top and bottom will be filled later - * (see set_wraparound_pointers, below). Initially we want the "above" - * pointers to duplicate the first actual data line. This only needs - * to happen in xbuffer[0]. - */ - for (i = 0; i < rgroup; i++) { - xbuf0[i - rgroup] = xbuf0[0]; - } - } -} - - -LOCAL(void) -set_wraparound_pointers (j_decompress_ptr cinfo) -/* Set up the "wraparound" pointers at top and bottom of the pointer lists. - * This changes the pointer list state from top-of-image to the normal state. - */ -{ - my_main_ptr mainp = (my_main_ptr) cinfo->main; - int ci, i, rgroup; - int M = cinfo->min_DCT_v_scaled_size; - jpeg_component_info *compptr; - JSAMPARRAY xbuf0, xbuf1; - - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - if (! compptr->component_needed) - continue; /* skip uninteresting component */ - rgroup = (compptr->v_samp_factor * compptr->DCT_v_scaled_size) / - cinfo->min_DCT_v_scaled_size; /* height of a row group of component */ - xbuf0 = mainp->xbuffer[0][ci]; - xbuf1 = mainp->xbuffer[1][ci]; - for (i = 0; i < rgroup; i++) { - xbuf0[i - rgroup] = xbuf0[rgroup*(M+1) + i]; - xbuf1[i - rgroup] = xbuf1[rgroup*(M+1) + i]; - xbuf0[rgroup*(M+2) + i] = xbuf0[i]; - xbuf1[rgroup*(M+2) + i] = xbuf1[i]; - } - } -} - - -LOCAL(void) -set_bottom_pointers (j_decompress_ptr cinfo) -/* Change the pointer lists to duplicate the last sample row at the bottom - * of the image. whichptr indicates which xbuffer holds the final iMCU row. - * Also sets rowgroups_avail to indicate number of nondummy row groups in row. - */ -{ - my_main_ptr mainp = (my_main_ptr) cinfo->main; - int ci, i, rgroup, iMCUheight, rows_left; - jpeg_component_info *compptr; - JSAMPARRAY xbuf; - - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - if (! compptr->component_needed) - continue; /* skip uninteresting component */ - /* Count sample rows in one iMCU row and in one row group */ - iMCUheight = compptr->v_samp_factor * compptr->DCT_v_scaled_size; - rgroup = iMCUheight / cinfo->min_DCT_v_scaled_size; - /* Count nondummy sample rows remaining for this component */ - rows_left = (int) (compptr->downsampled_height % (JDIMENSION) iMCUheight); - if (rows_left == 0) rows_left = iMCUheight; - /* Count nondummy row groups. Should get same answer for each component, - * so we need only do it once. - */ - if (ci == 0) { - mainp->rowgroups_avail = (JDIMENSION) ((rows_left-1) / rgroup + 1); - } - /* Duplicate the last real sample row rgroup*2 times; this pads out the - * last partial rowgroup and ensures at least one full rowgroup of context. - */ - xbuf = mainp->xbuffer[mainp->whichptr][ci]; - for (i = 0; i < rgroup * 2; i++) { - xbuf[rows_left + i] = xbuf[rows_left-1]; - } - } -} - - -/* - * Initialize for a processing pass. - */ - -METHODDEF(void) -start_pass_main (j_decompress_ptr cinfo, J_BUF_MODE pass_mode) -{ - my_main_ptr mainp = (my_main_ptr) cinfo->main; - - switch (pass_mode) { - case JBUF_PASS_THRU: - if (cinfo->upsample->need_context_rows) { - mainp->pub.process_data = process_data_context_main; - make_funny_pointers(cinfo); /* Create the xbuffer[] lists */ - mainp->whichptr = 0; /* Read first iMCU row into xbuffer[0] */ - mainp->context_state = CTX_PREPARE_FOR_IMCU; - mainp->iMCU_row_ctr = 0; - mainp->buffer_full = FALSE; /* Mark buffer empty */ - } else { - /* Simple case with no context needed */ - mainp->pub.process_data = process_data_simple_main; - mainp->rowgroup_ctr = mainp->rowgroups_avail; /* Mark buffer empty */ - } - break; -#ifdef QUANT_2PASS_SUPPORTED - case JBUF_CRANK_DEST: - /* For last pass of 2-pass quantization, just crank the postprocessor */ - mainp->pub.process_data = process_data_crank_post; - break; -#endif - default: - ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); - } -} - - -/* - * Process some data. - * This handles the simple case where no context is required. - */ - -METHODDEF(void) -process_data_simple_main (j_decompress_ptr cinfo, JSAMPARRAY output_buf, - JDIMENSION *out_row_ctr, JDIMENSION out_rows_avail) -{ - my_main_ptr mainp = (my_main_ptr) cinfo->main; - - /* Read input data if we haven't filled the main buffer yet */ - if (mainp->rowgroup_ctr >= mainp->rowgroups_avail) { - if (! (*cinfo->coef->decompress_data) (cinfo, mainp->buffer)) - return; /* suspension forced, can do nothing more */ - mainp->rowgroup_ctr = 0; /* OK, we have an iMCU row to work with */ - } - - /* Note: at the bottom of the image, we may pass extra garbage row groups - * to the postprocessor. The postprocessor has to check for bottom - * of image anyway (at row resolution), so no point in us doing it too. - */ - - /* Feed the postprocessor */ - (*cinfo->post->post_process_data) (cinfo, mainp->buffer, - &mainp->rowgroup_ctr, mainp->rowgroups_avail, - output_buf, out_row_ctr, out_rows_avail); -} - - -/* - * Process some data. - * This handles the case where context rows must be provided. - */ - -METHODDEF(void) -process_data_context_main (j_decompress_ptr cinfo, JSAMPARRAY output_buf, - JDIMENSION *out_row_ctr, JDIMENSION out_rows_avail) -{ - my_main_ptr mainp = (my_main_ptr) cinfo->main; - - /* Read input data if we haven't filled the main buffer yet */ - if (! mainp->buffer_full) { - if (! (*cinfo->coef->decompress_data) (cinfo, - mainp->xbuffer[mainp->whichptr])) - return; /* suspension forced, can do nothing more */ - mainp->buffer_full = TRUE; /* OK, we have an iMCU row to work with */ - mainp->iMCU_row_ctr++; /* count rows received */ - } - - /* Postprocessor typically will not swallow all the input data it is handed - * in one call (due to filling the output buffer first). Must be prepared - * to exit and restart. This switch lets us keep track of how far we got. - * Note that each case falls through to the next on successful completion. - */ - switch (mainp->context_state) { - case CTX_POSTPONED_ROW: - /* Call postprocessor using previously set pointers for postponed row */ - (*cinfo->post->post_process_data) (cinfo, mainp->xbuffer[mainp->whichptr], - &mainp->rowgroup_ctr, mainp->rowgroups_avail, - output_buf, out_row_ctr, out_rows_avail); - if (mainp->rowgroup_ctr < mainp->rowgroups_avail) - return; /* Need to suspend */ - mainp->context_state = CTX_PREPARE_FOR_IMCU; - if (*out_row_ctr >= out_rows_avail) - return; /* Postprocessor exactly filled output buf */ - /*FALLTHROUGH*/ - case CTX_PREPARE_FOR_IMCU: - /* Prepare to process first M-1 row groups of this iMCU row */ - mainp->rowgroup_ctr = 0; - mainp->rowgroups_avail = (JDIMENSION) (cinfo->min_DCT_v_scaled_size - 1); - /* Check for bottom of image: if so, tweak pointers to "duplicate" - * the last sample row, and adjust rowgroups_avail to ignore padding rows. - */ - if (mainp->iMCU_row_ctr == cinfo->total_iMCU_rows) - set_bottom_pointers(cinfo); - mainp->context_state = CTX_PROCESS_IMCU; - /*FALLTHROUGH*/ - case CTX_PROCESS_IMCU: - /* Call postprocessor using previously set pointers */ - (*cinfo->post->post_process_data) (cinfo, mainp->xbuffer[mainp->whichptr], - &mainp->rowgroup_ctr, mainp->rowgroups_avail, - output_buf, out_row_ctr, out_rows_avail); - if (mainp->rowgroup_ctr < mainp->rowgroups_avail) - return; /* Need to suspend */ - /* After the first iMCU, change wraparound pointers to normal state */ - if (mainp->iMCU_row_ctr == 1) - set_wraparound_pointers(cinfo); - /* Prepare to load new iMCU row using other xbuffer list */ - mainp->whichptr ^= 1; /* 0=>1 or 1=>0 */ - mainp->buffer_full = FALSE; - /* Still need to process last row group of this iMCU row, */ - /* which is saved at index M+1 of the other xbuffer */ - mainp->rowgroup_ctr = (JDIMENSION) (cinfo->min_DCT_v_scaled_size + 1); - mainp->rowgroups_avail = (JDIMENSION) (cinfo->min_DCT_v_scaled_size + 2); - mainp->context_state = CTX_POSTPONED_ROW; - } -} - - -/* - * Process some data. - * Final pass of two-pass quantization: just call the postprocessor. - * Source data will be the postprocessor controller's internal buffer. - */ - -#ifdef QUANT_2PASS_SUPPORTED - -METHODDEF(void) -process_data_crank_post (j_decompress_ptr cinfo, JSAMPARRAY output_buf, - JDIMENSION *out_row_ctr, JDIMENSION out_rows_avail) -{ - (*cinfo->post->post_process_data) (cinfo, (JSAMPIMAGE) NULL, - (JDIMENSION *) NULL, (JDIMENSION) 0, - output_buf, out_row_ctr, out_rows_avail); -} - -#endif /* QUANT_2PASS_SUPPORTED */ - - -/* - * Initialize main buffer controller. - */ - -GLOBAL(void) -jinit_d_main_controller (j_decompress_ptr cinfo, boolean need_full_buffer) -{ - my_main_ptr mainp; - int ci, rgroup, ngroups; - jpeg_component_info *compptr; - - mainp = (my_main_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_main_controller)); - cinfo->main = &mainp->pub; - mainp->pub.start_pass = start_pass_main; - - if (need_full_buffer) /* shouldn't happen */ - ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); - - /* Allocate the workspace. - * ngroups is the number of row groups we need. - */ - if (cinfo->upsample->need_context_rows) { - if (cinfo->min_DCT_v_scaled_size < 2) /* unsupported, see comments above */ - ERREXIT(cinfo, JERR_NOTIMPL); - alloc_funny_pointers(cinfo); /* Alloc space for xbuffer[] lists */ - ngroups = cinfo->min_DCT_v_scaled_size + 2; - } else { - /* There are always min_DCT_v_scaled_size row groups in an iMCU row. */ - ngroups = cinfo->min_DCT_v_scaled_size; - mainp->rowgroups_avail = (JDIMENSION) ngroups; - } - - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - if (! compptr->component_needed) - continue; /* skip uninteresting component */ - rgroup = (compptr->v_samp_factor * compptr->DCT_v_scaled_size) / - cinfo->min_DCT_v_scaled_size; /* height of a row group of component */ - mainp->buffer[ci] = (*cinfo->mem->alloc_sarray) - ((j_common_ptr) cinfo, JPOOL_IMAGE, - compptr->width_in_blocks * ((JDIMENSION) compptr->DCT_h_scaled_size), - (JDIMENSION) (rgroup * ngroups)); - } -} diff --git a/cmake/externals/libjpeg/jdmarker.c b/cmake/externals/libjpeg/jdmarker.c deleted file mode 100644 index c10fde605..000000000 --- a/cmake/externals/libjpeg/jdmarker.c +++ /dev/null @@ -1,1505 +0,0 @@ -/* - * jdmarker.c - * - * Copyright (C) 1991-1998, Thomas G. Lane. - * Modified 2009-2019 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains routines to decode JPEG datastream markers. - * Most of the complexity arises from our desire to support input - * suspension: if not all of the data for a marker is available, - * we must exit back to the application. On resumption, we reprocess - * the marker. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -typedef enum { /* JPEG marker codes */ - M_SOF0 = 0xc0, - M_SOF1 = 0xc1, - M_SOF2 = 0xc2, - M_SOF3 = 0xc3, - - M_SOF5 = 0xc5, - M_SOF6 = 0xc6, - M_SOF7 = 0xc7, - - M_JPG = 0xc8, - M_SOF9 = 0xc9, - M_SOF10 = 0xca, - M_SOF11 = 0xcb, - - M_SOF13 = 0xcd, - M_SOF14 = 0xce, - M_SOF15 = 0xcf, - - M_DHT = 0xc4, - - M_DAC = 0xcc, - - M_RST0 = 0xd0, - M_RST1 = 0xd1, - M_RST2 = 0xd2, - M_RST3 = 0xd3, - M_RST4 = 0xd4, - M_RST5 = 0xd5, - M_RST6 = 0xd6, - M_RST7 = 0xd7, - - M_SOI = 0xd8, - M_EOI = 0xd9, - M_SOS = 0xda, - M_DQT = 0xdb, - M_DNL = 0xdc, - M_DRI = 0xdd, - M_DHP = 0xde, - M_EXP = 0xdf, - - M_APP0 = 0xe0, - M_APP1 = 0xe1, - M_APP2 = 0xe2, - M_APP3 = 0xe3, - M_APP4 = 0xe4, - M_APP5 = 0xe5, - M_APP6 = 0xe6, - M_APP7 = 0xe7, - M_APP8 = 0xe8, - M_APP9 = 0xe9, - M_APP10 = 0xea, - M_APP11 = 0xeb, - M_APP12 = 0xec, - M_APP13 = 0xed, - M_APP14 = 0xee, - M_APP15 = 0xef, - - M_JPG0 = 0xf0, - M_JPG8 = 0xf8, - M_JPG13 = 0xfd, - M_COM = 0xfe, - - M_TEM = 0x01, - - M_ERROR = 0x100 -} JPEG_MARKER; - - -/* Private state */ - -typedef struct { - struct jpeg_marker_reader pub; /* public fields */ - - /* Application-overridable marker processing methods */ - jpeg_marker_parser_method process_COM; - jpeg_marker_parser_method process_APPn[16]; - - /* Limit on marker data length to save for each marker type */ - unsigned int length_limit_COM; - unsigned int length_limit_APPn[16]; - - /* Status of COM/APPn marker saving */ - jpeg_saved_marker_ptr cur_marker; /* NULL if not processing a marker */ - unsigned int bytes_read; /* data bytes read so far in marker */ - /* Note: cur_marker is not linked into marker_list until it's all read. */ -} my_marker_reader; - -typedef my_marker_reader * my_marker_ptr; - - -/* - * Macros for fetching data from the data source module. - * - * At all times, cinfo->src->next_input_byte and ->bytes_in_buffer reflect - * the current restart point; we update them only when we have reached a - * suitable place to restart if a suspension occurs. - */ - -/* Declare and initialize local copies of input pointer/count */ -#define INPUT_VARS(cinfo) \ - struct jpeg_source_mgr * datasrc = (cinfo)->src; \ - const JOCTET * next_input_byte = datasrc->next_input_byte; \ - size_t bytes_in_buffer = datasrc->bytes_in_buffer - -/* Unload the local copies --- do this only at a restart boundary */ -#define INPUT_SYNC(cinfo) \ - ( datasrc->next_input_byte = next_input_byte, \ - datasrc->bytes_in_buffer = bytes_in_buffer ) - -/* Reload the local copies --- used only in MAKE_BYTE_AVAIL */ -#define INPUT_RELOAD(cinfo) \ - ( next_input_byte = datasrc->next_input_byte, \ - bytes_in_buffer = datasrc->bytes_in_buffer ) - -/* Internal macro for INPUT_BYTE and INPUT_2BYTES: make a byte available. - * Note we do *not* do INPUT_SYNC before calling fill_input_buffer, - * but we must reload the local copies after a successful fill. - */ -#define MAKE_BYTE_AVAIL(cinfo,action) \ - if (bytes_in_buffer == 0) { \ - if (! (*datasrc->fill_input_buffer) (cinfo)) \ - { action; } \ - INPUT_RELOAD(cinfo); \ - } - -/* Read a byte into variable V. - * If must suspend, take the specified action (typically "return FALSE"). - */ -#define INPUT_BYTE(cinfo,V,action) \ - MAKESTMT( MAKE_BYTE_AVAIL(cinfo,action); \ - bytes_in_buffer--; \ - V = GETJOCTET(*next_input_byte++); ) - -/* As above, but read two bytes interpreted as an unsigned 16-bit integer. - * V should be declared unsigned int or perhaps INT32. - */ -#define INPUT_2BYTES(cinfo,V,action) \ - MAKESTMT( MAKE_BYTE_AVAIL(cinfo,action); \ - bytes_in_buffer--; \ - V = ((unsigned int) GETJOCTET(*next_input_byte++)) << 8; \ - MAKE_BYTE_AVAIL(cinfo,action); \ - bytes_in_buffer--; \ - V += GETJOCTET(*next_input_byte++); ) - - -/* - * Routines to process JPEG markers. - * - * Entry condition: JPEG marker itself has been read and its code saved - * in cinfo->unread_marker; input restart point is just after the marker. - * - * Exit: if return TRUE, have read and processed any parameters, and have - * updated the restart point to point after the parameters. - * If return FALSE, was forced to suspend before reaching end of - * marker parameters; restart point has not been moved. Same routine - * will be called again after application supplies more input data. - * - * This approach to suspension assumes that all of a marker's parameters - * can fit into a single input bufferload. This should hold for "normal" - * markers. Some COM/APPn markers might have large parameter segments - * that might not fit. If we are simply dropping such a marker, we use - * skip_input_data to get past it, and thereby put the problem on the - * source manager's shoulders. If we are saving the marker's contents - * into memory, we use a slightly different convention: when forced to - * suspend, the marker processor updates the restart point to the end of - * what it's consumed (ie, the end of the buffer) before returning FALSE. - * On resumption, cinfo->unread_marker still contains the marker code, - * but the data source will point to the next chunk of marker data. - * The marker processor must retain internal state to deal with this. - * - * Note that we don't bother to avoid duplicate trace messages if a - * suspension occurs within marker parameters. Other side effects - * require more care. - */ - - -LOCAL(boolean) -get_soi (j_decompress_ptr cinfo) -/* Process an SOI marker */ -{ - int i; - - TRACEMS(cinfo, 1, JTRC_SOI); - - if (cinfo->marker->saw_SOI) - ERREXIT(cinfo, JERR_SOI_DUPLICATE); - - /* Reset all parameters that are defined to be reset by SOI */ - - for (i = 0; i < NUM_ARITH_TBLS; i++) { - cinfo->arith_dc_L[i] = 0; - cinfo->arith_dc_U[i] = 1; - cinfo->arith_ac_K[i] = 5; - } - cinfo->restart_interval = 0; - - /* Set initial assumptions for colorspace etc */ - - cinfo->jpeg_color_space = JCS_UNKNOWN; - cinfo->color_transform = JCT_NONE; - cinfo->CCIR601_sampling = FALSE; /* Assume non-CCIR sampling??? */ - - cinfo->saw_JFIF_marker = FALSE; - cinfo->JFIF_major_version = 1; /* set default JFIF APP0 values */ - cinfo->JFIF_minor_version = 1; - cinfo->density_unit = 0; - cinfo->X_density = 1; - cinfo->Y_density = 1; - cinfo->saw_Adobe_marker = FALSE; - cinfo->Adobe_transform = 0; - - cinfo->marker->saw_SOI = TRUE; - - return TRUE; -} - - -LOCAL(boolean) -get_sof (j_decompress_ptr cinfo, boolean is_baseline, boolean is_prog, - boolean is_arith) -/* Process a SOFn marker */ -{ - INT32 length; - int c, ci, i; - jpeg_component_info * compptr; - INPUT_VARS(cinfo); - - cinfo->is_baseline = is_baseline; - cinfo->progressive_mode = is_prog; - cinfo->arith_code = is_arith; - - INPUT_2BYTES(cinfo, length, return FALSE); - - INPUT_BYTE(cinfo, cinfo->data_precision, return FALSE); - INPUT_2BYTES(cinfo, cinfo->image_height, return FALSE); - INPUT_2BYTES(cinfo, cinfo->image_width, return FALSE); - INPUT_BYTE(cinfo, cinfo->num_components, return FALSE); - - length -= 8; - - TRACEMS4(cinfo, 1, JTRC_SOF, cinfo->unread_marker, - (int) cinfo->image_width, (int) cinfo->image_height, - cinfo->num_components); - - if (cinfo->marker->saw_SOF) - ERREXIT(cinfo, JERR_SOF_DUPLICATE); - - /* We don't support files in which the image height is initially specified */ - /* as 0 and is later redefined by DNL. As long as we have to check that, */ - /* might as well have a general sanity check. */ - if (cinfo->image_height <= 0 || cinfo->image_width <= 0 || - cinfo->num_components <= 0) - ERREXIT(cinfo, JERR_EMPTY_IMAGE); - - if (length != (cinfo->num_components * 3)) - ERREXIT(cinfo, JERR_BAD_LENGTH); - - if (cinfo->comp_info == NULL) /* do only once, even if suspend */ - cinfo->comp_info = (jpeg_component_info *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, - cinfo->num_components * SIZEOF(jpeg_component_info)); - - for (ci = 0; ci < cinfo->num_components; ci++) { - INPUT_BYTE(cinfo, c, return FALSE); - /* Check to see whether component id has already been seen */ - /* (in violation of the spec, but unfortunately seen in some */ - /* files). If so, create "fake" component id equal to the */ - /* max id seen so far + 1. */ - for (i = 0, compptr = cinfo->comp_info; i < ci; i++, compptr++) { - if (c == compptr->component_id) { - compptr = cinfo->comp_info; - c = compptr->component_id; - compptr++; - for (i = 1; i < ci; i++, compptr++) { - if (compptr->component_id > c) c = compptr->component_id; - } - c++; - break; - } - } - compptr->component_id = c; - compptr->component_index = ci; - INPUT_BYTE(cinfo, c, return FALSE); - compptr->h_samp_factor = (c >> 4) & 15; - compptr->v_samp_factor = (c ) & 15; - INPUT_BYTE(cinfo, compptr->quant_tbl_no, return FALSE); - - TRACEMS4(cinfo, 1, JTRC_SOF_COMPONENT, - compptr->component_id, compptr->h_samp_factor, - compptr->v_samp_factor, compptr->quant_tbl_no); - } - - cinfo->marker->saw_SOF = TRUE; - - INPUT_SYNC(cinfo); - return TRUE; -} - - -LOCAL(boolean) -get_sos (j_decompress_ptr cinfo) -/* Process a SOS marker */ -{ - INT32 length; - int c, ci, i, n; - jpeg_component_info * compptr; - INPUT_VARS(cinfo); - - if (! cinfo->marker->saw_SOF) - ERREXITS(cinfo, JERR_SOF_BEFORE, "SOS"); - - INPUT_2BYTES(cinfo, length, return FALSE); - - INPUT_BYTE(cinfo, n, return FALSE); /* Number of components */ - - TRACEMS1(cinfo, 1, JTRC_SOS, n); - - if (length != (n * 2 + 6) || n > MAX_COMPS_IN_SCAN || - (n == 0 && !cinfo->progressive_mode)) - /* pseudo SOS marker only allowed in progressive mode */ - ERREXIT(cinfo, JERR_BAD_LENGTH); - - cinfo->comps_in_scan = n; - - /* Collect the component-spec parameters */ - - for (i = 0; i < n; i++) { - INPUT_BYTE(cinfo, c, return FALSE); - - /* Detect the case where component id's are not unique, and, if so, */ - /* create a fake component id using the same logic as in get_sof. */ - /* Note: This also ensures that all of the SOF components are */ - /* referenced in the single scan case, which prevents access to */ - /* uninitialized memory in later decoding stages. */ - for (ci = 0; ci < i; ci++) { - if (c == cinfo->cur_comp_info[ci]->component_id) { - c = cinfo->cur_comp_info[0]->component_id; - for (ci = 1; ci < i; ci++) { - compptr = cinfo->cur_comp_info[ci]; - if (compptr->component_id > c) c = compptr->component_id; - } - c++; - break; - } - } - - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - if (c == compptr->component_id) - goto id_found; - } - - ERREXIT1(cinfo, JERR_BAD_COMPONENT_ID, c); - - id_found: - - cinfo->cur_comp_info[i] = compptr; - INPUT_BYTE(cinfo, c, return FALSE); - compptr->dc_tbl_no = (c >> 4) & 15; - compptr->ac_tbl_no = (c ) & 15; - - TRACEMS3(cinfo, 1, JTRC_SOS_COMPONENT, compptr->component_id, - compptr->dc_tbl_no, compptr->ac_tbl_no); - } - - /* Collect the additional scan parameters Ss, Se, Ah/Al. */ - INPUT_BYTE(cinfo, c, return FALSE); - cinfo->Ss = c; - INPUT_BYTE(cinfo, c, return FALSE); - cinfo->Se = c; - INPUT_BYTE(cinfo, c, return FALSE); - cinfo->Ah = (c >> 4) & 15; - cinfo->Al = (c ) & 15; - - TRACEMS4(cinfo, 1, JTRC_SOS_PARAMS, cinfo->Ss, cinfo->Se, - cinfo->Ah, cinfo->Al); - - /* Prepare to scan data & restart markers */ - cinfo->marker->next_restart_num = 0; - - /* Count another (non-pseudo) SOS marker */ - if (n) cinfo->input_scan_number++; - - INPUT_SYNC(cinfo); - return TRUE; -} - - -#ifdef D_ARITH_CODING_SUPPORTED - -LOCAL(boolean) -get_dac (j_decompress_ptr cinfo) -/* Process a DAC marker */ -{ - INT32 length; - int index, val; - INPUT_VARS(cinfo); - - INPUT_2BYTES(cinfo, length, return FALSE); - length -= 2; - - while (length > 0) { - INPUT_BYTE(cinfo, index, return FALSE); - INPUT_BYTE(cinfo, val, return FALSE); - - length -= 2; - - TRACEMS2(cinfo, 1, JTRC_DAC, index, val); - - if (index < 0 || index >= (2*NUM_ARITH_TBLS)) - ERREXIT1(cinfo, JERR_DAC_INDEX, index); - - if (index >= NUM_ARITH_TBLS) { /* define AC table */ - cinfo->arith_ac_K[index-NUM_ARITH_TBLS] = (UINT8) val; - } else { /* define DC table */ - cinfo->arith_dc_L[index] = (UINT8) (val & 0x0F); - cinfo->arith_dc_U[index] = (UINT8) (val >> 4); - if (cinfo->arith_dc_L[index] > cinfo->arith_dc_U[index]) - ERREXIT1(cinfo, JERR_DAC_VALUE, val); - } - } - - if (length != 0) - ERREXIT(cinfo, JERR_BAD_LENGTH); - - INPUT_SYNC(cinfo); - return TRUE; -} - -#else /* ! D_ARITH_CODING_SUPPORTED */ - -#define get_dac(cinfo) skip_variable(cinfo) - -#endif /* D_ARITH_CODING_SUPPORTED */ - - -LOCAL(boolean) -get_dht (j_decompress_ptr cinfo) -/* Process a DHT marker */ -{ - INT32 length; - UINT8 bits[17]; - UINT8 huffval[256]; - int i, index, count; - JHUFF_TBL **htblptr; - INPUT_VARS(cinfo); - - INPUT_2BYTES(cinfo, length, return FALSE); - length -= 2; - - while (length > 16) { - INPUT_BYTE(cinfo, index, return FALSE); - - TRACEMS1(cinfo, 1, JTRC_DHT, index); - - bits[0] = 0; - count = 0; - for (i = 1; i <= 16; i++) { - INPUT_BYTE(cinfo, bits[i], return FALSE); - count += bits[i]; - } - - length -= 1 + 16; - - TRACEMS8(cinfo, 2, JTRC_HUFFBITS, - bits[1], bits[2], bits[3], bits[4], - bits[5], bits[6], bits[7], bits[8]); - TRACEMS8(cinfo, 2, JTRC_HUFFBITS, - bits[9], bits[10], bits[11], bits[12], - bits[13], bits[14], bits[15], bits[16]); - - /* Here we just do minimal validation of the counts to avoid walking - * off the end of our table space. jdhuff.c will check more carefully. - */ - if (count > 256 || ((INT32) count) > length) - ERREXIT(cinfo, JERR_BAD_HUFF_TABLE); - - for (i = 0; i < count; i++) - INPUT_BYTE(cinfo, huffval[i], return FALSE); - - length -= count; - - if (index & 0x10) { /* AC table definition */ - index -= 0x10; - htblptr = &cinfo->ac_huff_tbl_ptrs[index]; - } else { /* DC table definition */ - htblptr = &cinfo->dc_huff_tbl_ptrs[index]; - } - - if (index < 0 || index >= NUM_HUFF_TBLS) - ERREXIT1(cinfo, JERR_DHT_INDEX, index); - - if (*htblptr == NULL) - *htblptr = jpeg_alloc_huff_table((j_common_ptr) cinfo); - - MEMCOPY((*htblptr)->bits, bits, SIZEOF((*htblptr)->bits)); - if (count > 0) - MEMCOPY((*htblptr)->huffval, huffval, count * SIZEOF(UINT8)); - } - - if (length != 0) - ERREXIT(cinfo, JERR_BAD_LENGTH); - - INPUT_SYNC(cinfo); - return TRUE; -} - - -LOCAL(boolean) -get_dqt (j_decompress_ptr cinfo) -/* Process a DQT marker */ -{ - INT32 length, count, i; - int n, prec; - unsigned int tmp; - JQUANT_TBL *quant_ptr; - const int *natural_order; - INPUT_VARS(cinfo); - - INPUT_2BYTES(cinfo, length, return FALSE); - length -= 2; - - while (length > 0) { - length--; - INPUT_BYTE(cinfo, n, return FALSE); - prec = n >> 4; - n &= 0x0F; - - TRACEMS2(cinfo, 1, JTRC_DQT, n, prec); - - if (n >= NUM_QUANT_TBLS) - ERREXIT1(cinfo, JERR_DQT_INDEX, n); - - if (cinfo->quant_tbl_ptrs[n] == NULL) - cinfo->quant_tbl_ptrs[n] = jpeg_alloc_quant_table((j_common_ptr) cinfo); - quant_ptr = cinfo->quant_tbl_ptrs[n]; - - if (prec) { - if (length < DCTSIZE2 * 2) { - /* Initialize full table for safety. */ - for (i = 0; i < DCTSIZE2; i++) { - quant_ptr->quantval[i] = 1; - } - count = length >> 1; - } else - count = DCTSIZE2; - } else { - if (length < DCTSIZE2) { - /* Initialize full table for safety. */ - for (i = 0; i < DCTSIZE2; i++) { - quant_ptr->quantval[i] = 1; - } - count = length; - } else - count = DCTSIZE2; - } - - switch ((int) count) { - case (2*2): natural_order = jpeg_natural_order2; break; - case (3*3): natural_order = jpeg_natural_order3; break; - case (4*4): natural_order = jpeg_natural_order4; break; - case (5*5): natural_order = jpeg_natural_order5; break; - case (6*6): natural_order = jpeg_natural_order6; break; - case (7*7): natural_order = jpeg_natural_order7; break; - default: natural_order = jpeg_natural_order; - } - - for (i = 0; i < count; i++) { - if (prec) - INPUT_2BYTES(cinfo, tmp, return FALSE); - else - INPUT_BYTE(cinfo, tmp, return FALSE); - /* We convert the zigzag-order table to natural array order. */ - quant_ptr->quantval[natural_order[i]] = (UINT16) tmp; - } - - if (cinfo->err->trace_level >= 2) { - for (i = 0; i < DCTSIZE2; i += 8) { - TRACEMS8(cinfo, 2, JTRC_QUANTVALS, - quant_ptr->quantval[i], quant_ptr->quantval[i+1], - quant_ptr->quantval[i+2], quant_ptr->quantval[i+3], - quant_ptr->quantval[i+4], quant_ptr->quantval[i+5], - quant_ptr->quantval[i+6], quant_ptr->quantval[i+7]); - } - } - - length -= count; - if (prec) length -= count; - } - - if (length != 0) - ERREXIT(cinfo, JERR_BAD_LENGTH); - - INPUT_SYNC(cinfo); - return TRUE; -} - - -LOCAL(boolean) -get_dri (j_decompress_ptr cinfo) -/* Process a DRI marker */ -{ - INT32 length; - unsigned int tmp; - INPUT_VARS(cinfo); - - INPUT_2BYTES(cinfo, length, return FALSE); - - if (length != 4) - ERREXIT(cinfo, JERR_BAD_LENGTH); - - INPUT_2BYTES(cinfo, tmp, return FALSE); - - TRACEMS1(cinfo, 1, JTRC_DRI, tmp); - - cinfo->restart_interval = tmp; - - INPUT_SYNC(cinfo); - return TRUE; -} - - -LOCAL(boolean) -get_lse (j_decompress_ptr cinfo) -/* Process an LSE marker */ -{ - INT32 length; - unsigned int tmp; - int cid; - INPUT_VARS(cinfo); - - if (! cinfo->marker->saw_SOF) - ERREXITS(cinfo, JERR_SOF_BEFORE, "LSE"); - - if (cinfo->num_components < 3) goto bad; - - INPUT_2BYTES(cinfo, length, return FALSE); - - if (length != 24) - ERREXIT(cinfo, JERR_BAD_LENGTH); - - INPUT_BYTE(cinfo, tmp, return FALSE); - if (tmp != 0x0D) /* ID inverse transform specification */ - ERREXIT1(cinfo, JERR_UNKNOWN_MARKER, cinfo->unread_marker); - INPUT_2BYTES(cinfo, tmp, return FALSE); - if (tmp != MAXJSAMPLE) goto bad; /* MAXTRANS */ - INPUT_BYTE(cinfo, tmp, return FALSE); - if (tmp != 3) goto bad; /* Nt=3 */ - INPUT_BYTE(cinfo, cid, return FALSE); - if (cid != cinfo->comp_info[1].component_id) goto bad; - INPUT_BYTE(cinfo, cid, return FALSE); - if (cid != cinfo->comp_info[0].component_id) goto bad; - INPUT_BYTE(cinfo, cid, return FALSE); - if (cid != cinfo->comp_info[2].component_id) goto bad; - INPUT_BYTE(cinfo, tmp, return FALSE); - if (tmp != 0x80) goto bad; /* F1: CENTER1=1, NORM1=0 */ - INPUT_2BYTES(cinfo, tmp, return FALSE); - if (tmp != 0) goto bad; /* A(1,1)=0 */ - INPUT_2BYTES(cinfo, tmp, return FALSE); - if (tmp != 0) goto bad; /* A(1,2)=0 */ - INPUT_BYTE(cinfo, tmp, return FALSE); - if (tmp != 0) goto bad; /* F2: CENTER2=0, NORM2=0 */ - INPUT_2BYTES(cinfo, tmp, return FALSE); - if (tmp != 1) goto bad; /* A(2,1)=1 */ - INPUT_2BYTES(cinfo, tmp, return FALSE); - if (tmp != 0) goto bad; /* A(2,2)=0 */ - INPUT_BYTE(cinfo, tmp, return FALSE); - if (tmp != 0) goto bad; /* F3: CENTER3=0, NORM3=0 */ - INPUT_2BYTES(cinfo, tmp, return FALSE); - if (tmp != 1) goto bad; /* A(3,1)=1 */ - INPUT_2BYTES(cinfo, tmp, return FALSE); - if (tmp != 0) { /* A(3,2)=0 */ - bad: - ERREXIT(cinfo, JERR_CONVERSION_NOTIMPL); - } - - /* OK, valid transform that we can handle. */ - cinfo->color_transform = JCT_SUBTRACT_GREEN; - - INPUT_SYNC(cinfo); - return TRUE; -} - - -/* - * Routines for processing APPn and COM markers. - * These are either saved in memory or discarded, per application request. - * APP0 and APP14 are specially checked to see if they are - * JFIF and Adobe markers, respectively. - */ - -#define APP0_DATA_LEN 14 /* Length of interesting data in APP0 */ -#define APP14_DATA_LEN 12 /* Length of interesting data in APP14 */ -#define APPN_DATA_LEN 14 /* Must be the largest of the above!! */ - - -LOCAL(void) -examine_app0 (j_decompress_ptr cinfo, JOCTET FAR * data, - unsigned int datalen, INT32 remaining) -/* Examine first few bytes from an APP0. - * Take appropriate action if it is a JFIF marker. - * datalen is # of bytes at data[], remaining is length of rest of marker data. - */ -{ - INT32 totallen = (INT32) datalen + remaining; - - if (datalen >= APP0_DATA_LEN && - GETJOCTET(data[0]) == 0x4A && - GETJOCTET(data[1]) == 0x46 && - GETJOCTET(data[2]) == 0x49 && - GETJOCTET(data[3]) == 0x46 && - GETJOCTET(data[4]) == 0) { - /* Found JFIF APP0 marker: save info */ - cinfo->saw_JFIF_marker = TRUE; - cinfo->JFIF_major_version = GETJOCTET(data[5]); - cinfo->JFIF_minor_version = GETJOCTET(data[6]); - cinfo->density_unit = GETJOCTET(data[7]); - cinfo->X_density = (GETJOCTET(data[8]) << 8) + GETJOCTET(data[9]); - cinfo->Y_density = (GETJOCTET(data[10]) << 8) + GETJOCTET(data[11]); - /* Check version. - * Major version must be 1 or 2, anything else signals an incompatible - * change. - * (We used to treat this as an error, but now it's a nonfatal warning, - * because some bozo at Hijaak couldn't read the spec.) - * Minor version should be 0..2, but process anyway if newer. - */ - if (cinfo->JFIF_major_version != 1 && cinfo->JFIF_major_version != 2) - WARNMS2(cinfo, JWRN_JFIF_MAJOR, - cinfo->JFIF_major_version, cinfo->JFIF_minor_version); - /* Generate trace messages */ - TRACEMS5(cinfo, 1, JTRC_JFIF, - cinfo->JFIF_major_version, cinfo->JFIF_minor_version, - cinfo->X_density, cinfo->Y_density, cinfo->density_unit); - /* Validate thumbnail dimensions and issue appropriate messages */ - if (GETJOCTET(data[12]) | GETJOCTET(data[13])) - TRACEMS2(cinfo, 1, JTRC_JFIF_THUMBNAIL, - GETJOCTET(data[12]), GETJOCTET(data[13])); - totallen -= APP0_DATA_LEN; - if (totallen != - ((INT32)GETJOCTET(data[12]) * (INT32)GETJOCTET(data[13]) * (INT32) 3)) - TRACEMS1(cinfo, 1, JTRC_JFIF_BADTHUMBNAILSIZE, (int) totallen); - } else if (datalen >= 6 && - GETJOCTET(data[0]) == 0x4A && - GETJOCTET(data[1]) == 0x46 && - GETJOCTET(data[2]) == 0x58 && - GETJOCTET(data[3]) == 0x58 && - GETJOCTET(data[4]) == 0) { - /* Found JFIF "JFXX" extension APP0 marker */ - /* The library doesn't actually do anything with these, - * but we try to produce a helpful trace message. - */ - switch (GETJOCTET(data[5])) { - case 0x10: - TRACEMS1(cinfo, 1, JTRC_THUMB_JPEG, (int) totallen); - break; - case 0x11: - TRACEMS1(cinfo, 1, JTRC_THUMB_PALETTE, (int) totallen); - break; - case 0x13: - TRACEMS1(cinfo, 1, JTRC_THUMB_RGB, (int) totallen); - break; - default: - TRACEMS2(cinfo, 1, JTRC_JFIF_EXTENSION, - GETJOCTET(data[5]), (int) totallen); - } - } else { - /* Start of APP0 does not match "JFIF" or "JFXX", or too short */ - TRACEMS1(cinfo, 1, JTRC_APP0, (int) totallen); - } -} - - -LOCAL(void) -examine_app14 (j_decompress_ptr cinfo, JOCTET FAR * data, - unsigned int datalen, INT32 remaining) -/* Examine first few bytes from an APP14. - * Take appropriate action if it is an Adobe marker. - * datalen is # of bytes at data[], remaining is length of rest of marker data. - */ -{ - unsigned int version, flags0, flags1, transform; - - if (datalen >= APP14_DATA_LEN && - GETJOCTET(data[0]) == 0x41 && - GETJOCTET(data[1]) == 0x64 && - GETJOCTET(data[2]) == 0x6F && - GETJOCTET(data[3]) == 0x62 && - GETJOCTET(data[4]) == 0x65) { - /* Found Adobe APP14 marker */ - version = (GETJOCTET(data[5]) << 8) + GETJOCTET(data[6]); - flags0 = (GETJOCTET(data[7]) << 8) + GETJOCTET(data[8]); - flags1 = (GETJOCTET(data[9]) << 8) + GETJOCTET(data[10]); - transform = GETJOCTET(data[11]); - TRACEMS4(cinfo, 1, JTRC_ADOBE, version, flags0, flags1, transform); - cinfo->saw_Adobe_marker = TRUE; - cinfo->Adobe_transform = (UINT8) transform; - } else { - /* Start of APP14 does not match "Adobe", or too short */ - TRACEMS1(cinfo, 1, JTRC_APP14, (int) (datalen + remaining)); - } -} - - -METHODDEF(boolean) -get_interesting_appn (j_decompress_ptr cinfo) -/* Process an APP0 or APP14 marker without saving it */ -{ - INT32 length; - JOCTET b[APPN_DATA_LEN]; - unsigned int i, numtoread; - INPUT_VARS(cinfo); - - INPUT_2BYTES(cinfo, length, return FALSE); - length -= 2; - - /* get the interesting part of the marker data */ - if (length >= APPN_DATA_LEN) - numtoread = APPN_DATA_LEN; - else if (length > 0) - numtoread = (unsigned int) length; - else - numtoread = 0; - for (i = 0; i < numtoread; i++) - INPUT_BYTE(cinfo, b[i], return FALSE); - length -= numtoread; - - /* process it */ - switch (cinfo->unread_marker) { - case M_APP0: - examine_app0(cinfo, (JOCTET FAR *) b, numtoread, length); - break; - case M_APP14: - examine_app14(cinfo, (JOCTET FAR *) b, numtoread, length); - break; - default: - /* can't get here unless jpeg_save_markers chooses wrong processor */ - ERREXIT1(cinfo, JERR_UNKNOWN_MARKER, cinfo->unread_marker); - } - - /* skip any remaining data -- could be lots */ - INPUT_SYNC(cinfo); - if (length > 0) - (*cinfo->src->skip_input_data) (cinfo, (long) length); - - return TRUE; -} - - -#ifdef SAVE_MARKERS_SUPPORTED - -METHODDEF(boolean) -save_marker (j_decompress_ptr cinfo) -/* Save an APPn or COM marker into the marker list */ -{ - my_marker_ptr marker = (my_marker_ptr) cinfo->marker; - jpeg_saved_marker_ptr cur_marker = marker->cur_marker; - unsigned int bytes_read, data_length; - JOCTET FAR * data; - INT32 length = 0; - INPUT_VARS(cinfo); - - if (cur_marker == NULL) { - /* begin reading a marker */ - INPUT_2BYTES(cinfo, length, return FALSE); - length -= 2; - if (length >= 0) { /* watch out for bogus length word */ - /* figure out how much we want to save */ - unsigned int limit; - if (cinfo->unread_marker == (int) M_COM) - limit = marker->length_limit_COM; - else - limit = marker->length_limit_APPn[cinfo->unread_marker - (int) M_APP0]; - if ((unsigned int) length < limit) - limit = (unsigned int) length; - /* allocate and initialize the marker item */ - cur_marker = (jpeg_saved_marker_ptr) - (*cinfo->mem->alloc_large) ((j_common_ptr) cinfo, JPOOL_IMAGE, - SIZEOF(struct jpeg_marker_struct) + limit); - cur_marker->next = NULL; - cur_marker->marker = (UINT8) cinfo->unread_marker; - cur_marker->original_length = (unsigned int) length; - cur_marker->data_length = limit; - /* data area is just beyond the jpeg_marker_struct */ - data = cur_marker->data = (JOCTET FAR *) (cur_marker + 1); - marker->cur_marker = cur_marker; - marker->bytes_read = 0; - bytes_read = 0; - data_length = limit; - } else { - /* deal with bogus length word */ - bytes_read = data_length = 0; - data = NULL; - } - } else { - /* resume reading a marker */ - bytes_read = marker->bytes_read; - data_length = cur_marker->data_length; - data = cur_marker->data + bytes_read; - } - - while (bytes_read < data_length) { - INPUT_SYNC(cinfo); /* move the restart point to here */ - marker->bytes_read = bytes_read; - /* If there's not at least one byte in buffer, suspend */ - MAKE_BYTE_AVAIL(cinfo, return FALSE); - /* Copy bytes with reasonable rapidity */ - while (bytes_read < data_length && bytes_in_buffer > 0) { - *data++ = *next_input_byte++; - bytes_in_buffer--; - bytes_read++; - } - } - - /* Done reading what we want to read */ - if (cur_marker != NULL) { /* will be NULL if bogus length word */ - /* Add new marker to end of list */ - if (cinfo->marker_list == NULL) { - cinfo->marker_list = cur_marker; - } else { - jpeg_saved_marker_ptr prev = cinfo->marker_list; - while (prev->next != NULL) - prev = prev->next; - prev->next = cur_marker; - } - /* Reset pointer & calc remaining data length */ - data = cur_marker->data; - length = cur_marker->original_length - data_length; - } - /* Reset to initial state for next marker */ - marker->cur_marker = NULL; - - /* Process the marker if interesting; else just make a generic trace msg */ - switch (cinfo->unread_marker) { - case M_APP0: - examine_app0(cinfo, data, data_length, length); - break; - case M_APP14: - examine_app14(cinfo, data, data_length, length); - break; - default: - TRACEMS2(cinfo, 1, JTRC_MISC_MARKER, cinfo->unread_marker, - (int) (data_length + length)); - } - - /* skip any remaining data -- could be lots */ - INPUT_SYNC(cinfo); /* do before skip_input_data */ - if (length > 0) - (*cinfo->src->skip_input_data) (cinfo, (long) length); - - return TRUE; -} - -#endif /* SAVE_MARKERS_SUPPORTED */ - - -METHODDEF(boolean) -skip_variable (j_decompress_ptr cinfo) -/* Skip over an unknown or uninteresting variable-length marker */ -{ - INT32 length; - INPUT_VARS(cinfo); - - INPUT_2BYTES(cinfo, length, return FALSE); - length -= 2; - - TRACEMS2(cinfo, 1, JTRC_MISC_MARKER, cinfo->unread_marker, (int) length); - - INPUT_SYNC(cinfo); /* do before skip_input_data */ - if (length > 0) - (*cinfo->src->skip_input_data) (cinfo, (long) length); - - return TRUE; -} - - -/* - * Find the next JPEG marker, save it in cinfo->unread_marker. - * Returns FALSE if had to suspend before reaching a marker; - * in that case cinfo->unread_marker is unchanged. - * - * Note that the result might not be a valid marker code, - * but it will never be 0 or FF. - */ - -LOCAL(boolean) -next_marker (j_decompress_ptr cinfo) -{ - int c; - INPUT_VARS(cinfo); - - for (;;) { - INPUT_BYTE(cinfo, c, return FALSE); - /* Skip any non-FF bytes. - * This may look a bit inefficient, but it will not occur in a valid file. - * We sync after each discarded byte so that a suspending data source - * can discard the byte from its buffer. - */ - while (c != 0xFF) { - cinfo->marker->discarded_bytes++; - INPUT_SYNC(cinfo); - INPUT_BYTE(cinfo, c, return FALSE); - } - /* This loop swallows any duplicate FF bytes. Extra FFs are legal as - * pad bytes, so don't count them in discarded_bytes. We assume there - * will not be so many consecutive FF bytes as to overflow a suspending - * data source's input buffer. - */ - do { - INPUT_BYTE(cinfo, c, return FALSE); - } while (c == 0xFF); - if (c != 0) - break; /* found a valid marker, exit loop */ - /* Reach here if we found a stuffed-zero data sequence (FF/00). - * Discard it and loop back to try again. - */ - cinfo->marker->discarded_bytes += 2; - INPUT_SYNC(cinfo); - } - - if (cinfo->marker->discarded_bytes != 0) { - WARNMS2(cinfo, JWRN_EXTRANEOUS_DATA, cinfo->marker->discarded_bytes, c); - cinfo->marker->discarded_bytes = 0; - } - - cinfo->unread_marker = c; - - INPUT_SYNC(cinfo); - return TRUE; -} - - -LOCAL(boolean) -first_marker (j_decompress_ptr cinfo) -/* Like next_marker, but used to obtain the initial SOI marker. */ -/* For this marker, we do not allow preceding garbage or fill; otherwise, - * we might well scan an entire input file before realizing it ain't JPEG. - * If an application wants to process non-JFIF files, it must seek to the - * SOI before calling the JPEG library. - */ -{ - int c, c2; - INPUT_VARS(cinfo); - - INPUT_BYTE(cinfo, c, return FALSE); - INPUT_BYTE(cinfo, c2, return FALSE); - if (c != 0xFF || c2 != (int) M_SOI) - ERREXIT2(cinfo, JERR_NO_SOI, c, c2); - - cinfo->unread_marker = c2; - - INPUT_SYNC(cinfo); - return TRUE; -} - - -/* - * Read markers until SOS or EOI. - * - * Returns same codes as are defined for jpeg_consume_input: - * JPEG_SUSPENDED, JPEG_REACHED_SOS, or JPEG_REACHED_EOI. - * - * Note: This function may return a pseudo SOS marker (with zero - * component number) for treat by input controller's consume_input. - * consume_input itself should filter out (skip) the pseudo marker - * after processing for the caller. - */ - -METHODDEF(int) -read_markers (j_decompress_ptr cinfo) -{ - /* Outer loop repeats once for each marker. */ - for (;;) { - /* Collect the marker proper, unless we already did. */ - /* NB: first_marker() enforces the requirement that SOI appear first. */ - if (cinfo->unread_marker == 0) { - if (! cinfo->marker->saw_SOI) { - if (! first_marker(cinfo)) - return JPEG_SUSPENDED; - } else { - if (! next_marker(cinfo)) - return JPEG_SUSPENDED; - } - } - /* At this point cinfo->unread_marker contains the marker code and the - * input point is just past the marker proper, but before any parameters. - * A suspension will cause us to return with this state still true. - */ - switch (cinfo->unread_marker) { - case M_SOI: - if (! get_soi(cinfo)) - return JPEG_SUSPENDED; - break; - - case M_SOF0: /* Baseline */ - if (! get_sof(cinfo, TRUE, FALSE, FALSE)) - return JPEG_SUSPENDED; - break; - - case M_SOF1: /* Extended sequential, Huffman */ - if (! get_sof(cinfo, FALSE, FALSE, FALSE)) - return JPEG_SUSPENDED; - break; - - case M_SOF2: /* Progressive, Huffman */ - if (! get_sof(cinfo, FALSE, TRUE, FALSE)) - return JPEG_SUSPENDED; - break; - - case M_SOF9: /* Extended sequential, arithmetic */ - if (! get_sof(cinfo, FALSE, FALSE, TRUE)) - return JPEG_SUSPENDED; - break; - - case M_SOF10: /* Progressive, arithmetic */ - if (! get_sof(cinfo, FALSE, TRUE, TRUE)) - return JPEG_SUSPENDED; - break; - - /* Currently unsupported SOFn types */ - case M_SOF3: /* Lossless, Huffman */ - case M_SOF5: /* Differential sequential, Huffman */ - case M_SOF6: /* Differential progressive, Huffman */ - case M_SOF7: /* Differential lossless, Huffman */ - case M_JPG: /* Reserved for JPEG extensions */ - case M_SOF11: /* Lossless, arithmetic */ - case M_SOF13: /* Differential sequential, arithmetic */ - case M_SOF14: /* Differential progressive, arithmetic */ - case M_SOF15: /* Differential lossless, arithmetic */ - ERREXIT1(cinfo, JERR_SOF_UNSUPPORTED, cinfo->unread_marker); - break; - - case M_SOS: - if (! get_sos(cinfo)) - return JPEG_SUSPENDED; - cinfo->unread_marker = 0; /* processed the marker */ - return JPEG_REACHED_SOS; - - case M_EOI: - TRACEMS(cinfo, 1, JTRC_EOI); - cinfo->unread_marker = 0; /* processed the marker */ - return JPEG_REACHED_EOI; - - case M_DAC: - if (! get_dac(cinfo)) - return JPEG_SUSPENDED; - break; - - case M_DHT: - if (! get_dht(cinfo)) - return JPEG_SUSPENDED; - break; - - case M_DQT: - if (! get_dqt(cinfo)) - return JPEG_SUSPENDED; - break; - - case M_DRI: - if (! get_dri(cinfo)) - return JPEG_SUSPENDED; - break; - - case M_JPG8: - if (! get_lse(cinfo)) - return JPEG_SUSPENDED; - break; - - case M_APP0: - case M_APP1: - case M_APP2: - case M_APP3: - case M_APP4: - case M_APP5: - case M_APP6: - case M_APP7: - case M_APP8: - case M_APP9: - case M_APP10: - case M_APP11: - case M_APP12: - case M_APP13: - case M_APP14: - case M_APP15: - if (! (*((my_marker_ptr) cinfo->marker)->process_APPn[ - cinfo->unread_marker - (int) M_APP0]) (cinfo)) - return JPEG_SUSPENDED; - break; - - case M_COM: - if (! (*((my_marker_ptr) cinfo->marker)->process_COM) (cinfo)) - return JPEG_SUSPENDED; - break; - - case M_RST0: /* these are all parameterless */ - case M_RST1: - case M_RST2: - case M_RST3: - case M_RST4: - case M_RST5: - case M_RST6: - case M_RST7: - case M_TEM: - TRACEMS1(cinfo, 1, JTRC_PARMLESS_MARKER, cinfo->unread_marker); - break; - - case M_DNL: /* Ignore DNL ... perhaps the wrong thing */ - if (! skip_variable(cinfo)) - return JPEG_SUSPENDED; - break; - - default: /* must be DHP, EXP, JPGn, or RESn */ - /* For now, we treat the reserved markers as fatal errors since they are - * likely to be used to signal incompatible JPEG Part 3 extensions. - * Once the JPEG 3 version-number marker is well defined, this code - * ought to change! - */ - ERREXIT1(cinfo, JERR_UNKNOWN_MARKER, cinfo->unread_marker); - } - /* Successfully processed marker, so reset state variable */ - cinfo->unread_marker = 0; - } /* end loop */ -} - - -/* - * Read a restart marker, which is expected to appear next in the datastream; - * if the marker is not there, take appropriate recovery action. - * Returns FALSE if suspension is required. - * - * This is called by the entropy decoder after it has read an appropriate - * number of MCUs. cinfo->unread_marker may be nonzero if the entropy decoder - * has already read a marker from the data source. Under normal conditions - * cinfo->unread_marker will be reset to 0 before returning; if not reset, - * it holds a marker which the decoder will be unable to read past. - */ - -METHODDEF(boolean) -read_restart_marker (j_decompress_ptr cinfo) -{ - /* Obtain a marker unless we already did. */ - /* Note that next_marker will complain if it skips any data. */ - if (cinfo->unread_marker == 0) { - if (! next_marker(cinfo)) - return FALSE; - } - - if (cinfo->unread_marker == - ((int) M_RST0 + cinfo->marker->next_restart_num)) { - /* Normal case --- swallow the marker and let entropy decoder continue */ - TRACEMS1(cinfo, 3, JTRC_RST, cinfo->marker->next_restart_num); - cinfo->unread_marker = 0; - } else { - /* Uh-oh, the restart markers have been messed up. */ - /* Let the data source manager determine how to resync. */ - if (! (*cinfo->src->resync_to_restart) (cinfo, - cinfo->marker->next_restart_num)) - return FALSE; - } - - /* Update next-restart state */ - cinfo->marker->next_restart_num = (cinfo->marker->next_restart_num + 1) & 7; - - return TRUE; -} - - -/* - * This is the default resync_to_restart method for data source managers - * to use if they don't have any better approach. Some data source managers - * may be able to back up, or may have additional knowledge about the data - * which permits a more intelligent recovery strategy; such managers would - * presumably supply their own resync method. - * - * read_restart_marker calls resync_to_restart if it finds a marker other than - * the restart marker it was expecting. (This code is *not* used unless - * a nonzero restart interval has been declared.) cinfo->unread_marker is - * the marker code actually found (might be anything, except 0 or FF). - * The desired restart marker number (0..7) is passed as a parameter. - * This routine is supposed to apply whatever error recovery strategy seems - * appropriate in order to position the input stream to the next data segment. - * Note that cinfo->unread_marker is treated as a marker appearing before - * the current data-source input point; usually it should be reset to zero - * before returning. - * Returns FALSE if suspension is required. - * - * This implementation is substantially constrained by wanting to treat the - * input as a data stream; this means we can't back up. Therefore, we have - * only the following actions to work with: - * 1. Simply discard the marker and let the entropy decoder resume at next - * byte of file. - * 2. Read forward until we find another marker, discarding intervening - * data. (In theory we could look ahead within the current bufferload, - * without having to discard data if we don't find the desired marker. - * This idea is not implemented here, in part because it makes behavior - * dependent on buffer size and chance buffer-boundary positions.) - * 3. Leave the marker unread (by failing to zero cinfo->unread_marker). - * This will cause the entropy decoder to process an empty data segment, - * inserting dummy zeroes, and then we will reprocess the marker. - * - * #2 is appropriate if we think the desired marker lies ahead, while #3 is - * appropriate if the found marker is a future restart marker (indicating - * that we have missed the desired restart marker, probably because it got - * corrupted). - * We apply #2 or #3 if the found marker is a restart marker no more than - * two counts behind or ahead of the expected one. We also apply #2 if the - * found marker is not a legal JPEG marker code (it's certainly bogus data). - * If the found marker is a restart marker more than 2 counts away, we do #1 - * (too much risk that the marker is erroneous; with luck we will be able to - * resync at some future point). - * For any valid non-restart JPEG marker, we apply #3. This keeps us from - * overrunning the end of a scan. An implementation limited to single-scan - * files might find it better to apply #2 for markers other than EOI, since - * any other marker would have to be bogus data in that case. - */ - -GLOBAL(boolean) -jpeg_resync_to_restart (j_decompress_ptr cinfo, int desired) -{ - int marker = cinfo->unread_marker; - int action = 1; - - /* Always put up a warning. */ - WARNMS2(cinfo, JWRN_MUST_RESYNC, marker, desired); - - /* Outer loop handles repeated decision after scanning forward. */ - for (;;) { - if (marker < (int) M_SOF0) - action = 2; /* invalid marker */ - else if (marker < (int) M_RST0 || marker > (int) M_RST7) - action = 3; /* valid non-restart marker */ - else { - if (marker == ((int) M_RST0 + ((desired+1) & 7)) || - marker == ((int) M_RST0 + ((desired+2) & 7))) - action = 3; /* one of the next two expected restarts */ - else if (marker == ((int) M_RST0 + ((desired-1) & 7)) || - marker == ((int) M_RST0 + ((desired-2) & 7))) - action = 2; /* a prior restart, so advance */ - else - action = 1; /* desired restart or too far away */ - } - TRACEMS2(cinfo, 4, JTRC_RECOVERY_ACTION, marker, action); - switch (action) { - case 1: - /* Discard marker and let entropy decoder resume processing. */ - cinfo->unread_marker = 0; - return TRUE; - case 2: - /* Scan to the next marker, and repeat the decision loop. */ - if (! next_marker(cinfo)) - return FALSE; - marker = cinfo->unread_marker; - break; - case 3: - /* Return without advancing past this marker. */ - /* Entropy decoder will be forced to process an empty segment. */ - return TRUE; - } - } /* end loop */ -} - - -/* - * Reset marker processing state to begin a fresh datastream. - */ - -METHODDEF(void) -reset_marker_reader (j_decompress_ptr cinfo) -{ - my_marker_ptr marker = (my_marker_ptr) cinfo->marker; - - cinfo->comp_info = NULL; /* until allocated by get_sof */ - cinfo->input_scan_number = 0; /* no SOS seen yet */ - cinfo->unread_marker = 0; /* no pending marker */ - marker->pub.saw_SOI = FALSE; /* set internal state too */ - marker->pub.saw_SOF = FALSE; - marker->pub.discarded_bytes = 0; - marker->cur_marker = NULL; -} - - -/* - * Initialize the marker reader module. - * This is called only once, when the decompression object is created. - */ - -GLOBAL(void) -jinit_marker_reader (j_decompress_ptr cinfo) -{ - my_marker_ptr marker; - int i; - - /* Create subobject in permanent pool */ - marker = (my_marker_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_PERMANENT, SIZEOF(my_marker_reader)); - cinfo->marker = &marker->pub; - /* Initialize public method pointers */ - marker->pub.reset_marker_reader = reset_marker_reader; - marker->pub.read_markers = read_markers; - marker->pub.read_restart_marker = read_restart_marker; - /* Initialize COM/APPn processing. - * By default, we examine and then discard APP0 and APP14, - * but simply discard COM and all other APPn. - */ - marker->process_COM = skip_variable; - marker->length_limit_COM = 0; - for (i = 0; i < 16; i++) { - marker->process_APPn[i] = skip_variable; - marker->length_limit_APPn[i] = 0; - } - marker->process_APPn[0] = get_interesting_appn; - marker->process_APPn[14] = get_interesting_appn; - /* Reset marker processing state */ - reset_marker_reader(cinfo); -} - - -/* - * Control saving of COM and APPn markers into marker_list. - */ - -#ifdef SAVE_MARKERS_SUPPORTED - -GLOBAL(void) -jpeg_save_markers (j_decompress_ptr cinfo, int marker_code, - unsigned int length_limit) -{ - my_marker_ptr marker = (my_marker_ptr) cinfo->marker; - long maxlength; - jpeg_marker_parser_method processor; - - /* Length limit mustn't be larger than what we can allocate - * (should only be a concern in a 16-bit environment). - */ - maxlength = cinfo->mem->max_alloc_chunk - SIZEOF(struct jpeg_marker_struct); - if (((long) length_limit) > maxlength) - length_limit = (unsigned int) maxlength; - - /* Choose processor routine to use. - * APP0/APP14 have special requirements. - */ - if (length_limit) { - processor = save_marker; - /* If saving APP0/APP14, save at least enough for our internal use. */ - if (marker_code == (int) M_APP0 && length_limit < APP0_DATA_LEN) - length_limit = APP0_DATA_LEN; - else if (marker_code == (int) M_APP14 && length_limit < APP14_DATA_LEN) - length_limit = APP14_DATA_LEN; - } else { - processor = skip_variable; - /* If discarding APP0/APP14, use our regular on-the-fly processor. */ - if (marker_code == (int) M_APP0 || marker_code == (int) M_APP14) - processor = get_interesting_appn; - } - - if (marker_code == (int) M_COM) { - marker->process_COM = processor; - marker->length_limit_COM = length_limit; - } else if (marker_code >= (int) M_APP0 && marker_code <= (int) M_APP15) { - marker->process_APPn[marker_code - (int) M_APP0] = processor; - marker->length_limit_APPn[marker_code - (int) M_APP0] = length_limit; - } else - ERREXIT1(cinfo, JERR_UNKNOWN_MARKER, marker_code); -} - -#endif /* SAVE_MARKERS_SUPPORTED */ - - -/* - * Install a special processing method for COM or APPn markers. - */ - -GLOBAL(void) -jpeg_set_marker_processor (j_decompress_ptr cinfo, int marker_code, - jpeg_marker_parser_method routine) -{ - my_marker_ptr marker = (my_marker_ptr) cinfo->marker; - - if (marker_code == (int) M_COM) - marker->process_COM = routine; - else if (marker_code >= (int) M_APP0 && marker_code <= (int) M_APP15) - marker->process_APPn[marker_code - (int) M_APP0] = routine; - else - ERREXIT1(cinfo, JERR_UNKNOWN_MARKER, marker_code); -} diff --git a/cmake/externals/libjpeg/jdmaster.c b/cmake/externals/libjpeg/jdmaster.c deleted file mode 100644 index 3070b7bb4..000000000 --- a/cmake/externals/libjpeg/jdmaster.c +++ /dev/null @@ -1,532 +0,0 @@ -/* - * jdmaster.c - * - * Copyright (C) 1991-1997, Thomas G. Lane. - * Modified 2002-2020 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains master control logic for the JPEG decompressor. - * These routines are concerned with selecting the modules to be executed - * and with determining the number of passes and the work to be done in each - * pass. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* Private state */ - -typedef struct { - struct jpeg_decomp_master pub; /* public fields */ - - int pass_number; /* # of passes completed */ - - boolean using_merged_upsample; /* TRUE if using merged upsample/cconvert */ - - /* Saved references to initialized quantizer modules, - * in case we need to switch modes. - */ - struct jpeg_color_quantizer * quantizer_1pass; - struct jpeg_color_quantizer * quantizer_2pass; -} my_decomp_master; - -typedef my_decomp_master * my_master_ptr; - - -/* - * Determine whether merged upsample/color conversion should be used. - * CRUCIAL: this must match the actual capabilities of jdmerge.c! - */ - -LOCAL(boolean) -use_merged_upsample (j_decompress_ptr cinfo) -{ -#ifdef UPSAMPLE_MERGING_SUPPORTED - /* Merging is the equivalent of plain box-filter upsampling. */ - /* The following condition is only needed if fancy shall select - * a different upsampling method. In our current implementation - * fancy only affects the DCT scaling, thus we can use fancy - * upsampling and merged upsample simultaneously, in particular - * with scaled DCT sizes larger than the default DCTSIZE. - */ -#if 0 - if (cinfo->do_fancy_upsampling) - return FALSE; -#endif - if (cinfo->CCIR601_sampling) - return FALSE; - /* jdmerge.c only supports YCC=>RGB color conversion */ - if ((cinfo->jpeg_color_space != JCS_YCbCr && - cinfo->jpeg_color_space != JCS_BG_YCC) || - cinfo->num_components != 3 || - cinfo->out_color_space != JCS_RGB || - cinfo->out_color_components != RGB_PIXELSIZE || - cinfo->color_transform) - return FALSE; - /* and it only handles 2h1v or 2h2v sampling ratios */ - if (cinfo->comp_info[0].h_samp_factor != 2 || - cinfo->comp_info[1].h_samp_factor != 1 || - cinfo->comp_info[2].h_samp_factor != 1 || - cinfo->comp_info[0].v_samp_factor > 2 || - cinfo->comp_info[1].v_samp_factor != 1 || - cinfo->comp_info[2].v_samp_factor != 1) - return FALSE; - /* furthermore, it doesn't work if we've scaled the IDCTs differently */ - if (cinfo->comp_info[0].DCT_h_scaled_size != cinfo->min_DCT_h_scaled_size || - cinfo->comp_info[1].DCT_h_scaled_size != cinfo->min_DCT_h_scaled_size || - cinfo->comp_info[2].DCT_h_scaled_size != cinfo->min_DCT_h_scaled_size || - cinfo->comp_info[0].DCT_v_scaled_size != cinfo->min_DCT_v_scaled_size || - cinfo->comp_info[1].DCT_v_scaled_size != cinfo->min_DCT_v_scaled_size || - cinfo->comp_info[2].DCT_v_scaled_size != cinfo->min_DCT_v_scaled_size) - return FALSE; - /* ??? also need to test for upsample-time rescaling, when & if supported */ - return TRUE; /* by golly, it'll work... */ -#else - return FALSE; -#endif -} - - -/* - * Compute output image dimensions and related values. - * NOTE: this is exported for possible use by application. - * Hence it mustn't do anything that can't be done twice. - * Also note that it may be called before the master module is initialized! - */ - -GLOBAL(void) -jpeg_calc_output_dimensions (j_decompress_ptr cinfo) -/* Do computations that are needed before master selection phase. - * This function is used for full decompression. - */ -{ - int ci, i; - jpeg_component_info *compptr; - - /* Prevent application from calling me at wrong times */ - if (cinfo->global_state != DSTATE_READY) - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - - /* Compute core output image dimensions and DCT scaling choices. */ - jpeg_core_output_dimensions(cinfo); - -#ifdef IDCT_SCALING_SUPPORTED - - /* In selecting the actual DCT scaling for each component, we try to - * scale up the chroma components via IDCT scaling rather than upsampling. - * This saves time if the upsampler gets to use 1:1 scaling. - * Note this code adapts subsampling ratios which are powers of 2. - */ - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - int ssize = 1; - if (! cinfo->raw_data_out) - while (cinfo->min_DCT_h_scaled_size * ssize <= - (cinfo->do_fancy_upsampling ? DCTSIZE : DCTSIZE / 2) && - (cinfo->max_h_samp_factor % (compptr->h_samp_factor * ssize * 2)) == - 0) { - ssize = ssize * 2; - } - compptr->DCT_h_scaled_size = cinfo->min_DCT_h_scaled_size * ssize; - ssize = 1; - if (! cinfo->raw_data_out) - while (cinfo->min_DCT_v_scaled_size * ssize <= - (cinfo->do_fancy_upsampling ? DCTSIZE : DCTSIZE / 2) && - (cinfo->max_v_samp_factor % (compptr->v_samp_factor * ssize * 2)) == - 0) { - ssize = ssize * 2; - } - compptr->DCT_v_scaled_size = cinfo->min_DCT_v_scaled_size * ssize; - - /* We don't support IDCT ratios larger than 2. */ - if (compptr->DCT_h_scaled_size > compptr->DCT_v_scaled_size * 2) - compptr->DCT_h_scaled_size = compptr->DCT_v_scaled_size * 2; - else if (compptr->DCT_v_scaled_size > compptr->DCT_h_scaled_size * 2) - compptr->DCT_v_scaled_size = compptr->DCT_h_scaled_size * 2; - - /* Recompute downsampled dimensions of components; - * application needs to know these if using raw downsampled data. - */ - /* Size in samples, after IDCT scaling */ - compptr->downsampled_width = (JDIMENSION) - jdiv_round_up((long) cinfo->image_width * - (long) (compptr->h_samp_factor * compptr->DCT_h_scaled_size), - (long) (cinfo->max_h_samp_factor * cinfo->block_size)); - compptr->downsampled_height = (JDIMENSION) - jdiv_round_up((long) cinfo->image_height * - (long) (compptr->v_samp_factor * compptr->DCT_v_scaled_size), - (long) (cinfo->max_v_samp_factor * cinfo->block_size)); - } - -#endif /* IDCT_SCALING_SUPPORTED */ - - /* Report number of components in selected colorspace. */ - /* This should correspond to the actual code in the color conversion module. */ - switch (cinfo->out_color_space) { - case JCS_GRAYSCALE: - cinfo->out_color_components = 1; - break; - case JCS_RGB: - case JCS_BG_RGB: - cinfo->out_color_components = RGB_PIXELSIZE; - break; - default: /* YCCK <=> CMYK conversion or same colorspace as in file */ - i = 0; - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) - if (compptr->component_needed) - i++; /* count output color components */ - cinfo->out_color_components = i; - } - cinfo->output_components = (cinfo->quantize_colors ? 1 : - cinfo->out_color_components); - - /* See if upsampler will want to emit more than one row at a time */ - if (use_merged_upsample(cinfo)) - cinfo->rec_outbuf_height = cinfo->max_v_samp_factor; - else - cinfo->rec_outbuf_height = 1; -} - - -/* - * Several decompression processes need to range-limit values to the range - * 0..MAXJSAMPLE; the input value may fall somewhat outside this range - * due to noise introduced by quantization, roundoff error, etc. These - * processes are inner loops and need to be as fast as possible. On most - * machines, particularly CPUs with pipelines or instruction prefetch, - * a (subscript-check-less) C table lookup - * x = sample_range_limit[x]; - * is faster than explicit tests - * if (x < 0) x = 0; - * else if (x > MAXJSAMPLE) x = MAXJSAMPLE; - * These processes all use a common table prepared by the routine below. - * - * For most steps we can mathematically guarantee that the initial value - * of x is within 2*(MAXJSAMPLE+1) of the legal range, so a table running - * from -2*(MAXJSAMPLE+1) to 3*MAXJSAMPLE+2 is sufficient. But for the - * initial limiting step (just after the IDCT), a wildly out-of-range value - * is possible if the input data is corrupt. To avoid any chance of indexing - * off the end of memory and getting a bad-pointer trap, we perform the - * post-IDCT limiting thus: - * x = (sample_range_limit - SUBSET)[(x + CENTER) & MASK]; - * where MASK is 2 bits wider than legal sample data, ie 10 bits for 8-bit - * samples. Under normal circumstances this is more than enough range and - * a correct output will be generated; with bogus input data the mask will - * cause wraparound, and we will safely generate a bogus-but-in-range output. - * For the post-IDCT step, we want to convert the data from signed to unsigned - * representation by adding CENTERJSAMPLE at the same time that we limit it. - * This is accomplished with SUBSET = CENTER - CENTERJSAMPLE. - * - * Note that the table is allocated in near data space on PCs; it's small - * enough and used often enough to justify this. - */ - -LOCAL(void) -prepare_range_limit_table (j_decompress_ptr cinfo) -/* Allocate and fill in the sample_range_limit table */ -{ - JSAMPLE * table; - int i; - - table = (JSAMPLE *) (*cinfo->mem->alloc_small) ((j_common_ptr) cinfo, - JPOOL_IMAGE, (RANGE_CENTER * 2 + MAXJSAMPLE + 1) * SIZEOF(JSAMPLE)); - /* First segment of range limit table: limit[x] = 0 for x < 0 */ - MEMZERO(table, RANGE_CENTER * SIZEOF(JSAMPLE)); - table += RANGE_CENTER; /* allow negative subscripts of table */ - cinfo->sample_range_limit = table; - /* Main part of range limit table: limit[x] = x */ - for (i = 0; i <= MAXJSAMPLE; i++) - table[i] = (JSAMPLE) i; - /* End of range limit table: limit[x] = MAXJSAMPLE for x > MAXJSAMPLE */ - for (; i <= MAXJSAMPLE + RANGE_CENTER; i++) - table[i] = MAXJSAMPLE; -} - - -/* - * Master selection of decompression modules. - * This is done once at jpeg_start_decompress time. We determine - * which modules will be used and give them appropriate initialization calls. - * We also initialize the decompressor input side to begin consuming data. - * - * Since jpeg_read_header has finished, we know what is in the SOF - * and (first) SOS markers. We also have all the application parameter - * settings. - */ - -LOCAL(void) -master_selection (j_decompress_ptr cinfo) -{ - my_master_ptr master = (my_master_ptr) cinfo->master; - boolean use_c_buffer; - long samplesperrow; - JDIMENSION jd_samplesperrow; - - /* For now, precision must match compiled-in value... */ - if (cinfo->data_precision != BITS_IN_JSAMPLE) - ERREXIT1(cinfo, JERR_BAD_PRECISION, cinfo->data_precision); - - /* Initialize dimensions and other stuff */ - jpeg_calc_output_dimensions(cinfo); - prepare_range_limit_table(cinfo); - - /* Sanity check on image dimensions */ - if (cinfo->output_height <= 0 || cinfo->output_width <= 0 || - cinfo->out_color_components <= 0) - ERREXIT(cinfo, JERR_EMPTY_IMAGE); - - /* Width of an output scanline must be representable as JDIMENSION. */ - samplesperrow = (long) cinfo->output_width * (long) cinfo->out_color_components; - jd_samplesperrow = (JDIMENSION) samplesperrow; - if ((long) jd_samplesperrow != samplesperrow) - ERREXIT(cinfo, JERR_WIDTH_OVERFLOW); - - /* Initialize my private state */ - master->pass_number = 0; - master->using_merged_upsample = use_merged_upsample(cinfo); - - /* Color quantizer selection */ - master->quantizer_1pass = NULL; - master->quantizer_2pass = NULL; - /* No mode changes if not using buffered-image mode. */ - if (! cinfo->quantize_colors || ! cinfo->buffered_image) { - cinfo->enable_1pass_quant = FALSE; - cinfo->enable_external_quant = FALSE; - cinfo->enable_2pass_quant = FALSE; - } - if (cinfo->quantize_colors) { - if (cinfo->raw_data_out) - ERREXIT(cinfo, JERR_NOTIMPL); - /* 2-pass quantizer only works in 3-component color space. */ - if (cinfo->out_color_components != 3) { - cinfo->enable_1pass_quant = TRUE; - cinfo->enable_external_quant = FALSE; - cinfo->enable_2pass_quant = FALSE; - cinfo->colormap = NULL; - } else if (cinfo->colormap != NULL) { - cinfo->enable_external_quant = TRUE; - } else if (cinfo->two_pass_quantize) { - cinfo->enable_2pass_quant = TRUE; - } else { - cinfo->enable_1pass_quant = TRUE; - } - - if (cinfo->enable_1pass_quant) { -#ifdef QUANT_1PASS_SUPPORTED - jinit_1pass_quantizer(cinfo); - master->quantizer_1pass = cinfo->cquantize; -#else - ERREXIT(cinfo, JERR_NOT_COMPILED); -#endif - } - - /* We use the 2-pass code to map to external colormaps. */ - if (cinfo->enable_2pass_quant || cinfo->enable_external_quant) { -#ifdef QUANT_2PASS_SUPPORTED - jinit_2pass_quantizer(cinfo); - master->quantizer_2pass = cinfo->cquantize; -#else - ERREXIT(cinfo, JERR_NOT_COMPILED); -#endif - } - /* If both quantizers are initialized, the 2-pass one is left active; - * this is necessary for starting with quantization to an external map. - */ - } - - /* Post-processing: in particular, color conversion first */ - if (! cinfo->raw_data_out) { - if (master->using_merged_upsample) { -#ifdef UPSAMPLE_MERGING_SUPPORTED - jinit_merged_upsampler(cinfo); /* does color conversion too */ -#else - ERREXIT(cinfo, JERR_NOT_COMPILED); -#endif - } else { - jinit_color_deconverter(cinfo); - jinit_upsampler(cinfo); - } - jinit_d_post_controller(cinfo, cinfo->enable_2pass_quant); - } - /* Inverse DCT */ - jinit_inverse_dct(cinfo); - /* Entropy decoding: either Huffman or arithmetic coding. */ - if (cinfo->arith_code) - jinit_arith_decoder(cinfo); - else { - jinit_huff_decoder(cinfo); - } - - /* Initialize principal buffer controllers. */ - use_c_buffer = cinfo->inputctl->has_multiple_scans || cinfo->buffered_image; - jinit_d_coef_controller(cinfo, use_c_buffer); - - if (! cinfo->raw_data_out) - jinit_d_main_controller(cinfo, FALSE /* never need full buffer here */); - - /* We can now tell the memory manager to allocate virtual arrays. */ - (*cinfo->mem->realize_virt_arrays) ((j_common_ptr) cinfo); - - /* Initialize input side of decompressor to consume first scan. */ - (*cinfo->inputctl->start_input_pass) (cinfo); - -#ifdef D_MULTISCAN_FILES_SUPPORTED - /* If jpeg_start_decompress will read the whole file, initialize - * progress monitoring appropriately. The input step is counted - * as one pass. - */ - if (cinfo->progress != NULL && ! cinfo->buffered_image && - cinfo->inputctl->has_multiple_scans) { - int nscans; - /* Estimate number of scans to set pass_limit. */ - if (cinfo->progressive_mode) { - /* Arbitrarily estimate 2 interleaved DC scans + 3 AC scans/component. */ - nscans = 2 + 3 * cinfo->num_components; - } else { - /* For a nonprogressive multiscan file, estimate 1 scan per component. */ - nscans = cinfo->num_components; - } - cinfo->progress->pass_counter = 0L; - cinfo->progress->pass_limit = (long) cinfo->total_iMCU_rows * nscans; - cinfo->progress->completed_passes = 0; - cinfo->progress->total_passes = (cinfo->enable_2pass_quant ? 3 : 2); - /* Count the input pass as done */ - master->pass_number++; - } -#endif /* D_MULTISCAN_FILES_SUPPORTED */ -} - - -/* - * Per-pass setup. - * This is called at the beginning of each output pass. We determine which - * modules will be active during this pass and give them appropriate - * start_pass calls. We also set is_dummy_pass to indicate whether this - * is a "real" output pass or a dummy pass for color quantization. - * (In the latter case, jdapistd.c will crank the pass to completion.) - */ - -METHODDEF(void) -prepare_for_output_pass (j_decompress_ptr cinfo) -{ - my_master_ptr master = (my_master_ptr) cinfo->master; - - if (master->pub.is_dummy_pass) { -#ifdef QUANT_2PASS_SUPPORTED - /* Final pass of 2-pass quantization */ - master->pub.is_dummy_pass = FALSE; - (*cinfo->cquantize->start_pass) (cinfo, FALSE); - (*cinfo->post->start_pass) (cinfo, JBUF_CRANK_DEST); - (*cinfo->main->start_pass) (cinfo, JBUF_CRANK_DEST); -#else - ERREXIT(cinfo, JERR_NOT_COMPILED); -#endif /* QUANT_2PASS_SUPPORTED */ - } else { - if (cinfo->quantize_colors && cinfo->colormap == NULL) { - /* Select new quantization method */ - if (cinfo->two_pass_quantize && cinfo->enable_2pass_quant) { - cinfo->cquantize = master->quantizer_2pass; - master->pub.is_dummy_pass = TRUE; - } else if (cinfo->enable_1pass_quant) { - cinfo->cquantize = master->quantizer_1pass; - } else { - ERREXIT(cinfo, JERR_MODE_CHANGE); - } - } - (*cinfo->idct->start_pass) (cinfo); - (*cinfo->coef->start_output_pass) (cinfo); - if (! cinfo->raw_data_out) { - if (! master->using_merged_upsample) - (*cinfo->cconvert->start_pass) (cinfo); - (*cinfo->upsample->start_pass) (cinfo); - if (cinfo->quantize_colors) - (*cinfo->cquantize->start_pass) (cinfo, master->pub.is_dummy_pass); - (*cinfo->post->start_pass) (cinfo, - (master->pub.is_dummy_pass ? JBUF_SAVE_AND_PASS : JBUF_PASS_THRU)); - (*cinfo->main->start_pass) (cinfo, JBUF_PASS_THRU); - } - } - - /* Set up progress monitor's pass info if present */ - if (cinfo->progress != NULL) { - cinfo->progress->completed_passes = master->pass_number; - cinfo->progress->total_passes = master->pass_number + - (master->pub.is_dummy_pass ? 2 : 1); - /* In buffered-image mode, we assume one more output pass if EOI not - * yet reached, but no more passes if EOI has been reached. - */ - if (cinfo->buffered_image && ! cinfo->inputctl->eoi_reached) { - cinfo->progress->total_passes += (cinfo->enable_2pass_quant ? 2 : 1); - } - } -} - - -/* - * Finish up at end of an output pass. - */ - -METHODDEF(void) -finish_output_pass (j_decompress_ptr cinfo) -{ - my_master_ptr master = (my_master_ptr) cinfo->master; - - if (cinfo->quantize_colors) - (*cinfo->cquantize->finish_pass) (cinfo); - master->pass_number++; -} - - -#ifdef D_MULTISCAN_FILES_SUPPORTED - -/* - * Switch to a new external colormap between output passes. - */ - -GLOBAL(void) -jpeg_new_colormap (j_decompress_ptr cinfo) -{ - my_master_ptr master = (my_master_ptr) cinfo->master; - - /* Prevent application from calling me at wrong times */ - if (cinfo->global_state != DSTATE_BUFIMAGE) - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - - if (cinfo->quantize_colors && cinfo->enable_external_quant && - cinfo->colormap != NULL) { - /* Select 2-pass quantizer for external colormap use */ - cinfo->cquantize = master->quantizer_2pass; - /* Notify quantizer of colormap change */ - (*cinfo->cquantize->new_color_map) (cinfo); - master->pub.is_dummy_pass = FALSE; /* just in case */ - } else - ERREXIT(cinfo, JERR_MODE_CHANGE); -} - -#endif /* D_MULTISCAN_FILES_SUPPORTED */ - - -/* - * Initialize master decompression control and select active modules. - * This is performed at the start of jpeg_start_decompress. - */ - -GLOBAL(void) -jinit_master_decompress (j_decompress_ptr cinfo) -{ - my_master_ptr master; - - master = (my_master_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_decomp_master)); - cinfo->master = &master->pub; - master->pub.prepare_for_output_pass = prepare_for_output_pass; - master->pub.finish_output_pass = finish_output_pass; - - master->pub.is_dummy_pass = FALSE; - - master_selection(cinfo); -} diff --git a/cmake/externals/libjpeg/jdmerge.c b/cmake/externals/libjpeg/jdmerge.c deleted file mode 100644 index 0d16821be..000000000 --- a/cmake/externals/libjpeg/jdmerge.c +++ /dev/null @@ -1,437 +0,0 @@ -/* - * jdmerge.c - * - * Copyright (C) 1994-1996, Thomas G. Lane. - * Modified 2013-2022 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains code for merged upsampling/color conversion. - * - * This file combines functions from jdsample.c and jdcolor.c; - * read those files first to understand what's going on. - * - * When the chroma components are to be upsampled by simple replication - * (ie, box filtering), we can save some work in color conversion by - * calculating all the output pixels corresponding to a pair of chroma - * samples at one time. In the conversion equations - * R = Y + K1 * Cr - * G = Y + K2 * Cb + K3 * Cr - * B = Y + K4 * Cb - * only the Y term varies among the group of pixels corresponding to a pair - * of chroma samples, so the rest of the terms can be calculated just once. - * At typical sampling ratios, this eliminates half or three-quarters - * of the multiplications needed for color conversion. - * - * This file currently provides implementations for the following cases: - * YCC => RGB color conversion only (YCbCr or BG_YCC). - * Sampling ratios of 2h1v or 2h2v. - * No scaling needed at upsample time. - * Corner-aligned (non-CCIR601) sampling alignment. - * Other special cases could be added, but in most applications these - * are the only common cases. (For uncommon cases we fall back on - * the more general code in jdsample.c and jdcolor.c.) - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - -#ifdef UPSAMPLE_MERGING_SUPPORTED - - -#if RANGE_BITS < 2 - /* Deliberate syntax err */ - Sorry, this code requires 2 or more range extension bits. -#endif - - -/* Private subobject */ - -typedef struct { - struct jpeg_upsampler pub; /* public fields */ - - /* Pointer to routine to do actual upsampling/conversion of one row group */ - JMETHOD(void, upmethod, (j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION in_row_group_ctr, - JSAMPARRAY output_buf)); - - /* Private state for YCC->RGB conversion */ - int * Cr_r_tab; /* => table for Cr to R conversion */ - int * Cb_b_tab; /* => table for Cb to B conversion */ - INT32 * Cr_g_tab; /* => table for Cr to G conversion */ - INT32 * Cb_g_tab; /* => table for Cb to G conversion */ - - /* For 2:1 vertical sampling, we produce two output rows at a time. - * We need a "spare" row buffer to hold the second output row if the - * application provides just a one-row buffer; we also use the spare - * to discard the dummy last row if the image height is odd. - */ - JSAMPROW spare_row; - boolean spare_full; /* T if spare buffer is occupied */ - - JDIMENSION out_row_width; /* samples per output row */ - JDIMENSION rows_to_go; /* counts rows remaining in image */ -} my_upsampler; - -typedef my_upsampler * my_upsample_ptr; - -#define SCALEBITS 16 /* speediest right-shift on some machines */ -#define ONE_HALF ((INT32) 1 << (SCALEBITS-1)) -#define FIX(x) ((INT32) ((x) * (1L<RGB and BG_YCC->RGB colorspace conversion. - * This is taken directly from jdcolor.c; see that file for more info. - */ - -LOCAL(void) -build_ycc_rgb_table (j_decompress_ptr cinfo) -/* Normal case, sYCC */ -{ - my_upsample_ptr upsample = (my_upsample_ptr) cinfo->upsample; - int i; - INT32 x; - SHIFT_TEMPS - - upsample->Cr_r_tab = (int *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(int)); - upsample->Cb_b_tab = (int *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(int)); - upsample->Cr_g_tab = (INT32 *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(INT32)); - upsample->Cb_g_tab = (INT32 *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(INT32)); - - for (i = 0, x = -CENTERJSAMPLE; i <= MAXJSAMPLE; i++, x++) { - /* i is the actual input pixel value, in the range 0..MAXJSAMPLE */ - /* The Cb or Cr value we are thinking of is x = i - CENTERJSAMPLE */ - /* Cr=>R value is nearest int to 1.402 * x */ - upsample->Cr_r_tab[i] = (int) DESCALE(FIX(1.402) * x, SCALEBITS); - /* Cb=>B value is nearest int to 1.772 * x */ - upsample->Cb_b_tab[i] = (int) DESCALE(FIX(1.772) * x, SCALEBITS); - /* Cr=>G value is scaled-up -0.714136286 * x */ - upsample->Cr_g_tab[i] = (- FIX(0.714136286)) * x; - /* Cb=>G value is scaled-up -0.344136286 * x */ - /* We also add in ONE_HALF so that need not do it in inner loop */ - upsample->Cb_g_tab[i] = (- FIX(0.344136286)) * x + ONE_HALF; - } -} - - -LOCAL(void) -build_bg_ycc_rgb_table (j_decompress_ptr cinfo) -/* Wide gamut case, bg-sYCC */ -{ - my_upsample_ptr upsample = (my_upsample_ptr) cinfo->upsample; - int i; - INT32 x; - SHIFT_TEMPS - - upsample->Cr_r_tab = (int *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(int)); - upsample->Cb_b_tab = (int *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(int)); - upsample->Cr_g_tab = (INT32 *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(INT32)); - upsample->Cb_g_tab = (INT32 *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE+1) * SIZEOF(INT32)); - - for (i = 0, x = -CENTERJSAMPLE; i <= MAXJSAMPLE; i++, x++) { - /* i is the actual input pixel value, in the range 0..MAXJSAMPLE */ - /* The Cb or Cr value we are thinking of is x = i - CENTERJSAMPLE */ - /* Cr=>R value is nearest int to 2.804 * x */ - upsample->Cr_r_tab[i] = (int) DESCALE(FIX(2.804) * x, SCALEBITS); - /* Cb=>B value is nearest int to 3.544 * x */ - upsample->Cb_b_tab[i] = (int) DESCALE(FIX(3.544) * x, SCALEBITS); - /* Cr=>G value is scaled-up -1.428272572 * x */ - upsample->Cr_g_tab[i] = (- FIX(1.428272572)) * x; - /* Cb=>G value is scaled-up -0.688272572 * x */ - /* We also add in ONE_HALF so that need not do it in inner loop */ - upsample->Cb_g_tab[i] = (- FIX(0.688272572)) * x + ONE_HALF; - } -} - - -/* - * Initialize for an upsampling pass. - */ - -METHODDEF(void) -start_pass_merged_upsample (j_decompress_ptr cinfo) -{ - my_upsample_ptr upsample = (my_upsample_ptr) cinfo->upsample; - - /* Mark the spare buffer empty */ - upsample->spare_full = FALSE; - /* Initialize total-height counter for detecting bottom of image */ - upsample->rows_to_go = cinfo->output_height; -} - - -/* - * Control routine to do upsampling (and color conversion). - * - * The control routine just handles the row buffering considerations. - */ - -METHODDEF(void) -merged_2v_upsample (j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION *in_row_group_ctr, - JDIMENSION in_row_groups_avail, - JSAMPARRAY output_buf, JDIMENSION *out_row_ctr, - JDIMENSION out_rows_avail) -/* 2:1 vertical sampling case: may need a spare row. */ -{ - my_upsample_ptr upsample = (my_upsample_ptr) cinfo->upsample; - JSAMPROW work_ptrs[2]; - JDIMENSION num_rows; /* number of rows returned to caller */ - - if (upsample->spare_full) { - /* If we have a spare row saved from a previous cycle, just return it. */ - jcopy_sample_rows(& upsample->spare_row, output_buf + *out_row_ctr, - 1, upsample->out_row_width); - num_rows = 1; - upsample->spare_full = FALSE; - } else { - /* Figure number of rows to return to caller. */ - num_rows = 2; - /* Not more than the distance to the end of the image. */ - if (num_rows > upsample->rows_to_go) - num_rows = upsample->rows_to_go; - /* And not more than what the client can accept: */ - out_rows_avail -= *out_row_ctr; - if (num_rows > out_rows_avail) - num_rows = out_rows_avail; - /* Create output pointer array for upsampler. */ - work_ptrs[0] = output_buf[*out_row_ctr]; - if (num_rows > 1) { - work_ptrs[1] = output_buf[*out_row_ctr + 1]; - } else { - work_ptrs[1] = upsample->spare_row; - upsample->spare_full = TRUE; - } - /* Now do the upsampling. */ - (*upsample->upmethod) (cinfo, input_buf, *in_row_group_ctr, work_ptrs); - } - - /* Adjust counts */ - *out_row_ctr += num_rows; - upsample->rows_to_go -= num_rows; - /* When the buffer is emptied, declare this input row group consumed */ - if (! upsample->spare_full) - (*in_row_group_ctr)++; -} - - -METHODDEF(void) -merged_1v_upsample (j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION *in_row_group_ctr, - JDIMENSION in_row_groups_avail, - JSAMPARRAY output_buf, JDIMENSION *out_row_ctr, - JDIMENSION out_rows_avail) -/* 1:1 vertical sampling case: much easier, never need a spare row. */ -{ - my_upsample_ptr upsample = (my_upsample_ptr) cinfo->upsample; - - /* Just do the upsampling. */ - (*upsample->upmethod) (cinfo, input_buf, *in_row_group_ctr, - output_buf + *out_row_ctr); - /* Adjust counts */ - (*out_row_ctr)++; - (*in_row_group_ctr)++; -} - - -/* - * These are the routines invoked by the control routines to do - * the actual upsampling/conversion. One row group is processed per call. - * - * Note: since we may be writing directly into application-supplied buffers, - * we have to be honest about the output width; we can't assume the buffer - * has been rounded up to an even width. - */ - - -/* - * Upsample and color convert for the case of 2:1 horizontal and 1:1 vertical. - */ - -METHODDEF(void) -h2v1_merged_upsample (j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION in_row_group_ctr, - JSAMPARRAY output_buf) -{ - my_upsample_ptr upsample = (my_upsample_ptr) cinfo->upsample; - register int y, cred, cgreen, cblue; - int cb, cr; - register JSAMPROW outptr; - JSAMPROW inptr0, inptr1, inptr2; - JDIMENSION col; - /* copy these pointers into registers if possible */ - register JSAMPLE * range_limit = cinfo->sample_range_limit; - int * Crrtab = upsample->Cr_r_tab; - int * Cbbtab = upsample->Cb_b_tab; - INT32 * Crgtab = upsample->Cr_g_tab; - INT32 * Cbgtab = upsample->Cb_g_tab; - SHIFT_TEMPS - - inptr0 = input_buf[0][in_row_group_ctr]; - inptr1 = input_buf[1][in_row_group_ctr]; - inptr2 = input_buf[2][in_row_group_ctr]; - outptr = output_buf[0]; - /* Loop for each pair of output pixels */ - for (col = cinfo->output_width >> 1; col > 0; col--) { - /* Do the chroma part of the calculation */ - cb = GETJSAMPLE(*inptr1++); - cr = GETJSAMPLE(*inptr2++); - cgreen = (int) RIGHT_SHIFT(Cbgtab[cb] + Crgtab[cr], SCALEBITS); - cblue = Cbbtab[cb]; - cred = Crrtab[cr]; - /* Fetch 2 Y values and emit 2 pixels */ - y = GETJSAMPLE(*inptr0++); - outptr[RGB_RED] = range_limit[y + cred]; - outptr[RGB_GREEN] = range_limit[y + cgreen]; - outptr[RGB_BLUE] = range_limit[y + cblue]; - outptr += RGB_PIXELSIZE; - y = GETJSAMPLE(*inptr0++); - outptr[RGB_RED] = range_limit[y + cred]; - outptr[RGB_GREEN] = range_limit[y + cgreen]; - outptr[RGB_BLUE] = range_limit[y + cblue]; - outptr += RGB_PIXELSIZE; - } - /* If image width is odd, do the last output column separately */ - if (cinfo->output_width & 1) { - y = GETJSAMPLE(*inptr0); - cb = GETJSAMPLE(*inptr1); - cr = GETJSAMPLE(*inptr2); - outptr[RGB_RED] = range_limit[y + Crrtab[cr]]; - outptr[RGB_GREEN] = range_limit[y + - ((int) RIGHT_SHIFT(Cbgtab[cb] + Crgtab[cr], - SCALEBITS))]; - outptr[RGB_BLUE] = range_limit[y + Cbbtab[cb]]; - } -} - - -/* - * Upsample and color convert for the case of 2:1 horizontal and 2:1 vertical. - */ - -METHODDEF(void) -h2v2_merged_upsample (j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION in_row_group_ctr, - JSAMPARRAY output_buf) -{ - my_upsample_ptr upsample = (my_upsample_ptr) cinfo->upsample; - register int y, cred, cgreen, cblue; - int cb, cr; - register JSAMPROW outptr0, outptr1; - JSAMPROW inptr00, inptr01, inptr1, inptr2; - JDIMENSION col; - /* copy these pointers into registers if possible */ - register JSAMPLE * range_limit = cinfo->sample_range_limit; - int * Crrtab = upsample->Cr_r_tab; - int * Cbbtab = upsample->Cb_b_tab; - INT32 * Crgtab = upsample->Cr_g_tab; - INT32 * Cbgtab = upsample->Cb_g_tab; - SHIFT_TEMPS - - inptr00 = input_buf[0][in_row_group_ctr*2]; - inptr01 = input_buf[0][in_row_group_ctr*2 + 1]; - inptr1 = input_buf[1][in_row_group_ctr]; - inptr2 = input_buf[2][in_row_group_ctr]; - outptr0 = output_buf[0]; - outptr1 = output_buf[1]; - /* Loop for each group of output pixels */ - for (col = cinfo->output_width >> 1; col > 0; col--) { - /* Do the chroma part of the calculation */ - cb = GETJSAMPLE(*inptr1++); - cr = GETJSAMPLE(*inptr2++); - cgreen = (int) RIGHT_SHIFT(Cbgtab[cb] + Crgtab[cr], SCALEBITS); - cblue = Cbbtab[cb]; - cred = Crrtab[cr]; - /* Fetch 4 Y values and emit 4 pixels */ - y = GETJSAMPLE(*inptr00++); - outptr0[RGB_RED] = range_limit[y + cred]; - outptr0[RGB_GREEN] = range_limit[y + cgreen]; - outptr0[RGB_BLUE] = range_limit[y + cblue]; - outptr0 += RGB_PIXELSIZE; - y = GETJSAMPLE(*inptr00++); - outptr0[RGB_RED] = range_limit[y + cred]; - outptr0[RGB_GREEN] = range_limit[y + cgreen]; - outptr0[RGB_BLUE] = range_limit[y + cblue]; - outptr0 += RGB_PIXELSIZE; - y = GETJSAMPLE(*inptr01++); - outptr1[RGB_RED] = range_limit[y + cred]; - outptr1[RGB_GREEN] = range_limit[y + cgreen]; - outptr1[RGB_BLUE] = range_limit[y + cblue]; - outptr1 += RGB_PIXELSIZE; - y = GETJSAMPLE(*inptr01++); - outptr1[RGB_RED] = range_limit[y + cred]; - outptr1[RGB_GREEN] = range_limit[y + cgreen]; - outptr1[RGB_BLUE] = range_limit[y + cblue]; - outptr1 += RGB_PIXELSIZE; - } - /* If image width is odd, do the last output column separately */ - if (cinfo->output_width & 1) { - cb = GETJSAMPLE(*inptr1); - cr = GETJSAMPLE(*inptr2); - cgreen = (int) RIGHT_SHIFT(Cbgtab[cb] + Crgtab[cr], SCALEBITS); - cblue = Cbbtab[cb]; - cred = Crrtab[cr]; - y = GETJSAMPLE(*inptr00); - outptr0[RGB_RED] = range_limit[y + cred]; - outptr0[RGB_GREEN] = range_limit[y + cgreen]; - outptr0[RGB_BLUE] = range_limit[y + cblue]; - y = GETJSAMPLE(*inptr01); - outptr1[RGB_RED] = range_limit[y + cred]; - outptr1[RGB_GREEN] = range_limit[y + cgreen]; - outptr1[RGB_BLUE] = range_limit[y + cblue]; - } -} - - -/* - * Module initialization routine for merged upsampling/color conversion. - * - * NB: this is called under the conditions determined by use_merged_upsample() - * in jdmaster.c. That routine MUST correspond to the actual capabilities - * of this module; no safety checks are made here. - */ - -GLOBAL(void) -jinit_merged_upsampler (j_decompress_ptr cinfo) -{ - my_upsample_ptr upsample; - - upsample = (my_upsample_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_upsampler)); - cinfo->upsample = &upsample->pub; - upsample->pub.start_pass = start_pass_merged_upsample; - upsample->pub.need_context_rows = FALSE; - - upsample->out_row_width = cinfo->output_width * cinfo->out_color_components; - - if (cinfo->max_v_samp_factor == 2) { - upsample->pub.upsample = merged_2v_upsample; - upsample->upmethod = h2v2_merged_upsample; - /* Allocate a spare row buffer */ - upsample->spare_row = (JSAMPROW) (*cinfo->mem->alloc_large) - ((j_common_ptr) cinfo, JPOOL_IMAGE, - (size_t) upsample->out_row_width * SIZEOF(JSAMPLE)); - } else { - upsample->pub.upsample = merged_1v_upsample; - upsample->upmethod = h2v1_merged_upsample; - /* No spare row needed */ - upsample->spare_row = NULL; - } - - if (cinfo->jpeg_color_space == JCS_BG_YCC) - build_bg_ycc_rgb_table(cinfo); - else - build_ycc_rgb_table(cinfo); -} - -#endif /* UPSAMPLE_MERGING_SUPPORTED */ diff --git a/cmake/externals/libjpeg/jdpostct.c b/cmake/externals/libjpeg/jdpostct.c deleted file mode 100644 index 571563d72..000000000 --- a/cmake/externals/libjpeg/jdpostct.c +++ /dev/null @@ -1,290 +0,0 @@ -/* - * jdpostct.c - * - * Copyright (C) 1994-1996, Thomas G. Lane. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains the decompression postprocessing controller. - * This controller manages the upsampling, color conversion, and color - * quantization/reduction steps; specifically, it controls the buffering - * between upsample/color conversion and color quantization/reduction. - * - * If no color quantization/reduction is required, then this module has no - * work to do, and it just hands off to the upsample/color conversion code. - * An integrated upsample/convert/quantize process would replace this module - * entirely. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* Private buffer controller object */ - -typedef struct { - struct jpeg_d_post_controller pub; /* public fields */ - - /* Color quantization source buffer: this holds output data from - * the upsample/color conversion step to be passed to the quantizer. - * For two-pass color quantization, we need a full-image buffer; - * for one-pass operation, a strip buffer is sufficient. - */ - jvirt_sarray_ptr whole_image; /* virtual array, or NULL if one-pass */ - JSAMPARRAY buffer; /* strip buffer, or current strip of virtual */ - JDIMENSION strip_height; /* buffer size in rows */ - /* for two-pass mode only: */ - JDIMENSION starting_row; /* row # of first row in current strip */ - JDIMENSION next_row; /* index of next row to fill/empty in strip */ -} my_post_controller; - -typedef my_post_controller * my_post_ptr; - - -/* Forward declarations */ -METHODDEF(void) post_process_1pass - JPP((j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION *in_row_group_ctr, - JDIMENSION in_row_groups_avail, - JSAMPARRAY output_buf, JDIMENSION *out_row_ctr, - JDIMENSION out_rows_avail)); -#ifdef QUANT_2PASS_SUPPORTED -METHODDEF(void) post_process_prepass - JPP((j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION *in_row_group_ctr, - JDIMENSION in_row_groups_avail, - JSAMPARRAY output_buf, JDIMENSION *out_row_ctr, - JDIMENSION out_rows_avail)); -METHODDEF(void) post_process_2pass - JPP((j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION *in_row_group_ctr, - JDIMENSION in_row_groups_avail, - JSAMPARRAY output_buf, JDIMENSION *out_row_ctr, - JDIMENSION out_rows_avail)); -#endif - - -/* - * Initialize for a processing pass. - */ - -METHODDEF(void) -start_pass_dpost (j_decompress_ptr cinfo, J_BUF_MODE pass_mode) -{ - my_post_ptr post = (my_post_ptr) cinfo->post; - - switch (pass_mode) { - case JBUF_PASS_THRU: - if (cinfo->quantize_colors) { - /* Single-pass processing with color quantization. */ - post->pub.post_process_data = post_process_1pass; - /* We could be doing buffered-image output before starting a 2-pass - * color quantization; in that case, jinit_d_post_controller did not - * allocate a strip buffer. Use the virtual-array buffer as workspace. - */ - if (post->buffer == NULL) { - post->buffer = (*cinfo->mem->access_virt_sarray) - ((j_common_ptr) cinfo, post->whole_image, - (JDIMENSION) 0, post->strip_height, TRUE); - } - } else { - /* For single-pass processing without color quantization, - * I have no work to do; just call the upsampler directly. - */ - post->pub.post_process_data = cinfo->upsample->upsample; - } - break; -#ifdef QUANT_2PASS_SUPPORTED - case JBUF_SAVE_AND_PASS: - /* First pass of 2-pass quantization */ - if (post->whole_image == NULL) - ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); - post->pub.post_process_data = post_process_prepass; - break; - case JBUF_CRANK_DEST: - /* Second pass of 2-pass quantization */ - if (post->whole_image == NULL) - ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); - post->pub.post_process_data = post_process_2pass; - break; -#endif /* QUANT_2PASS_SUPPORTED */ - default: - ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); - break; - } - post->starting_row = post->next_row = 0; -} - - -/* - * Process some data in the one-pass (strip buffer) case. - * This is used for color precision reduction as well as one-pass quantization. - */ - -METHODDEF(void) -post_process_1pass (j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION *in_row_group_ctr, - JDIMENSION in_row_groups_avail, - JSAMPARRAY output_buf, JDIMENSION *out_row_ctr, - JDIMENSION out_rows_avail) -{ - my_post_ptr post = (my_post_ptr) cinfo->post; - JDIMENSION num_rows, max_rows; - - /* Fill the buffer, but not more than what we can dump out in one go. */ - /* Note we rely on the upsampler to detect bottom of image. */ - max_rows = out_rows_avail - *out_row_ctr; - if (max_rows > post->strip_height) - max_rows = post->strip_height; - num_rows = 0; - (*cinfo->upsample->upsample) (cinfo, - input_buf, in_row_group_ctr, in_row_groups_avail, - post->buffer, &num_rows, max_rows); - /* Quantize and emit data. */ - (*cinfo->cquantize->color_quantize) (cinfo, - post->buffer, output_buf + *out_row_ctr, (int) num_rows); - *out_row_ctr += num_rows; -} - - -#ifdef QUANT_2PASS_SUPPORTED - -/* - * Process some data in the first pass of 2-pass quantization. - */ - -METHODDEF(void) -post_process_prepass (j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION *in_row_group_ctr, - JDIMENSION in_row_groups_avail, - JSAMPARRAY output_buf, JDIMENSION *out_row_ctr, - JDIMENSION out_rows_avail) -{ - my_post_ptr post = (my_post_ptr) cinfo->post; - JDIMENSION old_next_row, num_rows; - - /* Reposition virtual buffer if at start of strip. */ - if (post->next_row == 0) { - post->buffer = (*cinfo->mem->access_virt_sarray) - ((j_common_ptr) cinfo, post->whole_image, - post->starting_row, post->strip_height, TRUE); - } - - /* Upsample some data (up to a strip height's worth). */ - old_next_row = post->next_row; - (*cinfo->upsample->upsample) (cinfo, - input_buf, in_row_group_ctr, in_row_groups_avail, - post->buffer, &post->next_row, post->strip_height); - - /* Allow quantizer to scan new data. No data is emitted, */ - /* but we advance out_row_ctr so outer loop can tell when we're done. */ - if (post->next_row > old_next_row) { - num_rows = post->next_row - old_next_row; - (*cinfo->cquantize->color_quantize) (cinfo, post->buffer + old_next_row, - (JSAMPARRAY) NULL, (int) num_rows); - *out_row_ctr += num_rows; - } - - /* Advance if we filled the strip. */ - if (post->next_row >= post->strip_height) { - post->starting_row += post->strip_height; - post->next_row = 0; - } -} - - -/* - * Process some data in the second pass of 2-pass quantization. - */ - -METHODDEF(void) -post_process_2pass (j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION *in_row_group_ctr, - JDIMENSION in_row_groups_avail, - JSAMPARRAY output_buf, JDIMENSION *out_row_ctr, - JDIMENSION out_rows_avail) -{ - my_post_ptr post = (my_post_ptr) cinfo->post; - JDIMENSION num_rows, max_rows; - - /* Reposition virtual buffer if at start of strip. */ - if (post->next_row == 0) { - post->buffer = (*cinfo->mem->access_virt_sarray) - ((j_common_ptr) cinfo, post->whole_image, - post->starting_row, post->strip_height, FALSE); - } - - /* Determine number of rows to emit. */ - num_rows = post->strip_height - post->next_row; /* available in strip */ - max_rows = out_rows_avail - *out_row_ctr; /* available in output area */ - if (num_rows > max_rows) - num_rows = max_rows; - /* We have to check bottom of image here, can't depend on upsampler. */ - max_rows = cinfo->output_height - post->starting_row; - if (num_rows > max_rows) - num_rows = max_rows; - - /* Quantize and emit data. */ - (*cinfo->cquantize->color_quantize) (cinfo, - post->buffer + post->next_row, output_buf + *out_row_ctr, - (int) num_rows); - *out_row_ctr += num_rows; - - /* Advance if we filled the strip. */ - post->next_row += num_rows; - if (post->next_row >= post->strip_height) { - post->starting_row += post->strip_height; - post->next_row = 0; - } -} - -#endif /* QUANT_2PASS_SUPPORTED */ - - -/* - * Initialize postprocessing controller. - */ - -GLOBAL(void) -jinit_d_post_controller (j_decompress_ptr cinfo, boolean need_full_buffer) -{ - my_post_ptr post; - - post = (my_post_ptr) - (*cinfo->mem->alloc_small) ((j_common_ptr) cinfo, JPOOL_IMAGE, - SIZEOF(my_post_controller)); - cinfo->post = (struct jpeg_d_post_controller *) post; - post->pub.start_pass = start_pass_dpost; - post->whole_image = NULL; /* flag for no virtual arrays */ - post->buffer = NULL; /* flag for no strip buffer */ - - /* Create the quantization buffer, if needed */ - if (cinfo->quantize_colors) { - /* The buffer strip height is max_v_samp_factor, which is typically - * an efficient number of rows for upsampling to return. - * (In the presence of output rescaling, we might want to be smarter?) - */ - post->strip_height = (JDIMENSION) cinfo->max_v_samp_factor; - if (need_full_buffer) { - /* Two-pass color quantization: need full-image storage. */ - /* We round up the number of rows to a multiple of the strip height. */ -#ifdef QUANT_2PASS_SUPPORTED - post->whole_image = (*cinfo->mem->request_virt_sarray) - ((j_common_ptr) cinfo, JPOOL_IMAGE, FALSE, - cinfo->output_width * cinfo->out_color_components, - (JDIMENSION) jround_up((long) cinfo->output_height, - (long) post->strip_height), - post->strip_height); -#else - ERREXIT(cinfo, JERR_BAD_BUFFER_MODE); -#endif /* QUANT_2PASS_SUPPORTED */ - } else { - /* One-pass color quantization: just make a strip buffer. */ - post->buffer = (*cinfo->mem->alloc_sarray) - ((j_common_ptr) cinfo, JPOOL_IMAGE, - cinfo->output_width * cinfo->out_color_components, - post->strip_height); - } - } -} diff --git a/cmake/externals/libjpeg/jdsample.c b/cmake/externals/libjpeg/jdsample.c deleted file mode 100644 index 15afeafe3..000000000 --- a/cmake/externals/libjpeg/jdsample.c +++ /dev/null @@ -1,341 +0,0 @@ -/* - * jdsample.c - * - * Copyright (C) 1991-1996, Thomas G. Lane. - * Modified 2002-2020 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains upsampling routines. - * - * Upsampling input data is counted in "row groups". A row group - * is defined to be (v_samp_factor * DCT_v_scaled_size / min_DCT_v_scaled_size) - * sample rows of each component. Upsampling will normally produce - * max_v_samp_factor pixel rows from each row group (but this could vary - * if the upsampler is applying a scale factor of its own). - * - * An excellent reference for image resampling is - * Digital Image Warping, George Wolberg, 1990. - * Pub. by IEEE Computer Society Press, Los Alamitos, CA. ISBN 0-8186-8944-7. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* Pointer to routine to upsample a single component */ -typedef JMETHOD(void, upsample1_ptr, - (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JSAMPARRAY input_data, JSAMPIMAGE output_data_ptr)); - -/* Private subobject */ - -typedef struct { - struct jpeg_upsampler pub; /* public fields */ - - /* Color conversion buffer. When using separate upsampling and color - * conversion steps, this buffer holds one upsampled row group until it - * has been color converted and output. - * Note: we do not allocate any storage for component(s) which are full-size, - * ie do not need rescaling. The corresponding entry of color_buf[] is - * simply set to point to the input data array, thereby avoiding copying. - */ - JSAMPARRAY color_buf[MAX_COMPONENTS]; - - /* Per-component upsampling method pointers */ - upsample1_ptr methods[MAX_COMPONENTS]; - - int next_row_out; /* counts rows emitted from color_buf */ - JDIMENSION rows_to_go; /* counts rows remaining in image */ - - /* Height of an input row group for each component. */ - int rowgroup_height[MAX_COMPONENTS]; - - /* These arrays save pixel expansion factors so that int_expand need not - * recompute them each time. They are unused for other upsampling methods. - */ - UINT8 h_expand[MAX_COMPONENTS]; - UINT8 v_expand[MAX_COMPONENTS]; -} my_upsampler; - -typedef my_upsampler * my_upsample_ptr; - - -/* - * Initialize for an upsampling pass. - */ - -METHODDEF(void) -start_pass_upsample (j_decompress_ptr cinfo) -{ - my_upsample_ptr upsample = (my_upsample_ptr) cinfo->upsample; - - /* Mark the conversion buffer empty */ - upsample->next_row_out = cinfo->max_v_samp_factor; - /* Initialize total-height counter for detecting bottom of image */ - upsample->rows_to_go = cinfo->output_height; -} - - -/* - * Control routine to do upsampling (and color conversion). - * - * In this version we upsample each component independently. - * We upsample one row group into the conversion buffer, then apply - * color conversion a row at a time. - */ - -METHODDEF(void) -sep_upsample (j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION *in_row_group_ctr, - JDIMENSION in_row_groups_avail, - JSAMPARRAY output_buf, JDIMENSION *out_row_ctr, - JDIMENSION out_rows_avail) -{ - my_upsample_ptr upsample = (my_upsample_ptr) cinfo->upsample; - int ci; - jpeg_component_info * compptr; - JDIMENSION num_rows; - - /* Fill the conversion buffer, if it's empty */ - if (upsample->next_row_out >= cinfo->max_v_samp_factor) { - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - /* Don't bother to upsample an uninteresting component. */ - if (! compptr->component_needed) - continue; - /* Invoke per-component upsample method. Notice we pass a POINTER - * to color_buf[ci], so that fullsize_upsample can change it. - */ - (*upsample->methods[ci]) (cinfo, compptr, - input_buf[ci] + (*in_row_group_ctr * upsample->rowgroup_height[ci]), - upsample->color_buf + ci); - } - upsample->next_row_out = 0; - } - - /* Color-convert and emit rows */ - - /* How many we have in the buffer: */ - num_rows = (JDIMENSION) (cinfo->max_v_samp_factor - upsample->next_row_out); - /* Not more than the distance to the end of the image. Need this test - * in case the image height is not a multiple of max_v_samp_factor: - */ - if (num_rows > upsample->rows_to_go) - num_rows = upsample->rows_to_go; - /* And not more than what the client can accept: */ - out_rows_avail -= *out_row_ctr; - if (num_rows > out_rows_avail) - num_rows = out_rows_avail; - - (*cinfo->cconvert->color_convert) (cinfo, upsample->color_buf, - (JDIMENSION) upsample->next_row_out, - output_buf + *out_row_ctr, - (int) num_rows); - - /* Adjust counts */ - *out_row_ctr += num_rows; - upsample->rows_to_go -= num_rows; - upsample->next_row_out += num_rows; - /* When the buffer is emptied, declare this input row group consumed */ - if (upsample->next_row_out >= cinfo->max_v_samp_factor) - (*in_row_group_ctr)++; -} - - -/* - * These are the routines invoked by sep_upsample to upsample pixel values - * of a single component. One row group is processed per call. - */ - - -/* - * For full-size components, we just make color_buf[ci] point at the - * input buffer, and thus avoid copying any data. Note that this is - * safe only because sep_upsample doesn't declare the input row group - * "consumed" until we are done color converting and emitting it. - */ - -METHODDEF(void) -fullsize_upsample (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JSAMPARRAY input_data, JSAMPIMAGE output_data_ptr) -{ - *output_data_ptr = input_data; -} - - -/* - * This version handles any integral sampling ratios. - * This is not used for typical JPEG files, so it need not be fast. - * Nor, for that matter, is it particularly accurate: the algorithm is - * simple replication of the input pixel onto the corresponding output - * pixels. The hi-falutin sampling literature refers to this as a - * "box filter". A box filter tends to introduce visible artifacts, - * so if you are actually going to use 3:1 or 4:1 sampling ratios - * you would be well advised to improve this code. - */ - -METHODDEF(void) -int_upsample (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JSAMPARRAY input_data, JSAMPIMAGE output_data_ptr) -{ - my_upsample_ptr upsample = (my_upsample_ptr) cinfo->upsample; - JSAMPARRAY output_data, output_end; - register JSAMPROW inptr, outptr; - register JSAMPLE invalue; - register int h; - JSAMPROW outend; - int h_expand, v_expand; - - h_expand = upsample->h_expand[compptr->component_index]; - v_expand = upsample->v_expand[compptr->component_index]; - - output_data = *output_data_ptr; - output_end = output_data + cinfo->max_v_samp_factor; - for (; output_data < output_end; output_data += v_expand) { - /* Generate one output row with proper horizontal expansion */ - inptr = *input_data++; - outptr = *output_data; - outend = outptr + cinfo->output_width; - while (outptr < outend) { - invalue = *inptr++; /* don't need GETJSAMPLE() here */ - for (h = h_expand; h > 0; h--) { - *outptr++ = invalue; - } - } - /* Generate any additional output rows by duplicating the first one */ - if (v_expand > 1) { - jcopy_sample_rows(output_data, output_data + 1, - v_expand - 1, cinfo->output_width); - } - } -} - - -/* - * Fast processing for the common case of 2:1 horizontal and 1:1 vertical. - * It's still a box filter. - */ - -METHODDEF(void) -h2v1_upsample (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JSAMPARRAY input_data, JSAMPIMAGE output_data_ptr) -{ - JSAMPARRAY output_data = *output_data_ptr; - register JSAMPROW inptr, outptr; - register JSAMPLE invalue; - JSAMPROW outend; - int outrow; - - for (outrow = 0; outrow < cinfo->max_v_samp_factor; outrow++) { - inptr = input_data[outrow]; - outptr = output_data[outrow]; - outend = outptr + cinfo->output_width; - while (outptr < outend) { - invalue = *inptr++; /* don't need GETJSAMPLE() here */ - *outptr++ = invalue; - *outptr++ = invalue; - } - } -} - - -/* - * Fast processing for the common case of 2:1 horizontal and 2:1 vertical. - * It's still a box filter. - */ - -METHODDEF(void) -h2v2_upsample (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JSAMPARRAY input_data, JSAMPIMAGE output_data_ptr) -{ - JSAMPARRAY output_data, output_end; - register JSAMPROW inptr, outptr; - register JSAMPLE invalue; - JSAMPROW outend; - - output_data = *output_data_ptr; - output_end = output_data + cinfo->max_v_samp_factor; - for (; output_data < output_end; output_data += 2) { - inptr = *input_data++; - outptr = *output_data; - outend = outptr + cinfo->output_width; - while (outptr < outend) { - invalue = *inptr++; /* don't need GETJSAMPLE() here */ - *outptr++ = invalue; - *outptr++ = invalue; - } - jcopy_sample_rows(output_data, output_data + 1, - 1, cinfo->output_width); - } -} - - -/* - * Module initialization routine for upsampling. - */ - -GLOBAL(void) -jinit_upsampler (j_decompress_ptr cinfo) -{ - my_upsample_ptr upsample; - int ci; - jpeg_component_info * compptr; - int h_in_group, v_in_group, h_out_group, v_out_group; - - upsample = (my_upsample_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_upsampler)); - cinfo->upsample = &upsample->pub; - upsample->pub.start_pass = start_pass_upsample; - upsample->pub.upsample = sep_upsample; - upsample->pub.need_context_rows = FALSE; /* until we find out differently */ - - if (cinfo->CCIR601_sampling) /* this isn't supported */ - ERREXIT(cinfo, JERR_CCIR601_NOTIMPL); - - /* Verify we can handle the sampling factors, select per-component methods, - * and create storage as needed. - */ - for (ci = 0, compptr = cinfo->comp_info; ci < cinfo->num_components; - ci++, compptr++) { - /* Don't bother to upsample an uninteresting component. */ - if (! compptr->component_needed) - continue; - /* Compute size of an "input group" after IDCT scaling. This many samples - * are to be converted to max_h_samp_factor * max_v_samp_factor pixels. - */ - h_in_group = (compptr->h_samp_factor * compptr->DCT_h_scaled_size) / - cinfo->min_DCT_h_scaled_size; - v_in_group = (compptr->v_samp_factor * compptr->DCT_v_scaled_size) / - cinfo->min_DCT_v_scaled_size; - h_out_group = cinfo->max_h_samp_factor; - v_out_group = cinfo->max_v_samp_factor; - upsample->rowgroup_height[ci] = v_in_group; /* save for use later */ - if (h_in_group == h_out_group && v_in_group == v_out_group) { - /* Fullsize components can be processed without any work. */ - upsample->methods[ci] = fullsize_upsample; - continue; /* don't need to allocate buffer */ - } - if (h_in_group * 2 == h_out_group && v_in_group == v_out_group) { - /* Special case for 2h1v upsampling */ - upsample->methods[ci] = h2v1_upsample; - } else if (h_in_group * 2 == h_out_group && - v_in_group * 2 == v_out_group) { - /* Special case for 2h2v upsampling */ - upsample->methods[ci] = h2v2_upsample; - } else if ((h_out_group % h_in_group) == 0 && - (v_out_group % v_in_group) == 0) { - /* Generic integral-factors upsampling method */ - upsample->methods[ci] = int_upsample; - upsample->h_expand[ci] = (UINT8) (h_out_group / h_in_group); - upsample->v_expand[ci] = (UINT8) (v_out_group / v_in_group); - } else - ERREXIT(cinfo, JERR_FRACT_SAMPLE_NOTIMPL); - upsample->color_buf[ci] = (*cinfo->mem->alloc_sarray) - ((j_common_ptr) cinfo, JPOOL_IMAGE, - (JDIMENSION) jround_up((long) cinfo->output_width, - (long) cinfo->max_h_samp_factor), - (JDIMENSION) cinfo->max_v_samp_factor); - } -} diff --git a/cmake/externals/libjpeg/jdtrans.c b/cmake/externals/libjpeg/jdtrans.c deleted file mode 100644 index 22dd47fb5..000000000 --- a/cmake/externals/libjpeg/jdtrans.c +++ /dev/null @@ -1,140 +0,0 @@ -/* - * jdtrans.c - * - * Copyright (C) 1995-1997, Thomas G. Lane. - * Modified 2000-2009 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains library routines for transcoding decompression, - * that is, reading raw DCT coefficient arrays from an input JPEG file. - * The routines in jdapimin.c will also be needed by a transcoder. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* Forward declarations */ -LOCAL(void) transdecode_master_selection JPP((j_decompress_ptr cinfo)); - - -/* - * Read the coefficient arrays from a JPEG file. - * jpeg_read_header must be completed before calling this. - * - * The entire image is read into a set of virtual coefficient-block arrays, - * one per component. The return value is a pointer to the array of - * virtual-array descriptors. These can be manipulated directly via the - * JPEG memory manager, or handed off to jpeg_write_coefficients(). - * To release the memory occupied by the virtual arrays, call - * jpeg_finish_decompress() when done with the data. - * - * An alternative usage is to simply obtain access to the coefficient arrays - * during a buffered-image-mode decompression operation. This is allowed - * after any jpeg_finish_output() call. The arrays can be accessed until - * jpeg_finish_decompress() is called. (Note that any call to the library - * may reposition the arrays, so don't rely on access_virt_barray() results - * to stay valid across library calls.) - * - * Returns NULL if suspended. This case need be checked only if - * a suspending data source is used. - */ - -GLOBAL(jvirt_barray_ptr *) -jpeg_read_coefficients (j_decompress_ptr cinfo) -{ - if (cinfo->global_state == DSTATE_READY) { - /* First call: initialize active modules */ - transdecode_master_selection(cinfo); - cinfo->global_state = DSTATE_RDCOEFS; - } - if (cinfo->global_state == DSTATE_RDCOEFS) { - /* Absorb whole file into the coef buffer */ - for (;;) { - int retcode; - /* Call progress monitor hook if present */ - if (cinfo->progress != NULL) - (*cinfo->progress->progress_monitor) ((j_common_ptr) cinfo); - /* Absorb some more input */ - retcode = (*cinfo->inputctl->consume_input) (cinfo); - if (retcode == JPEG_SUSPENDED) - return NULL; - if (retcode == JPEG_REACHED_EOI) - break; - /* Advance progress counter if appropriate */ - if (cinfo->progress != NULL && - (retcode == JPEG_ROW_COMPLETED || retcode == JPEG_REACHED_SOS)) { - if (++cinfo->progress->pass_counter >= cinfo->progress->pass_limit) { - /* startup underestimated number of scans; ratchet up one scan */ - cinfo->progress->pass_limit += (long) cinfo->total_iMCU_rows; - } - } - } - /* Set state so that jpeg_finish_decompress does the right thing */ - cinfo->global_state = DSTATE_STOPPING; - } - /* At this point we should be in state DSTATE_STOPPING if being used - * standalone, or in state DSTATE_BUFIMAGE if being invoked to get access - * to the coefficients during a full buffered-image-mode decompression. - */ - if ((cinfo->global_state == DSTATE_STOPPING || - cinfo->global_state == DSTATE_BUFIMAGE) && cinfo->buffered_image) { - return cinfo->coef->coef_arrays; - } - /* Oops, improper usage */ - ERREXIT1(cinfo, JERR_BAD_STATE, cinfo->global_state); - return NULL; /* keep compiler happy */ -} - - -/* - * Master selection of decompression modules for transcoding. - * This substitutes for jdmaster.c's initialization of the full decompressor. - */ - -LOCAL(void) -transdecode_master_selection (j_decompress_ptr cinfo) -{ - /* This is effectively a buffered-image operation. */ - cinfo->buffered_image = TRUE; - - /* Compute output image dimensions and related values. */ - jpeg_core_output_dimensions(cinfo); - - /* Entropy decoding: either Huffman or arithmetic coding. */ - if (cinfo->arith_code) - jinit_arith_decoder(cinfo); - else { - jinit_huff_decoder(cinfo); - } - - /* Always get a full-image coefficient buffer. */ - jinit_d_coef_controller(cinfo, TRUE); - - /* We can now tell the memory manager to allocate virtual arrays. */ - (*cinfo->mem->realize_virt_arrays) ((j_common_ptr) cinfo); - - /* Initialize input side of decompressor to consume first scan. */ - (*cinfo->inputctl->start_input_pass) (cinfo); - - /* Initialize progress monitoring. */ - if (cinfo->progress != NULL) { - int nscans; - /* Estimate number of scans to set pass_limit. */ - if (cinfo->progressive_mode) { - /* Arbitrarily estimate 2 interleaved DC scans + 3 AC scans/component. */ - nscans = 2 + 3 * cinfo->num_components; - } else if (cinfo->inputctl->has_multiple_scans) { - /* For a nonprogressive multiscan file, estimate 1 scan per component. */ - nscans = cinfo->num_components; - } else { - nscans = 1; - } - cinfo->progress->pass_counter = 0L; - cinfo->progress->pass_limit = (long) cinfo->total_iMCU_rows * nscans; - cinfo->progress->completed_passes = 0; - cinfo->progress->total_passes = 1; - } -} diff --git a/cmake/externals/libjpeg/jerror.c b/cmake/externals/libjpeg/jerror.c deleted file mode 100644 index 7163af699..000000000 --- a/cmake/externals/libjpeg/jerror.c +++ /dev/null @@ -1,253 +0,0 @@ -/* - * jerror.c - * - * Copyright (C) 1991-1998, Thomas G. Lane. - * Modified 2012-2015 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains simple error-reporting and trace-message routines. - * These are suitable for Unix-like systems and others where writing to - * stderr is the right thing to do. Many applications will want to replace - * some or all of these routines. - * - * If you define USE_WINDOWS_MESSAGEBOX in jconfig.h or in the makefile, - * you get a Windows-specific hack to display error messages in a dialog box. - * It ain't much, but it beats dropping error messages into the bit bucket, - * which is what happens to output to stderr under most Windows C compilers. - * - * These routines are used by both the compression and decompression code. - */ - -#ifdef USE_WINDOWS_MESSAGEBOX -#include -#endif - -/* this is not a core library module, so it doesn't define JPEG_INTERNALS */ -#include "jinclude.h" -#include "jpeglib.h" -#include "jversion.h" -#include "jerror.h" - -#ifndef EXIT_FAILURE /* define exit() codes if not provided */ -#define EXIT_FAILURE 1 -#endif - - -/* - * Create the message string table. - * We do this from the master message list in jerror.h by re-reading - * jerror.h with a suitable definition for macro JMESSAGE. - * The message table is made an external symbol just in case any applications - * want to refer to it directly. - */ - -#ifdef NEED_SHORT_EXTERNAL_NAMES -#define jpeg_std_message_table jMsgTable -#endif - -#define JMESSAGE(code,string) string , - -const char * const jpeg_std_message_table[] = { -#include "jerror.h" - NULL -}; - - -/* - * Error exit handler: must not return to caller. - * - * Applications may override this if they want to get control back after - * an error. Typically one would longjmp somewhere instead of exiting. - * The setjmp buffer can be made a private field within an expanded error - * handler object. Note that the info needed to generate an error message - * is stored in the error object, so you can generate the message now or - * later, at your convenience. - * You should make sure that the JPEG object is cleaned up (with jpeg_abort - * or jpeg_destroy) at some point. - */ - -METHODDEF(noreturn_t) -error_exit (j_common_ptr cinfo) -{ - /* Always display the message */ - (*cinfo->err->output_message) (cinfo); - - /* Let the memory manager delete any temp files before we die */ - jpeg_destroy(cinfo); - - exit(EXIT_FAILURE); -} - - -/* - * Actual output of an error or trace message. - * Applications may override this method to send JPEG messages somewhere - * other than stderr. - * - * On Windows, printing to stderr is generally completely useless, - * so we provide optional code to produce an error-dialog popup. - * Most Windows applications will still prefer to override this routine, - * but if they don't, it'll do something at least marginally useful. - * - * NOTE: to use the library in an environment that doesn't support the - * C stdio library, you may have to delete the call to fprintf() entirely, - * not just not use this routine. - */ - -METHODDEF(void) -output_message (j_common_ptr cinfo) -{ - char buffer[JMSG_LENGTH_MAX]; - - /* Create the message */ - (*cinfo->err->format_message) (cinfo, buffer); - -#ifdef USE_WINDOWS_MESSAGEBOX - /* Display it in a message dialog box */ - MessageBox(GetActiveWindow(), buffer, "JPEG Library Error", - MB_OK | MB_ICONERROR); -#else - /* Send it to stderr, adding a newline */ - fprintf(stderr, "%s\n", buffer); -#endif -} - - -/* - * Decide whether to emit a trace or warning message. - * msg_level is one of: - * -1: recoverable corrupt-data warning, may want to abort. - * 0: important advisory messages (always display to user). - * 1: first level of tracing detail. - * 2,3,...: successively more detailed tracing messages. - * An application might override this method if it wanted to abort on warnings - * or change the policy about which messages to display. - */ - -METHODDEF(void) -emit_message (j_common_ptr cinfo, int msg_level) -{ - struct jpeg_error_mgr * err = cinfo->err; - - if (msg_level < 0) { - /* It's a warning message. Since corrupt files may generate many warnings, - * the policy implemented here is to show only the first warning, - * unless trace_level >= 3. - */ - if (err->num_warnings == 0 || err->trace_level >= 3) - (*err->output_message) (cinfo); - /* Always count warnings in num_warnings. */ - err->num_warnings++; - } else { - /* It's a trace message. Show it if trace_level >= msg_level. */ - if (err->trace_level >= msg_level) - (*err->output_message) (cinfo); - } -} - - -/* - * Format a message string for the most recent JPEG error or message. - * The message is stored into buffer, which should be at least JMSG_LENGTH_MAX - * characters. Note that no '\n' character is added to the string. - * Few applications should need to override this method. - */ - -METHODDEF(void) -format_message (j_common_ptr cinfo, char * buffer) -{ - struct jpeg_error_mgr * err = cinfo->err; - int msg_code = err->msg_code; - const char * msgtext = NULL; - const char * msgptr; - char ch; - boolean isstring; - - /* Look up message string in proper table */ - if (msg_code > 0 && msg_code <= err->last_jpeg_message) { - msgtext = err->jpeg_message_table[msg_code]; - } else if (err->addon_message_table != NULL && - msg_code >= err->first_addon_message && - msg_code <= err->last_addon_message) { - msgtext = err->addon_message_table[msg_code - err->first_addon_message]; - } - - /* Defend against bogus message number */ - if (msgtext == NULL) { - err->msg_parm.i[0] = msg_code; - msgtext = err->jpeg_message_table[0]; - } - - /* Check for string parameter, as indicated by %s in the message text */ - isstring = FALSE; - msgptr = msgtext; - while ((ch = *msgptr++) != '\0') { - if (ch == '%') { - if (*msgptr == 's') isstring = TRUE; - break; - } - } - - /* Format the message into the passed buffer */ - if (isstring) - sprintf(buffer, msgtext, err->msg_parm.s); - else - sprintf(buffer, msgtext, - err->msg_parm.i[0], err->msg_parm.i[1], - err->msg_parm.i[2], err->msg_parm.i[3], - err->msg_parm.i[4], err->msg_parm.i[5], - err->msg_parm.i[6], err->msg_parm.i[7]); -} - - -/* - * Reset error state variables at start of a new image. - * This is called during compression startup to reset trace/error - * processing to default state, without losing any application-specific - * method pointers. An application might possibly want to override - * this method if it has additional error processing state. - */ - -METHODDEF(void) -reset_error_mgr (j_common_ptr cinfo) -{ - cinfo->err->num_warnings = 0; - /* trace_level is not reset since it is an application-supplied parameter */ - cinfo->err->msg_code = 0; /* may be useful as a flag for "no error" */ -} - - -/* - * Fill in the standard error-handling methods in a jpeg_error_mgr object. - * Typical call is: - * struct jpeg_compress_struct cinfo; - * struct jpeg_error_mgr err; - * - * cinfo.err = jpeg_std_error(&err); - * after which the application may override some of the methods. - */ - -GLOBAL(struct jpeg_error_mgr *) -jpeg_std_error (struct jpeg_error_mgr * err) -{ - err->error_exit = error_exit; - err->emit_message = emit_message; - err->output_message = output_message; - err->format_message = format_message; - err->reset_error_mgr = reset_error_mgr; - - err->trace_level = 0; /* default = no tracing */ - err->num_warnings = 0; /* no warnings emitted yet */ - err->msg_code = 0; /* may be useful as a flag for "no error" */ - - /* Initialize message table pointers */ - err->jpeg_message_table = jpeg_std_message_table; - err->last_jpeg_message = (int) JMSG_LASTMSGCODE - 1; - - err->addon_message_table = NULL; - err->first_addon_message = 0; /* for safety */ - err->last_addon_message = 0; - - return err; -} diff --git a/cmake/externals/libjpeg/jerror.h b/cmake/externals/libjpeg/jerror.h deleted file mode 100644 index db608b9c0..000000000 --- a/cmake/externals/libjpeg/jerror.h +++ /dev/null @@ -1,304 +0,0 @@ -/* - * jerror.h - * - * Copyright (C) 1994-1997, Thomas G. Lane. - * Modified 1997-2018 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file defines the error and message codes for the JPEG library. - * Edit this file to add new codes, or to translate the message strings to - * some other language. - * A set of error-reporting macros are defined too. Some applications using - * the JPEG library may wish to include this file to get the error codes - * and/or the macros. - */ - -/* - * To define the enum list of message codes, include this file without - * defining macro JMESSAGE. To create a message string table, include it - * again with a suitable JMESSAGE definition (see jerror.c for an example). - */ -#ifndef JMESSAGE -#ifndef JERROR_H -/* First time through, define the enum list */ -#define JMAKE_ENUM_LIST -#else -/* Repeated inclusions of this file are no-ops unless JMESSAGE is defined */ -#define JMESSAGE(code,string) -#endif /* JERROR_H */ -#endif /* JMESSAGE */ - -#ifdef JMAKE_ENUM_LIST - -typedef enum { - -#define JMESSAGE(code,string) code , - -#endif /* JMAKE_ENUM_LIST */ - -JMESSAGE(JMSG_NOMESSAGE, "Bogus message code %d") /* Must be first entry! */ - -/* For maintenance convenience, list is alphabetical by message code name */ -JMESSAGE(JERR_BAD_ALIGN_TYPE, "ALIGN_TYPE is wrong, please fix") -JMESSAGE(JERR_BAD_ALLOC_CHUNK, "MAX_ALLOC_CHUNK is wrong, please fix") -JMESSAGE(JERR_BAD_BUFFER_MODE, "Bogus buffer control mode") -JMESSAGE(JERR_BAD_COMPONENT_ID, "Invalid component ID %d in SOS") -JMESSAGE(JERR_BAD_CROP_SPEC, "Invalid crop request") -JMESSAGE(JERR_BAD_DCT_COEF, "DCT coefficient out of range") -JMESSAGE(JERR_BAD_DCTSIZE, "DCT scaled block size %dx%d not supported") -JMESSAGE(JERR_BAD_DROP_SAMPLING, - "Component index %d: mismatching sampling ratio %d:%d, %d:%d, %c") -JMESSAGE(JERR_BAD_HUFF_TABLE, "Bogus Huffman table definition") -JMESSAGE(JERR_BAD_IN_COLORSPACE, "Bogus input colorspace") -JMESSAGE(JERR_BAD_J_COLORSPACE, "Bogus JPEG colorspace") -JMESSAGE(JERR_BAD_LENGTH, "Bogus marker length") -JMESSAGE(JERR_BAD_LIB_VERSION, - "Wrong JPEG library version: library is %d, caller expects %d") -JMESSAGE(JERR_BAD_MCU_SIZE, "Sampling factors too large for interleaved scan") -JMESSAGE(JERR_BAD_POOL_ID, "Invalid memory pool code %d") -JMESSAGE(JERR_BAD_PRECISION, "Unsupported JPEG data precision %d") -JMESSAGE(JERR_BAD_PROGRESSION, - "Invalid progressive parameters Ss=%d Se=%d Ah=%d Al=%d") -JMESSAGE(JERR_BAD_PROG_SCRIPT, - "Invalid progressive parameters at scan script entry %d") -JMESSAGE(JERR_BAD_SAMPLING, "Bogus sampling factors") -JMESSAGE(JERR_BAD_SCAN_SCRIPT, "Invalid scan script at entry %d") -JMESSAGE(JERR_BAD_STATE, "Improper call to JPEG library in state %d") -JMESSAGE(JERR_BAD_STRUCT_SIZE, - "JPEG parameter struct mismatch: library thinks size is %u, caller expects %u") -JMESSAGE(JERR_BAD_VIRTUAL_ACCESS, "Bogus virtual array access") -JMESSAGE(JERR_BUFFER_SIZE, "Buffer passed to JPEG library is too small") -JMESSAGE(JERR_CANT_SUSPEND, "Suspension not allowed here") -JMESSAGE(JERR_CCIR601_NOTIMPL, "CCIR601 sampling not implemented yet") -JMESSAGE(JERR_COMPONENT_COUNT, "Too many color components: %d, max %d") -JMESSAGE(JERR_CONVERSION_NOTIMPL, "Unsupported color conversion request") -JMESSAGE(JERR_DAC_INDEX, "Bogus DAC index %d") -JMESSAGE(JERR_DAC_VALUE, "Bogus DAC value 0x%x") -JMESSAGE(JERR_DHT_INDEX, "Bogus DHT index %d") -JMESSAGE(JERR_DQT_INDEX, "Bogus DQT index %d") -JMESSAGE(JERR_EMPTY_IMAGE, "Empty JPEG image (DNL not supported)") -JMESSAGE(JERR_EMS_READ, "Read from EMS failed") -JMESSAGE(JERR_EMS_WRITE, "Write to EMS failed") -JMESSAGE(JERR_EOI_EXPECTED, "Didn't expect more than one scan") -JMESSAGE(JERR_FILE_READ, "Input file read error") -JMESSAGE(JERR_FILE_WRITE, "Output file write error --- out of disk space?") -JMESSAGE(JERR_FRACT_SAMPLE_NOTIMPL, "Fractional sampling not implemented yet") -JMESSAGE(JERR_HUFF_CLEN_OUTOFBOUNDS, "Huffman code size table out of bounds") -JMESSAGE(JERR_HUFF_MISSING_CODE, "Missing Huffman code table entry") -JMESSAGE(JERR_IMAGE_TOO_BIG, "Maximum supported image dimension is %u pixels") -JMESSAGE(JERR_INPUT_EMPTY, "Empty input file") -JMESSAGE(JERR_INPUT_EOF, "Premature end of input file") -JMESSAGE(JERR_MISMATCHED_QUANT_TABLE, - "Cannot transcode due to multiple use of quantization table %d") -JMESSAGE(JERR_MISSING_DATA, "Scan script does not transmit all data") -JMESSAGE(JERR_MODE_CHANGE, "Invalid color quantization mode change") -JMESSAGE(JERR_NOTIMPL, "Not implemented yet") -JMESSAGE(JERR_NOT_COMPILED, "Requested feature was omitted at compile time") -JMESSAGE(JERR_NO_ARITH_TABLE, "Arithmetic table 0x%02x was not defined") -JMESSAGE(JERR_NO_BACKING_STORE, "Backing store not supported") -JMESSAGE(JERR_NO_HUFF_TABLE, "Huffman table 0x%02x was not defined") -JMESSAGE(JERR_NO_IMAGE, "JPEG datastream contains no image") -JMESSAGE(JERR_NO_QUANT_TABLE, "Quantization table 0x%02x was not defined") -JMESSAGE(JERR_NO_SOI, "Not a JPEG file: starts with 0x%02x 0x%02x") -JMESSAGE(JERR_OUT_OF_MEMORY, "Insufficient memory (case %d)") -JMESSAGE(JERR_QUANT_COMPONENTS, - "Cannot quantize more than %d color components") -JMESSAGE(JERR_QUANT_FEW_COLORS, "Cannot quantize to fewer than %d colors") -JMESSAGE(JERR_QUANT_MANY_COLORS, "Cannot quantize to more than %d colors") -JMESSAGE(JERR_SOF_BEFORE, "Invalid JPEG file structure: %s before SOF") -JMESSAGE(JERR_SOF_DUPLICATE, "Invalid JPEG file structure: two SOF markers") -JMESSAGE(JERR_SOF_NO_SOS, "Invalid JPEG file structure: missing SOS marker") -JMESSAGE(JERR_SOF_UNSUPPORTED, "Unsupported JPEG process: SOF type 0x%02x") -JMESSAGE(JERR_SOI_DUPLICATE, "Invalid JPEG file structure: two SOI markers") -JMESSAGE(JERR_TFILE_CREATE, "Failed to create temporary file %s") -JMESSAGE(JERR_TFILE_READ, "Read failed on temporary file") -JMESSAGE(JERR_TFILE_SEEK, "Seek failed on temporary file") -JMESSAGE(JERR_TFILE_WRITE, - "Write failed on temporary file --- out of disk space?") -JMESSAGE(JERR_TOO_LITTLE_DATA, "Application transferred too few scanlines") -JMESSAGE(JERR_UNKNOWN_MARKER, "Unsupported marker type 0x%02x") -JMESSAGE(JERR_VIRTUAL_BUG, "Virtual array controller messed up") -JMESSAGE(JERR_WIDTH_OVERFLOW, "Image too wide for this implementation") -JMESSAGE(JERR_XMS_READ, "Read from XMS failed") -JMESSAGE(JERR_XMS_WRITE, "Write to XMS failed") -JMESSAGE(JMSG_COPYRIGHT, JCOPYRIGHT) -JMESSAGE(JMSG_VERSION, JVERSION) -JMESSAGE(JTRC_16BIT_TABLES, - "Caution: quantization tables are too coarse for baseline JPEG") -JMESSAGE(JTRC_ADOBE, - "Adobe APP14 marker: version %d, flags 0x%04x 0x%04x, transform %d") -JMESSAGE(JTRC_APP0, "Unknown APP0 marker (not JFIF), length %u") -JMESSAGE(JTRC_APP14, "Unknown APP14 marker (not Adobe), length %u") -JMESSAGE(JTRC_DAC, "Define Arithmetic Table 0x%02x: 0x%02x") -JMESSAGE(JTRC_DHT, "Define Huffman Table 0x%02x") -JMESSAGE(JTRC_DQT, "Define Quantization Table %d precision %d") -JMESSAGE(JTRC_DRI, "Define Restart Interval %u") -JMESSAGE(JTRC_EMS_CLOSE, "Freed EMS handle %u") -JMESSAGE(JTRC_EMS_OPEN, "Obtained EMS handle %u") -JMESSAGE(JTRC_EOI, "End Of Image") -JMESSAGE(JTRC_HUFFBITS, " %3d %3d %3d %3d %3d %3d %3d %3d") -JMESSAGE(JTRC_JFIF, "JFIF APP0 marker: version %d.%02d, density %dx%d %d") -JMESSAGE(JTRC_JFIF_BADTHUMBNAILSIZE, - "Warning: thumbnail image size does not match data length %u") -JMESSAGE(JTRC_JFIF_EXTENSION, - "JFIF extension marker: type 0x%02x, length %u") -JMESSAGE(JTRC_JFIF_THUMBNAIL, " with %d x %d thumbnail image") -JMESSAGE(JTRC_MISC_MARKER, "Miscellaneous marker 0x%02x, length %u") -JMESSAGE(JTRC_PARMLESS_MARKER, "Unexpected marker 0x%02x") -JMESSAGE(JTRC_QUANTVALS, " %4u %4u %4u %4u %4u %4u %4u %4u") -JMESSAGE(JTRC_QUANT_3_NCOLORS, "Quantizing to %d = %d*%d*%d colors") -JMESSAGE(JTRC_QUANT_NCOLORS, "Quantizing to %d colors") -JMESSAGE(JTRC_QUANT_SELECTED, "Selected %d colors for quantization") -JMESSAGE(JTRC_RECOVERY_ACTION, "At marker 0x%02x, recovery action %d") -JMESSAGE(JTRC_RST, "RST%d") -JMESSAGE(JTRC_SMOOTH_NOTIMPL, - "Smoothing not supported with nonstandard sampling ratios") -JMESSAGE(JTRC_SOF, "Start Of Frame 0x%02x: width=%u, height=%u, components=%d") -JMESSAGE(JTRC_SOF_COMPONENT, " Component %d: %dhx%dv q=%d") -JMESSAGE(JTRC_SOI, "Start of Image") -JMESSAGE(JTRC_SOS, "Start Of Scan: %d components") -JMESSAGE(JTRC_SOS_COMPONENT, " Component %d: dc=%d ac=%d") -JMESSAGE(JTRC_SOS_PARAMS, " Ss=%d, Se=%d, Ah=%d, Al=%d") -JMESSAGE(JTRC_TFILE_CLOSE, "Closed temporary file %s") -JMESSAGE(JTRC_TFILE_OPEN, "Opened temporary file %s") -JMESSAGE(JTRC_THUMB_JPEG, - "JFIF extension marker: JPEG-compressed thumbnail image, length %u") -JMESSAGE(JTRC_THUMB_PALETTE, - "JFIF extension marker: palette thumbnail image, length %u") -JMESSAGE(JTRC_THUMB_RGB, - "JFIF extension marker: RGB thumbnail image, length %u") -JMESSAGE(JTRC_UNKNOWN_IDS, - "Unrecognized component IDs %d %d %d, assuming YCbCr") -JMESSAGE(JTRC_XMS_CLOSE, "Freed XMS handle %u") -JMESSAGE(JTRC_XMS_OPEN, "Obtained XMS handle %u") -JMESSAGE(JWRN_ADOBE_XFORM, "Unknown Adobe color transform code %d") -JMESSAGE(JWRN_ARITH_BAD_CODE, "Corrupt JPEG data: bad arithmetic code") -JMESSAGE(JWRN_BOGUS_PROGRESSION, - "Inconsistent progression sequence for component %d coefficient %d") -JMESSAGE(JWRN_EXTRANEOUS_DATA, - "Corrupt JPEG data: %u extraneous bytes before marker 0x%02x") -JMESSAGE(JWRN_HIT_MARKER, "Corrupt JPEG data: premature end of data segment") -JMESSAGE(JWRN_HUFF_BAD_CODE, "Corrupt JPEG data: bad Huffman code") -JMESSAGE(JWRN_JFIF_MAJOR, "Warning: unknown JFIF revision number %d.%02d") -JMESSAGE(JWRN_JPEG_EOF, "Premature end of JPEG file") -JMESSAGE(JWRN_MUST_RESYNC, - "Corrupt JPEG data: found marker 0x%02x instead of RST%d") -JMESSAGE(JWRN_NOT_SEQUENTIAL, "Invalid SOS parameters for sequential JPEG") -JMESSAGE(JWRN_TOO_MUCH_DATA, "Application transferred too many scanlines") - -#ifdef JMAKE_ENUM_LIST - - JMSG_LASTMSGCODE -} J_MESSAGE_CODE; - -#undef JMAKE_ENUM_LIST -#endif /* JMAKE_ENUM_LIST */ - -/* Zap JMESSAGE macro so that future re-inclusions do nothing by default */ -#undef JMESSAGE - - -#ifndef JERROR_H -#define JERROR_H - -/* Macros to simplify using the error and trace message stuff */ -/* The first parameter is either type of cinfo pointer */ - -/* Fatal errors (print message and exit) */ -#define ERREXIT(cinfo,code) \ - ((cinfo)->err->msg_code = (code), \ - (*(cinfo)->err->error_exit) ((j_common_ptr) (cinfo))) -#define ERREXIT1(cinfo,code,p1) \ - ((cinfo)->err->msg_code = (code), \ - (cinfo)->err->msg_parm.i[0] = (p1), \ - (*(cinfo)->err->error_exit) ((j_common_ptr) (cinfo))) -#define ERREXIT2(cinfo,code,p1,p2) \ - ((cinfo)->err->msg_code = (code), \ - (cinfo)->err->msg_parm.i[0] = (p1), \ - (cinfo)->err->msg_parm.i[1] = (p2), \ - (*(cinfo)->err->error_exit) ((j_common_ptr) (cinfo))) -#define ERREXIT3(cinfo,code,p1,p2,p3) \ - ((cinfo)->err->msg_code = (code), \ - (cinfo)->err->msg_parm.i[0] = (p1), \ - (cinfo)->err->msg_parm.i[1] = (p2), \ - (cinfo)->err->msg_parm.i[2] = (p3), \ - (*(cinfo)->err->error_exit) ((j_common_ptr) (cinfo))) -#define ERREXIT4(cinfo,code,p1,p2,p3,p4) \ - ((cinfo)->err->msg_code = (code), \ - (cinfo)->err->msg_parm.i[0] = (p1), \ - (cinfo)->err->msg_parm.i[1] = (p2), \ - (cinfo)->err->msg_parm.i[2] = (p3), \ - (cinfo)->err->msg_parm.i[3] = (p4), \ - (*(cinfo)->err->error_exit) ((j_common_ptr) (cinfo))) -#define ERREXIT6(cinfo,code,p1,p2,p3,p4,p5,p6) \ - ((cinfo)->err->msg_code = (code), \ - (cinfo)->err->msg_parm.i[0] = (p1), \ - (cinfo)->err->msg_parm.i[1] = (p2), \ - (cinfo)->err->msg_parm.i[2] = (p3), \ - (cinfo)->err->msg_parm.i[3] = (p4), \ - (cinfo)->err->msg_parm.i[4] = (p5), \ - (cinfo)->err->msg_parm.i[5] = (p6), \ - (*(cinfo)->err->error_exit) ((j_common_ptr) (cinfo))) -#define ERREXITS(cinfo,code,str) \ - ((cinfo)->err->msg_code = (code), \ - strncpy((cinfo)->err->msg_parm.s, (str), JMSG_STR_PARM_MAX), \ - (*(cinfo)->err->error_exit) ((j_common_ptr) (cinfo))) - -#define MAKESTMT(stuff) do { stuff } while (0) - -/* Nonfatal errors (we can keep going, but the data is probably corrupt) */ -#define WARNMS(cinfo,code) \ - ((cinfo)->err->msg_code = (code), \ - (*(cinfo)->err->emit_message) ((j_common_ptr) (cinfo), -1)) -#define WARNMS1(cinfo,code,p1) \ - ((cinfo)->err->msg_code = (code), \ - (cinfo)->err->msg_parm.i[0] = (p1), \ - (*(cinfo)->err->emit_message) ((j_common_ptr) (cinfo), -1)) -#define WARNMS2(cinfo,code,p1,p2) \ - ((cinfo)->err->msg_code = (code), \ - (cinfo)->err->msg_parm.i[0] = (p1), \ - (cinfo)->err->msg_parm.i[1] = (p2), \ - (*(cinfo)->err->emit_message) ((j_common_ptr) (cinfo), -1)) - -/* Informational/debugging messages */ -#define TRACEMS(cinfo,lvl,code) \ - ((cinfo)->err->msg_code = (code), \ - (*(cinfo)->err->emit_message) ((j_common_ptr) (cinfo), (lvl))) -#define TRACEMS1(cinfo,lvl,code,p1) \ - ((cinfo)->err->msg_code = (code), \ - (cinfo)->err->msg_parm.i[0] = (p1), \ - (*(cinfo)->err->emit_message) ((j_common_ptr) (cinfo), (lvl))) -#define TRACEMS2(cinfo,lvl,code,p1,p2) \ - ((cinfo)->err->msg_code = (code), \ - (cinfo)->err->msg_parm.i[0] = (p1), \ - (cinfo)->err->msg_parm.i[1] = (p2), \ - (*(cinfo)->err->emit_message) ((j_common_ptr) (cinfo), (lvl))) -#define TRACEMS3(cinfo,lvl,code,p1,p2,p3) \ - MAKESTMT(int * _mp = (cinfo)->err->msg_parm.i; \ - _mp[0] = (p1); _mp[1] = (p2); _mp[2] = (p3); \ - (cinfo)->err->msg_code = (code); \ - (*(cinfo)->err->emit_message) ((j_common_ptr) (cinfo), (lvl)); ) -#define TRACEMS4(cinfo,lvl,code,p1,p2,p3,p4) \ - MAKESTMT(int * _mp = (cinfo)->err->msg_parm.i; \ - _mp[0] = (p1); _mp[1] = (p2); _mp[2] = (p3); _mp[3] = (p4); \ - (cinfo)->err->msg_code = (code); \ - (*(cinfo)->err->emit_message) ((j_common_ptr) (cinfo), (lvl)); ) -#define TRACEMS5(cinfo,lvl,code,p1,p2,p3,p4,p5) \ - MAKESTMT(int * _mp = (cinfo)->err->msg_parm.i; \ - _mp[0] = (p1); _mp[1] = (p2); _mp[2] = (p3); _mp[3] = (p4); \ - _mp[4] = (p5); \ - (cinfo)->err->msg_code = (code); \ - (*(cinfo)->err->emit_message) ((j_common_ptr) (cinfo), (lvl)); ) -#define TRACEMS8(cinfo,lvl,code,p1,p2,p3,p4,p5,p6,p7,p8) \ - MAKESTMT(int * _mp = (cinfo)->err->msg_parm.i; \ - _mp[0] = (p1); _mp[1] = (p2); _mp[2] = (p3); _mp[3] = (p4); \ - _mp[4] = (p5); _mp[5] = (p6); _mp[6] = (p7); _mp[7] = (p8); \ - (cinfo)->err->msg_code = (code); \ - (*(cinfo)->err->emit_message) ((j_common_ptr) (cinfo), (lvl)); ) -#define TRACEMSS(cinfo,lvl,code,str) \ - ((cinfo)->err->msg_code = (code), \ - strncpy((cinfo)->err->msg_parm.s, (str), JMSG_STR_PARM_MAX), \ - (*(cinfo)->err->emit_message) ((j_common_ptr) (cinfo), (lvl))) - -#endif /* JERROR_H */ diff --git a/cmake/externals/libjpeg/jfdctflt.c b/cmake/externals/libjpeg/jfdctflt.c deleted file mode 100644 index 013f29e0c..000000000 --- a/cmake/externals/libjpeg/jfdctflt.c +++ /dev/null @@ -1,176 +0,0 @@ -/* - * jfdctflt.c - * - * Copyright (C) 1994-1996, Thomas G. Lane. - * Modified 2003-2017 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains a floating-point implementation of the - * forward DCT (Discrete Cosine Transform). - * - * This implementation should be more accurate than either of the integer - * DCT implementations. However, it may not give the same results on all - * machines because of differences in roundoff behavior. Speed will depend - * on the hardware's floating point capacity. - * - * A 2-D DCT can be done by 1-D DCT on each row followed by 1-D DCT - * on each column. Direct algorithms are also available, but they are - * much more complex and seem not to be any faster when reduced to code. - * - * This implementation is based on Arai, Agui, and Nakajima's algorithm for - * scaled DCT. Their original paper (Trans. IEICE E-71(11):1095) is in - * Japanese, but the algorithm is described in the Pennebaker & Mitchell - * JPEG textbook (see REFERENCES section in file README). The following code - * is based directly on figure 4-8 in P&M. - * While an 8-point DCT cannot be done in less than 11 multiplies, it is - * possible to arrange the computation so that many of the multiplies are - * simple scalings of the final outputs. These multiplies can then be - * folded into the multiplications or divisions by the JPEG quantization - * table entries. The AA&N method leaves only 5 multiplies and 29 adds - * to be done in the DCT itself. - * The primary disadvantage of this method is that with a fixed-point - * implementation, accuracy is lost due to imprecise representation of the - * scaled quantization values. However, that problem does not arise if - * we use floating point arithmetic. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" -#include "jdct.h" /* Private declarations for DCT subsystem */ - -#ifdef DCT_FLOAT_SUPPORTED - - -/* - * This module is specialized to the case DCTSIZE = 8. - */ - -#if DCTSIZE != 8 - Sorry, this code only copes with 8x8 DCT blocks. /* deliberate syntax err */ -#endif - - -/* - * Perform the forward DCT on one block of samples. - * - * cK represents cos(K*pi/16). - */ - -GLOBAL(void) -jpeg_fdct_float (FAST_FLOAT * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - FAST_FLOAT tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7; - FAST_FLOAT tmp10, tmp11, tmp12, tmp13; - FAST_FLOAT z1, z2, z3, z4, z5, z11, z13; - FAST_FLOAT *dataptr; - JSAMPROW elemptr; - int ctr; - - /* Pass 1: process rows. */ - - dataptr = data; - for (ctr = 0; ctr < DCTSIZE; ctr++) { - elemptr = sample_data[ctr] + start_col; - - /* Load data into workspace */ - tmp0 = (FAST_FLOAT) (GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[7])); - tmp7 = (FAST_FLOAT) (GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[7])); - tmp1 = (FAST_FLOAT) (GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[6])); - tmp6 = (FAST_FLOAT) (GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[6])); - tmp2 = (FAST_FLOAT) (GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[5])); - tmp5 = (FAST_FLOAT) (GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[5])); - tmp3 = (FAST_FLOAT) (GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[4])); - tmp4 = (FAST_FLOAT) (GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[4])); - - /* Even part */ - - tmp10 = tmp0 + tmp3; /* phase 2 */ - tmp13 = tmp0 - tmp3; - tmp11 = tmp1 + tmp2; - tmp12 = tmp1 - tmp2; - - /* Apply unsigned->signed conversion. */ - dataptr[0] = tmp10 + tmp11 - 8 * CENTERJSAMPLE; /* phase 3 */ - dataptr[4] = tmp10 - tmp11; - - z1 = (tmp12 + tmp13) * ((FAST_FLOAT) 0.707106781); /* c4 */ - dataptr[2] = tmp13 + z1; /* phase 5 */ - dataptr[6] = tmp13 - z1; - - /* Odd part */ - - tmp10 = tmp4 + tmp5; /* phase 2 */ - tmp11 = tmp5 + tmp6; - tmp12 = tmp6 + tmp7; - - /* The rotator is modified from fig 4-8 to avoid extra negations. */ - z5 = (tmp10 - tmp12) * ((FAST_FLOAT) 0.382683433); /* c6 */ - z2 = ((FAST_FLOAT) 0.541196100) * tmp10 + z5; /* c2-c6 */ - z4 = ((FAST_FLOAT) 1.306562965) * tmp12 + z5; /* c2+c6 */ - z3 = tmp11 * ((FAST_FLOAT) 0.707106781); /* c4 */ - - z11 = tmp7 + z3; /* phase 5 */ - z13 = tmp7 - z3; - - dataptr[5] = z13 + z2; /* phase 6 */ - dataptr[3] = z13 - z2; - dataptr[1] = z11 + z4; - dataptr[7] = z11 - z4; - - dataptr += DCTSIZE; /* advance pointer to next row */ - } - - /* Pass 2: process columns. */ - - dataptr = data; - for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { - tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*7]; - tmp7 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*7]; - tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*6]; - tmp6 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*6]; - tmp2 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*5]; - tmp5 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*5]; - tmp3 = dataptr[DCTSIZE*3] + dataptr[DCTSIZE*4]; - tmp4 = dataptr[DCTSIZE*3] - dataptr[DCTSIZE*4]; - - /* Even part */ - - tmp10 = tmp0 + tmp3; /* phase 2 */ - tmp13 = tmp0 - tmp3; - tmp11 = tmp1 + tmp2; - tmp12 = tmp1 - tmp2; - - dataptr[DCTSIZE*0] = tmp10 + tmp11; /* phase 3 */ - dataptr[DCTSIZE*4] = tmp10 - tmp11; - - z1 = (tmp12 + tmp13) * ((FAST_FLOAT) 0.707106781); /* c4 */ - dataptr[DCTSIZE*2] = tmp13 + z1; /* phase 5 */ - dataptr[DCTSIZE*6] = tmp13 - z1; - - /* Odd part */ - - tmp10 = tmp4 + tmp5; /* phase 2 */ - tmp11 = tmp5 + tmp6; - tmp12 = tmp6 + tmp7; - - /* The rotator is modified from fig 4-8 to avoid extra negations. */ - z5 = (tmp10 - tmp12) * ((FAST_FLOAT) 0.382683433); /* c6 */ - z2 = ((FAST_FLOAT) 0.541196100) * tmp10 + z5; /* c2-c6 */ - z4 = ((FAST_FLOAT) 1.306562965) * tmp12 + z5; /* c2+c6 */ - z3 = tmp11 * ((FAST_FLOAT) 0.707106781); /* c4 */ - - z11 = tmp7 + z3; /* phase 5 */ - z13 = tmp7 - z3; - - dataptr[DCTSIZE*5] = z13 + z2; /* phase 6 */ - dataptr[DCTSIZE*3] = z13 - z2; - dataptr[DCTSIZE*1] = z11 + z4; - dataptr[DCTSIZE*7] = z11 - z4; - - dataptr++; /* advance pointer to next column */ - } -} - -#endif /* DCT_FLOAT_SUPPORTED */ diff --git a/cmake/externals/libjpeg/jfdctfst.c b/cmake/externals/libjpeg/jfdctfst.c deleted file mode 100644 index 5e4e01722..000000000 --- a/cmake/externals/libjpeg/jfdctfst.c +++ /dev/null @@ -1,232 +0,0 @@ -/* - * jfdctfst.c - * - * Copyright (C) 1994-1996, Thomas G. Lane. - * Modified 2003-2017 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains a fast, not so accurate integer implementation of the - * forward DCT (Discrete Cosine Transform). - * - * A 2-D DCT can be done by 1-D DCT on each row followed by 1-D DCT - * on each column. Direct algorithms are also available, but they are - * much more complex and seem not to be any faster when reduced to code. - * - * This implementation is based on Arai, Agui, and Nakajima's algorithm for - * scaled DCT. Their original paper (Trans. IEICE E-71(11):1095) is in - * Japanese, but the algorithm is described in the Pennebaker & Mitchell - * JPEG textbook (see REFERENCES section in file README). The following code - * is based directly on figure 4-8 in P&M. - * While an 8-point DCT cannot be done in less than 11 multiplies, it is - * possible to arrange the computation so that many of the multiplies are - * simple scalings of the final outputs. These multiplies can then be - * folded into the multiplications or divisions by the JPEG quantization - * table entries. The AA&N method leaves only 5 multiplies and 29 adds - * to be done in the DCT itself. - * The primary disadvantage of this method is that with fixed-point math, - * accuracy is lost due to imprecise representation of the scaled - * quantization values. The smaller the quantization table entry, the less - * precise the scaled value, so this implementation does worse with high- - * quality-setting files than with low-quality ones. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" -#include "jdct.h" /* Private declarations for DCT subsystem */ - -#ifdef DCT_IFAST_SUPPORTED - - -/* - * This module is specialized to the case DCTSIZE = 8. - */ - -#if DCTSIZE != 8 - Sorry, this code only copes with 8x8 DCT blocks. /* deliberate syntax err */ -#endif - - -/* Scaling decisions are generally the same as in the LL&M algorithm; - * see jfdctint.c for more details. However, we choose to descale - * (right shift) multiplication products as soon as they are formed, - * rather than carrying additional fractional bits into subsequent additions. - * This compromises accuracy slightly, but it lets us save a few shifts. - * More importantly, 16-bit arithmetic is then adequate (for 8-bit samples) - * everywhere except in the multiplications proper; this saves a good deal - * of work on 16-bit-int machines. - * - * Again to save a few shifts, the intermediate results between pass 1 and - * pass 2 are not upscaled, but are represented only to integral precision. - * - * A final compromise is to represent the multiplicative constants to only - * 8 fractional bits, rather than 13. This saves some shifting work on some - * machines, and may also reduce the cost of multiplication (since there - * are fewer one-bits in the constants). - */ - -#define CONST_BITS 8 - - -/* Some C compilers fail to reduce "FIX(constant)" at compile time, thus - * causing a lot of useless floating-point operations at run time. - * To get around this we use the following pre-calculated constants. - * If you change CONST_BITS you may want to add appropriate values. - * (With a reasonable C compiler, you can just rely on the FIX() macro...) - */ - -#if CONST_BITS == 8 -#define FIX_0_382683433 ((INT32) 98) /* FIX(0.382683433) */ -#define FIX_0_541196100 ((INT32) 139) /* FIX(0.541196100) */ -#define FIX_0_707106781 ((INT32) 181) /* FIX(0.707106781) */ -#define FIX_1_306562965 ((INT32) 334) /* FIX(1.306562965) */ -#else -#define FIX_0_382683433 FIX(0.382683433) -#define FIX_0_541196100 FIX(0.541196100) -#define FIX_0_707106781 FIX(0.707106781) -#define FIX_1_306562965 FIX(1.306562965) -#endif - - -/* We can gain a little more speed, with a further compromise in accuracy, - * by omitting the addition in a descaling shift. This yields an incorrectly - * rounded result half the time... - */ - -#ifndef USE_ACCURATE_ROUNDING -#undef DESCALE -#define DESCALE(x,n) RIGHT_SHIFT(x, n) -#endif - - -/* Multiply a DCTELEM variable by an INT32 constant, and immediately - * descale to yield a DCTELEM result. - */ - -#define MULTIPLY(var,const) ((DCTELEM) DESCALE((var) * (const), CONST_BITS)) - - -/* - * Perform the forward DCT on one block of samples. - * - * cK represents cos(K*pi/16). - */ - -GLOBAL(void) -jpeg_fdct_ifast (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - DCTELEM tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7; - DCTELEM tmp10, tmp11, tmp12, tmp13; - DCTELEM z1, z2, z3, z4, z5, z11, z13; - DCTELEM *dataptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pass 1: process rows. */ - - dataptr = data; - for (ctr = 0; ctr < DCTSIZE; ctr++) { - elemptr = sample_data[ctr] + start_col; - - /* Load data into workspace */ - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[7]); - tmp7 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[7]); - tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[6]); - tmp6 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[6]); - tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[5]); - tmp5 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[5]); - tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[4]); - tmp4 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[4]); - - /* Even part */ - - tmp10 = tmp0 + tmp3; /* phase 2 */ - tmp13 = tmp0 - tmp3; - tmp11 = tmp1 + tmp2; - tmp12 = tmp1 - tmp2; - - /* Apply unsigned->signed conversion. */ - dataptr[0] = tmp10 + tmp11 - 8 * CENTERJSAMPLE; /* phase 3 */ - dataptr[4] = tmp10 - tmp11; - - z1 = MULTIPLY(tmp12 + tmp13, FIX_0_707106781); /* c4 */ - dataptr[2] = tmp13 + z1; /* phase 5 */ - dataptr[6] = tmp13 - z1; - - /* Odd part */ - - tmp10 = tmp4 + tmp5; /* phase 2 */ - tmp11 = tmp5 + tmp6; - tmp12 = tmp6 + tmp7; - - /* The rotator is modified from fig 4-8 to avoid extra negations. */ - z5 = MULTIPLY(tmp10 - tmp12, FIX_0_382683433); /* c6 */ - z2 = MULTIPLY(tmp10, FIX_0_541196100) + z5; /* c2-c6 */ - z4 = MULTIPLY(tmp12, FIX_1_306562965) + z5; /* c2+c6 */ - z3 = MULTIPLY(tmp11, FIX_0_707106781); /* c4 */ - - z11 = tmp7 + z3; /* phase 5 */ - z13 = tmp7 - z3; - - dataptr[5] = z13 + z2; /* phase 6 */ - dataptr[3] = z13 - z2; - dataptr[1] = z11 + z4; - dataptr[7] = z11 - z4; - - dataptr += DCTSIZE; /* advance pointer to next row */ - } - - /* Pass 2: process columns. */ - - dataptr = data; - for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { - tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*7]; - tmp7 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*7]; - tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*6]; - tmp6 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*6]; - tmp2 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*5]; - tmp5 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*5]; - tmp3 = dataptr[DCTSIZE*3] + dataptr[DCTSIZE*4]; - tmp4 = dataptr[DCTSIZE*3] - dataptr[DCTSIZE*4]; - - /* Even part */ - - tmp10 = tmp0 + tmp3; /* phase 2 */ - tmp13 = tmp0 - tmp3; - tmp11 = tmp1 + tmp2; - tmp12 = tmp1 - tmp2; - - dataptr[DCTSIZE*0] = tmp10 + tmp11; /* phase 3 */ - dataptr[DCTSIZE*4] = tmp10 - tmp11; - - z1 = MULTIPLY(tmp12 + tmp13, FIX_0_707106781); /* c4 */ - dataptr[DCTSIZE*2] = tmp13 + z1; /* phase 5 */ - dataptr[DCTSIZE*6] = tmp13 - z1; - - /* Odd part */ - - tmp10 = tmp4 + tmp5; /* phase 2 */ - tmp11 = tmp5 + tmp6; - tmp12 = tmp6 + tmp7; - - /* The rotator is modified from fig 4-8 to avoid extra negations. */ - z5 = MULTIPLY(tmp10 - tmp12, FIX_0_382683433); /* c6 */ - z2 = MULTIPLY(tmp10, FIX_0_541196100) + z5; /* c2-c6 */ - z4 = MULTIPLY(tmp12, FIX_1_306562965) + z5; /* c2+c6 */ - z3 = MULTIPLY(tmp11, FIX_0_707106781); /* c4 */ - - z11 = tmp7 + z3; /* phase 5 */ - z13 = tmp7 - z3; - - dataptr[DCTSIZE*5] = z13 + z2; /* phase 6 */ - dataptr[DCTSIZE*3] = z13 - z2; - dataptr[DCTSIZE*1] = z11 + z4; - dataptr[DCTSIZE*7] = z11 - z4; - - dataptr++; /* advance pointer to next column */ - } -} - -#endif /* DCT_IFAST_SUPPORTED */ diff --git a/cmake/externals/libjpeg/jfdctint.c b/cmake/externals/libjpeg/jfdctint.c deleted file mode 100644 index 05df4750b..000000000 --- a/cmake/externals/libjpeg/jfdctint.c +++ /dev/null @@ -1,4415 +0,0 @@ -/* - * jfdctint.c - * - * Copyright (C) 1991-1996, Thomas G. Lane. - * Modification developed 2003-2018 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains a slow-but-accurate integer implementation of the - * forward DCT (Discrete Cosine Transform). - * - * A 2-D DCT can be done by 1-D DCT on each row followed by 1-D DCT - * on each column. Direct algorithms are also available, but they are - * much more complex and seem not to be any faster when reduced to code. - * - * This implementation is based on an algorithm described in - * C. Loeffler, A. Ligtenberg and G. Moschytz, "Practical Fast 1-D DCT - * Algorithms with 11 Multiplications", Proc. Int'l. Conf. on Acoustics, - * Speech, and Signal Processing 1989 (ICASSP '89), pp. 988-991. - * The primary algorithm described there uses 11 multiplies and 29 adds. - * We use their alternate method with 12 multiplies and 32 adds. - * The advantage of this method is that no data path contains more than one - * multiplication; this allows a very simple and accurate implementation in - * scaled fixed-point arithmetic, with a minimal number of shifts. - * - * We also provide FDCT routines with various input sample block sizes for - * direct resolution reduction or enlargement and for direct resolving the - * common 2x1 and 1x2 subsampling cases without additional resampling: NxN - * (N=1...16), 2NxN, and Nx2N (N=1...8) pixels for one 8x8 output DCT block. - * - * For N<8 we fill the remaining block coefficients with zero. - * For N>8 we apply a partial N-point FDCT on the input samples, computing - * just the lower 8 frequency coefficients and discarding the rest. - * - * We must scale the output coefficients of the N-point FDCT appropriately - * to the standard 8-point FDCT level by 8/N per 1-D pass. This scaling - * is folded into the constant multipliers (pass 2) and/or final/initial - * shifting. - * - * CAUTION: We rely on the FIX() macro except for the N=1,2,4,8 cases - * since there would be too many additional constants to pre-calculate. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" -#include "jdct.h" /* Private declarations for DCT subsystem */ - -#ifdef DCT_ISLOW_SUPPORTED - - -/* - * This module is specialized to the case DCTSIZE = 8. - */ - -#if DCTSIZE != 8 - Sorry, this code only copes with 8x8 DCT blocks. /* deliberate syntax err */ -#endif - - -/* - * The poop on this scaling stuff is as follows: - * - * Each 1-D DCT step produces outputs which are a factor of sqrt(N) - * larger than the true DCT outputs. The final outputs are therefore - * a factor of N larger than desired; since N=8 this can be cured by - * a simple right shift at the end of the algorithm. The advantage of - * this arrangement is that we save two multiplications per 1-D DCT, - * because the y0 and y4 outputs need not be divided by sqrt(N). - * In the IJG code, this factor of 8 is removed by the quantization step - * (in jcdctmgr.c), NOT in this module. - * - * We have to do addition and subtraction of the integer inputs, which - * is no problem, and multiplication by fractional constants, which is - * a problem to do in integer arithmetic. We multiply all the constants - * by CONST_SCALE and convert them to integer constants (thus retaining - * CONST_BITS bits of precision in the constants). After doing a - * multiplication we have to divide the product by CONST_SCALE, with proper - * rounding, to produce the correct output. This division can be done - * cheaply as a right shift of CONST_BITS bits. We postpone shifting - * as long as possible so that partial sums can be added together with - * full fractional precision. - * - * The outputs of the first pass are scaled up by PASS1_BITS bits so that - * they are represented to better-than-integral precision. These outputs - * require BITS_IN_JSAMPLE + PASS1_BITS + 3 bits; this fits in a 16-bit word - * with the recommended scaling. (For 12-bit sample data, the intermediate - * array is INT32 anyway.) - * - * To avoid overflow of the 32-bit intermediate results in pass 2, we must - * have BITS_IN_JSAMPLE + CONST_BITS + PASS1_BITS <= 26. Error analysis - * shows that the values given below are the most effective. - */ - -#if BITS_IN_JSAMPLE == 8 -#define CONST_BITS 13 -#define PASS1_BITS 2 -#else -#define CONST_BITS 13 -#define PASS1_BITS 1 /* lose a little precision to avoid overflow */ -#endif - -/* Some C compilers fail to reduce "FIX(constant)" at compile time, thus - * causing a lot of useless floating-point operations at run time. - * To get around this we use the following pre-calculated constants. - * If you change CONST_BITS you may want to add appropriate values. - * (With a reasonable C compiler, you can just rely on the FIX() macro...) - */ - -#if CONST_BITS == 13 -#define FIX_0_298631336 ((INT32) 2446) /* FIX(0.298631336) */ -#define FIX_0_390180644 ((INT32) 3196) /* FIX(0.390180644) */ -#define FIX_0_541196100 ((INT32) 4433) /* FIX(0.541196100) */ -#define FIX_0_765366865 ((INT32) 6270) /* FIX(0.765366865) */ -#define FIX_0_899976223 ((INT32) 7373) /* FIX(0.899976223) */ -#define FIX_1_175875602 ((INT32) 9633) /* FIX(1.175875602) */ -#define FIX_1_501321110 ((INT32) 12299) /* FIX(1.501321110) */ -#define FIX_1_847759065 ((INT32) 15137) /* FIX(1.847759065) */ -#define FIX_1_961570560 ((INT32) 16069) /* FIX(1.961570560) */ -#define FIX_2_053119869 ((INT32) 16819) /* FIX(2.053119869) */ -#define FIX_2_562915447 ((INT32) 20995) /* FIX(2.562915447) */ -#define FIX_3_072711026 ((INT32) 25172) /* FIX(3.072711026) */ -#else -#define FIX_0_298631336 FIX(0.298631336) -#define FIX_0_390180644 FIX(0.390180644) -#define FIX_0_541196100 FIX(0.541196100) -#define FIX_0_765366865 FIX(0.765366865) -#define FIX_0_899976223 FIX(0.899976223) -#define FIX_1_175875602 FIX(1.175875602) -#define FIX_1_501321110 FIX(1.501321110) -#define FIX_1_847759065 FIX(1.847759065) -#define FIX_1_961570560 FIX(1.961570560) -#define FIX_2_053119869 FIX(2.053119869) -#define FIX_2_562915447 FIX(2.562915447) -#define FIX_3_072711026 FIX(3.072711026) -#endif - - -/* Multiply an INT32 variable by an INT32 constant to yield an INT32 result. - * For 8-bit samples with the recommended scaling, all the variable - * and constant values involved are no more than 16 bits wide, so a - * 16x16->32 bit multiply can be used instead of a full 32x32 multiply. - * For 12-bit samples, a full 32-bit multiplication will be needed. - */ - -#if BITS_IN_JSAMPLE == 8 -#define MULTIPLY(var,const) MULTIPLY16C16(var,const) -#else -#define MULTIPLY(var,const) ((var) * (const)) -#endif - - -/* - * Perform the forward DCT on one block of samples. - */ - -GLOBAL(void) -jpeg_fdct_islow (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3; - INT32 tmp10, tmp11, tmp12, tmp13; - INT32 z1; - DCTELEM *dataptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT; - * furthermore, we scale the results by 2**PASS1_BITS. - * cK represents sqrt(2) * cos(K*pi/16). - */ - - dataptr = data; - for (ctr = 0; ctr < DCTSIZE; ctr++) { - elemptr = sample_data[ctr] + start_col; - - /* Even part per LL&M figure 1 --- note that published figure is faulty; - * rotator "c1" should be "c6". - */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[7]); - tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[6]); - tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[5]); - tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[4]); - - tmp10 = tmp0 + tmp3; - tmp12 = tmp0 - tmp3; - tmp11 = tmp1 + tmp2; - tmp13 = tmp1 - tmp2; - - tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[7]); - tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[6]); - tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[5]); - tmp3 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[4]); - - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) ((tmp10 + tmp11 - 8 * CENTERJSAMPLE) << PASS1_BITS); - dataptr[4] = (DCTELEM) ((tmp10 - tmp11) << PASS1_BITS); - - z1 = MULTIPLY(tmp12 + tmp13, FIX_0_541196100); /* c6 */ - /* Add fudge factor here for final descale. */ - z1 += ONE << (CONST_BITS-PASS1_BITS-1); - - dataptr[2] = (DCTELEM) - RIGHT_SHIFT(z1 + MULTIPLY(tmp12, FIX_0_765366865), /* c2-c6 */ - CONST_BITS-PASS1_BITS); - dataptr[6] = (DCTELEM) - RIGHT_SHIFT(z1 - MULTIPLY(tmp13, FIX_1_847759065), /* c2+c6 */ - CONST_BITS-PASS1_BITS); - - /* Odd part per figure 8 --- note paper omits factor of sqrt(2). - * i0..i3 in the paper are tmp0..tmp3 here. - */ - - tmp12 = tmp0 + tmp2; - tmp13 = tmp1 + tmp3; - - z1 = MULTIPLY(tmp12 + tmp13, FIX_1_175875602); /* c3 */ - /* Add fudge factor here for final descale. */ - z1 += ONE << (CONST_BITS-PASS1_BITS-1); - - tmp12 = MULTIPLY(tmp12, - FIX_0_390180644); /* -c3+c5 */ - tmp13 = MULTIPLY(tmp13, - FIX_1_961570560); /* -c3-c5 */ - tmp12 += z1; - tmp13 += z1; - - z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ - tmp0 = MULTIPLY(tmp0, FIX_1_501321110); /* c1+c3-c5-c7 */ - tmp3 = MULTIPLY(tmp3, FIX_0_298631336); /* -c1+c3+c5-c7 */ - tmp0 += z1 + tmp12; - tmp3 += z1 + tmp13; - - z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ - tmp1 = MULTIPLY(tmp1, FIX_3_072711026); /* c1+c3+c5-c7 */ - tmp2 = MULTIPLY(tmp2, FIX_2_053119869); /* c1+c3-c5+c7 */ - tmp1 += z1 + tmp13; - tmp2 += z1 + tmp12; - - dataptr[1] = (DCTELEM) RIGHT_SHIFT(tmp0, CONST_BITS-PASS1_BITS); - dataptr[3] = (DCTELEM) RIGHT_SHIFT(tmp1, CONST_BITS-PASS1_BITS); - dataptr[5] = (DCTELEM) RIGHT_SHIFT(tmp2, CONST_BITS-PASS1_BITS); - dataptr[7] = (DCTELEM) RIGHT_SHIFT(tmp3, CONST_BITS-PASS1_BITS); - - dataptr += DCTSIZE; /* advance pointer to next row */ - } - - /* Pass 2: process columns. - * We remove the PASS1_BITS scaling, but leave the results scaled up - * by an overall factor of 8. - * cK represents sqrt(2) * cos(K*pi/16). - */ - - dataptr = data; - for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { - /* Even part per LL&M figure 1 --- note that published figure is faulty; - * rotator "c1" should be "c6". - */ - - tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*7]; - tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*6]; - tmp2 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*5]; - tmp3 = dataptr[DCTSIZE*3] + dataptr[DCTSIZE*4]; - - /* Add fudge factor here for final descale. */ - tmp10 = tmp0 + tmp3 + (ONE << (PASS1_BITS-1)); - tmp12 = tmp0 - tmp3; - tmp11 = tmp1 + tmp2; - tmp13 = tmp1 - tmp2; - - tmp0 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*7]; - tmp1 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*6]; - tmp2 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*5]; - tmp3 = dataptr[DCTSIZE*3] - dataptr[DCTSIZE*4]; - - dataptr[DCTSIZE*0] = (DCTELEM) RIGHT_SHIFT(tmp10 + tmp11, PASS1_BITS); - dataptr[DCTSIZE*4] = (DCTELEM) RIGHT_SHIFT(tmp10 - tmp11, PASS1_BITS); - - z1 = MULTIPLY(tmp12 + tmp13, FIX_0_541196100); /* c6 */ - /* Add fudge factor here for final descale. */ - z1 += ONE << (CONST_BITS+PASS1_BITS-1); - - dataptr[DCTSIZE*2] = (DCTELEM) - RIGHT_SHIFT(z1 + MULTIPLY(tmp12, FIX_0_765366865), /* c2-c6 */ - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*6] = (DCTELEM) - RIGHT_SHIFT(z1 - MULTIPLY(tmp13, FIX_1_847759065), /* c2+c6 */ - CONST_BITS+PASS1_BITS); - - /* Odd part per figure 8 --- note paper omits factor of sqrt(2). - * i0..i3 in the paper are tmp0..tmp3 here. - */ - - tmp12 = tmp0 + tmp2; - tmp13 = tmp1 + tmp3; - - z1 = MULTIPLY(tmp12 + tmp13, FIX_1_175875602); /* c3 */ - /* Add fudge factor here for final descale. */ - z1 += ONE << (CONST_BITS+PASS1_BITS-1); - - tmp12 = MULTIPLY(tmp12, - FIX_0_390180644); /* -c3+c5 */ - tmp13 = MULTIPLY(tmp13, - FIX_1_961570560); /* -c3-c5 */ - tmp12 += z1; - tmp13 += z1; - - z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ - tmp0 = MULTIPLY(tmp0, FIX_1_501321110); /* c1+c3-c5-c7 */ - tmp3 = MULTIPLY(tmp3, FIX_0_298631336); /* -c1+c3+c5-c7 */ - tmp0 += z1 + tmp12; - tmp3 += z1 + tmp13; - - z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ - tmp1 = MULTIPLY(tmp1, FIX_3_072711026); /* c1+c3+c5-c7 */ - tmp2 = MULTIPLY(tmp2, FIX_2_053119869); /* c1+c3-c5+c7 */ - tmp1 += z1 + tmp13; - tmp2 += z1 + tmp12; - - dataptr[DCTSIZE*1] = (DCTELEM) RIGHT_SHIFT(tmp0, CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*3] = (DCTELEM) RIGHT_SHIFT(tmp1, CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*5] = (DCTELEM) RIGHT_SHIFT(tmp2, CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*7] = (DCTELEM) RIGHT_SHIFT(tmp3, CONST_BITS+PASS1_BITS); - - dataptr++; /* advance pointer to next column */ - } -} - -#ifdef DCT_SCALING_SUPPORTED - - -/* - * Perform the forward DCT on a 7x7 sample block. - */ - -GLOBAL(void) -jpeg_fdct_7x7 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3; - INT32 tmp10, tmp11, tmp12; - INT32 z1, z2, z3; - DCTELEM *dataptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pre-zero output coefficient block. */ - MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT; - * furthermore, we scale the results by 2**PASS1_BITS. - * cK represents sqrt(2) * cos(K*pi/14). - */ - - dataptr = data; - for (ctr = 0; ctr < 7; ctr++) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[6]); - tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[5]); - tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[4]); - tmp3 = GETJSAMPLE(elemptr[3]); - - tmp10 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[6]); - tmp11 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[5]); - tmp12 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[4]); - - z1 = tmp0 + tmp2; - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) - ((z1 + tmp1 + tmp3 - 7 * CENTERJSAMPLE) << PASS1_BITS); - tmp3 += tmp3; - z1 -= tmp3; - z1 -= tmp3; - z1 = MULTIPLY(z1, FIX(0.353553391)); /* (c2+c6-c4)/2 */ - z2 = MULTIPLY(tmp0 - tmp2, FIX(0.920609002)); /* (c2+c4-c6)/2 */ - z3 = MULTIPLY(tmp1 - tmp2, FIX(0.314692123)); /* c6 */ - dataptr[2] = (DCTELEM) DESCALE(z1 + z2 + z3, CONST_BITS-PASS1_BITS); - z1 -= z2; - z2 = MULTIPLY(tmp0 - tmp1, FIX(0.881747734)); /* c4 */ - dataptr[4] = (DCTELEM) - DESCALE(z2 + z3 - MULTIPLY(tmp1 - tmp3, FIX(0.707106781)), /* c2+c6-c4 */ - CONST_BITS-PASS1_BITS); - dataptr[6] = (DCTELEM) DESCALE(z1 + z2, CONST_BITS-PASS1_BITS); - - /* Odd part */ - - tmp1 = MULTIPLY(tmp10 + tmp11, FIX(0.935414347)); /* (c3+c1-c5)/2 */ - tmp2 = MULTIPLY(tmp10 - tmp11, FIX(0.170262339)); /* (c3+c5-c1)/2 */ - tmp0 = tmp1 - tmp2; - tmp1 += tmp2; - tmp2 = MULTIPLY(tmp11 + tmp12, - FIX(1.378756276)); /* -c1 */ - tmp1 += tmp2; - tmp3 = MULTIPLY(tmp10 + tmp12, FIX(0.613604268)); /* c5 */ - tmp0 += tmp3; - tmp2 += tmp3 + MULTIPLY(tmp12, FIX(1.870828693)); /* c3+c1-c5 */ - - dataptr[1] = (DCTELEM) DESCALE(tmp0, CONST_BITS-PASS1_BITS); - dataptr[3] = (DCTELEM) DESCALE(tmp1, CONST_BITS-PASS1_BITS); - dataptr[5] = (DCTELEM) DESCALE(tmp2, CONST_BITS-PASS1_BITS); - - dataptr += DCTSIZE; /* advance pointer to next row */ - } - - /* Pass 2: process columns. - * We remove the PASS1_BITS scaling, but leave the results scaled up - * by an overall factor of 8. - * We must also scale the output by (8/7)**2 = 64/49, which we fold - * into the constant multipliers: - * cK now represents sqrt(2) * cos(K*pi/14) * 64/49. - */ - - dataptr = data; - for (ctr = 0; ctr < 7; ctr++) { - /* Even part */ - - tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*6]; - tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*5]; - tmp2 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*4]; - tmp3 = dataptr[DCTSIZE*3]; - - tmp10 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*6]; - tmp11 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*5]; - tmp12 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*4]; - - z1 = tmp0 + tmp2; - dataptr[DCTSIZE*0] = (DCTELEM) - DESCALE(MULTIPLY(z1 + tmp1 + tmp3, FIX(1.306122449)), /* 64/49 */ - CONST_BITS+PASS1_BITS); - tmp3 += tmp3; - z1 -= tmp3; - z1 -= tmp3; - z1 = MULTIPLY(z1, FIX(0.461784020)); /* (c2+c6-c4)/2 */ - z2 = MULTIPLY(tmp0 - tmp2, FIX(1.202428084)); /* (c2+c4-c6)/2 */ - z3 = MULTIPLY(tmp1 - tmp2, FIX(0.411026446)); /* c6 */ - dataptr[DCTSIZE*2] = (DCTELEM) DESCALE(z1 + z2 + z3, CONST_BITS+PASS1_BITS); - z1 -= z2; - z2 = MULTIPLY(tmp0 - tmp1, FIX(1.151670509)); /* c4 */ - dataptr[DCTSIZE*4] = (DCTELEM) - DESCALE(z2 + z3 - MULTIPLY(tmp1 - tmp3, FIX(0.923568041)), /* c2+c6-c4 */ - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*6] = (DCTELEM) DESCALE(z1 + z2, CONST_BITS+PASS1_BITS); - - /* Odd part */ - - tmp1 = MULTIPLY(tmp10 + tmp11, FIX(1.221765677)); /* (c3+c1-c5)/2 */ - tmp2 = MULTIPLY(tmp10 - tmp11, FIX(0.222383464)); /* (c3+c5-c1)/2 */ - tmp0 = tmp1 - tmp2; - tmp1 += tmp2; - tmp2 = MULTIPLY(tmp11 + tmp12, - FIX(1.800824523)); /* -c1 */ - tmp1 += tmp2; - tmp3 = MULTIPLY(tmp10 + tmp12, FIX(0.801442310)); /* c5 */ - tmp0 += tmp3; - tmp2 += tmp3 + MULTIPLY(tmp12, FIX(2.443531355)); /* c3+c1-c5 */ - - dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp0, CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp1, CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp2, CONST_BITS+PASS1_BITS); - - dataptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 6x6 sample block. - */ - -GLOBAL(void) -jpeg_fdct_6x6 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2; - INT32 tmp10, tmp11, tmp12; - DCTELEM *dataptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pre-zero output coefficient block. */ - MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT; - * furthermore, we scale the results by 2**PASS1_BITS. - * cK represents sqrt(2) * cos(K*pi/12). - */ - - dataptr = data; - for (ctr = 0; ctr < 6; ctr++) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[5]); - tmp11 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[4]); - tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[3]); - - tmp10 = tmp0 + tmp2; - tmp12 = tmp0 - tmp2; - - tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[5]); - tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[4]); - tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[3]); - - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) - ((tmp10 + tmp11 - 6 * CENTERJSAMPLE) << PASS1_BITS); - dataptr[2] = (DCTELEM) - DESCALE(MULTIPLY(tmp12, FIX(1.224744871)), /* c2 */ - CONST_BITS-PASS1_BITS); - dataptr[4] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp11 - tmp11, FIX(0.707106781)), /* c4 */ - CONST_BITS-PASS1_BITS); - - /* Odd part */ - - tmp10 = DESCALE(MULTIPLY(tmp0 + tmp2, FIX(0.366025404)), /* c5 */ - CONST_BITS-PASS1_BITS); - - dataptr[1] = (DCTELEM) (tmp10 + ((tmp0 + tmp1) << PASS1_BITS)); - dataptr[3] = (DCTELEM) ((tmp0 - tmp1 - tmp2) << PASS1_BITS); - dataptr[5] = (DCTELEM) (tmp10 + ((tmp2 - tmp1) << PASS1_BITS)); - - dataptr += DCTSIZE; /* advance pointer to next row */ - } - - /* Pass 2: process columns. - * We remove the PASS1_BITS scaling, but leave the results scaled up - * by an overall factor of 8. - * We must also scale the output by (8/6)**2 = 16/9, which we fold - * into the constant multipliers: - * cK now represents sqrt(2) * cos(K*pi/12) * 16/9. - */ - - dataptr = data; - for (ctr = 0; ctr < 6; ctr++) { - /* Even part */ - - tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*5]; - tmp11 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*4]; - tmp2 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*3]; - - tmp10 = tmp0 + tmp2; - tmp12 = tmp0 - tmp2; - - tmp0 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*5]; - tmp1 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*4]; - tmp2 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*3]; - - dataptr[DCTSIZE*0] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 + tmp11, FIX(1.777777778)), /* 16/9 */ - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*2] = (DCTELEM) - DESCALE(MULTIPLY(tmp12, FIX(2.177324216)), /* c2 */ - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*4] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp11 - tmp11, FIX(1.257078722)), /* c4 */ - CONST_BITS+PASS1_BITS); - - /* Odd part */ - - tmp10 = MULTIPLY(tmp0 + tmp2, FIX(0.650711829)); /* c5 */ - - dataptr[DCTSIZE*1] = (DCTELEM) - DESCALE(tmp10 + MULTIPLY(tmp0 + tmp1, FIX(1.777777778)), /* 16/9 */ - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*3] = (DCTELEM) - DESCALE(MULTIPLY(tmp0 - tmp1 - tmp2, FIX(1.777777778)), /* 16/9 */ - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*5] = (DCTELEM) - DESCALE(tmp10 + MULTIPLY(tmp2 - tmp1, FIX(1.777777778)), /* 16/9 */ - CONST_BITS+PASS1_BITS); - - dataptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 5x5 sample block. - */ - -GLOBAL(void) -jpeg_fdct_5x5 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2; - INT32 tmp10, tmp11; - DCTELEM *dataptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pre-zero output coefficient block. */ - MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT; - * furthermore, we scale the results by 2**PASS1_BITS. - * We scale the results further by 2 as part of output adaption - * scaling for different DCT size. - * cK represents sqrt(2) * cos(K*pi/10). - */ - - dataptr = data; - for (ctr = 0; ctr < 5; ctr++) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[4]); - tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[3]); - tmp2 = GETJSAMPLE(elemptr[2]); - - tmp10 = tmp0 + tmp1; - tmp11 = tmp0 - tmp1; - - tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[4]); - tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[3]); - - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) - ((tmp10 + tmp2 - 5 * CENTERJSAMPLE) << (PASS1_BITS+1)); - tmp11 = MULTIPLY(tmp11, FIX(0.790569415)); /* (c2+c4)/2 */ - tmp10 -= tmp2 << 2; - tmp10 = MULTIPLY(tmp10, FIX(0.353553391)); /* (c2-c4)/2 */ - dataptr[2] = (DCTELEM) DESCALE(tmp11 + tmp10, CONST_BITS-PASS1_BITS-1); - dataptr[4] = (DCTELEM) DESCALE(tmp11 - tmp10, CONST_BITS-PASS1_BITS-1); - - /* Odd part */ - - tmp10 = MULTIPLY(tmp0 + tmp1, FIX(0.831253876)); /* c3 */ - - dataptr[1] = (DCTELEM) - DESCALE(tmp10 + MULTIPLY(tmp0, FIX(0.513743148)), /* c1-c3 */ - CONST_BITS-PASS1_BITS-1); - dataptr[3] = (DCTELEM) - DESCALE(tmp10 - MULTIPLY(tmp1, FIX(2.176250899)), /* c1+c3 */ - CONST_BITS-PASS1_BITS-1); - - dataptr += DCTSIZE; /* advance pointer to next row */ - } - - /* Pass 2: process columns. - * We remove the PASS1_BITS scaling, but leave the results scaled up - * by an overall factor of 8. - * We must also scale the output by (8/5)**2 = 64/25, which we partially - * fold into the constant multipliers (other part was done in pass 1): - * cK now represents sqrt(2) * cos(K*pi/10) * 32/25. - */ - - dataptr = data; - for (ctr = 0; ctr < 5; ctr++) { - /* Even part */ - - tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*4]; - tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*3]; - tmp2 = dataptr[DCTSIZE*2]; - - tmp10 = tmp0 + tmp1; - tmp11 = tmp0 - tmp1; - - tmp0 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*4]; - tmp1 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*3]; - - dataptr[DCTSIZE*0] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 + tmp2, FIX(1.28)), /* 32/25 */ - CONST_BITS+PASS1_BITS); - tmp11 = MULTIPLY(tmp11, FIX(1.011928851)); /* (c2+c4)/2 */ - tmp10 -= tmp2 << 2; - tmp10 = MULTIPLY(tmp10, FIX(0.452548340)); /* (c2-c4)/2 */ - dataptr[DCTSIZE*2] = (DCTELEM) DESCALE(tmp11 + tmp10, CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*4] = (DCTELEM) DESCALE(tmp11 - tmp10, CONST_BITS+PASS1_BITS); - - /* Odd part */ - - tmp10 = MULTIPLY(tmp0 + tmp1, FIX(1.064004961)); /* c3 */ - - dataptr[DCTSIZE*1] = (DCTELEM) - DESCALE(tmp10 + MULTIPLY(tmp0, FIX(0.657591230)), /* c1-c3 */ - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*3] = (DCTELEM) - DESCALE(tmp10 - MULTIPLY(tmp1, FIX(2.785601151)), /* c1+c3 */ - CONST_BITS+PASS1_BITS); - - dataptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 4x4 sample block. - */ - -GLOBAL(void) -jpeg_fdct_4x4 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1; - INT32 tmp10, tmp11; - DCTELEM *dataptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pre-zero output coefficient block. */ - MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT; - * furthermore, we scale the results by 2**PASS1_BITS. - * We must also scale the output by (8/4)**2 = 2**2, which we add here. - * cK represents sqrt(2) * cos(K*pi/16) [refers to 8-point FDCT]. - */ - - dataptr = data; - for (ctr = 0; ctr < 4; ctr++) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[3]); - tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[2]); - - tmp10 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[3]); - tmp11 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[2]); - - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) - ((tmp0 + tmp1 - 4 * CENTERJSAMPLE) << (PASS1_BITS+2)); - dataptr[2] = (DCTELEM) ((tmp0 - tmp1) << (PASS1_BITS+2)); - - /* Odd part */ - - tmp0 = MULTIPLY(tmp10 + tmp11, FIX_0_541196100); /* c6 */ - /* Add fudge factor here for final descale. */ - tmp0 += ONE << (CONST_BITS-PASS1_BITS-3); - - dataptr[1] = (DCTELEM) - RIGHT_SHIFT(tmp0 + MULTIPLY(tmp10, FIX_0_765366865), /* c2-c6 */ - CONST_BITS-PASS1_BITS-2); - dataptr[3] = (DCTELEM) - RIGHT_SHIFT(tmp0 - MULTIPLY(tmp11, FIX_1_847759065), /* c2+c6 */ - CONST_BITS-PASS1_BITS-2); - - dataptr += DCTSIZE; /* advance pointer to next row */ - } - - /* Pass 2: process columns. - * We remove the PASS1_BITS scaling, but leave the results scaled up - * by an overall factor of 8. - * cK represents sqrt(2) * cos(K*pi/16) [refers to 8-point FDCT]. - */ - - dataptr = data; - for (ctr = 0; ctr < 4; ctr++) { - /* Even part */ - - /* Add fudge factor here for final descale. */ - tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*3] + (ONE << (PASS1_BITS-1)); - tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*2]; - - tmp10 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*3]; - tmp11 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*2]; - - dataptr[DCTSIZE*0] = (DCTELEM) RIGHT_SHIFT(tmp0 + tmp1, PASS1_BITS); - dataptr[DCTSIZE*2] = (DCTELEM) RIGHT_SHIFT(tmp0 - tmp1, PASS1_BITS); - - /* Odd part */ - - tmp0 = MULTIPLY(tmp10 + tmp11, FIX_0_541196100); /* c6 */ - /* Add fudge factor here for final descale. */ - tmp0 += ONE << (CONST_BITS+PASS1_BITS-1); - - dataptr[DCTSIZE*1] = (DCTELEM) - RIGHT_SHIFT(tmp0 + MULTIPLY(tmp10, FIX_0_765366865), /* c2-c6 */ - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*3] = (DCTELEM) - RIGHT_SHIFT(tmp0 - MULTIPLY(tmp11, FIX_1_847759065), /* c2+c6 */ - CONST_BITS+PASS1_BITS); - - dataptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 3x3 sample block. - */ - -GLOBAL(void) -jpeg_fdct_3x3 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2; - DCTELEM *dataptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pre-zero output coefficient block. */ - MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT; - * furthermore, we scale the results by 2**PASS1_BITS. - * We scale the results further by 2**2 as part of output adaption - * scaling for different DCT size. - * cK represents sqrt(2) * cos(K*pi/6). - */ - - dataptr = data; - for (ctr = 0; ctr < 3; ctr++) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[2]); - tmp1 = GETJSAMPLE(elemptr[1]); - - tmp2 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[2]); - - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) - ((tmp0 + tmp1 - 3 * CENTERJSAMPLE) << (PASS1_BITS+2)); - dataptr[2] = (DCTELEM) - DESCALE(MULTIPLY(tmp0 - tmp1 - tmp1, FIX(0.707106781)), /* c2 */ - CONST_BITS-PASS1_BITS-2); - - /* Odd part */ - - dataptr[1] = (DCTELEM) - DESCALE(MULTIPLY(tmp2, FIX(1.224744871)), /* c1 */ - CONST_BITS-PASS1_BITS-2); - - dataptr += DCTSIZE; /* advance pointer to next row */ - } - - /* Pass 2: process columns. - * We remove the PASS1_BITS scaling, but leave the results scaled up - * by an overall factor of 8. - * We must also scale the output by (8/3)**2 = 64/9, which we partially - * fold into the constant multipliers (other part was done in pass 1): - * cK now represents sqrt(2) * cos(K*pi/6) * 16/9. - */ - - dataptr = data; - for (ctr = 0; ctr < 3; ctr++) { - /* Even part */ - - tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*2]; - tmp1 = dataptr[DCTSIZE*1]; - - tmp2 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*2]; - - dataptr[DCTSIZE*0] = (DCTELEM) - DESCALE(MULTIPLY(tmp0 + tmp1, FIX(1.777777778)), /* 16/9 */ - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*2] = (DCTELEM) - DESCALE(MULTIPLY(tmp0 - tmp1 - tmp1, FIX(1.257078722)), /* c2 */ - CONST_BITS+PASS1_BITS); - - /* Odd part */ - - dataptr[DCTSIZE*1] = (DCTELEM) - DESCALE(MULTIPLY(tmp2, FIX(2.177324216)), /* c1 */ - CONST_BITS+PASS1_BITS); - - dataptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 2x2 sample block. - */ - -GLOBAL(void) -jpeg_fdct_2x2 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - DCTELEM tmp0, tmp1, tmp2, tmp3; - JSAMPROW elemptr; - - /* Pre-zero output coefficient block. */ - MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT. - */ - - /* Row 0 */ - elemptr = sample_data[0] + start_col; - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[1]); - tmp1 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[1]); - - /* Row 1 */ - elemptr = sample_data[1] + start_col; - - tmp2 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[1]); - tmp3 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[1]); - - /* Pass 2: process columns. - * We leave the results scaled up by an overall factor of 8. - * We must also scale the output by (8/2)**2 = 2**4. - */ - - /* Column 0 */ - /* Apply unsigned->signed conversion. */ - data[DCTSIZE*0] = (tmp0 + tmp2 - 4 * CENTERJSAMPLE) << 4; - data[DCTSIZE*1] = (tmp0 - tmp2) << 4; - - /* Column 1 */ - data[DCTSIZE*0+1] = (tmp1 + tmp3) << 4; - data[DCTSIZE*1+1] = (tmp1 - tmp3) << 4; -} - - -/* - * Perform the forward DCT on a 1x1 sample block. - */ - -GLOBAL(void) -jpeg_fdct_1x1 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - DCTELEM dcval; - - /* Pre-zero output coefficient block. */ - MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); - - dcval = GETJSAMPLE(sample_data[0][start_col]); - - /* We leave the result scaled up by an overall factor of 8. */ - /* We must also scale the output by (8/1)**2 = 2**6. */ - /* Apply unsigned->signed conversion. */ - data[0] = (dcval - CENTERJSAMPLE) << 6; -} - - -/* - * Perform the forward DCT on a 9x9 sample block. - */ - -GLOBAL(void) -jpeg_fdct_9x9 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3, tmp4; - INT32 tmp10, tmp11, tmp12, tmp13; - INT32 z1, z2; - DCTELEM workspace[8]; - DCTELEM *dataptr; - DCTELEM *wsptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT; - * we scale the results further by 2 as part of output adaption - * scaling for different DCT size. - * cK represents sqrt(2) * cos(K*pi/18). - */ - - dataptr = data; - ctr = 0; - for (;;) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[8]); - tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[7]); - tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[6]); - tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[5]); - tmp4 = GETJSAMPLE(elemptr[4]); - - tmp10 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[8]); - tmp11 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[7]); - tmp12 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[6]); - tmp13 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[5]); - - z1 = tmp0 + tmp2 + tmp3; - z2 = tmp1 + tmp4; - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) ((z1 + z2 - 9 * CENTERJSAMPLE) << 1); - dataptr[6] = (DCTELEM) - DESCALE(MULTIPLY(z1 - z2 - z2, FIX(0.707106781)), /* c6 */ - CONST_BITS-1); - z1 = MULTIPLY(tmp0 - tmp2, FIX(1.328926049)); /* c2 */ - z2 = MULTIPLY(tmp1 - tmp4 - tmp4, FIX(0.707106781)); /* c6 */ - dataptr[2] = (DCTELEM) - DESCALE(MULTIPLY(tmp2 - tmp3, FIX(1.083350441)) /* c4 */ - + z1 + z2, CONST_BITS-1); - dataptr[4] = (DCTELEM) - DESCALE(MULTIPLY(tmp3 - tmp0, FIX(0.245575608)) /* c8 */ - + z1 - z2, CONST_BITS-1); - - /* Odd part */ - - dataptr[3] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp12 - tmp13, FIX(1.224744871)), /* c3 */ - CONST_BITS-1); - - tmp11 = MULTIPLY(tmp11, FIX(1.224744871)); /* c3 */ - tmp0 = MULTIPLY(tmp10 + tmp12, FIX(0.909038955)); /* c5 */ - tmp1 = MULTIPLY(tmp10 + tmp13, FIX(0.483689525)); /* c7 */ - - dataptr[1] = (DCTELEM) DESCALE(tmp11 + tmp0 + tmp1, CONST_BITS-1); - - tmp2 = MULTIPLY(tmp12 - tmp13, FIX(1.392728481)); /* c1 */ - - dataptr[5] = (DCTELEM) DESCALE(tmp0 - tmp11 - tmp2, CONST_BITS-1); - dataptr[7] = (DCTELEM) DESCALE(tmp1 - tmp11 + tmp2, CONST_BITS-1); - - ctr++; - - if (ctr != DCTSIZE) { - if (ctr == 9) - break; /* Done. */ - dataptr += DCTSIZE; /* advance pointer to next row */ - } else - dataptr = workspace; /* switch pointer to extended workspace */ - } - - /* Pass 2: process columns. - * We leave the results scaled up by an overall factor of 8. - * We must also scale the output by (8/9)**2 = 64/81, which we partially - * fold into the constant multipliers and final/initial shifting: - * cK now represents sqrt(2) * cos(K*pi/18) * 128/81. - */ - - dataptr = data; - wsptr = workspace; - for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { - /* Even part */ - - tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*0]; - tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*7]; - tmp2 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*6]; - tmp3 = dataptr[DCTSIZE*3] + dataptr[DCTSIZE*5]; - tmp4 = dataptr[DCTSIZE*4]; - - tmp10 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*0]; - tmp11 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*7]; - tmp12 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*6]; - tmp13 = dataptr[DCTSIZE*3] - dataptr[DCTSIZE*5]; - - z1 = tmp0 + tmp2 + tmp3; - z2 = tmp1 + tmp4; - dataptr[DCTSIZE*0] = (DCTELEM) - DESCALE(MULTIPLY(z1 + z2, FIX(1.580246914)), /* 128/81 */ - CONST_BITS+2); - dataptr[DCTSIZE*6] = (DCTELEM) - DESCALE(MULTIPLY(z1 - z2 - z2, FIX(1.117403309)), /* c6 */ - CONST_BITS+2); - z1 = MULTIPLY(tmp0 - tmp2, FIX(2.100031287)); /* c2 */ - z2 = MULTIPLY(tmp1 - tmp4 - tmp4, FIX(1.117403309)); /* c6 */ - dataptr[DCTSIZE*2] = (DCTELEM) - DESCALE(MULTIPLY(tmp2 - tmp3, FIX(1.711961190)) /* c4 */ - + z1 + z2, CONST_BITS+2); - dataptr[DCTSIZE*4] = (DCTELEM) - DESCALE(MULTIPLY(tmp3 - tmp0, FIX(0.388070096)) /* c8 */ - + z1 - z2, CONST_BITS+2); - - /* Odd part */ - - dataptr[DCTSIZE*3] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp12 - tmp13, FIX(1.935399303)), /* c3 */ - CONST_BITS+2); - - tmp11 = MULTIPLY(tmp11, FIX(1.935399303)); /* c3 */ - tmp0 = MULTIPLY(tmp10 + tmp12, FIX(1.436506004)); /* c5 */ - tmp1 = MULTIPLY(tmp10 + tmp13, FIX(0.764348879)); /* c7 */ - - dataptr[DCTSIZE*1] = (DCTELEM) - DESCALE(tmp11 + tmp0 + tmp1, CONST_BITS+2); - - tmp2 = MULTIPLY(tmp12 - tmp13, FIX(2.200854883)); /* c1 */ - - dataptr[DCTSIZE*5] = (DCTELEM) - DESCALE(tmp0 - tmp11 - tmp2, CONST_BITS+2); - dataptr[DCTSIZE*7] = (DCTELEM) - DESCALE(tmp1 - tmp11 + tmp2, CONST_BITS+2); - - dataptr++; /* advance pointer to next column */ - wsptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 10x10 sample block. - */ - -GLOBAL(void) -jpeg_fdct_10x10 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3, tmp4; - INT32 tmp10, tmp11, tmp12, tmp13, tmp14; - DCTELEM workspace[8*2]; - DCTELEM *dataptr; - DCTELEM *wsptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT; - * we scale the results further by 2 as part of output adaption - * scaling for different DCT size. - * cK represents sqrt(2) * cos(K*pi/20). - */ - - dataptr = data; - ctr = 0; - for (;;) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[9]); - tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[8]); - tmp12 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[7]); - tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[6]); - tmp4 = GETJSAMPLE(elemptr[4]) + GETJSAMPLE(elemptr[5]); - - tmp10 = tmp0 + tmp4; - tmp13 = tmp0 - tmp4; - tmp11 = tmp1 + tmp3; - tmp14 = tmp1 - tmp3; - - tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[9]); - tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[8]); - tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[7]); - tmp3 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[6]); - tmp4 = GETJSAMPLE(elemptr[4]) - GETJSAMPLE(elemptr[5]); - - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) - ((tmp10 + tmp11 + tmp12 - 10 * CENTERJSAMPLE) << 1); - tmp12 += tmp12; - dataptr[4] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp12, FIX(1.144122806)) - /* c4 */ - MULTIPLY(tmp11 - tmp12, FIX(0.437016024)), /* c8 */ - CONST_BITS-1); - tmp10 = MULTIPLY(tmp13 + tmp14, FIX(0.831253876)); /* c6 */ - dataptr[2] = (DCTELEM) - DESCALE(tmp10 + MULTIPLY(tmp13, FIX(0.513743148)), /* c2-c6 */ - CONST_BITS-1); - dataptr[6] = (DCTELEM) - DESCALE(tmp10 - MULTIPLY(tmp14, FIX(2.176250899)), /* c2+c6 */ - CONST_BITS-1); - - /* Odd part */ - - tmp10 = tmp0 + tmp4; - tmp11 = tmp1 - tmp3; - dataptr[5] = (DCTELEM) ((tmp10 - tmp11 - tmp2) << 1); - tmp2 <<= CONST_BITS; - dataptr[1] = (DCTELEM) - DESCALE(MULTIPLY(tmp0, FIX(1.396802247)) + /* c1 */ - MULTIPLY(tmp1, FIX(1.260073511)) + tmp2 + /* c3 */ - MULTIPLY(tmp3, FIX(0.642039522)) + /* c7 */ - MULTIPLY(tmp4, FIX(0.221231742)), /* c9 */ - CONST_BITS-1); - tmp12 = MULTIPLY(tmp0 - tmp4, FIX(0.951056516)) - /* (c3+c7)/2 */ - MULTIPLY(tmp1 + tmp3, FIX(0.587785252)); /* (c1-c9)/2 */ - tmp13 = MULTIPLY(tmp10 + tmp11, FIX(0.309016994)) + /* (c3-c7)/2 */ - (tmp11 << (CONST_BITS - 1)) - tmp2; - dataptr[3] = (DCTELEM) DESCALE(tmp12 + tmp13, CONST_BITS-1); - dataptr[7] = (DCTELEM) DESCALE(tmp12 - tmp13, CONST_BITS-1); - - ctr++; - - if (ctr != DCTSIZE) { - if (ctr == 10) - break; /* Done. */ - dataptr += DCTSIZE; /* advance pointer to next row */ - } else - dataptr = workspace; /* switch pointer to extended workspace */ - } - - /* Pass 2: process columns. - * We leave the results scaled up by an overall factor of 8. - * We must also scale the output by (8/10)**2 = 16/25, which we partially - * fold into the constant multipliers and final/initial shifting: - * cK now represents sqrt(2) * cos(K*pi/20) * 32/25. - */ - - dataptr = data; - wsptr = workspace; - for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { - /* Even part */ - - tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*1]; - tmp1 = dataptr[DCTSIZE*1] + wsptr[DCTSIZE*0]; - tmp12 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*7]; - tmp3 = dataptr[DCTSIZE*3] + dataptr[DCTSIZE*6]; - tmp4 = dataptr[DCTSIZE*4] + dataptr[DCTSIZE*5]; - - tmp10 = tmp0 + tmp4; - tmp13 = tmp0 - tmp4; - tmp11 = tmp1 + tmp3; - tmp14 = tmp1 - tmp3; - - tmp0 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*1]; - tmp1 = dataptr[DCTSIZE*1] - wsptr[DCTSIZE*0]; - tmp2 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*7]; - tmp3 = dataptr[DCTSIZE*3] - dataptr[DCTSIZE*6]; - tmp4 = dataptr[DCTSIZE*4] - dataptr[DCTSIZE*5]; - - dataptr[DCTSIZE*0] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 + tmp11 + tmp12, FIX(1.28)), /* 32/25 */ - CONST_BITS+2); - tmp12 += tmp12; - dataptr[DCTSIZE*4] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp12, FIX(1.464477191)) - /* c4 */ - MULTIPLY(tmp11 - tmp12, FIX(0.559380511)), /* c8 */ - CONST_BITS+2); - tmp10 = MULTIPLY(tmp13 + tmp14, FIX(1.064004961)); /* c6 */ - dataptr[DCTSIZE*2] = (DCTELEM) - DESCALE(tmp10 + MULTIPLY(tmp13, FIX(0.657591230)), /* c2-c6 */ - CONST_BITS+2); - dataptr[DCTSIZE*6] = (DCTELEM) - DESCALE(tmp10 - MULTIPLY(tmp14, FIX(2.785601151)), /* c2+c6 */ - CONST_BITS+2); - - /* Odd part */ - - tmp10 = tmp0 + tmp4; - tmp11 = tmp1 - tmp3; - dataptr[DCTSIZE*5] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp11 - tmp2, FIX(1.28)), /* 32/25 */ - CONST_BITS+2); - tmp2 = MULTIPLY(tmp2, FIX(1.28)); /* 32/25 */ - dataptr[DCTSIZE*1] = (DCTELEM) - DESCALE(MULTIPLY(tmp0, FIX(1.787906876)) + /* c1 */ - MULTIPLY(tmp1, FIX(1.612894094)) + tmp2 + /* c3 */ - MULTIPLY(tmp3, FIX(0.821810588)) + /* c7 */ - MULTIPLY(tmp4, FIX(0.283176630)), /* c9 */ - CONST_BITS+2); - tmp12 = MULTIPLY(tmp0 - tmp4, FIX(1.217352341)) - /* (c3+c7)/2 */ - MULTIPLY(tmp1 + tmp3, FIX(0.752365123)); /* (c1-c9)/2 */ - tmp13 = MULTIPLY(tmp10 + tmp11, FIX(0.395541753)) + /* (c3-c7)/2 */ - MULTIPLY(tmp11, FIX(0.64)) - tmp2; /* 16/25 */ - dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp12 + tmp13, CONST_BITS+2); - dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp12 - tmp13, CONST_BITS+2); - - dataptr++; /* advance pointer to next column */ - wsptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on an 11x11 sample block. - */ - -GLOBAL(void) -jpeg_fdct_11x11 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5; - INT32 tmp10, tmp11, tmp12, tmp13, tmp14; - INT32 z1, z2, z3; - DCTELEM workspace[8*3]; - DCTELEM *dataptr; - DCTELEM *wsptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT; - * we scale the results further by 2 as part of output adaption - * scaling for different DCT size. - * cK represents sqrt(2) * cos(K*pi/22). - */ - - dataptr = data; - ctr = 0; - for (;;) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[10]); - tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[9]); - tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[8]); - tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[7]); - tmp4 = GETJSAMPLE(elemptr[4]) + GETJSAMPLE(elemptr[6]); - tmp5 = GETJSAMPLE(elemptr[5]); - - tmp10 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[10]); - tmp11 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[9]); - tmp12 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[8]); - tmp13 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[7]); - tmp14 = GETJSAMPLE(elemptr[4]) - GETJSAMPLE(elemptr[6]); - - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) - ((tmp0 + tmp1 + tmp2 + tmp3 + tmp4 + tmp5 - 11 * CENTERJSAMPLE) << 1); - tmp5 += tmp5; - tmp0 -= tmp5; - tmp1 -= tmp5; - tmp2 -= tmp5; - tmp3 -= tmp5; - tmp4 -= tmp5; - z1 = MULTIPLY(tmp0 + tmp3, FIX(1.356927976)) + /* c2 */ - MULTIPLY(tmp2 + tmp4, FIX(0.201263574)); /* c10 */ - z2 = MULTIPLY(tmp1 - tmp3, FIX(0.926112931)); /* c6 */ - z3 = MULTIPLY(tmp0 - tmp1, FIX(1.189712156)); /* c4 */ - dataptr[2] = (DCTELEM) - DESCALE(z1 + z2 - MULTIPLY(tmp3, FIX(1.018300590)) /* c2+c8-c6 */ - - MULTIPLY(tmp4, FIX(1.390975730)), /* c4+c10 */ - CONST_BITS-1); - dataptr[4] = (DCTELEM) - DESCALE(z2 + z3 + MULTIPLY(tmp1, FIX(0.062335650)) /* c4-c6-c10 */ - - MULTIPLY(tmp2, FIX(1.356927976)) /* c2 */ - + MULTIPLY(tmp4, FIX(0.587485545)), /* c8 */ - CONST_BITS-1); - dataptr[6] = (DCTELEM) - DESCALE(z1 + z3 - MULTIPLY(tmp0, FIX(1.620527200)) /* c2+c4-c6 */ - - MULTIPLY(tmp2, FIX(0.788749120)), /* c8+c10 */ - CONST_BITS-1); - - /* Odd part */ - - tmp1 = MULTIPLY(tmp10 + tmp11, FIX(1.286413905)); /* c3 */ - tmp2 = MULTIPLY(tmp10 + tmp12, FIX(1.068791298)); /* c5 */ - tmp3 = MULTIPLY(tmp10 + tmp13, FIX(0.764581576)); /* c7 */ - tmp0 = tmp1 + tmp2 + tmp3 - MULTIPLY(tmp10, FIX(1.719967871)) /* c7+c5+c3-c1 */ - + MULTIPLY(tmp14, FIX(0.398430003)); /* c9 */ - tmp4 = MULTIPLY(tmp11 + tmp12, - FIX(0.764581576)); /* -c7 */ - tmp5 = MULTIPLY(tmp11 + tmp13, - FIX(1.399818907)); /* -c1 */ - tmp1 += tmp4 + tmp5 + MULTIPLY(tmp11, FIX(1.276416582)) /* c9+c7+c1-c3 */ - - MULTIPLY(tmp14, FIX(1.068791298)); /* c5 */ - tmp10 = MULTIPLY(tmp12 + tmp13, FIX(0.398430003)); /* c9 */ - tmp2 += tmp4 + tmp10 - MULTIPLY(tmp12, FIX(1.989053629)) /* c9+c5+c3-c7 */ - + MULTIPLY(tmp14, FIX(1.399818907)); /* c1 */ - tmp3 += tmp5 + tmp10 + MULTIPLY(tmp13, FIX(1.305598626)) /* c1+c5-c9-c7 */ - - MULTIPLY(tmp14, FIX(1.286413905)); /* c3 */ - - dataptr[1] = (DCTELEM) DESCALE(tmp0, CONST_BITS-1); - dataptr[3] = (DCTELEM) DESCALE(tmp1, CONST_BITS-1); - dataptr[5] = (DCTELEM) DESCALE(tmp2, CONST_BITS-1); - dataptr[7] = (DCTELEM) DESCALE(tmp3, CONST_BITS-1); - - ctr++; - - if (ctr != DCTSIZE) { - if (ctr == 11) - break; /* Done. */ - dataptr += DCTSIZE; /* advance pointer to next row */ - } else - dataptr = workspace; /* switch pointer to extended workspace */ - } - - /* Pass 2: process columns. - * We leave the results scaled up by an overall factor of 8. - * We must also scale the output by (8/11)**2 = 64/121, which we partially - * fold into the constant multipliers and final/initial shifting: - * cK now represents sqrt(2) * cos(K*pi/22) * 128/121. - */ - - dataptr = data; - wsptr = workspace; - for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { - /* Even part */ - - tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*2]; - tmp1 = dataptr[DCTSIZE*1] + wsptr[DCTSIZE*1]; - tmp2 = dataptr[DCTSIZE*2] + wsptr[DCTSIZE*0]; - tmp3 = dataptr[DCTSIZE*3] + dataptr[DCTSIZE*7]; - tmp4 = dataptr[DCTSIZE*4] + dataptr[DCTSIZE*6]; - tmp5 = dataptr[DCTSIZE*5]; - - tmp10 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*2]; - tmp11 = dataptr[DCTSIZE*1] - wsptr[DCTSIZE*1]; - tmp12 = dataptr[DCTSIZE*2] - wsptr[DCTSIZE*0]; - tmp13 = dataptr[DCTSIZE*3] - dataptr[DCTSIZE*7]; - tmp14 = dataptr[DCTSIZE*4] - dataptr[DCTSIZE*6]; - - dataptr[DCTSIZE*0] = (DCTELEM) - DESCALE(MULTIPLY(tmp0 + tmp1 + tmp2 + tmp3 + tmp4 + tmp5, - FIX(1.057851240)), /* 128/121 */ - CONST_BITS+2); - tmp5 += tmp5; - tmp0 -= tmp5; - tmp1 -= tmp5; - tmp2 -= tmp5; - tmp3 -= tmp5; - tmp4 -= tmp5; - z1 = MULTIPLY(tmp0 + tmp3, FIX(1.435427942)) + /* c2 */ - MULTIPLY(tmp2 + tmp4, FIX(0.212906922)); /* c10 */ - z2 = MULTIPLY(tmp1 - tmp3, FIX(0.979689713)); /* c6 */ - z3 = MULTIPLY(tmp0 - tmp1, FIX(1.258538479)); /* c4 */ - dataptr[DCTSIZE*2] = (DCTELEM) - DESCALE(z1 + z2 - MULTIPLY(tmp3, FIX(1.077210542)) /* c2+c8-c6 */ - - MULTIPLY(tmp4, FIX(1.471445400)), /* c4+c10 */ - CONST_BITS+2); - dataptr[DCTSIZE*4] = (DCTELEM) - DESCALE(z2 + z3 + MULTIPLY(tmp1, FIX(0.065941844)) /* c4-c6-c10 */ - - MULTIPLY(tmp2, FIX(1.435427942)) /* c2 */ - + MULTIPLY(tmp4, FIX(0.621472312)), /* c8 */ - CONST_BITS+2); - dataptr[DCTSIZE*6] = (DCTELEM) - DESCALE(z1 + z3 - MULTIPLY(tmp0, FIX(1.714276708)) /* c2+c4-c6 */ - - MULTIPLY(tmp2, FIX(0.834379234)), /* c8+c10 */ - CONST_BITS+2); - - /* Odd part */ - - tmp1 = MULTIPLY(tmp10 + tmp11, FIX(1.360834544)); /* c3 */ - tmp2 = MULTIPLY(tmp10 + tmp12, FIX(1.130622199)); /* c5 */ - tmp3 = MULTIPLY(tmp10 + tmp13, FIX(0.808813568)); /* c7 */ - tmp0 = tmp1 + tmp2 + tmp3 - MULTIPLY(tmp10, FIX(1.819470145)) /* c7+c5+c3-c1 */ - + MULTIPLY(tmp14, FIX(0.421479672)); /* c9 */ - tmp4 = MULTIPLY(tmp11 + tmp12, - FIX(0.808813568)); /* -c7 */ - tmp5 = MULTIPLY(tmp11 + tmp13, - FIX(1.480800167)); /* -c1 */ - tmp1 += tmp4 + tmp5 + MULTIPLY(tmp11, FIX(1.350258864)) /* c9+c7+c1-c3 */ - - MULTIPLY(tmp14, FIX(1.130622199)); /* c5 */ - tmp10 = MULTIPLY(tmp12 + tmp13, FIX(0.421479672)); /* c9 */ - tmp2 += tmp4 + tmp10 - MULTIPLY(tmp12, FIX(2.104122847)) /* c9+c5+c3-c7 */ - + MULTIPLY(tmp14, FIX(1.480800167)); /* c1 */ - tmp3 += tmp5 + tmp10 + MULTIPLY(tmp13, FIX(1.381129125)) /* c1+c5-c9-c7 */ - - MULTIPLY(tmp14, FIX(1.360834544)); /* c3 */ - - dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp0, CONST_BITS+2); - dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp1, CONST_BITS+2); - dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp2, CONST_BITS+2); - dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp3, CONST_BITS+2); - - dataptr++; /* advance pointer to next column */ - wsptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 12x12 sample block. - */ - -GLOBAL(void) -jpeg_fdct_12x12 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5; - INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15; - DCTELEM workspace[8*4]; - DCTELEM *dataptr; - DCTELEM *wsptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT. - * cK represents sqrt(2) * cos(K*pi/24). - */ - - dataptr = data; - ctr = 0; - for (;;) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[11]); - tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[10]); - tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[9]); - tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[8]); - tmp4 = GETJSAMPLE(elemptr[4]) + GETJSAMPLE(elemptr[7]); - tmp5 = GETJSAMPLE(elemptr[5]) + GETJSAMPLE(elemptr[6]); - - tmp10 = tmp0 + tmp5; - tmp13 = tmp0 - tmp5; - tmp11 = tmp1 + tmp4; - tmp14 = tmp1 - tmp4; - tmp12 = tmp2 + tmp3; - tmp15 = tmp2 - tmp3; - - tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[11]); - tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[10]); - tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[9]); - tmp3 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[8]); - tmp4 = GETJSAMPLE(elemptr[4]) - GETJSAMPLE(elemptr[7]); - tmp5 = GETJSAMPLE(elemptr[5]) - GETJSAMPLE(elemptr[6]); - - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) (tmp10 + tmp11 + tmp12 - 12 * CENTERJSAMPLE); - dataptr[6] = (DCTELEM) (tmp13 - tmp14 - tmp15); - dataptr[4] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp12, FIX(1.224744871)), /* c4 */ - CONST_BITS); - dataptr[2] = (DCTELEM) - DESCALE(tmp14 - tmp15 + MULTIPLY(tmp13 + tmp15, FIX(1.366025404)), /* c2 */ - CONST_BITS); - - /* Odd part */ - - tmp10 = MULTIPLY(tmp1 + tmp4, FIX_0_541196100); /* c9 */ - tmp14 = tmp10 + MULTIPLY(tmp1, FIX_0_765366865); /* c3-c9 */ - tmp15 = tmp10 - MULTIPLY(tmp4, FIX_1_847759065); /* c3+c9 */ - tmp12 = MULTIPLY(tmp0 + tmp2, FIX(1.121971054)); /* c5 */ - tmp13 = MULTIPLY(tmp0 + tmp3, FIX(0.860918669)); /* c7 */ - tmp10 = tmp12 + tmp13 + tmp14 - MULTIPLY(tmp0, FIX(0.580774953)) /* c5+c7-c1 */ - + MULTIPLY(tmp5, FIX(0.184591911)); /* c11 */ - tmp11 = MULTIPLY(tmp2 + tmp3, - FIX(0.184591911)); /* -c11 */ - tmp12 += tmp11 - tmp15 - MULTIPLY(tmp2, FIX(2.339493912)) /* c1+c5-c11 */ - + MULTIPLY(tmp5, FIX(0.860918669)); /* c7 */ - tmp13 += tmp11 - tmp14 + MULTIPLY(tmp3, FIX(0.725788011)) /* c1+c11-c7 */ - - MULTIPLY(tmp5, FIX(1.121971054)); /* c5 */ - tmp11 = tmp15 + MULTIPLY(tmp0 - tmp3, FIX(1.306562965)) /* c3 */ - - MULTIPLY(tmp2 + tmp5, FIX_0_541196100); /* c9 */ - - dataptr[1] = (DCTELEM) DESCALE(tmp10, CONST_BITS); - dataptr[3] = (DCTELEM) DESCALE(tmp11, CONST_BITS); - dataptr[5] = (DCTELEM) DESCALE(tmp12, CONST_BITS); - dataptr[7] = (DCTELEM) DESCALE(tmp13, CONST_BITS); - - ctr++; - - if (ctr != DCTSIZE) { - if (ctr == 12) - break; /* Done. */ - dataptr += DCTSIZE; /* advance pointer to next row */ - } else - dataptr = workspace; /* switch pointer to extended workspace */ - } - - /* Pass 2: process columns. - * We leave the results scaled up by an overall factor of 8. - * We must also scale the output by (8/12)**2 = 4/9, which we partially - * fold into the constant multipliers and final shifting: - * cK now represents sqrt(2) * cos(K*pi/24) * 8/9. - */ - - dataptr = data; - wsptr = workspace; - for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { - /* Even part */ - - tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*3]; - tmp1 = dataptr[DCTSIZE*1] + wsptr[DCTSIZE*2]; - tmp2 = dataptr[DCTSIZE*2] + wsptr[DCTSIZE*1]; - tmp3 = dataptr[DCTSIZE*3] + wsptr[DCTSIZE*0]; - tmp4 = dataptr[DCTSIZE*4] + dataptr[DCTSIZE*7]; - tmp5 = dataptr[DCTSIZE*5] + dataptr[DCTSIZE*6]; - - tmp10 = tmp0 + tmp5; - tmp13 = tmp0 - tmp5; - tmp11 = tmp1 + tmp4; - tmp14 = tmp1 - tmp4; - tmp12 = tmp2 + tmp3; - tmp15 = tmp2 - tmp3; - - tmp0 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*3]; - tmp1 = dataptr[DCTSIZE*1] - wsptr[DCTSIZE*2]; - tmp2 = dataptr[DCTSIZE*2] - wsptr[DCTSIZE*1]; - tmp3 = dataptr[DCTSIZE*3] - wsptr[DCTSIZE*0]; - tmp4 = dataptr[DCTSIZE*4] - dataptr[DCTSIZE*7]; - tmp5 = dataptr[DCTSIZE*5] - dataptr[DCTSIZE*6]; - - dataptr[DCTSIZE*0] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 + tmp11 + tmp12, FIX(0.888888889)), /* 8/9 */ - CONST_BITS+1); - dataptr[DCTSIZE*6] = (DCTELEM) - DESCALE(MULTIPLY(tmp13 - tmp14 - tmp15, FIX(0.888888889)), /* 8/9 */ - CONST_BITS+1); - dataptr[DCTSIZE*4] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp12, FIX(1.088662108)), /* c4 */ - CONST_BITS+1); - dataptr[DCTSIZE*2] = (DCTELEM) - DESCALE(MULTIPLY(tmp14 - tmp15, FIX(0.888888889)) + /* 8/9 */ - MULTIPLY(tmp13 + tmp15, FIX(1.214244803)), /* c2 */ - CONST_BITS+1); - - /* Odd part */ - - tmp10 = MULTIPLY(tmp1 + tmp4, FIX(0.481063200)); /* c9 */ - tmp14 = tmp10 + MULTIPLY(tmp1, FIX(0.680326102)); /* c3-c9 */ - tmp15 = tmp10 - MULTIPLY(tmp4, FIX(1.642452502)); /* c3+c9 */ - tmp12 = MULTIPLY(tmp0 + tmp2, FIX(0.997307603)); /* c5 */ - tmp13 = MULTIPLY(tmp0 + tmp3, FIX(0.765261039)); /* c7 */ - tmp10 = tmp12 + tmp13 + tmp14 - MULTIPLY(tmp0, FIX(0.516244403)) /* c5+c7-c1 */ - + MULTIPLY(tmp5, FIX(0.164081699)); /* c11 */ - tmp11 = MULTIPLY(tmp2 + tmp3, - FIX(0.164081699)); /* -c11 */ - tmp12 += tmp11 - tmp15 - MULTIPLY(tmp2, FIX(2.079550144)) /* c1+c5-c11 */ - + MULTIPLY(tmp5, FIX(0.765261039)); /* c7 */ - tmp13 += tmp11 - tmp14 + MULTIPLY(tmp3, FIX(0.645144899)) /* c1+c11-c7 */ - - MULTIPLY(tmp5, FIX(0.997307603)); /* c5 */ - tmp11 = tmp15 + MULTIPLY(tmp0 - tmp3, FIX(1.161389302)) /* c3 */ - - MULTIPLY(tmp2 + tmp5, FIX(0.481063200)); /* c9 */ - - dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp10, CONST_BITS+1); - dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp11, CONST_BITS+1); - dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp12, CONST_BITS+1); - dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp13, CONST_BITS+1); - - dataptr++; /* advance pointer to next column */ - wsptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 13x13 sample block. - */ - -GLOBAL(void) -jpeg_fdct_13x13 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6; - INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15; - INT32 z1, z2; - DCTELEM workspace[8*5]; - DCTELEM *dataptr; - DCTELEM *wsptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT. - * cK represents sqrt(2) * cos(K*pi/26). - */ - - dataptr = data; - ctr = 0; - for (;;) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[12]); - tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[11]); - tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[10]); - tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[9]); - tmp4 = GETJSAMPLE(elemptr[4]) + GETJSAMPLE(elemptr[8]); - tmp5 = GETJSAMPLE(elemptr[5]) + GETJSAMPLE(elemptr[7]); - tmp6 = GETJSAMPLE(elemptr[6]); - - tmp10 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[12]); - tmp11 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[11]); - tmp12 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[10]); - tmp13 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[9]); - tmp14 = GETJSAMPLE(elemptr[4]) - GETJSAMPLE(elemptr[8]); - tmp15 = GETJSAMPLE(elemptr[5]) - GETJSAMPLE(elemptr[7]); - - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) - (tmp0 + tmp1 + tmp2 + tmp3 + tmp4 + tmp5 + tmp6 - 13 * CENTERJSAMPLE); - tmp6 += tmp6; - tmp0 -= tmp6; - tmp1 -= tmp6; - tmp2 -= tmp6; - tmp3 -= tmp6; - tmp4 -= tmp6; - tmp5 -= tmp6; - dataptr[2] = (DCTELEM) - DESCALE(MULTIPLY(tmp0, FIX(1.373119086)) + /* c2 */ - MULTIPLY(tmp1, FIX(1.058554052)) + /* c6 */ - MULTIPLY(tmp2, FIX(0.501487041)) - /* c10 */ - MULTIPLY(tmp3, FIX(0.170464608)) - /* c12 */ - MULTIPLY(tmp4, FIX(0.803364869)) - /* c8 */ - MULTIPLY(tmp5, FIX(1.252223920)), /* c4 */ - CONST_BITS); - z1 = MULTIPLY(tmp0 - tmp2, FIX(1.155388986)) - /* (c4+c6)/2 */ - MULTIPLY(tmp3 - tmp4, FIX(0.435816023)) - /* (c2-c10)/2 */ - MULTIPLY(tmp1 - tmp5, FIX(0.316450131)); /* (c8-c12)/2 */ - z2 = MULTIPLY(tmp0 + tmp2, FIX(0.096834934)) - /* (c4-c6)/2 */ - MULTIPLY(tmp3 + tmp4, FIX(0.937303064)) + /* (c2+c10)/2 */ - MULTIPLY(tmp1 + tmp5, FIX(0.486914739)); /* (c8+c12)/2 */ - - dataptr[4] = (DCTELEM) DESCALE(z1 + z2, CONST_BITS); - dataptr[6] = (DCTELEM) DESCALE(z1 - z2, CONST_BITS); - - /* Odd part */ - - tmp1 = MULTIPLY(tmp10 + tmp11, FIX(1.322312651)); /* c3 */ - tmp2 = MULTIPLY(tmp10 + tmp12, FIX(1.163874945)); /* c5 */ - tmp3 = MULTIPLY(tmp10 + tmp13, FIX(0.937797057)) + /* c7 */ - MULTIPLY(tmp14 + tmp15, FIX(0.338443458)); /* c11 */ - tmp0 = tmp1 + tmp2 + tmp3 - - MULTIPLY(tmp10, FIX(2.020082300)) + /* c3+c5+c7-c1 */ - MULTIPLY(tmp14, FIX(0.318774355)); /* c9-c11 */ - tmp4 = MULTIPLY(tmp14 - tmp15, FIX(0.937797057)) - /* c7 */ - MULTIPLY(tmp11 + tmp12, FIX(0.338443458)); /* c11 */ - tmp5 = MULTIPLY(tmp11 + tmp13, - FIX(1.163874945)); /* -c5 */ - tmp1 += tmp4 + tmp5 + - MULTIPLY(tmp11, FIX(0.837223564)) - /* c5+c9+c11-c3 */ - MULTIPLY(tmp14, FIX(2.341699410)); /* c1+c7 */ - tmp6 = MULTIPLY(tmp12 + tmp13, - FIX(0.657217813)); /* -c9 */ - tmp2 += tmp4 + tmp6 - - MULTIPLY(tmp12, FIX(1.572116027)) + /* c1+c5-c9-c11 */ - MULTIPLY(tmp15, FIX(2.260109708)); /* c3+c7 */ - tmp3 += tmp5 + tmp6 + - MULTIPLY(tmp13, FIX(2.205608352)) - /* c3+c5+c9-c7 */ - MULTIPLY(tmp15, FIX(1.742345811)); /* c1+c11 */ - - dataptr[1] = (DCTELEM) DESCALE(tmp0, CONST_BITS); - dataptr[3] = (DCTELEM) DESCALE(tmp1, CONST_BITS); - dataptr[5] = (DCTELEM) DESCALE(tmp2, CONST_BITS); - dataptr[7] = (DCTELEM) DESCALE(tmp3, CONST_BITS); - - ctr++; - - if (ctr != DCTSIZE) { - if (ctr == 13) - break; /* Done. */ - dataptr += DCTSIZE; /* advance pointer to next row */ - } else - dataptr = workspace; /* switch pointer to extended workspace */ - } - - /* Pass 2: process columns. - * We leave the results scaled up by an overall factor of 8. - * We must also scale the output by (8/13)**2 = 64/169, which we partially - * fold into the constant multipliers and final shifting: - * cK now represents sqrt(2) * cos(K*pi/26) * 128/169. - */ - - dataptr = data; - wsptr = workspace; - for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { - /* Even part */ - - tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*4]; - tmp1 = dataptr[DCTSIZE*1] + wsptr[DCTSIZE*3]; - tmp2 = dataptr[DCTSIZE*2] + wsptr[DCTSIZE*2]; - tmp3 = dataptr[DCTSIZE*3] + wsptr[DCTSIZE*1]; - tmp4 = dataptr[DCTSIZE*4] + wsptr[DCTSIZE*0]; - tmp5 = dataptr[DCTSIZE*5] + dataptr[DCTSIZE*7]; - tmp6 = dataptr[DCTSIZE*6]; - - tmp10 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*4]; - tmp11 = dataptr[DCTSIZE*1] - wsptr[DCTSIZE*3]; - tmp12 = dataptr[DCTSIZE*2] - wsptr[DCTSIZE*2]; - tmp13 = dataptr[DCTSIZE*3] - wsptr[DCTSIZE*1]; - tmp14 = dataptr[DCTSIZE*4] - wsptr[DCTSIZE*0]; - tmp15 = dataptr[DCTSIZE*5] - dataptr[DCTSIZE*7]; - - dataptr[DCTSIZE*0] = (DCTELEM) - DESCALE(MULTIPLY(tmp0 + tmp1 + tmp2 + tmp3 + tmp4 + tmp5 + tmp6, - FIX(0.757396450)), /* 128/169 */ - CONST_BITS+1); - tmp6 += tmp6; - tmp0 -= tmp6; - tmp1 -= tmp6; - tmp2 -= tmp6; - tmp3 -= tmp6; - tmp4 -= tmp6; - tmp5 -= tmp6; - dataptr[DCTSIZE*2] = (DCTELEM) - DESCALE(MULTIPLY(tmp0, FIX(1.039995521)) + /* c2 */ - MULTIPLY(tmp1, FIX(0.801745081)) + /* c6 */ - MULTIPLY(tmp2, FIX(0.379824504)) - /* c10 */ - MULTIPLY(tmp3, FIX(0.129109289)) - /* c12 */ - MULTIPLY(tmp4, FIX(0.608465700)) - /* c8 */ - MULTIPLY(tmp5, FIX(0.948429952)), /* c4 */ - CONST_BITS+1); - z1 = MULTIPLY(tmp0 - tmp2, FIX(0.875087516)) - /* (c4+c6)/2 */ - MULTIPLY(tmp3 - tmp4, FIX(0.330085509)) - /* (c2-c10)/2 */ - MULTIPLY(tmp1 - tmp5, FIX(0.239678205)); /* (c8-c12)/2 */ - z2 = MULTIPLY(tmp0 + tmp2, FIX(0.073342435)) - /* (c4-c6)/2 */ - MULTIPLY(tmp3 + tmp4, FIX(0.709910013)) + /* (c2+c10)/2 */ - MULTIPLY(tmp1 + tmp5, FIX(0.368787494)); /* (c8+c12)/2 */ - - dataptr[DCTSIZE*4] = (DCTELEM) DESCALE(z1 + z2, CONST_BITS+1); - dataptr[DCTSIZE*6] = (DCTELEM) DESCALE(z1 - z2, CONST_BITS+1); - - /* Odd part */ - - tmp1 = MULTIPLY(tmp10 + tmp11, FIX(1.001514908)); /* c3 */ - tmp2 = MULTIPLY(tmp10 + tmp12, FIX(0.881514751)); /* c5 */ - tmp3 = MULTIPLY(tmp10 + tmp13, FIX(0.710284161)) + /* c7 */ - MULTIPLY(tmp14 + tmp15, FIX(0.256335874)); /* c11 */ - tmp0 = tmp1 + tmp2 + tmp3 - - MULTIPLY(tmp10, FIX(1.530003162)) + /* c3+c5+c7-c1 */ - MULTIPLY(tmp14, FIX(0.241438564)); /* c9-c11 */ - tmp4 = MULTIPLY(tmp14 - tmp15, FIX(0.710284161)) - /* c7 */ - MULTIPLY(tmp11 + tmp12, FIX(0.256335874)); /* c11 */ - tmp5 = MULTIPLY(tmp11 + tmp13, - FIX(0.881514751)); /* -c5 */ - tmp1 += tmp4 + tmp5 + - MULTIPLY(tmp11, FIX(0.634110155)) - /* c5+c9+c11-c3 */ - MULTIPLY(tmp14, FIX(1.773594819)); /* c1+c7 */ - tmp6 = MULTIPLY(tmp12 + tmp13, - FIX(0.497774438)); /* -c9 */ - tmp2 += tmp4 + tmp6 - - MULTIPLY(tmp12, FIX(1.190715098)) + /* c1+c5-c9-c11 */ - MULTIPLY(tmp15, FIX(1.711799069)); /* c3+c7 */ - tmp3 += tmp5 + tmp6 + - MULTIPLY(tmp13, FIX(1.670519935)) - /* c3+c5+c9-c7 */ - MULTIPLY(tmp15, FIX(1.319646532)); /* c1+c11 */ - - dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp0, CONST_BITS+1); - dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp1, CONST_BITS+1); - dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp2, CONST_BITS+1); - dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp3, CONST_BITS+1); - - dataptr++; /* advance pointer to next column */ - wsptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 14x14 sample block. - */ - -GLOBAL(void) -jpeg_fdct_14x14 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6; - INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15, tmp16; - DCTELEM workspace[8*6]; - DCTELEM *dataptr; - DCTELEM *wsptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT. - * cK represents sqrt(2) * cos(K*pi/28). - */ - - dataptr = data; - ctr = 0; - for (;;) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[13]); - tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[12]); - tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[11]); - tmp13 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[10]); - tmp4 = GETJSAMPLE(elemptr[4]) + GETJSAMPLE(elemptr[9]); - tmp5 = GETJSAMPLE(elemptr[5]) + GETJSAMPLE(elemptr[8]); - tmp6 = GETJSAMPLE(elemptr[6]) + GETJSAMPLE(elemptr[7]); - - tmp10 = tmp0 + tmp6; - tmp14 = tmp0 - tmp6; - tmp11 = tmp1 + tmp5; - tmp15 = tmp1 - tmp5; - tmp12 = tmp2 + tmp4; - tmp16 = tmp2 - tmp4; - - tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[13]); - tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[12]); - tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[11]); - tmp3 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[10]); - tmp4 = GETJSAMPLE(elemptr[4]) - GETJSAMPLE(elemptr[9]); - tmp5 = GETJSAMPLE(elemptr[5]) - GETJSAMPLE(elemptr[8]); - tmp6 = GETJSAMPLE(elemptr[6]) - GETJSAMPLE(elemptr[7]); - - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) - (tmp10 + tmp11 + tmp12 + tmp13 - 14 * CENTERJSAMPLE); - tmp13 += tmp13; - dataptr[4] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp13, FIX(1.274162392)) + /* c4 */ - MULTIPLY(tmp11 - tmp13, FIX(0.314692123)) - /* c12 */ - MULTIPLY(tmp12 - tmp13, FIX(0.881747734)), /* c8 */ - CONST_BITS); - - tmp10 = MULTIPLY(tmp14 + tmp15, FIX(1.105676686)); /* c6 */ - - dataptr[2] = (DCTELEM) - DESCALE(tmp10 + MULTIPLY(tmp14, FIX(0.273079590)) /* c2-c6 */ - + MULTIPLY(tmp16, FIX(0.613604268)), /* c10 */ - CONST_BITS); - dataptr[6] = (DCTELEM) - DESCALE(tmp10 - MULTIPLY(tmp15, FIX(1.719280954)) /* c6+c10 */ - - MULTIPLY(tmp16, FIX(1.378756276)), /* c2 */ - CONST_BITS); - - /* Odd part */ - - tmp10 = tmp1 + tmp2; - tmp11 = tmp5 - tmp4; - dataptr[7] = (DCTELEM) (tmp0 - tmp10 + tmp3 - tmp11 - tmp6); - tmp3 <<= CONST_BITS; - tmp10 = MULTIPLY(tmp10, - FIX(0.158341681)); /* -c13 */ - tmp11 = MULTIPLY(tmp11, FIX(1.405321284)); /* c1 */ - tmp10 += tmp11 - tmp3; - tmp11 = MULTIPLY(tmp0 + tmp2, FIX(1.197448846)) + /* c5 */ - MULTIPLY(tmp4 + tmp6, FIX(0.752406978)); /* c9 */ - dataptr[5] = (DCTELEM) - DESCALE(tmp10 + tmp11 - MULTIPLY(tmp2, FIX(2.373959773)) /* c3+c5-c13 */ - + MULTIPLY(tmp4, FIX(1.119999435)), /* c1+c11-c9 */ - CONST_BITS); - tmp12 = MULTIPLY(tmp0 + tmp1, FIX(1.334852607)) + /* c3 */ - MULTIPLY(tmp5 - tmp6, FIX(0.467085129)); /* c11 */ - dataptr[3] = (DCTELEM) - DESCALE(tmp10 + tmp12 - MULTIPLY(tmp1, FIX(0.424103948)) /* c3-c9-c13 */ - - MULTIPLY(tmp5, FIX(3.069855259)), /* c1+c5+c11 */ - CONST_BITS); - dataptr[1] = (DCTELEM) - DESCALE(tmp11 + tmp12 + tmp3 + tmp6 - - MULTIPLY(tmp0 + tmp6, FIX(1.126980169)), /* c3+c5-c1 */ - CONST_BITS); - - ctr++; - - if (ctr != DCTSIZE) { - if (ctr == 14) - break; /* Done. */ - dataptr += DCTSIZE; /* advance pointer to next row */ - } else - dataptr = workspace; /* switch pointer to extended workspace */ - } - - /* Pass 2: process columns. - * We leave the results scaled up by an overall factor of 8. - * We must also scale the output by (8/14)**2 = 16/49, which we partially - * fold into the constant multipliers and final shifting: - * cK now represents sqrt(2) * cos(K*pi/28) * 32/49. - */ - - dataptr = data; - wsptr = workspace; - for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { - /* Even part */ - - tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*5]; - tmp1 = dataptr[DCTSIZE*1] + wsptr[DCTSIZE*4]; - tmp2 = dataptr[DCTSIZE*2] + wsptr[DCTSIZE*3]; - tmp13 = dataptr[DCTSIZE*3] + wsptr[DCTSIZE*2]; - tmp4 = dataptr[DCTSIZE*4] + wsptr[DCTSIZE*1]; - tmp5 = dataptr[DCTSIZE*5] + wsptr[DCTSIZE*0]; - tmp6 = dataptr[DCTSIZE*6] + dataptr[DCTSIZE*7]; - - tmp10 = tmp0 + tmp6; - tmp14 = tmp0 - tmp6; - tmp11 = tmp1 + tmp5; - tmp15 = tmp1 - tmp5; - tmp12 = tmp2 + tmp4; - tmp16 = tmp2 - tmp4; - - tmp0 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*5]; - tmp1 = dataptr[DCTSIZE*1] - wsptr[DCTSIZE*4]; - tmp2 = dataptr[DCTSIZE*2] - wsptr[DCTSIZE*3]; - tmp3 = dataptr[DCTSIZE*3] - wsptr[DCTSIZE*2]; - tmp4 = dataptr[DCTSIZE*4] - wsptr[DCTSIZE*1]; - tmp5 = dataptr[DCTSIZE*5] - wsptr[DCTSIZE*0]; - tmp6 = dataptr[DCTSIZE*6] - dataptr[DCTSIZE*7]; - - dataptr[DCTSIZE*0] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 + tmp11 + tmp12 + tmp13, - FIX(0.653061224)), /* 32/49 */ - CONST_BITS+1); - tmp13 += tmp13; - dataptr[DCTSIZE*4] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp13, FIX(0.832106052)) + /* c4 */ - MULTIPLY(tmp11 - tmp13, FIX(0.205513223)) - /* c12 */ - MULTIPLY(tmp12 - tmp13, FIX(0.575835255)), /* c8 */ - CONST_BITS+1); - - tmp10 = MULTIPLY(tmp14 + tmp15, FIX(0.722074570)); /* c6 */ - - dataptr[DCTSIZE*2] = (DCTELEM) - DESCALE(tmp10 + MULTIPLY(tmp14, FIX(0.178337691)) /* c2-c6 */ - + MULTIPLY(tmp16, FIX(0.400721155)), /* c10 */ - CONST_BITS+1); - dataptr[DCTSIZE*6] = (DCTELEM) - DESCALE(tmp10 - MULTIPLY(tmp15, FIX(1.122795725)) /* c6+c10 */ - - MULTIPLY(tmp16, FIX(0.900412262)), /* c2 */ - CONST_BITS+1); - - /* Odd part */ - - tmp10 = tmp1 + tmp2; - tmp11 = tmp5 - tmp4; - dataptr[DCTSIZE*7] = (DCTELEM) - DESCALE(MULTIPLY(tmp0 - tmp10 + tmp3 - tmp11 - tmp6, - FIX(0.653061224)), /* 32/49 */ - CONST_BITS+1); - tmp3 = MULTIPLY(tmp3 , FIX(0.653061224)); /* 32/49 */ - tmp10 = MULTIPLY(tmp10, - FIX(0.103406812)); /* -c13 */ - tmp11 = MULTIPLY(tmp11, FIX(0.917760839)); /* c1 */ - tmp10 += tmp11 - tmp3; - tmp11 = MULTIPLY(tmp0 + tmp2, FIX(0.782007410)) + /* c5 */ - MULTIPLY(tmp4 + tmp6, FIX(0.491367823)); /* c9 */ - dataptr[DCTSIZE*5] = (DCTELEM) - DESCALE(tmp10 + tmp11 - MULTIPLY(tmp2, FIX(1.550341076)) /* c3+c5-c13 */ - + MULTIPLY(tmp4, FIX(0.731428202)), /* c1+c11-c9 */ - CONST_BITS+1); - tmp12 = MULTIPLY(tmp0 + tmp1, FIX(0.871740478)) + /* c3 */ - MULTIPLY(tmp5 - tmp6, FIX(0.305035186)); /* c11 */ - dataptr[DCTSIZE*3] = (DCTELEM) - DESCALE(tmp10 + tmp12 - MULTIPLY(tmp1, FIX(0.276965844)) /* c3-c9-c13 */ - - MULTIPLY(tmp5, FIX(2.004803435)), /* c1+c5+c11 */ - CONST_BITS+1); - dataptr[DCTSIZE*1] = (DCTELEM) - DESCALE(tmp11 + tmp12 + tmp3 - - MULTIPLY(tmp0, FIX(0.735987049)) /* c3+c5-c1 */ - - MULTIPLY(tmp6, FIX(0.082925825)), /* c9-c11-c13 */ - CONST_BITS+1); - - dataptr++; /* advance pointer to next column */ - wsptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 15x15 sample block. - */ - -GLOBAL(void) -jpeg_fdct_15x15 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7; - INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15, tmp16; - INT32 z1, z2, z3; - DCTELEM workspace[8*7]; - DCTELEM *dataptr; - DCTELEM *wsptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT. - * cK represents sqrt(2) * cos(K*pi/30). - */ - - dataptr = data; - ctr = 0; - for (;;) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[14]); - tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[13]); - tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[12]); - tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[11]); - tmp4 = GETJSAMPLE(elemptr[4]) + GETJSAMPLE(elemptr[10]); - tmp5 = GETJSAMPLE(elemptr[5]) + GETJSAMPLE(elemptr[9]); - tmp6 = GETJSAMPLE(elemptr[6]) + GETJSAMPLE(elemptr[8]); - tmp7 = GETJSAMPLE(elemptr[7]); - - tmp10 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[14]); - tmp11 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[13]); - tmp12 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[12]); - tmp13 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[11]); - tmp14 = GETJSAMPLE(elemptr[4]) - GETJSAMPLE(elemptr[10]); - tmp15 = GETJSAMPLE(elemptr[5]) - GETJSAMPLE(elemptr[9]); - tmp16 = GETJSAMPLE(elemptr[6]) - GETJSAMPLE(elemptr[8]); - - z1 = tmp0 + tmp4 + tmp5; - z2 = tmp1 + tmp3 + tmp6; - z3 = tmp2 + tmp7; - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) (z1 + z2 + z3 - 15 * CENTERJSAMPLE); - z3 += z3; - dataptr[6] = (DCTELEM) - DESCALE(MULTIPLY(z1 - z3, FIX(1.144122806)) - /* c6 */ - MULTIPLY(z2 - z3, FIX(0.437016024)), /* c12 */ - CONST_BITS); - tmp2 += ((tmp1 + tmp4) >> 1) - tmp7 - tmp7; - z1 = MULTIPLY(tmp3 - tmp2, FIX(1.531135173)) - /* c2+c14 */ - MULTIPLY(tmp6 - tmp2, FIX(2.238241955)); /* c4+c8 */ - z2 = MULTIPLY(tmp5 - tmp2, FIX(0.798468008)) - /* c8-c14 */ - MULTIPLY(tmp0 - tmp2, FIX(0.091361227)); /* c2-c4 */ - z3 = MULTIPLY(tmp0 - tmp3, FIX(1.383309603)) + /* c2 */ - MULTIPLY(tmp6 - tmp5, FIX(0.946293579)) + /* c8 */ - MULTIPLY(tmp1 - tmp4, FIX(0.790569415)); /* (c6+c12)/2 */ - - dataptr[2] = (DCTELEM) DESCALE(z1 + z3, CONST_BITS); - dataptr[4] = (DCTELEM) DESCALE(z2 + z3, CONST_BITS); - - /* Odd part */ - - tmp2 = MULTIPLY(tmp10 - tmp12 - tmp13 + tmp15 + tmp16, - FIX(1.224744871)); /* c5 */ - tmp1 = MULTIPLY(tmp10 - tmp14 - tmp15, FIX(1.344997024)) + /* c3 */ - MULTIPLY(tmp11 - tmp13 - tmp16, FIX(0.831253876)); /* c9 */ - tmp12 = MULTIPLY(tmp12, FIX(1.224744871)); /* c5 */ - tmp4 = MULTIPLY(tmp10 - tmp16, FIX(1.406466353)) + /* c1 */ - MULTIPLY(tmp11 + tmp14, FIX(1.344997024)) + /* c3 */ - MULTIPLY(tmp13 + tmp15, FIX(0.575212477)); /* c11 */ - tmp0 = MULTIPLY(tmp13, FIX(0.475753014)) - /* c7-c11 */ - MULTIPLY(tmp14, FIX(0.513743148)) + /* c3-c9 */ - MULTIPLY(tmp16, FIX(1.700497885)) + tmp4 + tmp12; /* c1+c13 */ - tmp3 = MULTIPLY(tmp10, - FIX(0.355500862)) - /* -(c1-c7) */ - MULTIPLY(tmp11, FIX(2.176250899)) - /* c3+c9 */ - MULTIPLY(tmp15, FIX(0.869244010)) + tmp4 - tmp12; /* c11+c13 */ - - dataptr[1] = (DCTELEM) DESCALE(tmp0, CONST_BITS); - dataptr[3] = (DCTELEM) DESCALE(tmp1, CONST_BITS); - dataptr[5] = (DCTELEM) DESCALE(tmp2, CONST_BITS); - dataptr[7] = (DCTELEM) DESCALE(tmp3, CONST_BITS); - - ctr++; - - if (ctr != DCTSIZE) { - if (ctr == 15) - break; /* Done. */ - dataptr += DCTSIZE; /* advance pointer to next row */ - } else - dataptr = workspace; /* switch pointer to extended workspace */ - } - - /* Pass 2: process columns. - * We leave the results scaled up by an overall factor of 8. - * We must also scale the output by (8/15)**2 = 64/225, which we partially - * fold into the constant multipliers and final shifting: - * cK now represents sqrt(2) * cos(K*pi/30) * 256/225. - */ - - dataptr = data; - wsptr = workspace; - for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { - /* Even part */ - - tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*6]; - tmp1 = dataptr[DCTSIZE*1] + wsptr[DCTSIZE*5]; - tmp2 = dataptr[DCTSIZE*2] + wsptr[DCTSIZE*4]; - tmp3 = dataptr[DCTSIZE*3] + wsptr[DCTSIZE*3]; - tmp4 = dataptr[DCTSIZE*4] + wsptr[DCTSIZE*2]; - tmp5 = dataptr[DCTSIZE*5] + wsptr[DCTSIZE*1]; - tmp6 = dataptr[DCTSIZE*6] + wsptr[DCTSIZE*0]; - tmp7 = dataptr[DCTSIZE*7]; - - tmp10 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*6]; - tmp11 = dataptr[DCTSIZE*1] - wsptr[DCTSIZE*5]; - tmp12 = dataptr[DCTSIZE*2] - wsptr[DCTSIZE*4]; - tmp13 = dataptr[DCTSIZE*3] - wsptr[DCTSIZE*3]; - tmp14 = dataptr[DCTSIZE*4] - wsptr[DCTSIZE*2]; - tmp15 = dataptr[DCTSIZE*5] - wsptr[DCTSIZE*1]; - tmp16 = dataptr[DCTSIZE*6] - wsptr[DCTSIZE*0]; - - z1 = tmp0 + tmp4 + tmp5; - z2 = tmp1 + tmp3 + tmp6; - z3 = tmp2 + tmp7; - dataptr[DCTSIZE*0] = (DCTELEM) - DESCALE(MULTIPLY(z1 + z2 + z3, FIX(1.137777778)), /* 256/225 */ - CONST_BITS+2); - z3 += z3; - dataptr[DCTSIZE*6] = (DCTELEM) - DESCALE(MULTIPLY(z1 - z3, FIX(1.301757503)) - /* c6 */ - MULTIPLY(z2 - z3, FIX(0.497227121)), /* c12 */ - CONST_BITS+2); - tmp2 += ((tmp1 + tmp4) >> 1) - tmp7 - tmp7; - z1 = MULTIPLY(tmp3 - tmp2, FIX(1.742091575)) - /* c2+c14 */ - MULTIPLY(tmp6 - tmp2, FIX(2.546621957)); /* c4+c8 */ - z2 = MULTIPLY(tmp5 - tmp2, FIX(0.908479156)) - /* c8-c14 */ - MULTIPLY(tmp0 - tmp2, FIX(0.103948774)); /* c2-c4 */ - z3 = MULTIPLY(tmp0 - tmp3, FIX(1.573898926)) + /* c2 */ - MULTIPLY(tmp6 - tmp5, FIX(1.076671805)) + /* c8 */ - MULTIPLY(tmp1 - tmp4, FIX(0.899492312)); /* (c6+c12)/2 */ - - dataptr[DCTSIZE*2] = (DCTELEM) DESCALE(z1 + z3, CONST_BITS+2); - dataptr[DCTSIZE*4] = (DCTELEM) DESCALE(z2 + z3, CONST_BITS+2); - - /* Odd part */ - - tmp2 = MULTIPLY(tmp10 - tmp12 - tmp13 + tmp15 + tmp16, - FIX(1.393487498)); /* c5 */ - tmp1 = MULTIPLY(tmp10 - tmp14 - tmp15, FIX(1.530307725)) + /* c3 */ - MULTIPLY(tmp11 - tmp13 - tmp16, FIX(0.945782187)); /* c9 */ - tmp12 = MULTIPLY(tmp12, FIX(1.393487498)); /* c5 */ - tmp4 = MULTIPLY(tmp10 - tmp16, FIX(1.600246161)) + /* c1 */ - MULTIPLY(tmp11 + tmp14, FIX(1.530307725)) + /* c3 */ - MULTIPLY(tmp13 + tmp15, FIX(0.654463974)); /* c11 */ - tmp0 = MULTIPLY(tmp13, FIX(0.541301207)) - /* c7-c11 */ - MULTIPLY(tmp14, FIX(0.584525538)) + /* c3-c9 */ - MULTIPLY(tmp16, FIX(1.934788705)) + tmp4 + tmp12; /* c1+c13 */ - tmp3 = MULTIPLY(tmp10, - FIX(0.404480980)) - /* -(c1-c7) */ - MULTIPLY(tmp11, FIX(2.476089912)) - /* c3+c9 */ - MULTIPLY(tmp15, FIX(0.989006518)) + tmp4 - tmp12; /* c11+c13 */ - - dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp0, CONST_BITS+2); - dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp1, CONST_BITS+2); - dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp2, CONST_BITS+2); - dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp3, CONST_BITS+2); - - dataptr++; /* advance pointer to next column */ - wsptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 16x16 sample block. - */ - -GLOBAL(void) -jpeg_fdct_16x16 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7; - INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15, tmp16, tmp17; - DCTELEM workspace[DCTSIZE2]; - DCTELEM *dataptr; - DCTELEM *wsptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT; - * furthermore, we scale the results by 2**PASS1_BITS. - * cK represents sqrt(2) * cos(K*pi/32). - */ - - dataptr = data; - ctr = 0; - for (;;) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[15]); - tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[14]); - tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[13]); - tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[12]); - tmp4 = GETJSAMPLE(elemptr[4]) + GETJSAMPLE(elemptr[11]); - tmp5 = GETJSAMPLE(elemptr[5]) + GETJSAMPLE(elemptr[10]); - tmp6 = GETJSAMPLE(elemptr[6]) + GETJSAMPLE(elemptr[9]); - tmp7 = GETJSAMPLE(elemptr[7]) + GETJSAMPLE(elemptr[8]); - - tmp10 = tmp0 + tmp7; - tmp14 = tmp0 - tmp7; - tmp11 = tmp1 + tmp6; - tmp15 = tmp1 - tmp6; - tmp12 = tmp2 + tmp5; - tmp16 = tmp2 - tmp5; - tmp13 = tmp3 + tmp4; - tmp17 = tmp3 - tmp4; - - tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[15]); - tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[14]); - tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[13]); - tmp3 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[12]); - tmp4 = GETJSAMPLE(elemptr[4]) - GETJSAMPLE(elemptr[11]); - tmp5 = GETJSAMPLE(elemptr[5]) - GETJSAMPLE(elemptr[10]); - tmp6 = GETJSAMPLE(elemptr[6]) - GETJSAMPLE(elemptr[9]); - tmp7 = GETJSAMPLE(elemptr[7]) - GETJSAMPLE(elemptr[8]); - - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) - ((tmp10 + tmp11 + tmp12 + tmp13 - 16 * CENTERJSAMPLE) << PASS1_BITS); - dataptr[4] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp13, FIX(1.306562965)) + /* c4[16] = c2[8] */ - MULTIPLY(tmp11 - tmp12, FIX_0_541196100), /* c12[16] = c6[8] */ - CONST_BITS-PASS1_BITS); - - tmp10 = MULTIPLY(tmp17 - tmp15, FIX(0.275899379)) + /* c14[16] = c7[8] */ - MULTIPLY(tmp14 - tmp16, FIX(1.387039845)); /* c2[16] = c1[8] */ - - dataptr[2] = (DCTELEM) - DESCALE(tmp10 + MULTIPLY(tmp15, FIX(1.451774982)) /* c6+c14 */ - + MULTIPLY(tmp16, FIX(2.172734804)), /* c2+c10 */ - CONST_BITS-PASS1_BITS); - dataptr[6] = (DCTELEM) - DESCALE(tmp10 - MULTIPLY(tmp14, FIX(0.211164243)) /* c2-c6 */ - - MULTIPLY(tmp17, FIX(1.061594338)), /* c10+c14 */ - CONST_BITS-PASS1_BITS); - - /* Odd part */ - - tmp11 = MULTIPLY(tmp0 + tmp1, FIX(1.353318001)) + /* c3 */ - MULTIPLY(tmp6 - tmp7, FIX(0.410524528)); /* c13 */ - tmp12 = MULTIPLY(tmp0 + tmp2, FIX(1.247225013)) + /* c5 */ - MULTIPLY(tmp5 + tmp7, FIX(0.666655658)); /* c11 */ - tmp13 = MULTIPLY(tmp0 + tmp3, FIX(1.093201867)) + /* c7 */ - MULTIPLY(tmp4 - tmp7, FIX(0.897167586)); /* c9 */ - tmp14 = MULTIPLY(tmp1 + tmp2, FIX(0.138617169)) + /* c15 */ - MULTIPLY(tmp6 - tmp5, FIX(1.407403738)); /* c1 */ - tmp15 = MULTIPLY(tmp1 + tmp3, - FIX(0.666655658)) + /* -c11 */ - MULTIPLY(tmp4 + tmp6, - FIX(1.247225013)); /* -c5 */ - tmp16 = MULTIPLY(tmp2 + tmp3, - FIX(1.353318001)) + /* -c3 */ - MULTIPLY(tmp5 - tmp4, FIX(0.410524528)); /* c13 */ - tmp10 = tmp11 + tmp12 + tmp13 - - MULTIPLY(tmp0, FIX(2.286341144)) + /* c7+c5+c3-c1 */ - MULTIPLY(tmp7, FIX(0.779653625)); /* c15+c13-c11+c9 */ - tmp11 += tmp14 + tmp15 + MULTIPLY(tmp1, FIX(0.071888074)) /* c9-c3-c15+c11 */ - - MULTIPLY(tmp6, FIX(1.663905119)); /* c7+c13+c1-c5 */ - tmp12 += tmp14 + tmp16 - MULTIPLY(tmp2, FIX(1.125726048)) /* c7+c5+c15-c3 */ - + MULTIPLY(tmp5, FIX(1.227391138)); /* c9-c11+c1-c13 */ - tmp13 += tmp15 + tmp16 + MULTIPLY(tmp3, FIX(1.065388962)) /* c15+c3+c11-c7 */ - + MULTIPLY(tmp4, FIX(2.167985692)); /* c1+c13+c5-c9 */ - - dataptr[1] = (DCTELEM) DESCALE(tmp10, CONST_BITS-PASS1_BITS); - dataptr[3] = (DCTELEM) DESCALE(tmp11, CONST_BITS-PASS1_BITS); - dataptr[5] = (DCTELEM) DESCALE(tmp12, CONST_BITS-PASS1_BITS); - dataptr[7] = (DCTELEM) DESCALE(tmp13, CONST_BITS-PASS1_BITS); - - ctr++; - - if (ctr != DCTSIZE) { - if (ctr == DCTSIZE * 2) - break; /* Done. */ - dataptr += DCTSIZE; /* advance pointer to next row */ - } else - dataptr = workspace; /* switch pointer to extended workspace */ - } - - /* Pass 2: process columns. - * We remove the PASS1_BITS scaling, but leave the results scaled up - * by an overall factor of 8. - * We must also scale the output by (8/16)**2 = 1/2**2. - * cK represents sqrt(2) * cos(K*pi/32). - */ - - dataptr = data; - wsptr = workspace; - for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { - /* Even part */ - - tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*7]; - tmp1 = dataptr[DCTSIZE*1] + wsptr[DCTSIZE*6]; - tmp2 = dataptr[DCTSIZE*2] + wsptr[DCTSIZE*5]; - tmp3 = dataptr[DCTSIZE*3] + wsptr[DCTSIZE*4]; - tmp4 = dataptr[DCTSIZE*4] + wsptr[DCTSIZE*3]; - tmp5 = dataptr[DCTSIZE*5] + wsptr[DCTSIZE*2]; - tmp6 = dataptr[DCTSIZE*6] + wsptr[DCTSIZE*1]; - tmp7 = dataptr[DCTSIZE*7] + wsptr[DCTSIZE*0]; - - tmp10 = tmp0 + tmp7; - tmp14 = tmp0 - tmp7; - tmp11 = tmp1 + tmp6; - tmp15 = tmp1 - tmp6; - tmp12 = tmp2 + tmp5; - tmp16 = tmp2 - tmp5; - tmp13 = tmp3 + tmp4; - tmp17 = tmp3 - tmp4; - - tmp0 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*7]; - tmp1 = dataptr[DCTSIZE*1] - wsptr[DCTSIZE*6]; - tmp2 = dataptr[DCTSIZE*2] - wsptr[DCTSIZE*5]; - tmp3 = dataptr[DCTSIZE*3] - wsptr[DCTSIZE*4]; - tmp4 = dataptr[DCTSIZE*4] - wsptr[DCTSIZE*3]; - tmp5 = dataptr[DCTSIZE*5] - wsptr[DCTSIZE*2]; - tmp6 = dataptr[DCTSIZE*6] - wsptr[DCTSIZE*1]; - tmp7 = dataptr[DCTSIZE*7] - wsptr[DCTSIZE*0]; - - dataptr[DCTSIZE*0] = (DCTELEM) - DESCALE(tmp10 + tmp11 + tmp12 + tmp13, PASS1_BITS+2); - dataptr[DCTSIZE*4] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp13, FIX(1.306562965)) + /* c4[16] = c2[8] */ - MULTIPLY(tmp11 - tmp12, FIX_0_541196100), /* c12[16] = c6[8] */ - CONST_BITS+PASS1_BITS+2); - - tmp10 = MULTIPLY(tmp17 - tmp15, FIX(0.275899379)) + /* c14[16] = c7[8] */ - MULTIPLY(tmp14 - tmp16, FIX(1.387039845)); /* c2[16] = c1[8] */ - - dataptr[DCTSIZE*2] = (DCTELEM) - DESCALE(tmp10 + MULTIPLY(tmp15, FIX(1.451774982)) /* c6+c14 */ - + MULTIPLY(tmp16, FIX(2.172734804)), /* c2+10 */ - CONST_BITS+PASS1_BITS+2); - dataptr[DCTSIZE*6] = (DCTELEM) - DESCALE(tmp10 - MULTIPLY(tmp14, FIX(0.211164243)) /* c2-c6 */ - - MULTIPLY(tmp17, FIX(1.061594338)), /* c10+c14 */ - CONST_BITS+PASS1_BITS+2); - - /* Odd part */ - - tmp11 = MULTIPLY(tmp0 + tmp1, FIX(1.353318001)) + /* c3 */ - MULTIPLY(tmp6 - tmp7, FIX(0.410524528)); /* c13 */ - tmp12 = MULTIPLY(tmp0 + tmp2, FIX(1.247225013)) + /* c5 */ - MULTIPLY(tmp5 + tmp7, FIX(0.666655658)); /* c11 */ - tmp13 = MULTIPLY(tmp0 + tmp3, FIX(1.093201867)) + /* c7 */ - MULTIPLY(tmp4 - tmp7, FIX(0.897167586)); /* c9 */ - tmp14 = MULTIPLY(tmp1 + tmp2, FIX(0.138617169)) + /* c15 */ - MULTIPLY(tmp6 - tmp5, FIX(1.407403738)); /* c1 */ - tmp15 = MULTIPLY(tmp1 + tmp3, - FIX(0.666655658)) + /* -c11 */ - MULTIPLY(tmp4 + tmp6, - FIX(1.247225013)); /* -c5 */ - tmp16 = MULTIPLY(tmp2 + tmp3, - FIX(1.353318001)) + /* -c3 */ - MULTIPLY(tmp5 - tmp4, FIX(0.410524528)); /* c13 */ - tmp10 = tmp11 + tmp12 + tmp13 - - MULTIPLY(tmp0, FIX(2.286341144)) + /* c7+c5+c3-c1 */ - MULTIPLY(tmp7, FIX(0.779653625)); /* c15+c13-c11+c9 */ - tmp11 += tmp14 + tmp15 + MULTIPLY(tmp1, FIX(0.071888074)) /* c9-c3-c15+c11 */ - - MULTIPLY(tmp6, FIX(1.663905119)); /* c7+c13+c1-c5 */ - tmp12 += tmp14 + tmp16 - MULTIPLY(tmp2, FIX(1.125726048)) /* c7+c5+c15-c3 */ - + MULTIPLY(tmp5, FIX(1.227391138)); /* c9-c11+c1-c13 */ - tmp13 += tmp15 + tmp16 + MULTIPLY(tmp3, FIX(1.065388962)) /* c15+c3+c11-c7 */ - + MULTIPLY(tmp4, FIX(2.167985692)); /* c1+c13+c5-c9 */ - - dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp10, CONST_BITS+PASS1_BITS+2); - dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp11, CONST_BITS+PASS1_BITS+2); - dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp12, CONST_BITS+PASS1_BITS+2); - dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp13, CONST_BITS+PASS1_BITS+2); - - dataptr++; /* advance pointer to next column */ - wsptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 16x8 sample block. - * - * 16-point FDCT in pass 1 (rows), 8-point in pass 2 (columns). - */ - -GLOBAL(void) -jpeg_fdct_16x8 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7; - INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15, tmp16, tmp17; - INT32 z1; - DCTELEM *dataptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT; - * furthermore, we scale the results by 2**PASS1_BITS. - * 16-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/32). - */ - - dataptr = data; - ctr = 0; - for (ctr = 0; ctr < DCTSIZE; ctr++) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[15]); - tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[14]); - tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[13]); - tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[12]); - tmp4 = GETJSAMPLE(elemptr[4]) + GETJSAMPLE(elemptr[11]); - tmp5 = GETJSAMPLE(elemptr[5]) + GETJSAMPLE(elemptr[10]); - tmp6 = GETJSAMPLE(elemptr[6]) + GETJSAMPLE(elemptr[9]); - tmp7 = GETJSAMPLE(elemptr[7]) + GETJSAMPLE(elemptr[8]); - - tmp10 = tmp0 + tmp7; - tmp14 = tmp0 - tmp7; - tmp11 = tmp1 + tmp6; - tmp15 = tmp1 - tmp6; - tmp12 = tmp2 + tmp5; - tmp16 = tmp2 - tmp5; - tmp13 = tmp3 + tmp4; - tmp17 = tmp3 - tmp4; - - tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[15]); - tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[14]); - tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[13]); - tmp3 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[12]); - tmp4 = GETJSAMPLE(elemptr[4]) - GETJSAMPLE(elemptr[11]); - tmp5 = GETJSAMPLE(elemptr[5]) - GETJSAMPLE(elemptr[10]); - tmp6 = GETJSAMPLE(elemptr[6]) - GETJSAMPLE(elemptr[9]); - tmp7 = GETJSAMPLE(elemptr[7]) - GETJSAMPLE(elemptr[8]); - - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) - ((tmp10 + tmp11 + tmp12 + tmp13 - 16 * CENTERJSAMPLE) << PASS1_BITS); - dataptr[4] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp13, FIX(1.306562965)) + /* c4[16] = c2[8] */ - MULTIPLY(tmp11 - tmp12, FIX_0_541196100), /* c12[16] = c6[8] */ - CONST_BITS-PASS1_BITS); - - tmp10 = MULTIPLY(tmp17 - tmp15, FIX(0.275899379)) + /* c14[16] = c7[8] */ - MULTIPLY(tmp14 - tmp16, FIX(1.387039845)); /* c2[16] = c1[8] */ - - dataptr[2] = (DCTELEM) - DESCALE(tmp10 + MULTIPLY(tmp15, FIX(1.451774982)) /* c6+c14 */ - + MULTIPLY(tmp16, FIX(2.172734804)), /* c2+c10 */ - CONST_BITS-PASS1_BITS); - dataptr[6] = (DCTELEM) - DESCALE(tmp10 - MULTIPLY(tmp14, FIX(0.211164243)) /* c2-c6 */ - - MULTIPLY(tmp17, FIX(1.061594338)), /* c10+c14 */ - CONST_BITS-PASS1_BITS); - - /* Odd part */ - - tmp11 = MULTIPLY(tmp0 + tmp1, FIX(1.353318001)) + /* c3 */ - MULTIPLY(tmp6 - tmp7, FIX(0.410524528)); /* c13 */ - tmp12 = MULTIPLY(tmp0 + tmp2, FIX(1.247225013)) + /* c5 */ - MULTIPLY(tmp5 + tmp7, FIX(0.666655658)); /* c11 */ - tmp13 = MULTIPLY(tmp0 + tmp3, FIX(1.093201867)) + /* c7 */ - MULTIPLY(tmp4 - tmp7, FIX(0.897167586)); /* c9 */ - tmp14 = MULTIPLY(tmp1 + tmp2, FIX(0.138617169)) + /* c15 */ - MULTIPLY(tmp6 - tmp5, FIX(1.407403738)); /* c1 */ - tmp15 = MULTIPLY(tmp1 + tmp3, - FIX(0.666655658)) + /* -c11 */ - MULTIPLY(tmp4 + tmp6, - FIX(1.247225013)); /* -c5 */ - tmp16 = MULTIPLY(tmp2 + tmp3, - FIX(1.353318001)) + /* -c3 */ - MULTIPLY(tmp5 - tmp4, FIX(0.410524528)); /* c13 */ - tmp10 = tmp11 + tmp12 + tmp13 - - MULTIPLY(tmp0, FIX(2.286341144)) + /* c7+c5+c3-c1 */ - MULTIPLY(tmp7, FIX(0.779653625)); /* c15+c13-c11+c9 */ - tmp11 += tmp14 + tmp15 + MULTIPLY(tmp1, FIX(0.071888074)) /* c9-c3-c15+c11 */ - - MULTIPLY(tmp6, FIX(1.663905119)); /* c7+c13+c1-c5 */ - tmp12 += tmp14 + tmp16 - MULTIPLY(tmp2, FIX(1.125726048)) /* c7+c5+c15-c3 */ - + MULTIPLY(tmp5, FIX(1.227391138)); /* c9-c11+c1-c13 */ - tmp13 += tmp15 + tmp16 + MULTIPLY(tmp3, FIX(1.065388962)) /* c15+c3+c11-c7 */ - + MULTIPLY(tmp4, FIX(2.167985692)); /* c1+c13+c5-c9 */ - - dataptr[1] = (DCTELEM) DESCALE(tmp10, CONST_BITS-PASS1_BITS); - dataptr[3] = (DCTELEM) DESCALE(tmp11, CONST_BITS-PASS1_BITS); - dataptr[5] = (DCTELEM) DESCALE(tmp12, CONST_BITS-PASS1_BITS); - dataptr[7] = (DCTELEM) DESCALE(tmp13, CONST_BITS-PASS1_BITS); - - dataptr += DCTSIZE; /* advance pointer to next row */ - } - - /* Pass 2: process columns. - * We remove the PASS1_BITS scaling, but leave the results scaled up - * by an overall factor of 8. - * We must also scale the output by 8/16 = 1/2. - * 8-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/16). - */ - - dataptr = data; - for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { - /* Even part per LL&M figure 1 --- note that published figure is faulty; - * rotator "c1" should be "c6". - */ - - tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*7]; - tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*6]; - tmp2 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*5]; - tmp3 = dataptr[DCTSIZE*3] + dataptr[DCTSIZE*4]; - - tmp10 = tmp0 + tmp3; - tmp12 = tmp0 - tmp3; - tmp11 = tmp1 + tmp2; - tmp13 = tmp1 - tmp2; - - tmp0 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*7]; - tmp1 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*6]; - tmp2 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*5]; - tmp3 = dataptr[DCTSIZE*3] - dataptr[DCTSIZE*4]; - - dataptr[DCTSIZE*0] = (DCTELEM) DESCALE(tmp10 + tmp11, PASS1_BITS+1); - dataptr[DCTSIZE*4] = (DCTELEM) DESCALE(tmp10 - tmp11, PASS1_BITS+1); - - z1 = MULTIPLY(tmp12 + tmp13, FIX_0_541196100); /* c6 */ - dataptr[DCTSIZE*2] = (DCTELEM) - DESCALE(z1 + MULTIPLY(tmp12, FIX_0_765366865), /* c2-c6 */ - CONST_BITS+PASS1_BITS+1); - dataptr[DCTSIZE*6] = (DCTELEM) - DESCALE(z1 - MULTIPLY(tmp13, FIX_1_847759065), /* c2+c6 */ - CONST_BITS+PASS1_BITS+1); - - /* Odd part per figure 8 --- note paper omits factor of sqrt(2). - * i0..i3 in the paper are tmp0..tmp3 here. - */ - - tmp12 = tmp0 + tmp2; - tmp13 = tmp1 + tmp3; - - z1 = MULTIPLY(tmp12 + tmp13, FIX_1_175875602); /* c3 */ - tmp12 = MULTIPLY(tmp12, - FIX_0_390180644); /* -c3+c5 */ - tmp13 = MULTIPLY(tmp13, - FIX_1_961570560); /* -c3-c5 */ - tmp12 += z1; - tmp13 += z1; - - z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ - tmp0 = MULTIPLY(tmp0, FIX_1_501321110); /* c1+c3-c5-c7 */ - tmp3 = MULTIPLY(tmp3, FIX_0_298631336); /* -c1+c3+c5-c7 */ - tmp0 += z1 + tmp12; - tmp3 += z1 + tmp13; - - z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ - tmp1 = MULTIPLY(tmp1, FIX_3_072711026); /* c1+c3+c5-c7 */ - tmp2 = MULTIPLY(tmp2, FIX_2_053119869); /* c1+c3-c5+c7 */ - tmp1 += z1 + tmp13; - tmp2 += z1 + tmp12; - - dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp0, CONST_BITS+PASS1_BITS+1); - dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp1, CONST_BITS+PASS1_BITS+1); - dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp2, CONST_BITS+PASS1_BITS+1); - dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp3, CONST_BITS+PASS1_BITS+1); - - dataptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 14x7 sample block. - * - * 14-point FDCT in pass 1 (rows), 7-point in pass 2 (columns). - */ - -GLOBAL(void) -jpeg_fdct_14x7 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6; - INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15, tmp16; - INT32 z1, z2, z3; - DCTELEM *dataptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Zero bottom row of output coefficient block. */ - MEMZERO(&data[DCTSIZE*7], SIZEOF(DCTELEM) * DCTSIZE); - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT; - * furthermore, we scale the results by 2**PASS1_BITS. - * 14-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/28). - */ - - dataptr = data; - for (ctr = 0; ctr < 7; ctr++) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[13]); - tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[12]); - tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[11]); - tmp13 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[10]); - tmp4 = GETJSAMPLE(elemptr[4]) + GETJSAMPLE(elemptr[9]); - tmp5 = GETJSAMPLE(elemptr[5]) + GETJSAMPLE(elemptr[8]); - tmp6 = GETJSAMPLE(elemptr[6]) + GETJSAMPLE(elemptr[7]); - - tmp10 = tmp0 + tmp6; - tmp14 = tmp0 - tmp6; - tmp11 = tmp1 + tmp5; - tmp15 = tmp1 - tmp5; - tmp12 = tmp2 + tmp4; - tmp16 = tmp2 - tmp4; - - tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[13]); - tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[12]); - tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[11]); - tmp3 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[10]); - tmp4 = GETJSAMPLE(elemptr[4]) - GETJSAMPLE(elemptr[9]); - tmp5 = GETJSAMPLE(elemptr[5]) - GETJSAMPLE(elemptr[8]); - tmp6 = GETJSAMPLE(elemptr[6]) - GETJSAMPLE(elemptr[7]); - - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) - ((tmp10 + tmp11 + tmp12 + tmp13 - 14 * CENTERJSAMPLE) << PASS1_BITS); - tmp13 += tmp13; - dataptr[4] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp13, FIX(1.274162392)) + /* c4 */ - MULTIPLY(tmp11 - tmp13, FIX(0.314692123)) - /* c12 */ - MULTIPLY(tmp12 - tmp13, FIX(0.881747734)), /* c8 */ - CONST_BITS-PASS1_BITS); - - tmp10 = MULTIPLY(tmp14 + tmp15, FIX(1.105676686)); /* c6 */ - - dataptr[2] = (DCTELEM) - DESCALE(tmp10 + MULTIPLY(tmp14, FIX(0.273079590)) /* c2-c6 */ - + MULTIPLY(tmp16, FIX(0.613604268)), /* c10 */ - CONST_BITS-PASS1_BITS); - dataptr[6] = (DCTELEM) - DESCALE(tmp10 - MULTIPLY(tmp15, FIX(1.719280954)) /* c6+c10 */ - - MULTIPLY(tmp16, FIX(1.378756276)), /* c2 */ - CONST_BITS-PASS1_BITS); - - /* Odd part */ - - tmp10 = tmp1 + tmp2; - tmp11 = tmp5 - tmp4; - dataptr[7] = (DCTELEM) ((tmp0 - tmp10 + tmp3 - tmp11 - tmp6) << PASS1_BITS); - tmp3 <<= CONST_BITS; - tmp10 = MULTIPLY(tmp10, - FIX(0.158341681)); /* -c13 */ - tmp11 = MULTIPLY(tmp11, FIX(1.405321284)); /* c1 */ - tmp10 += tmp11 - tmp3; - tmp11 = MULTIPLY(tmp0 + tmp2, FIX(1.197448846)) + /* c5 */ - MULTIPLY(tmp4 + tmp6, FIX(0.752406978)); /* c9 */ - dataptr[5] = (DCTELEM) - DESCALE(tmp10 + tmp11 - MULTIPLY(tmp2, FIX(2.373959773)) /* c3+c5-c13 */ - + MULTIPLY(tmp4, FIX(1.119999435)), /* c1+c11-c9 */ - CONST_BITS-PASS1_BITS); - tmp12 = MULTIPLY(tmp0 + tmp1, FIX(1.334852607)) + /* c3 */ - MULTIPLY(tmp5 - tmp6, FIX(0.467085129)); /* c11 */ - dataptr[3] = (DCTELEM) - DESCALE(tmp10 + tmp12 - MULTIPLY(tmp1, FIX(0.424103948)) /* c3-c9-c13 */ - - MULTIPLY(tmp5, FIX(3.069855259)), /* c1+c5+c11 */ - CONST_BITS-PASS1_BITS); - dataptr[1] = (DCTELEM) - DESCALE(tmp11 + tmp12 + tmp3 + tmp6 - - MULTIPLY(tmp0 + tmp6, FIX(1.126980169)), /* c3+c5-c1 */ - CONST_BITS-PASS1_BITS); - - dataptr += DCTSIZE; /* advance pointer to next row */ - } - - /* Pass 2: process columns. - * We remove the PASS1_BITS scaling, but leave the results scaled up - * by an overall factor of 8. - * We must also scale the output by (8/14)*(8/7) = 32/49, which we - * partially fold into the constant multipliers and final shifting: - * 7-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/14) * 64/49. - */ - - dataptr = data; - for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { - /* Even part */ - - tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*6]; - tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*5]; - tmp2 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*4]; - tmp3 = dataptr[DCTSIZE*3]; - - tmp10 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*6]; - tmp11 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*5]; - tmp12 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*4]; - - z1 = tmp0 + tmp2; - dataptr[DCTSIZE*0] = (DCTELEM) - DESCALE(MULTIPLY(z1 + tmp1 + tmp3, FIX(1.306122449)), /* 64/49 */ - CONST_BITS+PASS1_BITS+1); - tmp3 += tmp3; - z1 -= tmp3; - z1 -= tmp3; - z1 = MULTIPLY(z1, FIX(0.461784020)); /* (c2+c6-c4)/2 */ - z2 = MULTIPLY(tmp0 - tmp2, FIX(1.202428084)); /* (c2+c4-c6)/2 */ - z3 = MULTIPLY(tmp1 - tmp2, FIX(0.411026446)); /* c6 */ - dataptr[DCTSIZE*2] = (DCTELEM) DESCALE(z1 + z2 + z3, CONST_BITS+PASS1_BITS+1); - z1 -= z2; - z2 = MULTIPLY(tmp0 - tmp1, FIX(1.151670509)); /* c4 */ - dataptr[DCTSIZE*4] = (DCTELEM) - DESCALE(z2 + z3 - MULTIPLY(tmp1 - tmp3, FIX(0.923568041)), /* c2+c6-c4 */ - CONST_BITS+PASS1_BITS+1); - dataptr[DCTSIZE*6] = (DCTELEM) DESCALE(z1 + z2, CONST_BITS+PASS1_BITS+1); - - /* Odd part */ - - tmp1 = MULTIPLY(tmp10 + tmp11, FIX(1.221765677)); /* (c3+c1-c5)/2 */ - tmp2 = MULTIPLY(tmp10 - tmp11, FIX(0.222383464)); /* (c3+c5-c1)/2 */ - tmp0 = tmp1 - tmp2; - tmp1 += tmp2; - tmp2 = MULTIPLY(tmp11 + tmp12, - FIX(1.800824523)); /* -c1 */ - tmp1 += tmp2; - tmp3 = MULTIPLY(tmp10 + tmp12, FIX(0.801442310)); /* c5 */ - tmp0 += tmp3; - tmp2 += tmp3 + MULTIPLY(tmp12, FIX(2.443531355)); /* c3+c1-c5 */ - - dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp0, CONST_BITS+PASS1_BITS+1); - dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp1, CONST_BITS+PASS1_BITS+1); - dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp2, CONST_BITS+PASS1_BITS+1); - - dataptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 12x6 sample block. - * - * 12-point FDCT in pass 1 (rows), 6-point in pass 2 (columns). - */ - -GLOBAL(void) -jpeg_fdct_12x6 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5; - INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15; - DCTELEM *dataptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Zero 2 bottom rows of output coefficient block. */ - MEMZERO(&data[DCTSIZE*6], SIZEOF(DCTELEM) * DCTSIZE * 2); - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT; - * furthermore, we scale the results by 2**PASS1_BITS. - * 12-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/24). - */ - - dataptr = data; - for (ctr = 0; ctr < 6; ctr++) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[11]); - tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[10]); - tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[9]); - tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[8]); - tmp4 = GETJSAMPLE(elemptr[4]) + GETJSAMPLE(elemptr[7]); - tmp5 = GETJSAMPLE(elemptr[5]) + GETJSAMPLE(elemptr[6]); - - tmp10 = tmp0 + tmp5; - tmp13 = tmp0 - tmp5; - tmp11 = tmp1 + tmp4; - tmp14 = tmp1 - tmp4; - tmp12 = tmp2 + tmp3; - tmp15 = tmp2 - tmp3; - - tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[11]); - tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[10]); - tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[9]); - tmp3 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[8]); - tmp4 = GETJSAMPLE(elemptr[4]) - GETJSAMPLE(elemptr[7]); - tmp5 = GETJSAMPLE(elemptr[5]) - GETJSAMPLE(elemptr[6]); - - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) - ((tmp10 + tmp11 + tmp12 - 12 * CENTERJSAMPLE) << PASS1_BITS); - dataptr[6] = (DCTELEM) ((tmp13 - tmp14 - tmp15) << PASS1_BITS); - dataptr[4] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp12, FIX(1.224744871)), /* c4 */ - CONST_BITS-PASS1_BITS); - dataptr[2] = (DCTELEM) - DESCALE(tmp14 - tmp15 + MULTIPLY(tmp13 + tmp15, FIX(1.366025404)), /* c2 */ - CONST_BITS-PASS1_BITS); - - /* Odd part */ - - tmp10 = MULTIPLY(tmp1 + tmp4, FIX_0_541196100); /* c9 */ - tmp14 = tmp10 + MULTIPLY(tmp1, FIX_0_765366865); /* c3-c9 */ - tmp15 = tmp10 - MULTIPLY(tmp4, FIX_1_847759065); /* c3+c9 */ - tmp12 = MULTIPLY(tmp0 + tmp2, FIX(1.121971054)); /* c5 */ - tmp13 = MULTIPLY(tmp0 + tmp3, FIX(0.860918669)); /* c7 */ - tmp10 = tmp12 + tmp13 + tmp14 - MULTIPLY(tmp0, FIX(0.580774953)) /* c5+c7-c1 */ - + MULTIPLY(tmp5, FIX(0.184591911)); /* c11 */ - tmp11 = MULTIPLY(tmp2 + tmp3, - FIX(0.184591911)); /* -c11 */ - tmp12 += tmp11 - tmp15 - MULTIPLY(tmp2, FIX(2.339493912)) /* c1+c5-c11 */ - + MULTIPLY(tmp5, FIX(0.860918669)); /* c7 */ - tmp13 += tmp11 - tmp14 + MULTIPLY(tmp3, FIX(0.725788011)) /* c1+c11-c7 */ - - MULTIPLY(tmp5, FIX(1.121971054)); /* c5 */ - tmp11 = tmp15 + MULTIPLY(tmp0 - tmp3, FIX(1.306562965)) /* c3 */ - - MULTIPLY(tmp2 + tmp5, FIX_0_541196100); /* c9 */ - - dataptr[1] = (DCTELEM) DESCALE(tmp10, CONST_BITS-PASS1_BITS); - dataptr[3] = (DCTELEM) DESCALE(tmp11, CONST_BITS-PASS1_BITS); - dataptr[5] = (DCTELEM) DESCALE(tmp12, CONST_BITS-PASS1_BITS); - dataptr[7] = (DCTELEM) DESCALE(tmp13, CONST_BITS-PASS1_BITS); - - dataptr += DCTSIZE; /* advance pointer to next row */ - } - - /* Pass 2: process columns. - * We remove the PASS1_BITS scaling, but leave the results scaled up - * by an overall factor of 8. - * We must also scale the output by (8/12)*(8/6) = 8/9, which we - * partially fold into the constant multipliers and final shifting: - * 6-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/12) * 16/9. - */ - - dataptr = data; - for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { - /* Even part */ - - tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*5]; - tmp11 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*4]; - tmp2 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*3]; - - tmp10 = tmp0 + tmp2; - tmp12 = tmp0 - tmp2; - - tmp0 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*5]; - tmp1 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*4]; - tmp2 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*3]; - - dataptr[DCTSIZE*0] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 + tmp11, FIX(1.777777778)), /* 16/9 */ - CONST_BITS+PASS1_BITS+1); - dataptr[DCTSIZE*2] = (DCTELEM) - DESCALE(MULTIPLY(tmp12, FIX(2.177324216)), /* c2 */ - CONST_BITS+PASS1_BITS+1); - dataptr[DCTSIZE*4] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp11 - tmp11, FIX(1.257078722)), /* c4 */ - CONST_BITS+PASS1_BITS+1); - - /* Odd part */ - - tmp10 = MULTIPLY(tmp0 + tmp2, FIX(0.650711829)); /* c5 */ - - dataptr[DCTSIZE*1] = (DCTELEM) - DESCALE(tmp10 + MULTIPLY(tmp0 + tmp1, FIX(1.777777778)), /* 16/9 */ - CONST_BITS+PASS1_BITS+1); - dataptr[DCTSIZE*3] = (DCTELEM) - DESCALE(MULTIPLY(tmp0 - tmp1 - tmp2, FIX(1.777777778)), /* 16/9 */ - CONST_BITS+PASS1_BITS+1); - dataptr[DCTSIZE*5] = (DCTELEM) - DESCALE(tmp10 + MULTIPLY(tmp2 - tmp1, FIX(1.777777778)), /* 16/9 */ - CONST_BITS+PASS1_BITS+1); - - dataptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 10x5 sample block. - * - * 10-point FDCT in pass 1 (rows), 5-point in pass 2 (columns). - */ - -GLOBAL(void) -jpeg_fdct_10x5 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3, tmp4; - INT32 tmp10, tmp11, tmp12, tmp13, tmp14; - DCTELEM *dataptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Zero 3 bottom rows of output coefficient block. */ - MEMZERO(&data[DCTSIZE*5], SIZEOF(DCTELEM) * DCTSIZE * 3); - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT; - * furthermore, we scale the results by 2**PASS1_BITS. - * 10-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/20). - */ - - dataptr = data; - for (ctr = 0; ctr < 5; ctr++) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[9]); - tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[8]); - tmp12 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[7]); - tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[6]); - tmp4 = GETJSAMPLE(elemptr[4]) + GETJSAMPLE(elemptr[5]); - - tmp10 = tmp0 + tmp4; - tmp13 = tmp0 - tmp4; - tmp11 = tmp1 + tmp3; - tmp14 = tmp1 - tmp3; - - tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[9]); - tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[8]); - tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[7]); - tmp3 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[6]); - tmp4 = GETJSAMPLE(elemptr[4]) - GETJSAMPLE(elemptr[5]); - - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) - ((tmp10 + tmp11 + tmp12 - 10 * CENTERJSAMPLE) << PASS1_BITS); - tmp12 += tmp12; - dataptr[4] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp12, FIX(1.144122806)) - /* c4 */ - MULTIPLY(tmp11 - tmp12, FIX(0.437016024)), /* c8 */ - CONST_BITS-PASS1_BITS); - tmp10 = MULTIPLY(tmp13 + tmp14, FIX(0.831253876)); /* c6 */ - dataptr[2] = (DCTELEM) - DESCALE(tmp10 + MULTIPLY(tmp13, FIX(0.513743148)), /* c2-c6 */ - CONST_BITS-PASS1_BITS); - dataptr[6] = (DCTELEM) - DESCALE(tmp10 - MULTIPLY(tmp14, FIX(2.176250899)), /* c2+c6 */ - CONST_BITS-PASS1_BITS); - - /* Odd part */ - - tmp10 = tmp0 + tmp4; - tmp11 = tmp1 - tmp3; - dataptr[5] = (DCTELEM) ((tmp10 - tmp11 - tmp2) << PASS1_BITS); - tmp2 <<= CONST_BITS; - dataptr[1] = (DCTELEM) - DESCALE(MULTIPLY(tmp0, FIX(1.396802247)) + /* c1 */ - MULTIPLY(tmp1, FIX(1.260073511)) + tmp2 + /* c3 */ - MULTIPLY(tmp3, FIX(0.642039522)) + /* c7 */ - MULTIPLY(tmp4, FIX(0.221231742)), /* c9 */ - CONST_BITS-PASS1_BITS); - tmp12 = MULTIPLY(tmp0 - tmp4, FIX(0.951056516)) - /* (c3+c7)/2 */ - MULTIPLY(tmp1 + tmp3, FIX(0.587785252)); /* (c1-c9)/2 */ - tmp13 = MULTIPLY(tmp10 + tmp11, FIX(0.309016994)) + /* (c3-c7)/2 */ - (tmp11 << (CONST_BITS - 1)) - tmp2; - dataptr[3] = (DCTELEM) DESCALE(tmp12 + tmp13, CONST_BITS-PASS1_BITS); - dataptr[7] = (DCTELEM) DESCALE(tmp12 - tmp13, CONST_BITS-PASS1_BITS); - - dataptr += DCTSIZE; /* advance pointer to next row */ - } - - /* Pass 2: process columns. - * We remove the PASS1_BITS scaling, but leave the results scaled up - * by an overall factor of 8. - * We must also scale the output by (8/10)*(8/5) = 32/25, which we - * fold into the constant multipliers: - * 5-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/10) * 32/25. - */ - - dataptr = data; - for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { - /* Even part */ - - tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*4]; - tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*3]; - tmp2 = dataptr[DCTSIZE*2]; - - tmp10 = tmp0 + tmp1; - tmp11 = tmp0 - tmp1; - - tmp0 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*4]; - tmp1 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*3]; - - dataptr[DCTSIZE*0] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 + tmp2, FIX(1.28)), /* 32/25 */ - CONST_BITS+PASS1_BITS); - tmp11 = MULTIPLY(tmp11, FIX(1.011928851)); /* (c2+c4)/2 */ - tmp10 -= tmp2 << 2; - tmp10 = MULTIPLY(tmp10, FIX(0.452548340)); /* (c2-c4)/2 */ - dataptr[DCTSIZE*2] = (DCTELEM) DESCALE(tmp11 + tmp10, CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*4] = (DCTELEM) DESCALE(tmp11 - tmp10, CONST_BITS+PASS1_BITS); - - /* Odd part */ - - tmp10 = MULTIPLY(tmp0 + tmp1, FIX(1.064004961)); /* c3 */ - - dataptr[DCTSIZE*1] = (DCTELEM) - DESCALE(tmp10 + MULTIPLY(tmp0, FIX(0.657591230)), /* c1-c3 */ - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*3] = (DCTELEM) - DESCALE(tmp10 - MULTIPLY(tmp1, FIX(2.785601151)), /* c1+c3 */ - CONST_BITS+PASS1_BITS); - - dataptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on an 8x4 sample block. - * - * 8-point FDCT in pass 1 (rows), 4-point in pass 2 (columns). - */ - -GLOBAL(void) -jpeg_fdct_8x4 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3; - INT32 tmp10, tmp11, tmp12, tmp13; - INT32 z1; - DCTELEM *dataptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Zero 4 bottom rows of output coefficient block. */ - MEMZERO(&data[DCTSIZE*4], SIZEOF(DCTELEM) * DCTSIZE * 4); - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT; - * furthermore, we scale the results by 2**PASS1_BITS. - * We must also scale the output by 8/4 = 2, which we add here. - * 8-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/16). - */ - - dataptr = data; - for (ctr = 0; ctr < 4; ctr++) { - elemptr = sample_data[ctr] + start_col; - - /* Even part per LL&M figure 1 --- note that published figure is faulty; - * rotator "c1" should be "c6". - */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[7]); - tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[6]); - tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[5]); - tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[4]); - - tmp10 = tmp0 + tmp3; - tmp12 = tmp0 - tmp3; - tmp11 = tmp1 + tmp2; - tmp13 = tmp1 - tmp2; - - tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[7]); - tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[6]); - tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[5]); - tmp3 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[4]); - - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) - ((tmp10 + tmp11 - 8 * CENTERJSAMPLE) << (PASS1_BITS+1)); - dataptr[4] = (DCTELEM) ((tmp10 - tmp11) << (PASS1_BITS+1)); - - z1 = MULTIPLY(tmp12 + tmp13, FIX_0_541196100); /* c6 */ - /* Add fudge factor here for final descale. */ - z1 += ONE << (CONST_BITS-PASS1_BITS-2); - - dataptr[2] = (DCTELEM) - RIGHT_SHIFT(z1 + MULTIPLY(tmp12, FIX_0_765366865), /* c2-c6 */ - CONST_BITS-PASS1_BITS-1); - dataptr[6] = (DCTELEM) - RIGHT_SHIFT(z1 - MULTIPLY(tmp13, FIX_1_847759065), /* c2+c6 */ - CONST_BITS-PASS1_BITS-1); - - /* Odd part per figure 8 --- note paper omits factor of sqrt(2). - * i0..i3 in the paper are tmp0..tmp3 here. - */ - - tmp12 = tmp0 + tmp2; - tmp13 = tmp1 + tmp3; - - z1 = MULTIPLY(tmp12 + tmp13, FIX_1_175875602); /* c3 */ - /* Add fudge factor here for final descale. */ - z1 += ONE << (CONST_BITS-PASS1_BITS-2); - - tmp12 = MULTIPLY(tmp12, - FIX_0_390180644); /* -c3+c5 */ - tmp13 = MULTIPLY(tmp13, - FIX_1_961570560); /* -c3-c5 */ - tmp12 += z1; - tmp13 += z1; - - z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ - tmp0 = MULTIPLY(tmp0, FIX_1_501321110); /* c1+c3-c5-c7 */ - tmp3 = MULTIPLY(tmp3, FIX_0_298631336); /* -c1+c3+c5-c7 */ - tmp0 += z1 + tmp12; - tmp3 += z1 + tmp13; - - z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ - tmp1 = MULTIPLY(tmp1, FIX_3_072711026); /* c1+c3+c5-c7 */ - tmp2 = MULTIPLY(tmp2, FIX_2_053119869); /* c1+c3-c5+c7 */ - tmp1 += z1 + tmp13; - tmp2 += z1 + tmp12; - - dataptr[1] = (DCTELEM) RIGHT_SHIFT(tmp0, CONST_BITS-PASS1_BITS-1); - dataptr[3] = (DCTELEM) RIGHT_SHIFT(tmp1, CONST_BITS-PASS1_BITS-1); - dataptr[5] = (DCTELEM) RIGHT_SHIFT(tmp2, CONST_BITS-PASS1_BITS-1); - dataptr[7] = (DCTELEM) RIGHT_SHIFT(tmp3, CONST_BITS-PASS1_BITS-1); - - dataptr += DCTSIZE; /* advance pointer to next row */ - } - - /* Pass 2: process columns. - * We remove the PASS1_BITS scaling, but leave the results scaled up - * by an overall factor of 8. - * 4-point FDCT kernel, - * cK represents sqrt(2) * cos(K*pi/16) [refers to 8-point FDCT]. - */ - - dataptr = data; - for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { - /* Even part */ - - /* Add fudge factor here for final descale. */ - tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*3] + (ONE << (PASS1_BITS-1)); - tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*2]; - - tmp10 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*3]; - tmp11 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*2]; - - dataptr[DCTSIZE*0] = (DCTELEM) RIGHT_SHIFT(tmp0 + tmp1, PASS1_BITS); - dataptr[DCTSIZE*2] = (DCTELEM) RIGHT_SHIFT(tmp0 - tmp1, PASS1_BITS); - - /* Odd part */ - - tmp0 = MULTIPLY(tmp10 + tmp11, FIX_0_541196100); /* c6 */ - /* Add fudge factor here for final descale. */ - tmp0 += ONE << (CONST_BITS+PASS1_BITS-1); - - dataptr[DCTSIZE*1] = (DCTELEM) - RIGHT_SHIFT(tmp0 + MULTIPLY(tmp10, FIX_0_765366865), /* c2-c6 */ - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*3] = (DCTELEM) - RIGHT_SHIFT(tmp0 - MULTIPLY(tmp11, FIX_1_847759065), /* c2+c6 */ - CONST_BITS+PASS1_BITS); - - dataptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 6x3 sample block. - * - * 6-point FDCT in pass 1 (rows), 3-point in pass 2 (columns). - */ - -GLOBAL(void) -jpeg_fdct_6x3 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2; - INT32 tmp10, tmp11, tmp12; - DCTELEM *dataptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pre-zero output coefficient block. */ - MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT; - * furthermore, we scale the results by 2**PASS1_BITS. - * We scale the results further by 2 as part of output adaption - * scaling for different DCT size. - * 6-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/12). - */ - - dataptr = data; - for (ctr = 0; ctr < 3; ctr++) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[5]); - tmp11 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[4]); - tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[3]); - - tmp10 = tmp0 + tmp2; - tmp12 = tmp0 - tmp2; - - tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[5]); - tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[4]); - tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[3]); - - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) - ((tmp10 + tmp11 - 6 * CENTERJSAMPLE) << (PASS1_BITS+1)); - dataptr[2] = (DCTELEM) - DESCALE(MULTIPLY(tmp12, FIX(1.224744871)), /* c2 */ - CONST_BITS-PASS1_BITS-1); - dataptr[4] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp11 - tmp11, FIX(0.707106781)), /* c4 */ - CONST_BITS-PASS1_BITS-1); - - /* Odd part */ - - tmp10 = DESCALE(MULTIPLY(tmp0 + tmp2, FIX(0.366025404)), /* c5 */ - CONST_BITS-PASS1_BITS-1); - - dataptr[1] = (DCTELEM) (tmp10 + ((tmp0 + tmp1) << (PASS1_BITS+1))); - dataptr[3] = (DCTELEM) ((tmp0 - tmp1 - tmp2) << (PASS1_BITS+1)); - dataptr[5] = (DCTELEM) (tmp10 + ((tmp2 - tmp1) << (PASS1_BITS+1))); - - dataptr += DCTSIZE; /* advance pointer to next row */ - } - - /* Pass 2: process columns. - * We remove the PASS1_BITS scaling, but leave the results scaled up - * by an overall factor of 8. - * We must also scale the output by (8/6)*(8/3) = 32/9, which we partially - * fold into the constant multipliers (other part was done in pass 1): - * 3-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/6) * 16/9. - */ - - dataptr = data; - for (ctr = 0; ctr < 6; ctr++) { - /* Even part */ - - tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*2]; - tmp1 = dataptr[DCTSIZE*1]; - - tmp2 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*2]; - - dataptr[DCTSIZE*0] = (DCTELEM) - DESCALE(MULTIPLY(tmp0 + tmp1, FIX(1.777777778)), /* 16/9 */ - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*2] = (DCTELEM) - DESCALE(MULTIPLY(tmp0 - tmp1 - tmp1, FIX(1.257078722)), /* c2 */ - CONST_BITS+PASS1_BITS); - - /* Odd part */ - - dataptr[DCTSIZE*1] = (DCTELEM) - DESCALE(MULTIPLY(tmp2, FIX(2.177324216)), /* c1 */ - CONST_BITS+PASS1_BITS); - - dataptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 4x2 sample block. - * - * 4-point FDCT in pass 1 (rows), 2-point in pass 2 (columns). - */ - -GLOBAL(void) -jpeg_fdct_4x2 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - DCTELEM tmp0, tmp2, tmp10, tmp12, tmp4, tmp5; - INT32 tmp1, tmp3, tmp11, tmp13; - INT32 z1, z2, z3; - JSAMPROW elemptr; - SHIFT_TEMPS - - /* Pre-zero output coefficient block. */ - MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT. - * 4-point FDCT kernel, - * cK represents sqrt(2) * cos(K*pi/16) [refers to 8-point FDCT]. - */ - - /* Row 0 */ - elemptr = sample_data[0] + start_col; - - /* Even part */ - - tmp4 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[3]); - tmp5 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[2]); - - tmp0 = tmp4 + tmp5; - tmp2 = tmp4 - tmp5; - - /* Odd part */ - - z2 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[3]); - z3 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[2]); - - z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ - /* Add fudge factor here for final descale. */ - z1 += ONE << (CONST_BITS-3-1); - tmp1 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ - tmp3 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ - - /* Row 1 */ - elemptr = sample_data[1] + start_col; - - /* Even part */ - - tmp4 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[3]); - tmp5 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[2]); - - tmp10 = tmp4 + tmp5; - tmp12 = tmp4 - tmp5; - - /* Odd part */ - - z2 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[3]); - z3 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[2]); - - z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ - tmp11 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ - tmp13 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ - - /* Pass 2: process columns. - * We leave the results scaled up by an overall factor of 8. - * We must also scale the output by (8/4)*(8/2) = 2**3. - */ - - /* Column 0 */ - /* Apply unsigned->signed conversion. */ - data[DCTSIZE*0] = (tmp0 + tmp10 - 8 * CENTERJSAMPLE) << 3; - data[DCTSIZE*1] = (tmp0 - tmp10) << 3; - - /* Column 1 */ - data[DCTSIZE*0+1] = (DCTELEM) RIGHT_SHIFT(tmp1 + tmp11, CONST_BITS-3); - data[DCTSIZE*1+1] = (DCTELEM) RIGHT_SHIFT(tmp1 - tmp11, CONST_BITS-3); - - /* Column 2 */ - data[DCTSIZE*0+2] = (tmp2 + tmp12) << 3; - data[DCTSIZE*1+2] = (tmp2 - tmp12) << 3; - - /* Column 3 */ - data[DCTSIZE*0+3] = (DCTELEM) RIGHT_SHIFT(tmp3 + tmp13, CONST_BITS-3); - data[DCTSIZE*1+3] = (DCTELEM) RIGHT_SHIFT(tmp3 - tmp13, CONST_BITS-3); -} - - -/* - * Perform the forward DCT on a 2x1 sample block. - * - * 2-point FDCT in pass 1 (rows), 1-point in pass 2 (columns). - */ - -GLOBAL(void) -jpeg_fdct_2x1 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - DCTELEM tmp0, tmp1; - JSAMPROW elemptr; - - /* Pre-zero output coefficient block. */ - MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); - - elemptr = sample_data[0] + start_col; - - tmp0 = GETJSAMPLE(elemptr[0]); - tmp1 = GETJSAMPLE(elemptr[1]); - - /* We leave the results scaled up by an overall factor of 8. - * We must also scale the output by (8/2)*(8/1) = 2**5. - */ - - /* Even part */ - - /* Apply unsigned->signed conversion. */ - data[0] = (tmp0 + tmp1 - 2 * CENTERJSAMPLE) << 5; - - /* Odd part */ - - data[1] = (tmp0 - tmp1) << 5; -} - - -/* - * Perform the forward DCT on an 8x16 sample block. - * - * 8-point FDCT in pass 1 (rows), 16-point in pass 2 (columns). - */ - -GLOBAL(void) -jpeg_fdct_8x16 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7; - INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15, tmp16, tmp17; - INT32 z1; - DCTELEM workspace[DCTSIZE2]; - DCTELEM *dataptr; - DCTELEM *wsptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT; - * furthermore, we scale the results by 2**PASS1_BITS. - * 8-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/16). - */ - - dataptr = data; - ctr = 0; - for (;;) { - elemptr = sample_data[ctr] + start_col; - - /* Even part per LL&M figure 1 --- note that published figure is faulty; - * rotator "c1" should be "c6". - */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[7]); - tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[6]); - tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[5]); - tmp3 = GETJSAMPLE(elemptr[3]) + GETJSAMPLE(elemptr[4]); - - tmp10 = tmp0 + tmp3; - tmp12 = tmp0 - tmp3; - tmp11 = tmp1 + tmp2; - tmp13 = tmp1 - tmp2; - - tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[7]); - tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[6]); - tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[5]); - tmp3 = GETJSAMPLE(elemptr[3]) - GETJSAMPLE(elemptr[4]); - - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) ((tmp10 + tmp11 - 8 * CENTERJSAMPLE) << PASS1_BITS); - dataptr[4] = (DCTELEM) ((tmp10 - tmp11) << PASS1_BITS); - - z1 = MULTIPLY(tmp12 + tmp13, FIX_0_541196100); /* c6 */ - dataptr[2] = (DCTELEM) - DESCALE(z1 + MULTIPLY(tmp12, FIX_0_765366865), /* c2-c6 */ - CONST_BITS-PASS1_BITS); - dataptr[6] = (DCTELEM) - DESCALE(z1 - MULTIPLY(tmp13, FIX_1_847759065), /* c2+c6 */ - CONST_BITS-PASS1_BITS); - - /* Odd part per figure 8 --- note paper omits factor of sqrt(2). - * i0..i3 in the paper are tmp0..tmp3 here. - */ - - tmp12 = tmp0 + tmp2; - tmp13 = tmp1 + tmp3; - - z1 = MULTIPLY(tmp12 + tmp13, FIX_1_175875602); /* c3 */ - tmp12 = MULTIPLY(tmp12, - FIX_0_390180644); /* -c3+c5 */ - tmp13 = MULTIPLY(tmp13, - FIX_1_961570560); /* -c3-c5 */ - tmp12 += z1; - tmp13 += z1; - - z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ - tmp0 = MULTIPLY(tmp0, FIX_1_501321110); /* c1+c3-c5-c7 */ - tmp3 = MULTIPLY(tmp3, FIX_0_298631336); /* -c1+c3+c5-c7 */ - tmp0 += z1 + tmp12; - tmp3 += z1 + tmp13; - - z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ - tmp1 = MULTIPLY(tmp1, FIX_3_072711026); /* c1+c3+c5-c7 */ - tmp2 = MULTIPLY(tmp2, FIX_2_053119869); /* c1+c3-c5+c7 */ - tmp1 += z1 + tmp13; - tmp2 += z1 + tmp12; - - dataptr[1] = (DCTELEM) DESCALE(tmp0, CONST_BITS-PASS1_BITS); - dataptr[3] = (DCTELEM) DESCALE(tmp1, CONST_BITS-PASS1_BITS); - dataptr[5] = (DCTELEM) DESCALE(tmp2, CONST_BITS-PASS1_BITS); - dataptr[7] = (DCTELEM) DESCALE(tmp3, CONST_BITS-PASS1_BITS); - - ctr++; - - if (ctr != DCTSIZE) { - if (ctr == DCTSIZE * 2) - break; /* Done. */ - dataptr += DCTSIZE; /* advance pointer to next row */ - } else - dataptr = workspace; /* switch pointer to extended workspace */ - } - - /* Pass 2: process columns. - * We remove the PASS1_BITS scaling, but leave the results scaled up - * by an overall factor of 8. - * We must also scale the output by 8/16 = 1/2. - * 16-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/32). - */ - - dataptr = data; - wsptr = workspace; - for (ctr = DCTSIZE-1; ctr >= 0; ctr--) { - /* Even part */ - - tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*7]; - tmp1 = dataptr[DCTSIZE*1] + wsptr[DCTSIZE*6]; - tmp2 = dataptr[DCTSIZE*2] + wsptr[DCTSIZE*5]; - tmp3 = dataptr[DCTSIZE*3] + wsptr[DCTSIZE*4]; - tmp4 = dataptr[DCTSIZE*4] + wsptr[DCTSIZE*3]; - tmp5 = dataptr[DCTSIZE*5] + wsptr[DCTSIZE*2]; - tmp6 = dataptr[DCTSIZE*6] + wsptr[DCTSIZE*1]; - tmp7 = dataptr[DCTSIZE*7] + wsptr[DCTSIZE*0]; - - tmp10 = tmp0 + tmp7; - tmp14 = tmp0 - tmp7; - tmp11 = tmp1 + tmp6; - tmp15 = tmp1 - tmp6; - tmp12 = tmp2 + tmp5; - tmp16 = tmp2 - tmp5; - tmp13 = tmp3 + tmp4; - tmp17 = tmp3 - tmp4; - - tmp0 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*7]; - tmp1 = dataptr[DCTSIZE*1] - wsptr[DCTSIZE*6]; - tmp2 = dataptr[DCTSIZE*2] - wsptr[DCTSIZE*5]; - tmp3 = dataptr[DCTSIZE*3] - wsptr[DCTSIZE*4]; - tmp4 = dataptr[DCTSIZE*4] - wsptr[DCTSIZE*3]; - tmp5 = dataptr[DCTSIZE*5] - wsptr[DCTSIZE*2]; - tmp6 = dataptr[DCTSIZE*6] - wsptr[DCTSIZE*1]; - tmp7 = dataptr[DCTSIZE*7] - wsptr[DCTSIZE*0]; - - dataptr[DCTSIZE*0] = (DCTELEM) - DESCALE(tmp10 + tmp11 + tmp12 + tmp13, PASS1_BITS+1); - dataptr[DCTSIZE*4] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp13, FIX(1.306562965)) + /* c4[16] = c2[8] */ - MULTIPLY(tmp11 - tmp12, FIX_0_541196100), /* c12[16] = c6[8] */ - CONST_BITS+PASS1_BITS+1); - - tmp10 = MULTIPLY(tmp17 - tmp15, FIX(0.275899379)) + /* c14[16] = c7[8] */ - MULTIPLY(tmp14 - tmp16, FIX(1.387039845)); /* c2[16] = c1[8] */ - - dataptr[DCTSIZE*2] = (DCTELEM) - DESCALE(tmp10 + MULTIPLY(tmp15, FIX(1.451774982)) /* c6+c14 */ - + MULTIPLY(tmp16, FIX(2.172734804)), /* c2+c10 */ - CONST_BITS+PASS1_BITS+1); - dataptr[DCTSIZE*6] = (DCTELEM) - DESCALE(tmp10 - MULTIPLY(tmp14, FIX(0.211164243)) /* c2-c6 */ - - MULTIPLY(tmp17, FIX(1.061594338)), /* c10+c14 */ - CONST_BITS+PASS1_BITS+1); - - /* Odd part */ - - tmp11 = MULTIPLY(tmp0 + tmp1, FIX(1.353318001)) + /* c3 */ - MULTIPLY(tmp6 - tmp7, FIX(0.410524528)); /* c13 */ - tmp12 = MULTIPLY(tmp0 + tmp2, FIX(1.247225013)) + /* c5 */ - MULTIPLY(tmp5 + tmp7, FIX(0.666655658)); /* c11 */ - tmp13 = MULTIPLY(tmp0 + tmp3, FIX(1.093201867)) + /* c7 */ - MULTIPLY(tmp4 - tmp7, FIX(0.897167586)); /* c9 */ - tmp14 = MULTIPLY(tmp1 + tmp2, FIX(0.138617169)) + /* c15 */ - MULTIPLY(tmp6 - tmp5, FIX(1.407403738)); /* c1 */ - tmp15 = MULTIPLY(tmp1 + tmp3, - FIX(0.666655658)) + /* -c11 */ - MULTIPLY(tmp4 + tmp6, - FIX(1.247225013)); /* -c5 */ - tmp16 = MULTIPLY(tmp2 + tmp3, - FIX(1.353318001)) + /* -c3 */ - MULTIPLY(tmp5 - tmp4, FIX(0.410524528)); /* c13 */ - tmp10 = tmp11 + tmp12 + tmp13 - - MULTIPLY(tmp0, FIX(2.286341144)) + /* c7+c5+c3-c1 */ - MULTIPLY(tmp7, FIX(0.779653625)); /* c15+c13-c11+c9 */ - tmp11 += tmp14 + tmp15 + MULTIPLY(tmp1, FIX(0.071888074)) /* c9-c3-c15+c11 */ - - MULTIPLY(tmp6, FIX(1.663905119)); /* c7+c13+c1-c5 */ - tmp12 += tmp14 + tmp16 - MULTIPLY(tmp2, FIX(1.125726048)) /* c7+c5+c15-c3 */ - + MULTIPLY(tmp5, FIX(1.227391138)); /* c9-c11+c1-c13 */ - tmp13 += tmp15 + tmp16 + MULTIPLY(tmp3, FIX(1.065388962)) /* c15+c3+c11-c7 */ - + MULTIPLY(tmp4, FIX(2.167985692)); /* c1+c13+c5-c9 */ - - dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp10, CONST_BITS+PASS1_BITS+1); - dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp11, CONST_BITS+PASS1_BITS+1); - dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp12, CONST_BITS+PASS1_BITS+1); - dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp13, CONST_BITS+PASS1_BITS+1); - - dataptr++; /* advance pointer to next column */ - wsptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 7x14 sample block. - * - * 7-point FDCT in pass 1 (rows), 14-point in pass 2 (columns). - */ - -GLOBAL(void) -jpeg_fdct_7x14 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6; - INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15, tmp16; - INT32 z1, z2, z3; - DCTELEM workspace[8*6]; - DCTELEM *dataptr; - DCTELEM *wsptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pre-zero output coefficient block. */ - MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT; - * furthermore, we scale the results by 2**PASS1_BITS. - * 7-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/14). - */ - - dataptr = data; - ctr = 0; - for (;;) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[6]); - tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[5]); - tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[4]); - tmp3 = GETJSAMPLE(elemptr[3]); - - tmp10 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[6]); - tmp11 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[5]); - tmp12 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[4]); - - z1 = tmp0 + tmp2; - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) - ((z1 + tmp1 + tmp3 - 7 * CENTERJSAMPLE) << PASS1_BITS); - tmp3 += tmp3; - z1 -= tmp3; - z1 -= tmp3; - z1 = MULTIPLY(z1, FIX(0.353553391)); /* (c2+c6-c4)/2 */ - z2 = MULTIPLY(tmp0 - tmp2, FIX(0.920609002)); /* (c2+c4-c6)/2 */ - z3 = MULTIPLY(tmp1 - tmp2, FIX(0.314692123)); /* c6 */ - dataptr[2] = (DCTELEM) DESCALE(z1 + z2 + z3, CONST_BITS-PASS1_BITS); - z1 -= z2; - z2 = MULTIPLY(tmp0 - tmp1, FIX(0.881747734)); /* c4 */ - dataptr[4] = (DCTELEM) - DESCALE(z2 + z3 - MULTIPLY(tmp1 - tmp3, FIX(0.707106781)), /* c2+c6-c4 */ - CONST_BITS-PASS1_BITS); - dataptr[6] = (DCTELEM) DESCALE(z1 + z2, CONST_BITS-PASS1_BITS); - - /* Odd part */ - - tmp1 = MULTIPLY(tmp10 + tmp11, FIX(0.935414347)); /* (c3+c1-c5)/2 */ - tmp2 = MULTIPLY(tmp10 - tmp11, FIX(0.170262339)); /* (c3+c5-c1)/2 */ - tmp0 = tmp1 - tmp2; - tmp1 += tmp2; - tmp2 = MULTIPLY(tmp11 + tmp12, - FIX(1.378756276)); /* -c1 */ - tmp1 += tmp2; - tmp3 = MULTIPLY(tmp10 + tmp12, FIX(0.613604268)); /* c5 */ - tmp0 += tmp3; - tmp2 += tmp3 + MULTIPLY(tmp12, FIX(1.870828693)); /* c3+c1-c5 */ - - dataptr[1] = (DCTELEM) DESCALE(tmp0, CONST_BITS-PASS1_BITS); - dataptr[3] = (DCTELEM) DESCALE(tmp1, CONST_BITS-PASS1_BITS); - dataptr[5] = (DCTELEM) DESCALE(tmp2, CONST_BITS-PASS1_BITS); - - ctr++; - - if (ctr != DCTSIZE) { - if (ctr == 14) - break; /* Done. */ - dataptr += DCTSIZE; /* advance pointer to next row */ - } else - dataptr = workspace; /* switch pointer to extended workspace */ - } - - /* Pass 2: process columns. - * We remove the PASS1_BITS scaling, but leave the results scaled up - * by an overall factor of 8. - * We must also scale the output by (8/7)*(8/14) = 32/49, which we - * fold into the constant multipliers: - * 14-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/28) * 32/49. - */ - - dataptr = data; - wsptr = workspace; - for (ctr = 0; ctr < 7; ctr++) { - /* Even part */ - - tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*5]; - tmp1 = dataptr[DCTSIZE*1] + wsptr[DCTSIZE*4]; - tmp2 = dataptr[DCTSIZE*2] + wsptr[DCTSIZE*3]; - tmp13 = dataptr[DCTSIZE*3] + wsptr[DCTSIZE*2]; - tmp4 = dataptr[DCTSIZE*4] + wsptr[DCTSIZE*1]; - tmp5 = dataptr[DCTSIZE*5] + wsptr[DCTSIZE*0]; - tmp6 = dataptr[DCTSIZE*6] + dataptr[DCTSIZE*7]; - - tmp10 = tmp0 + tmp6; - tmp14 = tmp0 - tmp6; - tmp11 = tmp1 + tmp5; - tmp15 = tmp1 - tmp5; - tmp12 = tmp2 + tmp4; - tmp16 = tmp2 - tmp4; - - tmp0 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*5]; - tmp1 = dataptr[DCTSIZE*1] - wsptr[DCTSIZE*4]; - tmp2 = dataptr[DCTSIZE*2] - wsptr[DCTSIZE*3]; - tmp3 = dataptr[DCTSIZE*3] - wsptr[DCTSIZE*2]; - tmp4 = dataptr[DCTSIZE*4] - wsptr[DCTSIZE*1]; - tmp5 = dataptr[DCTSIZE*5] - wsptr[DCTSIZE*0]; - tmp6 = dataptr[DCTSIZE*6] - dataptr[DCTSIZE*7]; - - dataptr[DCTSIZE*0] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 + tmp11 + tmp12 + tmp13, - FIX(0.653061224)), /* 32/49 */ - CONST_BITS+PASS1_BITS); - tmp13 += tmp13; - dataptr[DCTSIZE*4] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp13, FIX(0.832106052)) + /* c4 */ - MULTIPLY(tmp11 - tmp13, FIX(0.205513223)) - /* c12 */ - MULTIPLY(tmp12 - tmp13, FIX(0.575835255)), /* c8 */ - CONST_BITS+PASS1_BITS); - - tmp10 = MULTIPLY(tmp14 + tmp15, FIX(0.722074570)); /* c6 */ - - dataptr[DCTSIZE*2] = (DCTELEM) - DESCALE(tmp10 + MULTIPLY(tmp14, FIX(0.178337691)) /* c2-c6 */ - + MULTIPLY(tmp16, FIX(0.400721155)), /* c10 */ - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*6] = (DCTELEM) - DESCALE(tmp10 - MULTIPLY(tmp15, FIX(1.122795725)) /* c6+c10 */ - - MULTIPLY(tmp16, FIX(0.900412262)), /* c2 */ - CONST_BITS+PASS1_BITS); - - /* Odd part */ - - tmp10 = tmp1 + tmp2; - tmp11 = tmp5 - tmp4; - dataptr[DCTSIZE*7] = (DCTELEM) - DESCALE(MULTIPLY(tmp0 - tmp10 + tmp3 - tmp11 - tmp6, - FIX(0.653061224)), /* 32/49 */ - CONST_BITS+PASS1_BITS); - tmp3 = MULTIPLY(tmp3 , FIX(0.653061224)); /* 32/49 */ - tmp10 = MULTIPLY(tmp10, - FIX(0.103406812)); /* -c13 */ - tmp11 = MULTIPLY(tmp11, FIX(0.917760839)); /* c1 */ - tmp10 += tmp11 - tmp3; - tmp11 = MULTIPLY(tmp0 + tmp2, FIX(0.782007410)) + /* c5 */ - MULTIPLY(tmp4 + tmp6, FIX(0.491367823)); /* c9 */ - dataptr[DCTSIZE*5] = (DCTELEM) - DESCALE(tmp10 + tmp11 - MULTIPLY(tmp2, FIX(1.550341076)) /* c3+c5-c13 */ - + MULTIPLY(tmp4, FIX(0.731428202)), /* c1+c11-c9 */ - CONST_BITS+PASS1_BITS); - tmp12 = MULTIPLY(tmp0 + tmp1, FIX(0.871740478)) + /* c3 */ - MULTIPLY(tmp5 - tmp6, FIX(0.305035186)); /* c11 */ - dataptr[DCTSIZE*3] = (DCTELEM) - DESCALE(tmp10 + tmp12 - MULTIPLY(tmp1, FIX(0.276965844)) /* c3-c9-c13 */ - - MULTIPLY(tmp5, FIX(2.004803435)), /* c1+c5+c11 */ - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*1] = (DCTELEM) - DESCALE(tmp11 + tmp12 + tmp3 - - MULTIPLY(tmp0, FIX(0.735987049)) /* c3+c5-c1 */ - - MULTIPLY(tmp6, FIX(0.082925825)), /* c9-c11-c13 */ - CONST_BITS+PASS1_BITS); - - dataptr++; /* advance pointer to next column */ - wsptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 6x12 sample block. - * - * 6-point FDCT in pass 1 (rows), 12-point in pass 2 (columns). - */ - -GLOBAL(void) -jpeg_fdct_6x12 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3, tmp4, tmp5; - INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15; - DCTELEM workspace[8*4]; - DCTELEM *dataptr; - DCTELEM *wsptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pre-zero output coefficient block. */ - MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT; - * furthermore, we scale the results by 2**PASS1_BITS. - * 6-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/12). - */ - - dataptr = data; - ctr = 0; - for (;;) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[5]); - tmp11 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[4]); - tmp2 = GETJSAMPLE(elemptr[2]) + GETJSAMPLE(elemptr[3]); - - tmp10 = tmp0 + tmp2; - tmp12 = tmp0 - tmp2; - - tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[5]); - tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[4]); - tmp2 = GETJSAMPLE(elemptr[2]) - GETJSAMPLE(elemptr[3]); - - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) - ((tmp10 + tmp11 - 6 * CENTERJSAMPLE) << PASS1_BITS); - dataptr[2] = (DCTELEM) - DESCALE(MULTIPLY(tmp12, FIX(1.224744871)), /* c2 */ - CONST_BITS-PASS1_BITS); - dataptr[4] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp11 - tmp11, FIX(0.707106781)), /* c4 */ - CONST_BITS-PASS1_BITS); - - /* Odd part */ - - tmp10 = DESCALE(MULTIPLY(tmp0 + tmp2, FIX(0.366025404)), /* c5 */ - CONST_BITS-PASS1_BITS); - - dataptr[1] = (DCTELEM) (tmp10 + ((tmp0 + tmp1) << PASS1_BITS)); - dataptr[3] = (DCTELEM) ((tmp0 - tmp1 - tmp2) << PASS1_BITS); - dataptr[5] = (DCTELEM) (tmp10 + ((tmp2 - tmp1) << PASS1_BITS)); - - ctr++; - - if (ctr != DCTSIZE) { - if (ctr == 12) - break; /* Done. */ - dataptr += DCTSIZE; /* advance pointer to next row */ - } else - dataptr = workspace; /* switch pointer to extended workspace */ - } - - /* Pass 2: process columns. - * We remove the PASS1_BITS scaling, but leave the results scaled up - * by an overall factor of 8. - * We must also scale the output by (8/6)*(8/12) = 8/9, which we - * fold into the constant multipliers: - * 12-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/24) * 8/9. - */ - - dataptr = data; - wsptr = workspace; - for (ctr = 0; ctr < 6; ctr++) { - /* Even part */ - - tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*3]; - tmp1 = dataptr[DCTSIZE*1] + wsptr[DCTSIZE*2]; - tmp2 = dataptr[DCTSIZE*2] + wsptr[DCTSIZE*1]; - tmp3 = dataptr[DCTSIZE*3] + wsptr[DCTSIZE*0]; - tmp4 = dataptr[DCTSIZE*4] + dataptr[DCTSIZE*7]; - tmp5 = dataptr[DCTSIZE*5] + dataptr[DCTSIZE*6]; - - tmp10 = tmp0 + tmp5; - tmp13 = tmp0 - tmp5; - tmp11 = tmp1 + tmp4; - tmp14 = tmp1 - tmp4; - tmp12 = tmp2 + tmp3; - tmp15 = tmp2 - tmp3; - - tmp0 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*3]; - tmp1 = dataptr[DCTSIZE*1] - wsptr[DCTSIZE*2]; - tmp2 = dataptr[DCTSIZE*2] - wsptr[DCTSIZE*1]; - tmp3 = dataptr[DCTSIZE*3] - wsptr[DCTSIZE*0]; - tmp4 = dataptr[DCTSIZE*4] - dataptr[DCTSIZE*7]; - tmp5 = dataptr[DCTSIZE*5] - dataptr[DCTSIZE*6]; - - dataptr[DCTSIZE*0] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 + tmp11 + tmp12, FIX(0.888888889)), /* 8/9 */ - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*6] = (DCTELEM) - DESCALE(MULTIPLY(tmp13 - tmp14 - tmp15, FIX(0.888888889)), /* 8/9 */ - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*4] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp12, FIX(1.088662108)), /* c4 */ - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*2] = (DCTELEM) - DESCALE(MULTIPLY(tmp14 - tmp15, FIX(0.888888889)) + /* 8/9 */ - MULTIPLY(tmp13 + tmp15, FIX(1.214244803)), /* c2 */ - CONST_BITS+PASS1_BITS); - - /* Odd part */ - - tmp10 = MULTIPLY(tmp1 + tmp4, FIX(0.481063200)); /* c9 */ - tmp14 = tmp10 + MULTIPLY(tmp1, FIX(0.680326102)); /* c3-c9 */ - tmp15 = tmp10 - MULTIPLY(tmp4, FIX(1.642452502)); /* c3+c9 */ - tmp12 = MULTIPLY(tmp0 + tmp2, FIX(0.997307603)); /* c5 */ - tmp13 = MULTIPLY(tmp0 + tmp3, FIX(0.765261039)); /* c7 */ - tmp10 = tmp12 + tmp13 + tmp14 - MULTIPLY(tmp0, FIX(0.516244403)) /* c5+c7-c1 */ - + MULTIPLY(tmp5, FIX(0.164081699)); /* c11 */ - tmp11 = MULTIPLY(tmp2 + tmp3, - FIX(0.164081699)); /* -c11 */ - tmp12 += tmp11 - tmp15 - MULTIPLY(tmp2, FIX(2.079550144)) /* c1+c5-c11 */ - + MULTIPLY(tmp5, FIX(0.765261039)); /* c7 */ - tmp13 += tmp11 - tmp14 + MULTIPLY(tmp3, FIX(0.645144899)) /* c1+c11-c7 */ - - MULTIPLY(tmp5, FIX(0.997307603)); /* c5 */ - tmp11 = tmp15 + MULTIPLY(tmp0 - tmp3, FIX(1.161389302)) /* c3 */ - - MULTIPLY(tmp2 + tmp5, FIX(0.481063200)); /* c9 */ - - dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp10, CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp11, CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp12, CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp13, CONST_BITS+PASS1_BITS); - - dataptr++; /* advance pointer to next column */ - wsptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 5x10 sample block. - * - * 5-point FDCT in pass 1 (rows), 10-point in pass 2 (columns). - */ - -GLOBAL(void) -jpeg_fdct_5x10 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3, tmp4; - INT32 tmp10, tmp11, tmp12, tmp13, tmp14; - DCTELEM workspace[8*2]; - DCTELEM *dataptr; - DCTELEM *wsptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pre-zero output coefficient block. */ - MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT; - * furthermore, we scale the results by 2**PASS1_BITS. - * 5-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/10). - */ - - dataptr = data; - ctr = 0; - for (;;) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[4]); - tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[3]); - tmp2 = GETJSAMPLE(elemptr[2]); - - tmp10 = tmp0 + tmp1; - tmp11 = tmp0 - tmp1; - - tmp0 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[4]); - tmp1 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[3]); - - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) - ((tmp10 + tmp2 - 5 * CENTERJSAMPLE) << PASS1_BITS); - tmp11 = MULTIPLY(tmp11, FIX(0.790569415)); /* (c2+c4)/2 */ - tmp10 -= tmp2 << 2; - tmp10 = MULTIPLY(tmp10, FIX(0.353553391)); /* (c2-c4)/2 */ - dataptr[2] = (DCTELEM) DESCALE(tmp11 + tmp10, CONST_BITS-PASS1_BITS); - dataptr[4] = (DCTELEM) DESCALE(tmp11 - tmp10, CONST_BITS-PASS1_BITS); - - /* Odd part */ - - tmp10 = MULTIPLY(tmp0 + tmp1, FIX(0.831253876)); /* c3 */ - - dataptr[1] = (DCTELEM) - DESCALE(tmp10 + MULTIPLY(tmp0, FIX(0.513743148)), /* c1-c3 */ - CONST_BITS-PASS1_BITS); - dataptr[3] = (DCTELEM) - DESCALE(tmp10 - MULTIPLY(tmp1, FIX(2.176250899)), /* c1+c3 */ - CONST_BITS-PASS1_BITS); - - ctr++; - - if (ctr != DCTSIZE) { - if (ctr == 10) - break; /* Done. */ - dataptr += DCTSIZE; /* advance pointer to next row */ - } else - dataptr = workspace; /* switch pointer to extended workspace */ - } - - /* Pass 2: process columns. - * We remove the PASS1_BITS scaling, but leave the results scaled up - * by an overall factor of 8. - * We must also scale the output by (8/5)*(8/10) = 32/25, which we - * fold into the constant multipliers: - * 10-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/20) * 32/25. - */ - - dataptr = data; - wsptr = workspace; - for (ctr = 0; ctr < 5; ctr++) { - /* Even part */ - - tmp0 = dataptr[DCTSIZE*0] + wsptr[DCTSIZE*1]; - tmp1 = dataptr[DCTSIZE*1] + wsptr[DCTSIZE*0]; - tmp12 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*7]; - tmp3 = dataptr[DCTSIZE*3] + dataptr[DCTSIZE*6]; - tmp4 = dataptr[DCTSIZE*4] + dataptr[DCTSIZE*5]; - - tmp10 = tmp0 + tmp4; - tmp13 = tmp0 - tmp4; - tmp11 = tmp1 + tmp3; - tmp14 = tmp1 - tmp3; - - tmp0 = dataptr[DCTSIZE*0] - wsptr[DCTSIZE*1]; - tmp1 = dataptr[DCTSIZE*1] - wsptr[DCTSIZE*0]; - tmp2 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*7]; - tmp3 = dataptr[DCTSIZE*3] - dataptr[DCTSIZE*6]; - tmp4 = dataptr[DCTSIZE*4] - dataptr[DCTSIZE*5]; - - dataptr[DCTSIZE*0] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 + tmp11 + tmp12, FIX(1.28)), /* 32/25 */ - CONST_BITS+PASS1_BITS); - tmp12 += tmp12; - dataptr[DCTSIZE*4] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp12, FIX(1.464477191)) - /* c4 */ - MULTIPLY(tmp11 - tmp12, FIX(0.559380511)), /* c8 */ - CONST_BITS+PASS1_BITS); - tmp10 = MULTIPLY(tmp13 + tmp14, FIX(1.064004961)); /* c6 */ - dataptr[DCTSIZE*2] = (DCTELEM) - DESCALE(tmp10 + MULTIPLY(tmp13, FIX(0.657591230)), /* c2-c6 */ - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*6] = (DCTELEM) - DESCALE(tmp10 - MULTIPLY(tmp14, FIX(2.785601151)), /* c2+c6 */ - CONST_BITS+PASS1_BITS); - - /* Odd part */ - - tmp10 = tmp0 + tmp4; - tmp11 = tmp1 - tmp3; - dataptr[DCTSIZE*5] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp11 - tmp2, FIX(1.28)), /* 32/25 */ - CONST_BITS+PASS1_BITS); - tmp2 = MULTIPLY(tmp2, FIX(1.28)); /* 32/25 */ - dataptr[DCTSIZE*1] = (DCTELEM) - DESCALE(MULTIPLY(tmp0, FIX(1.787906876)) + /* c1 */ - MULTIPLY(tmp1, FIX(1.612894094)) + tmp2 + /* c3 */ - MULTIPLY(tmp3, FIX(0.821810588)) + /* c7 */ - MULTIPLY(tmp4, FIX(0.283176630)), /* c9 */ - CONST_BITS+PASS1_BITS); - tmp12 = MULTIPLY(tmp0 - tmp4, FIX(1.217352341)) - /* (c3+c7)/2 */ - MULTIPLY(tmp1 + tmp3, FIX(0.752365123)); /* (c1-c9)/2 */ - tmp13 = MULTIPLY(tmp10 + tmp11, FIX(0.395541753)) + /* (c3-c7)/2 */ - MULTIPLY(tmp11, FIX(0.64)) - tmp2; /* 16/25 */ - dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp12 + tmp13, CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp12 - tmp13, CONST_BITS+PASS1_BITS); - - dataptr++; /* advance pointer to next column */ - wsptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 4x8 sample block. - * - * 4-point FDCT in pass 1 (rows), 8-point in pass 2 (columns). - */ - -GLOBAL(void) -jpeg_fdct_4x8 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3; - INT32 tmp10, tmp11, tmp12, tmp13; - INT32 z1; - DCTELEM *dataptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pre-zero output coefficient block. */ - MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT; - * furthermore, we scale the results by 2**PASS1_BITS. - * We must also scale the output by 8/4 = 2, which we add here. - * 4-point FDCT kernel, - * cK represents sqrt(2) * cos(K*pi/16) [refers to 8-point FDCT]. - */ - - dataptr = data; - for (ctr = 0; ctr < DCTSIZE; ctr++) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[3]); - tmp1 = GETJSAMPLE(elemptr[1]) + GETJSAMPLE(elemptr[2]); - - tmp10 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[3]); - tmp11 = GETJSAMPLE(elemptr[1]) - GETJSAMPLE(elemptr[2]); - - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) - ((tmp0 + tmp1 - 4 * CENTERJSAMPLE) << (PASS1_BITS+1)); - dataptr[2] = (DCTELEM) ((tmp0 - tmp1) << (PASS1_BITS+1)); - - /* Odd part */ - - tmp0 = MULTIPLY(tmp10 + tmp11, FIX_0_541196100); /* c6 */ - /* Add fudge factor here for final descale. */ - tmp0 += ONE << (CONST_BITS-PASS1_BITS-2); - - dataptr[1] = (DCTELEM) - RIGHT_SHIFT(tmp0 + MULTIPLY(tmp10, FIX_0_765366865), /* c2-c6 */ - CONST_BITS-PASS1_BITS-1); - dataptr[3] = (DCTELEM) - RIGHT_SHIFT(tmp0 - MULTIPLY(tmp11, FIX_1_847759065), /* c2+c6 */ - CONST_BITS-PASS1_BITS-1); - - dataptr += DCTSIZE; /* advance pointer to next row */ - } - - /* Pass 2: process columns. - * We remove the PASS1_BITS scaling, but leave the results scaled up - * by an overall factor of 8. - * 8-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/16). - */ - - dataptr = data; - for (ctr = 0; ctr < 4; ctr++) { - /* Even part per LL&M figure 1 --- note that published figure is faulty; - * rotator "c1" should be "c6". - */ - - tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*7]; - tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*6]; - tmp2 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*5]; - tmp3 = dataptr[DCTSIZE*3] + dataptr[DCTSIZE*4]; - - /* Add fudge factor here for final descale. */ - tmp10 = tmp0 + tmp3 + (ONE << (PASS1_BITS-1)); - tmp12 = tmp0 - tmp3; - tmp11 = tmp1 + tmp2; - tmp13 = tmp1 - tmp2; - - tmp0 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*7]; - tmp1 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*6]; - tmp2 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*5]; - tmp3 = dataptr[DCTSIZE*3] - dataptr[DCTSIZE*4]; - - dataptr[DCTSIZE*0] = (DCTELEM) RIGHT_SHIFT(tmp10 + tmp11, PASS1_BITS); - dataptr[DCTSIZE*4] = (DCTELEM) RIGHT_SHIFT(tmp10 - tmp11, PASS1_BITS); - - z1 = MULTIPLY(tmp12 + tmp13, FIX_0_541196100); /* c6 */ - /* Add fudge factor here for final descale. */ - z1 += ONE << (CONST_BITS+PASS1_BITS-1); - - dataptr[DCTSIZE*2] = (DCTELEM) - RIGHT_SHIFT(z1 + MULTIPLY(tmp12, FIX_0_765366865), /* c2-c6 */ - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*6] = (DCTELEM) - RIGHT_SHIFT(z1 - MULTIPLY(tmp13, FIX_1_847759065), /* c2+c6 */ - CONST_BITS+PASS1_BITS); - - /* Odd part per figure 8 --- note paper omits factor of sqrt(2). - * i0..i3 in the paper are tmp0..tmp3 here. - */ - - tmp12 = tmp0 + tmp2; - tmp13 = tmp1 + tmp3; - - z1 = MULTIPLY(tmp12 + tmp13, FIX_1_175875602); /* c3 */ - /* Add fudge factor here for final descale. */ - z1 += ONE << (CONST_BITS+PASS1_BITS-1); - - tmp12 = MULTIPLY(tmp12, - FIX_0_390180644); /* -c3+c5 */ - tmp13 = MULTIPLY(tmp13, - FIX_1_961570560); /* -c3-c5 */ - tmp12 += z1; - tmp13 += z1; - - z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ - tmp0 = MULTIPLY(tmp0, FIX_1_501321110); /* c1+c3-c5-c7 */ - tmp3 = MULTIPLY(tmp3, FIX_0_298631336); /* -c1+c3+c5-c7 */ - tmp0 += z1 + tmp12; - tmp3 += z1 + tmp13; - - z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ - tmp1 = MULTIPLY(tmp1, FIX_3_072711026); /* c1+c3+c5-c7 */ - tmp2 = MULTIPLY(tmp2, FIX_2_053119869); /* c1+c3-c5+c7 */ - tmp1 += z1 + tmp13; - tmp2 += z1 + tmp12; - - dataptr[DCTSIZE*1] = (DCTELEM) RIGHT_SHIFT(tmp0, CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*3] = (DCTELEM) RIGHT_SHIFT(tmp1, CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*5] = (DCTELEM) RIGHT_SHIFT(tmp2, CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*7] = (DCTELEM) RIGHT_SHIFT(tmp3, CONST_BITS+PASS1_BITS); - - dataptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 3x6 sample block. - * - * 3-point FDCT in pass 1 (rows), 6-point in pass 2 (columns). - */ - -GLOBAL(void) -jpeg_fdct_3x6 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1, tmp2; - INT32 tmp10, tmp11, tmp12; - DCTELEM *dataptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pre-zero output coefficient block. */ - MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT; - * furthermore, we scale the results by 2**PASS1_BITS. - * We scale the results further by 2 as part of output adaption - * scaling for different DCT size. - * 3-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/6). - */ - - dataptr = data; - for (ctr = 0; ctr < 6; ctr++) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]) + GETJSAMPLE(elemptr[2]); - tmp1 = GETJSAMPLE(elemptr[1]); - - tmp2 = GETJSAMPLE(elemptr[0]) - GETJSAMPLE(elemptr[2]); - - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) - ((tmp0 + tmp1 - 3 * CENTERJSAMPLE) << (PASS1_BITS+1)); - dataptr[2] = (DCTELEM) - DESCALE(MULTIPLY(tmp0 - tmp1 - tmp1, FIX(0.707106781)), /* c2 */ - CONST_BITS-PASS1_BITS-1); - - /* Odd part */ - - dataptr[1] = (DCTELEM) - DESCALE(MULTIPLY(tmp2, FIX(1.224744871)), /* c1 */ - CONST_BITS-PASS1_BITS-1); - - dataptr += DCTSIZE; /* advance pointer to next row */ - } - - /* Pass 2: process columns. - * We remove the PASS1_BITS scaling, but leave the results scaled up - * by an overall factor of 8. - * We must also scale the output by (8/6)*(8/3) = 32/9, which we partially - * fold into the constant multipliers (other part was done in pass 1): - * 6-point FDCT kernel, cK represents sqrt(2) * cos(K*pi/12) * 16/9. - */ - - dataptr = data; - for (ctr = 0; ctr < 3; ctr++) { - /* Even part */ - - tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*5]; - tmp11 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*4]; - tmp2 = dataptr[DCTSIZE*2] + dataptr[DCTSIZE*3]; - - tmp10 = tmp0 + tmp2; - tmp12 = tmp0 - tmp2; - - tmp0 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*5]; - tmp1 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*4]; - tmp2 = dataptr[DCTSIZE*2] - dataptr[DCTSIZE*3]; - - dataptr[DCTSIZE*0] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 + tmp11, FIX(1.777777778)), /* 16/9 */ - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*2] = (DCTELEM) - DESCALE(MULTIPLY(tmp12, FIX(2.177324216)), /* c2 */ - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*4] = (DCTELEM) - DESCALE(MULTIPLY(tmp10 - tmp11 - tmp11, FIX(1.257078722)), /* c4 */ - CONST_BITS+PASS1_BITS); - - /* Odd part */ - - tmp10 = MULTIPLY(tmp0 + tmp2, FIX(0.650711829)); /* c5 */ - - dataptr[DCTSIZE*1] = (DCTELEM) - DESCALE(tmp10 + MULTIPLY(tmp0 + tmp1, FIX(1.777777778)), /* 16/9 */ - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*3] = (DCTELEM) - DESCALE(MULTIPLY(tmp0 - tmp1 - tmp2, FIX(1.777777778)), /* 16/9 */ - CONST_BITS+PASS1_BITS); - dataptr[DCTSIZE*5] = (DCTELEM) - DESCALE(tmp10 + MULTIPLY(tmp2 - tmp1, FIX(1.777777778)), /* 16/9 */ - CONST_BITS+PASS1_BITS); - - dataptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 2x4 sample block. - * - * 2-point FDCT in pass 1 (rows), 4-point in pass 2 (columns). - */ - -GLOBAL(void) -jpeg_fdct_2x4 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - INT32 tmp0, tmp1; - INT32 tmp10, tmp11; - DCTELEM *dataptr; - JSAMPROW elemptr; - int ctr; - SHIFT_TEMPS - - /* Pre-zero output coefficient block. */ - MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); - - /* Pass 1: process rows. - * Note results are scaled up by sqrt(8) compared to a true DCT. - */ - - dataptr = data; - for (ctr = 0; ctr < 4; ctr++) { - elemptr = sample_data[ctr] + start_col; - - /* Even part */ - - tmp0 = GETJSAMPLE(elemptr[0]); - tmp1 = GETJSAMPLE(elemptr[1]); - - /* Apply unsigned->signed conversion. */ - dataptr[0] = (DCTELEM) (tmp0 + tmp1 - 2 * CENTERJSAMPLE); - - /* Odd part */ - - dataptr[1] = (DCTELEM) (tmp0 - tmp1); - - dataptr += DCTSIZE; /* advance pointer to next row */ - } - - /* Pass 2: process columns. - * We leave the results scaled up by an overall factor of 8. - * We must also scale the output by (8/2)*(8/4) = 2**3. - * 4-point FDCT kernel, - * cK represents sqrt(2) * cos(K*pi/16) [refers to 8-point FDCT]. - */ - - dataptr = data; - for (ctr = 0; ctr < 2; ctr++) { - /* Even part */ - - tmp0 = dataptr[DCTSIZE*0] + dataptr[DCTSIZE*3]; - tmp1 = dataptr[DCTSIZE*1] + dataptr[DCTSIZE*2]; - - tmp10 = dataptr[DCTSIZE*0] - dataptr[DCTSIZE*3]; - tmp11 = dataptr[DCTSIZE*1] - dataptr[DCTSIZE*2]; - - dataptr[DCTSIZE*0] = (DCTELEM) ((tmp0 + tmp1) << 3); - dataptr[DCTSIZE*2] = (DCTELEM) ((tmp0 - tmp1) << 3); - - /* Odd part */ - - tmp0 = MULTIPLY(tmp10 + tmp11, FIX_0_541196100); /* c6 */ - /* Add fudge factor here for final descale. */ - tmp0 += ONE << (CONST_BITS-3-1); - - dataptr[DCTSIZE*1] = (DCTELEM) - RIGHT_SHIFT(tmp0 + MULTIPLY(tmp10, FIX_0_765366865), /* c2-c6 */ - CONST_BITS-3); - dataptr[DCTSIZE*3] = (DCTELEM) - RIGHT_SHIFT(tmp0 - MULTIPLY(tmp11, FIX_1_847759065), /* c2+c6 */ - CONST_BITS-3); - - dataptr++; /* advance pointer to next column */ - } -} - - -/* - * Perform the forward DCT on a 1x2 sample block. - * - * 1-point FDCT in pass 1 (rows), 2-point in pass 2 (columns). - */ - -GLOBAL(void) -jpeg_fdct_1x2 (DCTELEM * data, JSAMPARRAY sample_data, JDIMENSION start_col) -{ - DCTELEM tmp0, tmp1; - - /* Pre-zero output coefficient block. */ - MEMZERO(data, SIZEOF(DCTELEM) * DCTSIZE2); - - /* Pass 1: empty. */ - - /* Pass 2: process columns. - * We leave the results scaled up by an overall factor of 8. - * We must also scale the output by (8/1)*(8/2) = 2**5. - */ - - /* Even part */ - - tmp0 = GETJSAMPLE(sample_data[0][start_col]); - tmp1 = GETJSAMPLE(sample_data[1][start_col]); - - /* Apply unsigned->signed conversion. */ - data[DCTSIZE*0] = (tmp0 + tmp1 - 2 * CENTERJSAMPLE) << 5; - - /* Odd part */ - - data[DCTSIZE*1] = (tmp0 - tmp1) << 5; -} - -#endif /* DCT_SCALING_SUPPORTED */ -#endif /* DCT_ISLOW_SUPPORTED */ diff --git a/cmake/externals/libjpeg/jidctflt.c b/cmake/externals/libjpeg/jidctflt.c deleted file mode 100644 index e33a2b5e4..000000000 --- a/cmake/externals/libjpeg/jidctflt.c +++ /dev/null @@ -1,238 +0,0 @@ -/* - * jidctflt.c - * - * Copyright (C) 1994-1998, Thomas G. Lane. - * Modified 2010-2017 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains a floating-point implementation of the - * inverse DCT (Discrete Cosine Transform). In the IJG code, this routine - * must also perform dequantization of the input coefficients. - * - * This implementation should be more accurate than either of the integer - * IDCT implementations. However, it may not give the same results on all - * machines because of differences in roundoff behavior. Speed will depend - * on the hardware's floating point capacity. - * - * A 2-D IDCT can be done by 1-D IDCT on each column followed by 1-D IDCT - * on each row (or vice versa, but it's more convenient to emit a row at - * a time). Direct algorithms are also available, but they are much more - * complex and seem not to be any faster when reduced to code. - * - * This implementation is based on Arai, Agui, and Nakajima's algorithm for - * scaled DCT. Their original paper (Trans. IEICE E-71(11):1095) is in - * Japanese, but the algorithm is described in the Pennebaker & Mitchell - * JPEG textbook (see REFERENCES section in file README). The following code - * is based directly on figure 4-8 in P&M. - * While an 8-point DCT cannot be done in less than 11 multiplies, it is - * possible to arrange the computation so that many of the multiplies are - * simple scalings of the final outputs. These multiplies can then be - * folded into the multiplications or divisions by the JPEG quantization - * table entries. The AA&N method leaves only 5 multiplies and 29 adds - * to be done in the DCT itself. - * The primary disadvantage of this method is that with a fixed-point - * implementation, accuracy is lost due to imprecise representation of the - * scaled quantization values. However, that problem does not arise if - * we use floating point arithmetic. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" -#include "jdct.h" /* Private declarations for DCT subsystem */ - -#ifdef DCT_FLOAT_SUPPORTED - - -/* - * This module is specialized to the case DCTSIZE = 8. - */ - -#if DCTSIZE != 8 - Sorry, this code only copes with 8x8 DCT blocks. /* deliberate syntax err */ -#endif - - -/* Dequantize a coefficient by multiplying it by the multiplier-table - * entry; produce a float result. - */ - -#define DEQUANTIZE(coef,quantval) (((FAST_FLOAT) (coef)) * (quantval)) - - -/* - * Perform dequantization and inverse DCT on one block of coefficients. - * - * cK represents cos(K*pi/16). - */ - -GLOBAL(void) -jpeg_idct_float (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - FAST_FLOAT tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7; - FAST_FLOAT tmp10, tmp11, tmp12, tmp13; - FAST_FLOAT z5, z10, z11, z12, z13; - JCOEFPTR inptr; - FLOAT_MULT_TYPE * quantptr; - FAST_FLOAT * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - FAST_FLOAT workspace[DCTSIZE2]; /* buffers data between passes */ - - /* Pass 1: process columns from input, store into work array. */ - - inptr = coef_block; - quantptr = (FLOAT_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = DCTSIZE; ctr > 0; ctr--) { - /* Due to quantization, we will usually find that many of the input - * coefficients are zero, especially the AC terms. We can exploit this - * by short-circuiting the IDCT calculation for any column in which all - * the AC terms are zero. In that case each output is equal to the - * DC coefficient (with scale factor as needed). - * With typical images and quantization tables, half or more of the - * column DCT calculations can be simplified this way. - */ - - if (inptr[DCTSIZE*1] == 0 && inptr[DCTSIZE*2] == 0 && - inptr[DCTSIZE*3] == 0 && inptr[DCTSIZE*4] == 0 && - inptr[DCTSIZE*5] == 0 && inptr[DCTSIZE*6] == 0 && - inptr[DCTSIZE*7] == 0) { - /* AC terms all zero */ - FAST_FLOAT dcval = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - - wsptr[DCTSIZE*0] = dcval; - wsptr[DCTSIZE*1] = dcval; - wsptr[DCTSIZE*2] = dcval; - wsptr[DCTSIZE*3] = dcval; - wsptr[DCTSIZE*4] = dcval; - wsptr[DCTSIZE*5] = dcval; - wsptr[DCTSIZE*6] = dcval; - wsptr[DCTSIZE*7] = dcval; - - inptr++; /* advance pointers to next column */ - quantptr++; - wsptr++; - continue; - } - - /* Even part */ - - tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - tmp1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - tmp2 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - tmp3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); - - tmp10 = tmp0 + tmp2; /* phase 3 */ - tmp11 = tmp0 - tmp2; - - tmp13 = tmp1 + tmp3; /* phases 5-3 */ - tmp12 = (tmp1 - tmp3) * ((FAST_FLOAT) 1.414213562) - tmp13; /* 2*c4 */ - - tmp0 = tmp10 + tmp13; /* phase 2 */ - tmp3 = tmp10 - tmp13; - tmp1 = tmp11 + tmp12; - tmp2 = tmp11 - tmp12; - - /* Odd part */ - - tmp4 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - tmp5 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - tmp6 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); - tmp7 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); - - z13 = tmp6 + tmp5; /* phase 6 */ - z10 = tmp6 - tmp5; - z11 = tmp4 + tmp7; - z12 = tmp4 - tmp7; - - tmp7 = z11 + z13; /* phase 5 */ - tmp11 = (z11 - z13) * ((FAST_FLOAT) 1.414213562); /* 2*c4 */ - - z5 = (z10 + z12) * ((FAST_FLOAT) 1.847759065); /* 2*c2 */ - tmp10 = z5 - z12 * ((FAST_FLOAT) 1.082392200); /* 2*(c2-c6) */ - tmp12 = z5 - z10 * ((FAST_FLOAT) 2.613125930); /* 2*(c2+c6) */ - - tmp6 = tmp12 - tmp7; /* phase 2 */ - tmp5 = tmp11 - tmp6; - tmp4 = tmp10 - tmp5; - - wsptr[DCTSIZE*0] = tmp0 + tmp7; - wsptr[DCTSIZE*7] = tmp0 - tmp7; - wsptr[DCTSIZE*1] = tmp1 + tmp6; - wsptr[DCTSIZE*6] = tmp1 - tmp6; - wsptr[DCTSIZE*2] = tmp2 + tmp5; - wsptr[DCTSIZE*5] = tmp2 - tmp5; - wsptr[DCTSIZE*3] = tmp3 + tmp4; - wsptr[DCTSIZE*4] = tmp3 - tmp4; - - inptr++; /* advance pointers to next column */ - quantptr++; - wsptr++; - } - - /* Pass 2: process rows from work array, store into output array. */ - - wsptr = workspace; - for (ctr = 0; ctr < DCTSIZE; ctr++) { - outptr = output_buf[ctr] + output_col; - /* Rows of zeroes can be exploited in the same way as we did with columns. - * However, the column calculation has created many nonzero AC terms, so - * the simplification applies less often (typically 5% to 10% of the time). - * And testing floats for zero is relatively expensive, so we don't bother. - */ - - /* Even part */ - - /* Prepare range-limit and float->int conversion */ - z5 = wsptr[0] + (((FAST_FLOAT) RANGE_CENTER) + ((FAST_FLOAT) 0.5)); - tmp10 = z5 + wsptr[4]; - tmp11 = z5 - wsptr[4]; - - tmp13 = wsptr[2] + wsptr[6]; - tmp12 = (wsptr[2] - wsptr[6]) * - ((FAST_FLOAT) 1.414213562) - tmp13; /* 2*c4 */ - - tmp0 = tmp10 + tmp13; - tmp3 = tmp10 - tmp13; - tmp1 = tmp11 + tmp12; - tmp2 = tmp11 - tmp12; - - /* Odd part */ - - z13 = wsptr[5] + wsptr[3]; - z10 = wsptr[5] - wsptr[3]; - z11 = wsptr[1] + wsptr[7]; - z12 = wsptr[1] - wsptr[7]; - - tmp7 = z11 + z13; /* phase 5 */ - tmp11 = (z11 - z13) * ((FAST_FLOAT) 1.414213562); /* 2*c4 */ - - z5 = (z10 + z12) * ((FAST_FLOAT) 1.847759065); /* 2*c2 */ - tmp10 = z5 - z12 * ((FAST_FLOAT) 1.082392200); /* 2*(c2-c6) */ - tmp12 = z5 - z10 * ((FAST_FLOAT) 2.613125930); /* 2*(c2+c6) */ - - tmp6 = tmp12 - tmp7; /* phase 2 */ - tmp5 = tmp11 - tmp6; - tmp4 = tmp10 - tmp5; - - /* Final output stage: float->int conversion and range-limit */ - - outptr[0] = range_limit[(int) (tmp0 + tmp7) & RANGE_MASK]; - outptr[7] = range_limit[(int) (tmp0 - tmp7) & RANGE_MASK]; - outptr[1] = range_limit[(int) (tmp1 + tmp6) & RANGE_MASK]; - outptr[6] = range_limit[(int) (tmp1 - tmp6) & RANGE_MASK]; - outptr[2] = range_limit[(int) (tmp2 + tmp5) & RANGE_MASK]; - outptr[5] = range_limit[(int) (tmp2 - tmp5) & RANGE_MASK]; - outptr[3] = range_limit[(int) (tmp3 + tmp4) & RANGE_MASK]; - outptr[4] = range_limit[(int) (tmp3 - tmp4) & RANGE_MASK]; - - wsptr += DCTSIZE; /* advance pointer to next row */ - } -} - -#endif /* DCT_FLOAT_SUPPORTED */ diff --git a/cmake/externals/libjpeg/jidctfst.c b/cmake/externals/libjpeg/jidctfst.c deleted file mode 100644 index 1ac3e39cb..000000000 --- a/cmake/externals/libjpeg/jidctfst.c +++ /dev/null @@ -1,351 +0,0 @@ -/* - * jidctfst.c - * - * Copyright (C) 1994-1998, Thomas G. Lane. - * Modified 2015-2017 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains a fast, not so accurate integer implementation of the - * inverse DCT (Discrete Cosine Transform). In the IJG code, this routine - * must also perform dequantization of the input coefficients. - * - * A 2-D IDCT can be done by 1-D IDCT on each column followed by 1-D IDCT - * on each row (or vice versa, but it's more convenient to emit a row at - * a time). Direct algorithms are also available, but they are much more - * complex and seem not to be any faster when reduced to code. - * - * This implementation is based on Arai, Agui, and Nakajima's algorithm for - * scaled DCT. Their original paper (Trans. IEICE E-71(11):1095) is in - * Japanese, but the algorithm is described in the Pennebaker & Mitchell - * JPEG textbook (see REFERENCES section in file README). The following code - * is based directly on figure 4-8 in P&M. - * While an 8-point DCT cannot be done in less than 11 multiplies, it is - * possible to arrange the computation so that many of the multiplies are - * simple scalings of the final outputs. These multiplies can then be - * folded into the multiplications or divisions by the JPEG quantization - * table entries. The AA&N method leaves only 5 multiplies and 29 adds - * to be done in the DCT itself. - * The primary disadvantage of this method is that with fixed-point math, - * accuracy is lost due to imprecise representation of the scaled - * quantization values. The smaller the quantization table entry, the less - * precise the scaled value, so this implementation does worse with high- - * quality-setting files than with low-quality ones. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" -#include "jdct.h" /* Private declarations for DCT subsystem */ - -#ifdef DCT_IFAST_SUPPORTED - - -/* - * This module is specialized to the case DCTSIZE = 8. - */ - -#if DCTSIZE != 8 - Sorry, this code only copes with 8x8 DCT blocks. /* deliberate syntax err */ -#endif - - -/* Scaling decisions are generally the same as in the LL&M algorithm; - * see jidctint.c for more details. However, we choose to descale - * (right shift) multiplication products as soon as they are formed, - * rather than carrying additional fractional bits into subsequent additions. - * This compromises accuracy slightly, but it lets us save a few shifts. - * More importantly, 16-bit arithmetic is then adequate (for 8-bit samples) - * everywhere except in the multiplications proper; this saves a good deal - * of work on 16-bit-int machines. - * - * The dequantized coefficients are not integers because the AA&N scaling - * factors have been incorporated. We represent them scaled up by PASS1_BITS, - * so that the first and second IDCT rounds have the same input scaling. - * For 8-bit JSAMPLEs, we choose IFAST_SCALE_BITS = PASS1_BITS so as to - * avoid a descaling shift; this compromises accuracy rather drastically - * for small quantization table entries, but it saves a lot of shifts. - * For 12-bit JSAMPLEs, there's no hope of using 16x16 multiplies anyway, - * so we use a much larger scaling factor to preserve accuracy. - * - * A final compromise is to represent the multiplicative constants to only - * 8 fractional bits, rather than 13. This saves some shifting work on some - * machines, and may also reduce the cost of multiplication (since there - * are fewer one-bits in the constants). - */ - -#if BITS_IN_JSAMPLE == 8 -#define CONST_BITS 8 -#define PASS1_BITS 2 -#else -#define CONST_BITS 8 -#define PASS1_BITS 1 /* lose a little precision to avoid overflow */ -#endif - -/* Some C compilers fail to reduce "FIX(constant)" at compile time, thus - * causing a lot of useless floating-point operations at run time. - * To get around this we use the following pre-calculated constants. - * If you change CONST_BITS you may want to add appropriate values. - * (With a reasonable C compiler, you can just rely on the FIX() macro...) - */ - -#if CONST_BITS == 8 -#define FIX_1_082392200 ((INT32) 277) /* FIX(1.082392200) */ -#define FIX_1_414213562 ((INT32) 362) /* FIX(1.414213562) */ -#define FIX_1_847759065 ((INT32) 473) /* FIX(1.847759065) */ -#define FIX_2_613125930 ((INT32) 669) /* FIX(2.613125930) */ -#else -#define FIX_1_082392200 FIX(1.082392200) -#define FIX_1_414213562 FIX(1.414213562) -#define FIX_1_847759065 FIX(1.847759065) -#define FIX_2_613125930 FIX(2.613125930) -#endif - - -/* We can gain a little more speed, with a further compromise in accuracy, - * by omitting the addition in a descaling shift. This yields an incorrectly - * rounded result half the time... - */ - -#ifndef USE_ACCURATE_ROUNDING -#undef DESCALE -#define DESCALE(x,n) RIGHT_SHIFT(x, n) -#endif - - -/* Multiply a DCTELEM variable by an INT32 constant, and immediately - * descale to yield a DCTELEM result. - */ - -#define MULTIPLY(var,const) ((DCTELEM) DESCALE((var) * (const), CONST_BITS)) - - -/* Dequantize a coefficient by multiplying it by the multiplier-table - * entry; produce a DCTELEM result. For 8-bit data a 16x16->16 - * multiplication will do. For 12-bit data, the multiplier table is - * declared INT32, so a 32-bit multiply will be used. - */ - -#if BITS_IN_JSAMPLE == 8 -#define DEQUANTIZE(coef,quantval) (((IFAST_MULT_TYPE) (coef)) * (quantval)) -#else -#define DEQUANTIZE(coef,quantval) \ - DESCALE((coef)*(quantval), IFAST_SCALE_BITS-PASS1_BITS) -#endif - - -/* - * Perform dequantization and inverse DCT on one block of coefficients. - * - * cK represents cos(K*pi/16). - */ - -GLOBAL(void) -jpeg_idct_ifast (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - DCTELEM tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7; - DCTELEM tmp10, tmp11, tmp12, tmp13; - DCTELEM z5, z10, z11, z12, z13; - JCOEFPTR inptr; - IFAST_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[DCTSIZE2]; /* buffers data between passes */ - SHIFT_TEMPS /* for DESCALE */ - ISHIFT_TEMPS /* for IRIGHT_SHIFT */ - - /* Pass 1: process columns from input, store into work array. */ - - inptr = coef_block; - quantptr = (IFAST_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = DCTSIZE; ctr > 0; ctr--) { - /* Due to quantization, we will usually find that many of the input - * coefficients are zero, especially the AC terms. We can exploit this - * by short-circuiting the IDCT calculation for any column in which all - * the AC terms are zero. In that case each output is equal to the - * DC coefficient (with scale factor as needed). - * With typical images and quantization tables, half or more of the - * column DCT calculations can be simplified this way. - */ - - if (inptr[DCTSIZE*1] == 0 && inptr[DCTSIZE*2] == 0 && - inptr[DCTSIZE*3] == 0 && inptr[DCTSIZE*4] == 0 && - inptr[DCTSIZE*5] == 0 && inptr[DCTSIZE*6] == 0 && - inptr[DCTSIZE*7] == 0) { - /* AC terms all zero */ - int dcval = (int) DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - - wsptr[DCTSIZE*0] = dcval; - wsptr[DCTSIZE*1] = dcval; - wsptr[DCTSIZE*2] = dcval; - wsptr[DCTSIZE*3] = dcval; - wsptr[DCTSIZE*4] = dcval; - wsptr[DCTSIZE*5] = dcval; - wsptr[DCTSIZE*6] = dcval; - wsptr[DCTSIZE*7] = dcval; - - inptr++; /* advance pointers to next column */ - quantptr++; - wsptr++; - continue; - } - - /* Even part */ - - tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - tmp1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - tmp2 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - tmp3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); - - tmp10 = tmp0 + tmp2; /* phase 3 */ - tmp11 = tmp0 - tmp2; - - tmp13 = tmp1 + tmp3; /* phases 5-3 */ - tmp12 = MULTIPLY(tmp1 - tmp3, FIX_1_414213562) - tmp13; /* 2*c4 */ - - tmp0 = tmp10 + tmp13; /* phase 2 */ - tmp3 = tmp10 - tmp13; - tmp1 = tmp11 + tmp12; - tmp2 = tmp11 - tmp12; - - /* Odd part */ - - tmp4 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - tmp5 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - tmp6 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); - tmp7 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); - - z13 = tmp6 + tmp5; /* phase 6 */ - z10 = tmp6 - tmp5; - z11 = tmp4 + tmp7; - z12 = tmp4 - tmp7; - - tmp7 = z11 + z13; /* phase 5 */ - tmp11 = MULTIPLY(z11 - z13, FIX_1_414213562); /* 2*c4 */ - - z5 = MULTIPLY(z10 + z12, FIX_1_847759065); /* 2*c2 */ - tmp10 = z5 - MULTIPLY(z12, FIX_1_082392200); /* 2*(c2-c6) */ - tmp12 = z5 - MULTIPLY(z10, FIX_2_613125930); /* 2*(c2+c6) */ - - tmp6 = tmp12 - tmp7; /* phase 2 */ - tmp5 = tmp11 - tmp6; - tmp4 = tmp10 - tmp5; - - wsptr[DCTSIZE*0] = (int) (tmp0 + tmp7); - wsptr[DCTSIZE*7] = (int) (tmp0 - tmp7); - wsptr[DCTSIZE*1] = (int) (tmp1 + tmp6); - wsptr[DCTSIZE*6] = (int) (tmp1 - tmp6); - wsptr[DCTSIZE*2] = (int) (tmp2 + tmp5); - wsptr[DCTSIZE*5] = (int) (tmp2 - tmp5); - wsptr[DCTSIZE*3] = (int) (tmp3 + tmp4); - wsptr[DCTSIZE*4] = (int) (tmp3 - tmp4); - - inptr++; /* advance pointers to next column */ - quantptr++; - wsptr++; - } - - /* Pass 2: process rows from work array, store into output array. - * Note that we must descale the results by a factor of 8 == 2**3, - * and also undo the PASS1_BITS scaling. - */ - - wsptr = workspace; - for (ctr = 0; ctr < DCTSIZE; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Add range center and fudge factor for final descale and range-limit. */ - z5 = (DCTELEM) wsptr[0] + - ((((DCTELEM) RANGE_CENTER) << (PASS1_BITS+3)) + - (1 << (PASS1_BITS+2))); - - /* Rows of zeroes can be exploited in the same way as we did with columns. - * However, the column calculation has created many nonzero AC terms, so - * the simplification applies less often (typically 5% to 10% of the time). - * On machines with very fast multiplication, it's possible that the - * test takes more time than it's worth. In that case this section - * may be commented out. - */ - -#ifndef NO_ZERO_ROW_TEST - if (wsptr[1] == 0 && wsptr[2] == 0 && wsptr[3] == 0 && wsptr[4] == 0 && - wsptr[5] == 0 && wsptr[6] == 0 && wsptr[7] == 0) { - /* AC terms all zero */ - JSAMPLE dcval = range_limit[(int) IRIGHT_SHIFT(z5, PASS1_BITS+3) - & RANGE_MASK]; - - outptr[0] = dcval; - outptr[1] = dcval; - outptr[2] = dcval; - outptr[3] = dcval; - outptr[4] = dcval; - outptr[5] = dcval; - outptr[6] = dcval; - outptr[7] = dcval; - - wsptr += DCTSIZE; /* advance pointer to next row */ - continue; - } -#endif - - /* Even part */ - - tmp10 = z5 + (DCTELEM) wsptr[4]; - tmp11 = z5 - (DCTELEM) wsptr[4]; - - tmp13 = (DCTELEM) wsptr[2] + (DCTELEM) wsptr[6]; - tmp12 = MULTIPLY((DCTELEM) wsptr[2] - (DCTELEM) wsptr[6], - FIX_1_414213562) - tmp13; /* 2*c4 */ - - tmp0 = tmp10 + tmp13; - tmp3 = tmp10 - tmp13; - tmp1 = tmp11 + tmp12; - tmp2 = tmp11 - tmp12; - - /* Odd part */ - - z13 = (DCTELEM) wsptr[5] + (DCTELEM) wsptr[3]; - z10 = (DCTELEM) wsptr[5] - (DCTELEM) wsptr[3]; - z11 = (DCTELEM) wsptr[1] + (DCTELEM) wsptr[7]; - z12 = (DCTELEM) wsptr[1] - (DCTELEM) wsptr[7]; - - tmp7 = z11 + z13; /* phase 5 */ - tmp11 = MULTIPLY(z11 - z13, FIX_1_414213562); /* 2*c4 */ - - z5 = MULTIPLY(z10 + z12, FIX_1_847759065); /* 2*c2 */ - tmp10 = z5 - MULTIPLY(z12, FIX_1_082392200); /* 2*(c2-c6) */ - tmp12 = z5 - MULTIPLY(z10, FIX_2_613125930); /* 2*(c2+c6) */ - - tmp6 = tmp12 - tmp7; /* phase 2 */ - tmp5 = tmp11 - tmp6; - tmp4 = tmp10 - tmp5; - - /* Final output stage: scale down by a factor of 8 and range-limit */ - - outptr[0] = range_limit[(int) IRIGHT_SHIFT(tmp0 + tmp7, PASS1_BITS+3) - & RANGE_MASK]; - outptr[7] = range_limit[(int) IRIGHT_SHIFT(tmp0 - tmp7, PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) IRIGHT_SHIFT(tmp1 + tmp6, PASS1_BITS+3) - & RANGE_MASK]; - outptr[6] = range_limit[(int) IRIGHT_SHIFT(tmp1 - tmp6, PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) IRIGHT_SHIFT(tmp2 + tmp5, PASS1_BITS+3) - & RANGE_MASK]; - outptr[5] = range_limit[(int) IRIGHT_SHIFT(tmp2 - tmp5, PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) IRIGHT_SHIFT(tmp3 + tmp4, PASS1_BITS+3) - & RANGE_MASK]; - outptr[4] = range_limit[(int) IRIGHT_SHIFT(tmp3 - tmp4, PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += DCTSIZE; /* advance pointer to next row */ - } -} - -#endif /* DCT_IFAST_SUPPORTED */ diff --git a/cmake/externals/libjpeg/jidctint.c b/cmake/externals/libjpeg/jidctint.c deleted file mode 100644 index e30ec8c8a..000000000 --- a/cmake/externals/libjpeg/jidctint.c +++ /dev/null @@ -1,5240 +0,0 @@ -/* - * jidctint.c - * - * Copyright (C) 1991-1998, Thomas G. Lane. - * Modification developed 2002-2018 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains a slow-but-accurate integer implementation of the - * inverse DCT (Discrete Cosine Transform). In the IJG code, this routine - * must also perform dequantization of the input coefficients. - * - * A 2-D IDCT can be done by 1-D IDCT on each column followed by 1-D IDCT - * on each row (or vice versa, but it's more convenient to emit a row at - * a time). Direct algorithms are also available, but they are much more - * complex and seem not to be any faster when reduced to code. - * - * This implementation is based on an algorithm described in - * C. Loeffler, A. Ligtenberg and G. Moschytz, "Practical Fast 1-D DCT - * Algorithms with 11 Multiplications", Proc. Int'l. Conf. on Acoustics, - * Speech, and Signal Processing 1989 (ICASSP '89), pp. 988-991. - * The primary algorithm described there uses 11 multiplies and 29 adds. - * We use their alternate method with 12 multiplies and 32 adds. - * The advantage of this method is that no data path contains more than one - * multiplication; this allows a very simple and accurate implementation in - * scaled fixed-point arithmetic, with a minimal number of shifts. - * - * We also provide IDCT routines with various output sample block sizes for - * direct resolution reduction or enlargement and for direct resolving the - * common 2x1 and 1x2 subsampling cases without additional resampling: NxN - * (N=1...16), 2NxN, and Nx2N (N=1...8) pixels for one 8x8 input DCT block. - * - * For N<8 we simply take the corresponding low-frequency coefficients of - * the 8x8 input DCT block and apply an NxN point IDCT on the sub-block - * to yield the downscaled outputs. - * This can be seen as direct low-pass downsampling from the DCT domain - * point of view rather than the usual spatial domain point of view, - * yielding significant computational savings and results at least - * as good as common bilinear (averaging) spatial downsampling. - * - * For N>8 we apply a partial NxN IDCT on the 8 input coefficients as - * lower frequencies and higher frequencies assumed to be zero. - * It turns out that the computational effort is similar to the 8x8 IDCT - * regarding the output size. - * Furthermore, the scaling and descaling is the same for all IDCT sizes. - * - * CAUTION: We rely on the FIX() macro except for the N=1,2,4,8 cases - * since there would be too many additional constants to pre-calculate. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" -#include "jdct.h" /* Private declarations for DCT subsystem */ - -#ifdef DCT_ISLOW_SUPPORTED - - -/* - * This module is specialized to the case DCTSIZE = 8. - */ - -#if DCTSIZE != 8 - Sorry, this code only copes with 8x8 DCT blocks. /* deliberate syntax err */ -#endif - - -/* - * The poop on this scaling stuff is as follows: - * - * Each 1-D IDCT step produces outputs which are a factor of sqrt(N) - * larger than the true IDCT outputs. The final outputs are therefore - * a factor of N larger than desired; since N=8 this can be cured by - * a simple right shift at the end of the algorithm. The advantage of - * this arrangement is that we save two multiplications per 1-D IDCT, - * because the y0 and y4 inputs need not be divided by sqrt(N). - * - * We have to do addition and subtraction of the integer inputs, which - * is no problem, and multiplication by fractional constants, which is - * a problem to do in integer arithmetic. We multiply all the constants - * by CONST_SCALE and convert them to integer constants (thus retaining - * CONST_BITS bits of precision in the constants). After doing a - * multiplication we have to divide the product by CONST_SCALE, with proper - * rounding, to produce the correct output. This division can be done - * cheaply as a right shift of CONST_BITS bits. We postpone shifting - * as long as possible so that partial sums can be added together with - * full fractional precision. - * - * The outputs of the first pass are scaled up by PASS1_BITS bits so that - * they are represented to better-than-integral precision. These outputs - * require BITS_IN_JSAMPLE + PASS1_BITS + 3 bits; this fits in a 16-bit word - * with the recommended scaling. (To scale up 12-bit sample data further, an - * intermediate INT32 array would be needed.) - * - * To avoid overflow of the 32-bit intermediate results in pass 2, we must - * have BITS_IN_JSAMPLE + CONST_BITS + PASS1_BITS <= 26. Error analysis - * shows that the values given below are the most effective. - */ - -#if BITS_IN_JSAMPLE == 8 -#define CONST_BITS 13 -#define PASS1_BITS 2 -#else -#define CONST_BITS 13 -#define PASS1_BITS 1 /* lose a little precision to avoid overflow */ -#endif - -/* Some C compilers fail to reduce "FIX(constant)" at compile time, thus - * causing a lot of useless floating-point operations at run time. - * To get around this we use the following pre-calculated constants. - * If you change CONST_BITS you may want to add appropriate values. - * (With a reasonable C compiler, you can just rely on the FIX() macro...) - */ - -#if CONST_BITS == 13 -#define FIX_0_298631336 ((INT32) 2446) /* FIX(0.298631336) */ -#define FIX_0_390180644 ((INT32) 3196) /* FIX(0.390180644) */ -#define FIX_0_541196100 ((INT32) 4433) /* FIX(0.541196100) */ -#define FIX_0_765366865 ((INT32) 6270) /* FIX(0.765366865) */ -#define FIX_0_899976223 ((INT32) 7373) /* FIX(0.899976223) */ -#define FIX_1_175875602 ((INT32) 9633) /* FIX(1.175875602) */ -#define FIX_1_501321110 ((INT32) 12299) /* FIX(1.501321110) */ -#define FIX_1_847759065 ((INT32) 15137) /* FIX(1.847759065) */ -#define FIX_1_961570560 ((INT32) 16069) /* FIX(1.961570560) */ -#define FIX_2_053119869 ((INT32) 16819) /* FIX(2.053119869) */ -#define FIX_2_562915447 ((INT32) 20995) /* FIX(2.562915447) */ -#define FIX_3_072711026 ((INT32) 25172) /* FIX(3.072711026) */ -#else -#define FIX_0_298631336 FIX(0.298631336) -#define FIX_0_390180644 FIX(0.390180644) -#define FIX_0_541196100 FIX(0.541196100) -#define FIX_0_765366865 FIX(0.765366865) -#define FIX_0_899976223 FIX(0.899976223) -#define FIX_1_175875602 FIX(1.175875602) -#define FIX_1_501321110 FIX(1.501321110) -#define FIX_1_847759065 FIX(1.847759065) -#define FIX_1_961570560 FIX(1.961570560) -#define FIX_2_053119869 FIX(2.053119869) -#define FIX_2_562915447 FIX(2.562915447) -#define FIX_3_072711026 FIX(3.072711026) -#endif - - -/* Multiply an INT32 variable by an INT32 constant to yield an INT32 result. - * For 8-bit samples with the recommended scaling, all the variable - * and constant values involved are no more than 16 bits wide, so a - * 16x16->32 bit multiply can be used instead of a full 32x32 multiply. - * For 12-bit samples, a full 32-bit multiplication will be needed. - */ - -#if BITS_IN_JSAMPLE == 8 -#define MULTIPLY(var,const) MULTIPLY16C16(var,const) -#else -#define MULTIPLY(var,const) ((var) * (const)) -#endif - - -/* Dequantize a coefficient by multiplying it by the multiplier-table - * entry; produce an int result. In this module, both inputs and result - * are 16 bits or less, so either int or short multiply will work. - */ - -#define DEQUANTIZE(coef,quantval) (((ISLOW_MULT_TYPE) (coef)) * (quantval)) - - -/* - * Perform dequantization and inverse DCT on one block of coefficients. - * - * Optimized algorithm with 12 multiplications in the 1-D kernel. - * cK represents sqrt(2) * cos(K*pi/16). - */ - -GLOBAL(void) -jpeg_idct_islow (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3; - INT32 tmp10, tmp11, tmp12, tmp13; - INT32 z1, z2, z3; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[DCTSIZE2]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. - * Note results are scaled up by sqrt(8) compared to a true IDCT; - * furthermore, we scale the results by 2**PASS1_BITS. - */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = DCTSIZE; ctr > 0; ctr--) { - /* Due to quantization, we will usually find that many of the input - * coefficients are zero, especially the AC terms. We can exploit this - * by short-circuiting the IDCT calculation for any column in which all - * the AC terms are zero. In that case each output is equal to the - * DC coefficient (with scale factor as needed). - * With typical images and quantization tables, half or more of the - * column DCT calculations can be simplified this way. - */ - - if (inptr[DCTSIZE*1] == 0 && inptr[DCTSIZE*2] == 0 && - inptr[DCTSIZE*3] == 0 && inptr[DCTSIZE*4] == 0 && - inptr[DCTSIZE*5] == 0 && inptr[DCTSIZE*6] == 0 && - inptr[DCTSIZE*7] == 0) { - /* AC terms all zero */ - int dcval = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]) << PASS1_BITS; - - wsptr[DCTSIZE*0] = dcval; - wsptr[DCTSIZE*1] = dcval; - wsptr[DCTSIZE*2] = dcval; - wsptr[DCTSIZE*3] = dcval; - wsptr[DCTSIZE*4] = dcval; - wsptr[DCTSIZE*5] = dcval; - wsptr[DCTSIZE*6] = dcval; - wsptr[DCTSIZE*7] = dcval; - - inptr++; /* advance pointers to next column */ - quantptr++; - wsptr++; - continue; - } - - /* Even part: reverse the even part of the forward DCT. - * The rotator is c(-6). - */ - - z2 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - z3 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - z2 <<= CONST_BITS; - z3 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - z2 += ONE << (CONST_BITS-PASS1_BITS-1); - - tmp0 = z2 + z3; - tmp1 = z2 - z3; - - z2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - z3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); - - z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ - tmp2 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ - tmp3 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ - - tmp10 = tmp0 + tmp2; - tmp13 = tmp0 - tmp2; - tmp11 = tmp1 + tmp3; - tmp12 = tmp1 - tmp3; - - /* Odd part per figure 8; the matrix is unitary and hence its - * transpose is its inverse. i0..i3 are y7,y5,y3,y1 respectively. - */ - - tmp0 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); - tmp1 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); - tmp2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - tmp3 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - - z2 = tmp0 + tmp2; - z3 = tmp1 + tmp3; - - z1 = MULTIPLY(z2 + z3, FIX_1_175875602); /* c3 */ - z2 = MULTIPLY(z2, - FIX_1_961570560); /* -c3-c5 */ - z3 = MULTIPLY(z3, - FIX_0_390180644); /* -c3+c5 */ - z2 += z1; - z3 += z1; - - z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ - tmp0 = MULTIPLY(tmp0, FIX_0_298631336); /* -c1+c3+c5-c7 */ - tmp3 = MULTIPLY(tmp3, FIX_1_501321110); /* c1+c3-c5-c7 */ - tmp0 += z1 + z2; - tmp3 += z1 + z3; - - z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ - tmp1 = MULTIPLY(tmp1, FIX_2_053119869); /* c1+c3-c5+c7 */ - tmp2 = MULTIPLY(tmp2, FIX_3_072711026); /* c1+c3+c5-c7 */ - tmp1 += z1 + z3; - tmp2 += z1 + z2; - - /* Final output stage: inputs are tmp10..tmp13, tmp0..tmp3 */ - - wsptr[DCTSIZE*0] = (int) RIGHT_SHIFT(tmp10 + tmp3, CONST_BITS-PASS1_BITS); - wsptr[DCTSIZE*7] = (int) RIGHT_SHIFT(tmp10 - tmp3, CONST_BITS-PASS1_BITS); - wsptr[DCTSIZE*1] = (int) RIGHT_SHIFT(tmp11 + tmp2, CONST_BITS-PASS1_BITS); - wsptr[DCTSIZE*6] = (int) RIGHT_SHIFT(tmp11 - tmp2, CONST_BITS-PASS1_BITS); - wsptr[DCTSIZE*2] = (int) RIGHT_SHIFT(tmp12 + tmp1, CONST_BITS-PASS1_BITS); - wsptr[DCTSIZE*5] = (int) RIGHT_SHIFT(tmp12 - tmp1, CONST_BITS-PASS1_BITS); - wsptr[DCTSIZE*3] = (int) RIGHT_SHIFT(tmp13 + tmp0, CONST_BITS-PASS1_BITS); - wsptr[DCTSIZE*4] = (int) RIGHT_SHIFT(tmp13 - tmp0, CONST_BITS-PASS1_BITS); - - inptr++; /* advance pointers to next column */ - quantptr++; - wsptr++; - } - - /* Pass 2: process rows from work array, store into output array. - * Note that we must descale the results by a factor of 8 == 2**3, - * and also undo the PASS1_BITS scaling. - */ - - wsptr = workspace; - for (ctr = 0; ctr < DCTSIZE; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Add range center and fudge factor for final descale and range-limit. */ - z2 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - - /* Rows of zeroes can be exploited in the same way as we did with columns. - * However, the column calculation has created many nonzero AC terms, so - * the simplification applies less often (typically 5% to 10% of the time). - * On machines with very fast multiplication, it's possible that the - * test takes more time than it's worth. In that case this section - * may be commented out. - */ - -#ifndef NO_ZERO_ROW_TEST - if (wsptr[1] == 0 && wsptr[2] == 0 && wsptr[3] == 0 && wsptr[4] == 0 && - wsptr[5] == 0 && wsptr[6] == 0 && wsptr[7] == 0) { - /* AC terms all zero */ - JSAMPLE dcval = range_limit[(int) RIGHT_SHIFT(z2, PASS1_BITS+3) - & RANGE_MASK]; - - outptr[0] = dcval; - outptr[1] = dcval; - outptr[2] = dcval; - outptr[3] = dcval; - outptr[4] = dcval; - outptr[5] = dcval; - outptr[6] = dcval; - outptr[7] = dcval; - - wsptr += DCTSIZE; /* advance pointer to next row */ - continue; - } -#endif - - /* Even part: reverse the even part of the forward DCT. - * The rotator is c(-6). - */ - - z3 = (INT32) wsptr[4]; - - tmp0 = (z2 + z3) << CONST_BITS; - tmp1 = (z2 - z3) << CONST_BITS; - - z2 = (INT32) wsptr[2]; - z3 = (INT32) wsptr[6]; - - z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ - tmp2 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ - tmp3 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ - - tmp10 = tmp0 + tmp2; - tmp13 = tmp0 - tmp2; - tmp11 = tmp1 + tmp3; - tmp12 = tmp1 - tmp3; - - /* Odd part per figure 8; the matrix is unitary and hence its - * transpose is its inverse. i0..i3 are y7,y5,y3,y1 respectively. - */ - - tmp0 = (INT32) wsptr[7]; - tmp1 = (INT32) wsptr[5]; - tmp2 = (INT32) wsptr[3]; - tmp3 = (INT32) wsptr[1]; - - z2 = tmp0 + tmp2; - z3 = tmp1 + tmp3; - - z1 = MULTIPLY(z2 + z3, FIX_1_175875602); /* c3 */ - z2 = MULTIPLY(z2, - FIX_1_961570560); /* -c3-c5 */ - z3 = MULTIPLY(z3, - FIX_0_390180644); /* -c3+c5 */ - z2 += z1; - z3 += z1; - - z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ - tmp0 = MULTIPLY(tmp0, FIX_0_298631336); /* -c1+c3+c5-c7 */ - tmp3 = MULTIPLY(tmp3, FIX_1_501321110); /* c1+c3-c5-c7 */ - tmp0 += z1 + z2; - tmp3 += z1 + z3; - - z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ - tmp1 = MULTIPLY(tmp1, FIX_2_053119869); /* c1+c3-c5+c7 */ - tmp2 = MULTIPLY(tmp2, FIX_3_072711026); /* c1+c3+c5-c7 */ - tmp1 += z1 + z3; - tmp2 += z1 + z2; - - /* Final output stage: inputs are tmp10..tmp13, tmp0..tmp3 */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp3, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp3, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp11 + tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp11 - tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12 + tmp1, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp12 - tmp1, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp13 + tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp13 - tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += DCTSIZE; /* advance pointer to next row */ - } -} - -#ifdef IDCT_SCALING_SUPPORTED - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a reduced-size 7x7 output block. - * - * Optimized algorithm with 12 multiplications in the 1-D kernel. - * cK represents sqrt(2) * cos(K*pi/14). - */ - -GLOBAL(void) -jpeg_idct_7x7 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp0, tmp1, tmp2, tmp10, tmp11, tmp12, tmp13; - INT32 z1, z2, z3; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[7*7]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 7; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - tmp13 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - tmp13 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - tmp13 += ONE << (CONST_BITS-PASS1_BITS-1); - - z1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - z2 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - z3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); - - tmp10 = MULTIPLY(z2 - z3, FIX(0.881747734)); /* c4 */ - tmp12 = MULTIPLY(z1 - z2, FIX(0.314692123)); /* c6 */ - tmp11 = tmp10 + tmp12 + tmp13 - MULTIPLY(z2, FIX(1.841218003)); /* c2+c4-c6 */ - tmp0 = z1 + z3; - z2 -= tmp0; - tmp0 = MULTIPLY(tmp0, FIX(1.274162392)) + tmp13; /* c2 */ - tmp10 += tmp0 - MULTIPLY(z3, FIX(0.077722536)); /* c2-c4-c6 */ - tmp12 += tmp0 - MULTIPLY(z1, FIX(2.470602249)); /* c2+c4+c6 */ - tmp13 += MULTIPLY(z2, FIX(1.414213562)); /* c0 */ - - /* Odd part */ - - z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); - - tmp1 = MULTIPLY(z1 + z2, FIX(0.935414347)); /* (c3+c1-c5)/2 */ - tmp2 = MULTIPLY(z1 - z2, FIX(0.170262339)); /* (c3+c5-c1)/2 */ - tmp0 = tmp1 - tmp2; - tmp1 += tmp2; - tmp2 = MULTIPLY(z2 + z3, - FIX(1.378756276)); /* -c1 */ - tmp1 += tmp2; - z2 = MULTIPLY(z1 + z3, FIX(0.613604268)); /* c5 */ - tmp0 += z2; - tmp2 += z2 + MULTIPLY(z3, FIX(1.870828693)); /* c3+c1-c5 */ - - /* Final output stage */ - - wsptr[7*0] = (int) RIGHT_SHIFT(tmp10 + tmp0, CONST_BITS-PASS1_BITS); - wsptr[7*6] = (int) RIGHT_SHIFT(tmp10 - tmp0, CONST_BITS-PASS1_BITS); - wsptr[7*1] = (int) RIGHT_SHIFT(tmp11 + tmp1, CONST_BITS-PASS1_BITS); - wsptr[7*5] = (int) RIGHT_SHIFT(tmp11 - tmp1, CONST_BITS-PASS1_BITS); - wsptr[7*2] = (int) RIGHT_SHIFT(tmp12 + tmp2, CONST_BITS-PASS1_BITS); - wsptr[7*4] = (int) RIGHT_SHIFT(tmp12 - tmp2, CONST_BITS-PASS1_BITS); - wsptr[7*3] = (int) RIGHT_SHIFT(tmp13, CONST_BITS-PASS1_BITS); - } - - /* Pass 2: process 7 rows from work array, store into output array. */ - - wsptr = workspace; - for (ctr = 0; ctr < 7; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - tmp13 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - tmp13 <<= CONST_BITS; - - z1 = (INT32) wsptr[2]; - z2 = (INT32) wsptr[4]; - z3 = (INT32) wsptr[6]; - - tmp10 = MULTIPLY(z2 - z3, FIX(0.881747734)); /* c4 */ - tmp12 = MULTIPLY(z1 - z2, FIX(0.314692123)); /* c6 */ - tmp11 = tmp10 + tmp12 + tmp13 - MULTIPLY(z2, FIX(1.841218003)); /* c2+c4-c6 */ - tmp0 = z1 + z3; - z2 -= tmp0; - tmp0 = MULTIPLY(tmp0, FIX(1.274162392)) + tmp13; /* c2 */ - tmp10 += tmp0 - MULTIPLY(z3, FIX(0.077722536)); /* c2-c4-c6 */ - tmp12 += tmp0 - MULTIPLY(z1, FIX(2.470602249)); /* c2+c4+c6 */ - tmp13 += MULTIPLY(z2, FIX(1.414213562)); /* c0 */ - - /* Odd part */ - - z1 = (INT32) wsptr[1]; - z2 = (INT32) wsptr[3]; - z3 = (INT32) wsptr[5]; - - tmp1 = MULTIPLY(z1 + z2, FIX(0.935414347)); /* (c3+c1-c5)/2 */ - tmp2 = MULTIPLY(z1 - z2, FIX(0.170262339)); /* (c3+c5-c1)/2 */ - tmp0 = tmp1 - tmp2; - tmp1 += tmp2; - tmp2 = MULTIPLY(z2 + z3, - FIX(1.378756276)); /* -c1 */ - tmp1 += tmp2; - z2 = MULTIPLY(z1 + z3, FIX(0.613604268)); /* c5 */ - tmp0 += z2; - tmp2 += z2 + MULTIPLY(z3, FIX(1.870828693)); /* c3+c1-c5 */ - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp11 + tmp1, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp11 - tmp1, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12 + tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp12 - tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += 7; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a reduced-size 6x6 output block. - * - * Optimized algorithm with 3 multiplications in the 1-D kernel. - * cK represents sqrt(2) * cos(K*pi/12). - */ - -GLOBAL(void) -jpeg_idct_6x6 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp0, tmp1, tmp2, tmp10, tmp11, tmp12; - INT32 z1, z2, z3; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[6*6]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 6; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - tmp0 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - tmp0 += ONE << (CONST_BITS-PASS1_BITS-1); - tmp2 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - tmp10 = MULTIPLY(tmp2, FIX(0.707106781)); /* c4 */ - tmp1 = tmp0 + tmp10; - tmp11 = RIGHT_SHIFT(tmp0 - tmp10 - tmp10, CONST_BITS-PASS1_BITS); - tmp10 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - tmp0 = MULTIPLY(tmp10, FIX(1.224744871)); /* c2 */ - tmp10 = tmp1 + tmp0; - tmp12 = tmp1 - tmp0; - - /* Odd part */ - - z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); - tmp1 = MULTIPLY(z1 + z3, FIX(0.366025404)); /* c5 */ - tmp0 = tmp1 + ((z1 + z2) << CONST_BITS); - tmp2 = tmp1 + ((z3 - z2) << CONST_BITS); - tmp1 = (z1 - z2 - z3) << PASS1_BITS; - - /* Final output stage */ - - wsptr[6*0] = (int) RIGHT_SHIFT(tmp10 + tmp0, CONST_BITS-PASS1_BITS); - wsptr[6*5] = (int) RIGHT_SHIFT(tmp10 - tmp0, CONST_BITS-PASS1_BITS); - wsptr[6*1] = (int) (tmp11 + tmp1); - wsptr[6*4] = (int) (tmp11 - tmp1); - wsptr[6*2] = (int) RIGHT_SHIFT(tmp12 + tmp2, CONST_BITS-PASS1_BITS); - wsptr[6*3] = (int) RIGHT_SHIFT(tmp12 - tmp2, CONST_BITS-PASS1_BITS); - } - - /* Pass 2: process 6 rows from work array, store into output array. */ - - wsptr = workspace; - for (ctr = 0; ctr < 6; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - tmp0 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - tmp0 <<= CONST_BITS; - tmp2 = (INT32) wsptr[4]; - tmp10 = MULTIPLY(tmp2, FIX(0.707106781)); /* c4 */ - tmp1 = tmp0 + tmp10; - tmp11 = tmp0 - tmp10 - tmp10; - tmp10 = (INT32) wsptr[2]; - tmp0 = MULTIPLY(tmp10, FIX(1.224744871)); /* c2 */ - tmp10 = tmp1 + tmp0; - tmp12 = tmp1 - tmp0; - - /* Odd part */ - - z1 = (INT32) wsptr[1]; - z2 = (INT32) wsptr[3]; - z3 = (INT32) wsptr[5]; - tmp1 = MULTIPLY(z1 + z3, FIX(0.366025404)); /* c5 */ - tmp0 = tmp1 + ((z1 + z2) << CONST_BITS); - tmp2 = tmp1 + ((z3 - z2) << CONST_BITS); - tmp1 = (z1 - z2 - z3) << CONST_BITS; - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp11 + tmp1, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp11 - tmp1, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12 + tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp12 - tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += 6; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a reduced-size 5x5 output block. - * - * Optimized algorithm with 5 multiplications in the 1-D kernel. - * cK represents sqrt(2) * cos(K*pi/10). - */ - -GLOBAL(void) -jpeg_idct_5x5 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp0, tmp1, tmp10, tmp11, tmp12; - INT32 z1, z2, z3; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[5*5]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 5; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - tmp12 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - tmp12 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - tmp12 += ONE << (CONST_BITS-PASS1_BITS-1); - tmp0 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - tmp1 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - z1 = MULTIPLY(tmp0 + tmp1, FIX(0.790569415)); /* (c2+c4)/2 */ - z2 = MULTIPLY(tmp0 - tmp1, FIX(0.353553391)); /* (c2-c4)/2 */ - z3 = tmp12 + z2; - tmp10 = z3 + z1; - tmp11 = z3 - z1; - tmp12 -= z2 << 2; - - /* Odd part */ - - z2 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - z3 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - - z1 = MULTIPLY(z2 + z3, FIX(0.831253876)); /* c3 */ - tmp0 = z1 + MULTIPLY(z2, FIX(0.513743148)); /* c1-c3 */ - tmp1 = z1 - MULTIPLY(z3, FIX(2.176250899)); /* c1+c3 */ - - /* Final output stage */ - - wsptr[5*0] = (int) RIGHT_SHIFT(tmp10 + tmp0, CONST_BITS-PASS1_BITS); - wsptr[5*4] = (int) RIGHT_SHIFT(tmp10 - tmp0, CONST_BITS-PASS1_BITS); - wsptr[5*1] = (int) RIGHT_SHIFT(tmp11 + tmp1, CONST_BITS-PASS1_BITS); - wsptr[5*3] = (int) RIGHT_SHIFT(tmp11 - tmp1, CONST_BITS-PASS1_BITS); - wsptr[5*2] = (int) RIGHT_SHIFT(tmp12, CONST_BITS-PASS1_BITS); - } - - /* Pass 2: process 5 rows from work array, store into output array. */ - - wsptr = workspace; - for (ctr = 0; ctr < 5; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - tmp12 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - tmp12 <<= CONST_BITS; - tmp0 = (INT32) wsptr[2]; - tmp1 = (INT32) wsptr[4]; - z1 = MULTIPLY(tmp0 + tmp1, FIX(0.790569415)); /* (c2+c4)/2 */ - z2 = MULTIPLY(tmp0 - tmp1, FIX(0.353553391)); /* (c2-c4)/2 */ - z3 = tmp12 + z2; - tmp10 = z3 + z1; - tmp11 = z3 - z1; - tmp12 -= z2 << 2; - - /* Odd part */ - - z2 = (INT32) wsptr[1]; - z3 = (INT32) wsptr[3]; - - z1 = MULTIPLY(z2 + z3, FIX(0.831253876)); /* c3 */ - tmp0 = z1 + MULTIPLY(z2, FIX(0.513743148)); /* c1-c3 */ - tmp1 = z1 - MULTIPLY(z3, FIX(2.176250899)); /* c1+c3 */ - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp11 + tmp1, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp11 - tmp1, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += 5; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a reduced-size 4x4 output block. - * - * Optimized algorithm with 3 multiplications in the 1-D kernel. - * cK represents sqrt(2) * cos(K*pi/16) [refers to 8-point IDCT]. - */ - -GLOBAL(void) -jpeg_idct_4x4 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp0, tmp2, tmp10, tmp12; - INT32 z1, z2, z3; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[4*4]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 4; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - tmp2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - - tmp10 = (tmp0 + tmp2) << PASS1_BITS; - tmp12 = (tmp0 - tmp2) << PASS1_BITS; - - /* Odd part */ - /* Same rotation as in the even part of the 8x8 LL&M IDCT */ - - z2 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - z3 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - - z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ - /* Add fudge factor here for final descale. */ - z1 += ONE << (CONST_BITS-PASS1_BITS-1); - tmp0 = RIGHT_SHIFT(z1 + MULTIPLY(z2, FIX_0_765366865), /* c2-c6 */ - CONST_BITS-PASS1_BITS); - tmp2 = RIGHT_SHIFT(z1 - MULTIPLY(z3, FIX_1_847759065), /* c2+c6 */ - CONST_BITS-PASS1_BITS); - - /* Final output stage */ - - wsptr[4*0] = (int) (tmp10 + tmp0); - wsptr[4*3] = (int) (tmp10 - tmp0); - wsptr[4*1] = (int) (tmp12 + tmp2); - wsptr[4*2] = (int) (tmp12 - tmp2); - } - - /* Pass 2: process 4 rows from work array, store into output array. */ - - wsptr = workspace; - for (ctr = 0; ctr < 4; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - tmp0 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - tmp2 = (INT32) wsptr[2]; - - tmp10 = (tmp0 + tmp2) << CONST_BITS; - tmp12 = (tmp0 - tmp2) << CONST_BITS; - - /* Odd part */ - /* Same rotation as in the even part of the 8x8 LL&M IDCT */ - - z2 = (INT32) wsptr[1]; - z3 = (INT32) wsptr[3]; - - z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ - tmp0 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ - tmp2 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp12 + tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12 - tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += 4; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a reduced-size 3x3 output block. - * - * Optimized algorithm with 2 multiplications in the 1-D kernel. - * cK represents sqrt(2) * cos(K*pi/6). - */ - -GLOBAL(void) -jpeg_idct_3x3 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp0, tmp2, tmp10, tmp12; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[3*3]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 3; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - tmp0 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - tmp0 += ONE << (CONST_BITS-PASS1_BITS-1); - tmp2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - tmp12 = MULTIPLY(tmp2, FIX(0.707106781)); /* c2 */ - tmp10 = tmp0 + tmp12; - tmp2 = tmp0 - tmp12 - tmp12; - - /* Odd part */ - - tmp12 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - tmp0 = MULTIPLY(tmp12, FIX(1.224744871)); /* c1 */ - - /* Final output stage */ - - wsptr[3*0] = (int) RIGHT_SHIFT(tmp10 + tmp0, CONST_BITS-PASS1_BITS); - wsptr[3*2] = (int) RIGHT_SHIFT(tmp10 - tmp0, CONST_BITS-PASS1_BITS); - wsptr[3*1] = (int) RIGHT_SHIFT(tmp2, CONST_BITS-PASS1_BITS); - } - - /* Pass 2: process 3 rows from work array, store into output array. */ - - wsptr = workspace; - for (ctr = 0; ctr < 3; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - tmp0 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - tmp0 <<= CONST_BITS; - tmp2 = (INT32) wsptr[2]; - tmp12 = MULTIPLY(tmp2, FIX(0.707106781)); /* c2 */ - tmp10 = tmp0 + tmp12; - tmp2 = tmp0 - tmp12 - tmp12; - - /* Odd part */ - - tmp12 = (INT32) wsptr[1]; - tmp0 = MULTIPLY(tmp12, FIX(1.224744871)); /* c1 */ - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += 3; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a reduced-size 2x2 output block. - * - * Multiplication-less algorithm. - */ - -GLOBAL(void) -jpeg_idct_2x2 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - DCTELEM tmp0, tmp1, tmp2, tmp3, tmp4, tmp5; - ISLOW_MULT_TYPE * quantptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - ISHIFT_TEMPS - - /* Pass 1: process columns from input. */ - - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - - /* Column 0 */ - tmp4 = DEQUANTIZE(coef_block[DCTSIZE*0], quantptr[DCTSIZE*0]); - tmp5 = DEQUANTIZE(coef_block[DCTSIZE*1], quantptr[DCTSIZE*1]); - /* Add range center and fudge factor for final descale and range-limit. */ - tmp4 += (((DCTELEM) RANGE_CENTER) << 3) + (1 << 2); - - tmp0 = tmp4 + tmp5; - tmp2 = tmp4 - tmp5; - - /* Column 1 */ - tmp4 = DEQUANTIZE(coef_block[DCTSIZE*0+1], quantptr[DCTSIZE*0+1]); - tmp5 = DEQUANTIZE(coef_block[DCTSIZE*1+1], quantptr[DCTSIZE*1+1]); - - tmp1 = tmp4 + tmp5; - tmp3 = tmp4 - tmp5; - - /* Pass 2: process 2 rows, store into output array. */ - - /* Row 0 */ - outptr = output_buf[0] + output_col; - - outptr[0] = range_limit[(int) IRIGHT_SHIFT(tmp0 + tmp1, 3) & RANGE_MASK]; - outptr[1] = range_limit[(int) IRIGHT_SHIFT(tmp0 - tmp1, 3) & RANGE_MASK]; - - /* Row 1 */ - outptr = output_buf[1] + output_col; - - outptr[0] = range_limit[(int) IRIGHT_SHIFT(tmp2 + tmp3, 3) & RANGE_MASK]; - outptr[1] = range_limit[(int) IRIGHT_SHIFT(tmp2 - tmp3, 3) & RANGE_MASK]; -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a reduced-size 1x1 output block. - * - * We hardly need an inverse DCT routine for this: just take the - * average pixel value, which is one-eighth of the DC coefficient. - */ - -GLOBAL(void) -jpeg_idct_1x1 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - DCTELEM dcval; - ISLOW_MULT_TYPE * quantptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - ISHIFT_TEMPS - - /* 1x1 is trivial: just take the DC coefficient divided by 8. */ - - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - - dcval = DEQUANTIZE(coef_block[0], quantptr[0]); - /* Add range center and fudge factor for descale and range-limit. */ - dcval += (((DCTELEM) RANGE_CENTER) << 3) + (1 << 2); - - output_buf[0][output_col] = - range_limit[(int) IRIGHT_SHIFT(dcval, 3) & RANGE_MASK]; -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a 9x9 output block. - * - * Optimized algorithm with 10 multiplications in the 1-D kernel. - * cK represents sqrt(2) * cos(K*pi/18). - */ - -GLOBAL(void) -jpeg_idct_9x9 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3, tmp10, tmp11, tmp12, tmp13, tmp14; - INT32 z1, z2, z3, z4; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[8*9]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - tmp0 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - tmp0 += ONE << (CONST_BITS-PASS1_BITS-1); - - z1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - z2 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - z3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); - - tmp3 = MULTIPLY(z3, FIX(0.707106781)); /* c6 */ - tmp1 = tmp0 + tmp3; - tmp2 = tmp0 - tmp3 - tmp3; - - tmp0 = MULTIPLY(z1 - z2, FIX(0.707106781)); /* c6 */ - tmp11 = tmp2 + tmp0; - tmp14 = tmp2 - tmp0 - tmp0; - - tmp0 = MULTIPLY(z1 + z2, FIX(1.328926049)); /* c2 */ - tmp2 = MULTIPLY(z1, FIX(1.083350441)); /* c4 */ - tmp3 = MULTIPLY(z2, FIX(0.245575608)); /* c8 */ - - tmp10 = tmp1 + tmp0 - tmp3; - tmp12 = tmp1 - tmp0 + tmp2; - tmp13 = tmp1 - tmp2 + tmp3; - - /* Odd part */ - - z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); - z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); - - z2 = MULTIPLY(z2, - FIX(1.224744871)); /* -c3 */ - - tmp2 = MULTIPLY(z1 + z3, FIX(0.909038955)); /* c5 */ - tmp3 = MULTIPLY(z1 + z4, FIX(0.483689525)); /* c7 */ - tmp0 = tmp2 + tmp3 - z2; - tmp1 = MULTIPLY(z3 - z4, FIX(1.392728481)); /* c1 */ - tmp2 += z2 - tmp1; - tmp3 += z2 + tmp1; - tmp1 = MULTIPLY(z1 - z3 - z4, FIX(1.224744871)); /* c3 */ - - /* Final output stage */ - - wsptr[8*0] = (int) RIGHT_SHIFT(tmp10 + tmp0, CONST_BITS-PASS1_BITS); - wsptr[8*8] = (int) RIGHT_SHIFT(tmp10 - tmp0, CONST_BITS-PASS1_BITS); - wsptr[8*1] = (int) RIGHT_SHIFT(tmp11 + tmp1, CONST_BITS-PASS1_BITS); - wsptr[8*7] = (int) RIGHT_SHIFT(tmp11 - tmp1, CONST_BITS-PASS1_BITS); - wsptr[8*2] = (int) RIGHT_SHIFT(tmp12 + tmp2, CONST_BITS-PASS1_BITS); - wsptr[8*6] = (int) RIGHT_SHIFT(tmp12 - tmp2, CONST_BITS-PASS1_BITS); - wsptr[8*3] = (int) RIGHT_SHIFT(tmp13 + tmp3, CONST_BITS-PASS1_BITS); - wsptr[8*5] = (int) RIGHT_SHIFT(tmp13 - tmp3, CONST_BITS-PASS1_BITS); - wsptr[8*4] = (int) RIGHT_SHIFT(tmp14, CONST_BITS-PASS1_BITS); - } - - /* Pass 2: process 9 rows from work array, store into output array. */ - - wsptr = workspace; - for (ctr = 0; ctr < 9; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - tmp0 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - tmp0 <<= CONST_BITS; - - z1 = (INT32) wsptr[2]; - z2 = (INT32) wsptr[4]; - z3 = (INT32) wsptr[6]; - - tmp3 = MULTIPLY(z3, FIX(0.707106781)); /* c6 */ - tmp1 = tmp0 + tmp3; - tmp2 = tmp0 - tmp3 - tmp3; - - tmp0 = MULTIPLY(z1 - z2, FIX(0.707106781)); /* c6 */ - tmp11 = tmp2 + tmp0; - tmp14 = tmp2 - tmp0 - tmp0; - - tmp0 = MULTIPLY(z1 + z2, FIX(1.328926049)); /* c2 */ - tmp2 = MULTIPLY(z1, FIX(1.083350441)); /* c4 */ - tmp3 = MULTIPLY(z2, FIX(0.245575608)); /* c8 */ - - tmp10 = tmp1 + tmp0 - tmp3; - tmp12 = tmp1 - tmp0 + tmp2; - tmp13 = tmp1 - tmp2 + tmp3; - - /* Odd part */ - - z1 = (INT32) wsptr[1]; - z2 = (INT32) wsptr[3]; - z3 = (INT32) wsptr[5]; - z4 = (INT32) wsptr[7]; - - z2 = MULTIPLY(z2, - FIX(1.224744871)); /* -c3 */ - - tmp2 = MULTIPLY(z1 + z3, FIX(0.909038955)); /* c5 */ - tmp3 = MULTIPLY(z1 + z4, FIX(0.483689525)); /* c7 */ - tmp0 = tmp2 + tmp3 - z2; - tmp1 = MULTIPLY(z3 - z4, FIX(1.392728481)); /* c1 */ - tmp2 += z2 - tmp1; - tmp3 += z2 + tmp1; - tmp1 = MULTIPLY(z1 - z3 - z4, FIX(1.224744871)); /* c3 */ - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp11 + tmp1, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp11 - tmp1, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12 + tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp12 - tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp13 + tmp3, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp13 - tmp3, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp14, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += 8; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a 10x10 output block. - * - * Optimized algorithm with 12 multiplications in the 1-D kernel. - * cK represents sqrt(2) * cos(K*pi/20). - */ - -GLOBAL(void) -jpeg_idct_10x10 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp10, tmp11, tmp12, tmp13, tmp14; - INT32 tmp20, tmp21, tmp22, tmp23, tmp24; - INT32 z1, z2, z3, z4, z5; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[8*10]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - z3 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - z3 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - z3 += ONE << (CONST_BITS-PASS1_BITS-1); - z4 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - z1 = MULTIPLY(z4, FIX(1.144122806)); /* c4 */ - z2 = MULTIPLY(z4, FIX(0.437016024)); /* c8 */ - tmp10 = z3 + z1; - tmp11 = z3 - z2; - - tmp22 = RIGHT_SHIFT(z3 - ((z1 - z2) << 1), /* c0 = (c4-c8)*2 */ - CONST_BITS-PASS1_BITS); - - z2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - z3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); - - z1 = MULTIPLY(z2 + z3, FIX(0.831253876)); /* c6 */ - tmp12 = z1 + MULTIPLY(z2, FIX(0.513743148)); /* c2-c6 */ - tmp13 = z1 - MULTIPLY(z3, FIX(2.176250899)); /* c2+c6 */ - - tmp20 = tmp10 + tmp12; - tmp24 = tmp10 - tmp12; - tmp21 = tmp11 + tmp13; - tmp23 = tmp11 - tmp13; - - /* Odd part */ - - z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); - z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); - - tmp11 = z2 + z4; - tmp13 = z2 - z4; - - tmp12 = MULTIPLY(tmp13, FIX(0.309016994)); /* (c3-c7)/2 */ - z5 = z3 << CONST_BITS; - - z2 = MULTIPLY(tmp11, FIX(0.951056516)); /* (c3+c7)/2 */ - z4 = z5 + tmp12; - - tmp10 = MULTIPLY(z1, FIX(1.396802247)) + z2 + z4; /* c1 */ - tmp14 = MULTIPLY(z1, FIX(0.221231742)) - z2 + z4; /* c9 */ - - z2 = MULTIPLY(tmp11, FIX(0.587785252)); /* (c1-c9)/2 */ - z4 = z5 - tmp12 - (tmp13 << (CONST_BITS - 1)); - - tmp12 = (z1 - tmp13 - z3) << PASS1_BITS; - - tmp11 = MULTIPLY(z1, FIX(1.260073511)) - z2 - z4; /* c3 */ - tmp13 = MULTIPLY(z1, FIX(0.642039522)) - z2 + z4; /* c7 */ - - /* Final output stage */ - - wsptr[8*0] = (int) RIGHT_SHIFT(tmp20 + tmp10, CONST_BITS-PASS1_BITS); - wsptr[8*9] = (int) RIGHT_SHIFT(tmp20 - tmp10, CONST_BITS-PASS1_BITS); - wsptr[8*1] = (int) RIGHT_SHIFT(tmp21 + tmp11, CONST_BITS-PASS1_BITS); - wsptr[8*8] = (int) RIGHT_SHIFT(tmp21 - tmp11, CONST_BITS-PASS1_BITS); - wsptr[8*2] = (int) (tmp22 + tmp12); - wsptr[8*7] = (int) (tmp22 - tmp12); - wsptr[8*3] = (int) RIGHT_SHIFT(tmp23 + tmp13, CONST_BITS-PASS1_BITS); - wsptr[8*6] = (int) RIGHT_SHIFT(tmp23 - tmp13, CONST_BITS-PASS1_BITS); - wsptr[8*4] = (int) RIGHT_SHIFT(tmp24 + tmp14, CONST_BITS-PASS1_BITS); - wsptr[8*5] = (int) RIGHT_SHIFT(tmp24 - tmp14, CONST_BITS-PASS1_BITS); - } - - /* Pass 2: process 10 rows from work array, store into output array. */ - - wsptr = workspace; - for (ctr = 0; ctr < 10; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - z3 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - z3 <<= CONST_BITS; - z4 = (INT32) wsptr[4]; - z1 = MULTIPLY(z4, FIX(1.144122806)); /* c4 */ - z2 = MULTIPLY(z4, FIX(0.437016024)); /* c8 */ - tmp10 = z3 + z1; - tmp11 = z3 - z2; - - tmp22 = z3 - ((z1 - z2) << 1); /* c0 = (c4-c8)*2 */ - - z2 = (INT32) wsptr[2]; - z3 = (INT32) wsptr[6]; - - z1 = MULTIPLY(z2 + z3, FIX(0.831253876)); /* c6 */ - tmp12 = z1 + MULTIPLY(z2, FIX(0.513743148)); /* c2-c6 */ - tmp13 = z1 - MULTIPLY(z3, FIX(2.176250899)); /* c2+c6 */ - - tmp20 = tmp10 + tmp12; - tmp24 = tmp10 - tmp12; - tmp21 = tmp11 + tmp13; - tmp23 = tmp11 - tmp13; - - /* Odd part */ - - z1 = (INT32) wsptr[1]; - z2 = (INT32) wsptr[3]; - z3 = (INT32) wsptr[5]; - z3 <<= CONST_BITS; - z4 = (INT32) wsptr[7]; - - tmp11 = z2 + z4; - tmp13 = z2 - z4; - - tmp12 = MULTIPLY(tmp13, FIX(0.309016994)); /* (c3-c7)/2 */ - - z2 = MULTIPLY(tmp11, FIX(0.951056516)); /* (c3+c7)/2 */ - z4 = z3 + tmp12; - - tmp10 = MULTIPLY(z1, FIX(1.396802247)) + z2 + z4; /* c1 */ - tmp14 = MULTIPLY(z1, FIX(0.221231742)) - z2 + z4; /* c9 */ - - z2 = MULTIPLY(tmp11, FIX(0.587785252)); /* (c1-c9)/2 */ - z4 = z3 - tmp12 - (tmp13 << (CONST_BITS - 1)); - - tmp12 = ((z1 - tmp13) << CONST_BITS) - z3; - - tmp11 = MULTIPLY(z1, FIX(1.260073511)) - z2 - z4; /* c3 */ - tmp13 = MULTIPLY(z1, FIX(0.642039522)) - z2 + z4; /* c7 */ - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[9] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23 + tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp23 - tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp24 + tmp14, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp24 - tmp14, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += 8; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing an 11x11 output block. - * - * Optimized algorithm with 24 multiplications in the 1-D kernel. - * cK represents sqrt(2) * cos(K*pi/22). - */ - -GLOBAL(void) -jpeg_idct_11x11 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp10, tmp11, tmp12, tmp13, tmp14; - INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25; - INT32 z1, z2, z3, z4; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[8*11]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - tmp10 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - tmp10 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - tmp10 += ONE << (CONST_BITS-PASS1_BITS-1); - - z1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - z2 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - z3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); - - tmp20 = MULTIPLY(z2 - z3, FIX(2.546640132)); /* c2+c4 */ - tmp23 = MULTIPLY(z2 - z1, FIX(0.430815045)); /* c2-c6 */ - z4 = z1 + z3; - tmp24 = MULTIPLY(z4, - FIX(1.155664402)); /* -(c2-c10) */ - z4 -= z2; - tmp25 = tmp10 + MULTIPLY(z4, FIX(1.356927976)); /* c2 */ - tmp21 = tmp20 + tmp23 + tmp25 - - MULTIPLY(z2, FIX(1.821790775)); /* c2+c4+c10-c6 */ - tmp20 += tmp25 + MULTIPLY(z3, FIX(2.115825087)); /* c4+c6 */ - tmp23 += tmp25 - MULTIPLY(z1, FIX(1.513598477)); /* c6+c8 */ - tmp24 += tmp25; - tmp22 = tmp24 - MULTIPLY(z3, FIX(0.788749120)); /* c8+c10 */ - tmp24 += MULTIPLY(z2, FIX(1.944413522)) - /* c2+c8 */ - MULTIPLY(z1, FIX(1.390975730)); /* c4+c10 */ - tmp25 = tmp10 - MULTIPLY(z4, FIX(1.414213562)); /* c0 */ - - /* Odd part */ - - z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); - z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); - - tmp11 = z1 + z2; - tmp14 = MULTIPLY(tmp11 + z3 + z4, FIX(0.398430003)); /* c9 */ - tmp11 = MULTIPLY(tmp11, FIX(0.887983902)); /* c3-c9 */ - tmp12 = MULTIPLY(z1 + z3, FIX(0.670361295)); /* c5-c9 */ - tmp13 = tmp14 + MULTIPLY(z1 + z4, FIX(0.366151574)); /* c7-c9 */ - tmp10 = tmp11 + tmp12 + tmp13 - - MULTIPLY(z1, FIX(0.923107866)); /* c7+c5+c3-c1-2*c9 */ - z1 = tmp14 - MULTIPLY(z2 + z3, FIX(1.163011579)); /* c7+c9 */ - tmp11 += z1 + MULTIPLY(z2, FIX(2.073276588)); /* c1+c7+3*c9-c3 */ - tmp12 += z1 - MULTIPLY(z3, FIX(1.192193623)); /* c3+c5-c7-c9 */ - z1 = MULTIPLY(z2 + z4, - FIX(1.798248910)); /* -(c1+c9) */ - tmp11 += z1; - tmp13 += z1 + MULTIPLY(z4, FIX(2.102458632)); /* c1+c5+c9-c7 */ - tmp14 += MULTIPLY(z2, - FIX(1.467221301)) + /* -(c5+c9) */ - MULTIPLY(z3, FIX(1.001388905)) - /* c1-c9 */ - MULTIPLY(z4, FIX(1.684843907)); /* c3+c9 */ - - /* Final output stage */ - - wsptr[8*0] = (int) RIGHT_SHIFT(tmp20 + tmp10, CONST_BITS-PASS1_BITS); - wsptr[8*10] = (int) RIGHT_SHIFT(tmp20 - tmp10, CONST_BITS-PASS1_BITS); - wsptr[8*1] = (int) RIGHT_SHIFT(tmp21 + tmp11, CONST_BITS-PASS1_BITS); - wsptr[8*9] = (int) RIGHT_SHIFT(tmp21 - tmp11, CONST_BITS-PASS1_BITS); - wsptr[8*2] = (int) RIGHT_SHIFT(tmp22 + tmp12, CONST_BITS-PASS1_BITS); - wsptr[8*8] = (int) RIGHT_SHIFT(tmp22 - tmp12, CONST_BITS-PASS1_BITS); - wsptr[8*3] = (int) RIGHT_SHIFT(tmp23 + tmp13, CONST_BITS-PASS1_BITS); - wsptr[8*7] = (int) RIGHT_SHIFT(tmp23 - tmp13, CONST_BITS-PASS1_BITS); - wsptr[8*4] = (int) RIGHT_SHIFT(tmp24 + tmp14, CONST_BITS-PASS1_BITS); - wsptr[8*6] = (int) RIGHT_SHIFT(tmp24 - tmp14, CONST_BITS-PASS1_BITS); - wsptr[8*5] = (int) RIGHT_SHIFT(tmp25, CONST_BITS-PASS1_BITS); - } - - /* Pass 2: process 11 rows from work array, store into output array. */ - - wsptr = workspace; - for (ctr = 0; ctr < 11; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - tmp10 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - tmp10 <<= CONST_BITS; - - z1 = (INT32) wsptr[2]; - z2 = (INT32) wsptr[4]; - z3 = (INT32) wsptr[6]; - - tmp20 = MULTIPLY(z2 - z3, FIX(2.546640132)); /* c2+c4 */ - tmp23 = MULTIPLY(z2 - z1, FIX(0.430815045)); /* c2-c6 */ - z4 = z1 + z3; - tmp24 = MULTIPLY(z4, - FIX(1.155664402)); /* -(c2-c10) */ - z4 -= z2; - tmp25 = tmp10 + MULTIPLY(z4, FIX(1.356927976)); /* c2 */ - tmp21 = tmp20 + tmp23 + tmp25 - - MULTIPLY(z2, FIX(1.821790775)); /* c2+c4+c10-c6 */ - tmp20 += tmp25 + MULTIPLY(z3, FIX(2.115825087)); /* c4+c6 */ - tmp23 += tmp25 - MULTIPLY(z1, FIX(1.513598477)); /* c6+c8 */ - tmp24 += tmp25; - tmp22 = tmp24 - MULTIPLY(z3, FIX(0.788749120)); /* c8+c10 */ - tmp24 += MULTIPLY(z2, FIX(1.944413522)) - /* c2+c8 */ - MULTIPLY(z1, FIX(1.390975730)); /* c4+c10 */ - tmp25 = tmp10 - MULTIPLY(z4, FIX(1.414213562)); /* c0 */ - - /* Odd part */ - - z1 = (INT32) wsptr[1]; - z2 = (INT32) wsptr[3]; - z3 = (INT32) wsptr[5]; - z4 = (INT32) wsptr[7]; - - tmp11 = z1 + z2; - tmp14 = MULTIPLY(tmp11 + z3 + z4, FIX(0.398430003)); /* c9 */ - tmp11 = MULTIPLY(tmp11, FIX(0.887983902)); /* c3-c9 */ - tmp12 = MULTIPLY(z1 + z3, FIX(0.670361295)); /* c5-c9 */ - tmp13 = tmp14 + MULTIPLY(z1 + z4, FIX(0.366151574)); /* c7-c9 */ - tmp10 = tmp11 + tmp12 + tmp13 - - MULTIPLY(z1, FIX(0.923107866)); /* c7+c5+c3-c1-2*c9 */ - z1 = tmp14 - MULTIPLY(z2 + z3, FIX(1.163011579)); /* c7+c9 */ - tmp11 += z1 + MULTIPLY(z2, FIX(2.073276588)); /* c1+c7+3*c9-c3 */ - tmp12 += z1 - MULTIPLY(z3, FIX(1.192193623)); /* c3+c5-c7-c9 */ - z1 = MULTIPLY(z2 + z4, - FIX(1.798248910)); /* -(c1+c9) */ - tmp11 += z1; - tmp13 += z1 + MULTIPLY(z4, FIX(2.102458632)); /* c1+c5+c9-c7 */ - tmp14 += MULTIPLY(z2, - FIX(1.467221301)) + /* -(c5+c9) */ - MULTIPLY(z3, FIX(1.001388905)) - /* c1-c9 */ - MULTIPLY(z4, FIX(1.684843907)); /* c3+c9 */ - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[10] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[9] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23 + tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp23 - tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp24 + tmp14, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp24 - tmp14, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp25, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += 8; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a 12x12 output block. - * - * Optimized algorithm with 15 multiplications in the 1-D kernel. - * cK represents sqrt(2) * cos(K*pi/24). - */ - -GLOBAL(void) -jpeg_idct_12x12 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15; - INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25; - INT32 z1, z2, z3, z4; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[8*12]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - z3 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - z3 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - z3 += ONE << (CONST_BITS-PASS1_BITS-1); - - z4 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - z4 = MULTIPLY(z4, FIX(1.224744871)); /* c4 */ - - tmp10 = z3 + z4; - tmp11 = z3 - z4; - - z1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - z4 = MULTIPLY(z1, FIX(1.366025404)); /* c2 */ - z1 <<= CONST_BITS; - z2 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); - z2 <<= CONST_BITS; - - tmp12 = z1 - z2; - - tmp21 = z3 + tmp12; - tmp24 = z3 - tmp12; - - tmp12 = z4 + z2; - - tmp20 = tmp10 + tmp12; - tmp25 = tmp10 - tmp12; - - tmp12 = z4 - z1 - z2; - - tmp22 = tmp11 + tmp12; - tmp23 = tmp11 - tmp12; - - /* Odd part */ - - z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); - z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); - - tmp11 = MULTIPLY(z2, FIX(1.306562965)); /* c3 */ - tmp14 = MULTIPLY(z2, - FIX_0_541196100); /* -c9 */ - - tmp10 = z1 + z3; - tmp15 = MULTIPLY(tmp10 + z4, FIX(0.860918669)); /* c7 */ - tmp12 = tmp15 + MULTIPLY(tmp10, FIX(0.261052384)); /* c5-c7 */ - tmp10 = tmp12 + tmp11 + MULTIPLY(z1, FIX(0.280143716)); /* c1-c5 */ - tmp13 = MULTIPLY(z3 + z4, - FIX(1.045510580)); /* -(c7+c11) */ - tmp12 += tmp13 + tmp14 - MULTIPLY(z3, FIX(1.478575242)); /* c1+c5-c7-c11 */ - tmp13 += tmp15 - tmp11 + MULTIPLY(z4, FIX(1.586706681)); /* c1+c11 */ - tmp15 += tmp14 - MULTIPLY(z1, FIX(0.676326758)) - /* c7-c11 */ - MULTIPLY(z4, FIX(1.982889723)); /* c5+c7 */ - - z1 -= z4; - z2 -= z3; - z3 = MULTIPLY(z1 + z2, FIX_0_541196100); /* c9 */ - tmp11 = z3 + MULTIPLY(z1, FIX_0_765366865); /* c3-c9 */ - tmp14 = z3 - MULTIPLY(z2, FIX_1_847759065); /* c3+c9 */ - - /* Final output stage */ - - wsptr[8*0] = (int) RIGHT_SHIFT(tmp20 + tmp10, CONST_BITS-PASS1_BITS); - wsptr[8*11] = (int) RIGHT_SHIFT(tmp20 - tmp10, CONST_BITS-PASS1_BITS); - wsptr[8*1] = (int) RIGHT_SHIFT(tmp21 + tmp11, CONST_BITS-PASS1_BITS); - wsptr[8*10] = (int) RIGHT_SHIFT(tmp21 - tmp11, CONST_BITS-PASS1_BITS); - wsptr[8*2] = (int) RIGHT_SHIFT(tmp22 + tmp12, CONST_BITS-PASS1_BITS); - wsptr[8*9] = (int) RIGHT_SHIFT(tmp22 - tmp12, CONST_BITS-PASS1_BITS); - wsptr[8*3] = (int) RIGHT_SHIFT(tmp23 + tmp13, CONST_BITS-PASS1_BITS); - wsptr[8*8] = (int) RIGHT_SHIFT(tmp23 - tmp13, CONST_BITS-PASS1_BITS); - wsptr[8*4] = (int) RIGHT_SHIFT(tmp24 + tmp14, CONST_BITS-PASS1_BITS); - wsptr[8*7] = (int) RIGHT_SHIFT(tmp24 - tmp14, CONST_BITS-PASS1_BITS); - wsptr[8*5] = (int) RIGHT_SHIFT(tmp25 + tmp15, CONST_BITS-PASS1_BITS); - wsptr[8*6] = (int) RIGHT_SHIFT(tmp25 - tmp15, CONST_BITS-PASS1_BITS); - } - - /* Pass 2: process 12 rows from work array, store into output array. */ - - wsptr = workspace; - for (ctr = 0; ctr < 12; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - z3 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - z3 <<= CONST_BITS; - - z4 = (INT32) wsptr[4]; - z4 = MULTIPLY(z4, FIX(1.224744871)); /* c4 */ - - tmp10 = z3 + z4; - tmp11 = z3 - z4; - - z1 = (INT32) wsptr[2]; - z4 = MULTIPLY(z1, FIX(1.366025404)); /* c2 */ - z1 <<= CONST_BITS; - z2 = (INT32) wsptr[6]; - z2 <<= CONST_BITS; - - tmp12 = z1 - z2; - - tmp21 = z3 + tmp12; - tmp24 = z3 - tmp12; - - tmp12 = z4 + z2; - - tmp20 = tmp10 + tmp12; - tmp25 = tmp10 - tmp12; - - tmp12 = z4 - z1 - z2; - - tmp22 = tmp11 + tmp12; - tmp23 = tmp11 - tmp12; - - /* Odd part */ - - z1 = (INT32) wsptr[1]; - z2 = (INT32) wsptr[3]; - z3 = (INT32) wsptr[5]; - z4 = (INT32) wsptr[7]; - - tmp11 = MULTIPLY(z2, FIX(1.306562965)); /* c3 */ - tmp14 = MULTIPLY(z2, - FIX_0_541196100); /* -c9 */ - - tmp10 = z1 + z3; - tmp15 = MULTIPLY(tmp10 + z4, FIX(0.860918669)); /* c7 */ - tmp12 = tmp15 + MULTIPLY(tmp10, FIX(0.261052384)); /* c5-c7 */ - tmp10 = tmp12 + tmp11 + MULTIPLY(z1, FIX(0.280143716)); /* c1-c5 */ - tmp13 = MULTIPLY(z3 + z4, - FIX(1.045510580)); /* -(c7+c11) */ - tmp12 += tmp13 + tmp14 - MULTIPLY(z3, FIX(1.478575242)); /* c1+c5-c7-c11 */ - tmp13 += tmp15 - tmp11 + MULTIPLY(z4, FIX(1.586706681)); /* c1+c11 */ - tmp15 += tmp14 - MULTIPLY(z1, FIX(0.676326758)) - /* c7-c11 */ - MULTIPLY(z4, FIX(1.982889723)); /* c5+c7 */ - - z1 -= z4; - z2 -= z3; - z3 = MULTIPLY(z1 + z2, FIX_0_541196100); /* c9 */ - tmp11 = z3 + MULTIPLY(z1, FIX_0_765366865); /* c3-c9 */ - tmp14 = z3 - MULTIPLY(z2, FIX_1_847759065); /* c3+c9 */ - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[11] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[10] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[9] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23 + tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp23 - tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp24 + tmp14, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp24 - tmp14, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp25 + tmp15, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp25 - tmp15, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += 8; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a 13x13 output block. - * - * Optimized algorithm with 29 multiplications in the 1-D kernel. - * cK represents sqrt(2) * cos(K*pi/26). - */ - -GLOBAL(void) -jpeg_idct_13x13 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15; - INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25, tmp26; - INT32 z1, z2, z3, z4; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[8*13]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - z1 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - z1 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - z1 += ONE << (CONST_BITS-PASS1_BITS-1); - - z2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - z3 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - z4 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); - - tmp10 = z3 + z4; - tmp11 = z3 - z4; - - tmp12 = MULTIPLY(tmp10, FIX(1.155388986)); /* (c4+c6)/2 */ - tmp13 = MULTIPLY(tmp11, FIX(0.096834934)) + z1; /* (c4-c6)/2 */ - - tmp20 = MULTIPLY(z2, FIX(1.373119086)) + tmp12 + tmp13; /* c2 */ - tmp22 = MULTIPLY(z2, FIX(0.501487041)) - tmp12 + tmp13; /* c10 */ - - tmp12 = MULTIPLY(tmp10, FIX(0.316450131)); /* (c8-c12)/2 */ - tmp13 = MULTIPLY(tmp11, FIX(0.486914739)) + z1; /* (c8+c12)/2 */ - - tmp21 = MULTIPLY(z2, FIX(1.058554052)) - tmp12 + tmp13; /* c6 */ - tmp25 = MULTIPLY(z2, - FIX(1.252223920)) + tmp12 + tmp13; /* c4 */ - - tmp12 = MULTIPLY(tmp10, FIX(0.435816023)); /* (c2-c10)/2 */ - tmp13 = MULTIPLY(tmp11, FIX(0.937303064)) - z1; /* (c2+c10)/2 */ - - tmp23 = MULTIPLY(z2, - FIX(0.170464608)) - tmp12 - tmp13; /* c12 */ - tmp24 = MULTIPLY(z2, - FIX(0.803364869)) + tmp12 - tmp13; /* c8 */ - - tmp26 = MULTIPLY(tmp11 - z2, FIX(1.414213562)) + z1; /* c0 */ - - /* Odd part */ - - z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); - z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); - - tmp11 = MULTIPLY(z1 + z2, FIX(1.322312651)); /* c3 */ - tmp12 = MULTIPLY(z1 + z3, FIX(1.163874945)); /* c5 */ - tmp15 = z1 + z4; - tmp13 = MULTIPLY(tmp15, FIX(0.937797057)); /* c7 */ - tmp10 = tmp11 + tmp12 + tmp13 - - MULTIPLY(z1, FIX(2.020082300)); /* c7+c5+c3-c1 */ - tmp14 = MULTIPLY(z2 + z3, - FIX(0.338443458)); /* -c11 */ - tmp11 += tmp14 + MULTIPLY(z2, FIX(0.837223564)); /* c5+c9+c11-c3 */ - tmp12 += tmp14 - MULTIPLY(z3, FIX(1.572116027)); /* c1+c5-c9-c11 */ - tmp14 = MULTIPLY(z2 + z4, - FIX(1.163874945)); /* -c5 */ - tmp11 += tmp14; - tmp13 += tmp14 + MULTIPLY(z4, FIX(2.205608352)); /* c3+c5+c9-c7 */ - tmp14 = MULTIPLY(z3 + z4, - FIX(0.657217813)); /* -c9 */ - tmp12 += tmp14; - tmp13 += tmp14; - tmp15 = MULTIPLY(tmp15, FIX(0.338443458)); /* c11 */ - tmp14 = tmp15 + MULTIPLY(z1, FIX(0.318774355)) - /* c9-c11 */ - MULTIPLY(z2, FIX(0.466105296)); /* c1-c7 */ - z1 = MULTIPLY(z3 - z2, FIX(0.937797057)); /* c7 */ - tmp14 += z1; - tmp15 += z1 + MULTIPLY(z3, FIX(0.384515595)) - /* c3-c7 */ - MULTIPLY(z4, FIX(1.742345811)); /* c1+c11 */ - - /* Final output stage */ - - wsptr[8*0] = (int) RIGHT_SHIFT(tmp20 + tmp10, CONST_BITS-PASS1_BITS); - wsptr[8*12] = (int) RIGHT_SHIFT(tmp20 - tmp10, CONST_BITS-PASS1_BITS); - wsptr[8*1] = (int) RIGHT_SHIFT(tmp21 + tmp11, CONST_BITS-PASS1_BITS); - wsptr[8*11] = (int) RIGHT_SHIFT(tmp21 - tmp11, CONST_BITS-PASS1_BITS); - wsptr[8*2] = (int) RIGHT_SHIFT(tmp22 + tmp12, CONST_BITS-PASS1_BITS); - wsptr[8*10] = (int) RIGHT_SHIFT(tmp22 - tmp12, CONST_BITS-PASS1_BITS); - wsptr[8*3] = (int) RIGHT_SHIFT(tmp23 + tmp13, CONST_BITS-PASS1_BITS); - wsptr[8*9] = (int) RIGHT_SHIFT(tmp23 - tmp13, CONST_BITS-PASS1_BITS); - wsptr[8*4] = (int) RIGHT_SHIFT(tmp24 + tmp14, CONST_BITS-PASS1_BITS); - wsptr[8*8] = (int) RIGHT_SHIFT(tmp24 - tmp14, CONST_BITS-PASS1_BITS); - wsptr[8*5] = (int) RIGHT_SHIFT(tmp25 + tmp15, CONST_BITS-PASS1_BITS); - wsptr[8*7] = (int) RIGHT_SHIFT(tmp25 - tmp15, CONST_BITS-PASS1_BITS); - wsptr[8*6] = (int) RIGHT_SHIFT(tmp26, CONST_BITS-PASS1_BITS); - } - - /* Pass 2: process 13 rows from work array, store into output array. */ - - wsptr = workspace; - for (ctr = 0; ctr < 13; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - z1 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - z1 <<= CONST_BITS; - - z2 = (INT32) wsptr[2]; - z3 = (INT32) wsptr[4]; - z4 = (INT32) wsptr[6]; - - tmp10 = z3 + z4; - tmp11 = z3 - z4; - - tmp12 = MULTIPLY(tmp10, FIX(1.155388986)); /* (c4+c6)/2 */ - tmp13 = MULTIPLY(tmp11, FIX(0.096834934)) + z1; /* (c4-c6)/2 */ - - tmp20 = MULTIPLY(z2, FIX(1.373119086)) + tmp12 + tmp13; /* c2 */ - tmp22 = MULTIPLY(z2, FIX(0.501487041)) - tmp12 + tmp13; /* c10 */ - - tmp12 = MULTIPLY(tmp10, FIX(0.316450131)); /* (c8-c12)/2 */ - tmp13 = MULTIPLY(tmp11, FIX(0.486914739)) + z1; /* (c8+c12)/2 */ - - tmp21 = MULTIPLY(z2, FIX(1.058554052)) - tmp12 + tmp13; /* c6 */ - tmp25 = MULTIPLY(z2, - FIX(1.252223920)) + tmp12 + tmp13; /* c4 */ - - tmp12 = MULTIPLY(tmp10, FIX(0.435816023)); /* (c2-c10)/2 */ - tmp13 = MULTIPLY(tmp11, FIX(0.937303064)) - z1; /* (c2+c10)/2 */ - - tmp23 = MULTIPLY(z2, - FIX(0.170464608)) - tmp12 - tmp13; /* c12 */ - tmp24 = MULTIPLY(z2, - FIX(0.803364869)) + tmp12 - tmp13; /* c8 */ - - tmp26 = MULTIPLY(tmp11 - z2, FIX(1.414213562)) + z1; /* c0 */ - - /* Odd part */ - - z1 = (INT32) wsptr[1]; - z2 = (INT32) wsptr[3]; - z3 = (INT32) wsptr[5]; - z4 = (INT32) wsptr[7]; - - tmp11 = MULTIPLY(z1 + z2, FIX(1.322312651)); /* c3 */ - tmp12 = MULTIPLY(z1 + z3, FIX(1.163874945)); /* c5 */ - tmp15 = z1 + z4; - tmp13 = MULTIPLY(tmp15, FIX(0.937797057)); /* c7 */ - tmp10 = tmp11 + tmp12 + tmp13 - - MULTIPLY(z1, FIX(2.020082300)); /* c7+c5+c3-c1 */ - tmp14 = MULTIPLY(z2 + z3, - FIX(0.338443458)); /* -c11 */ - tmp11 += tmp14 + MULTIPLY(z2, FIX(0.837223564)); /* c5+c9+c11-c3 */ - tmp12 += tmp14 - MULTIPLY(z3, FIX(1.572116027)); /* c1+c5-c9-c11 */ - tmp14 = MULTIPLY(z2 + z4, - FIX(1.163874945)); /* -c5 */ - tmp11 += tmp14; - tmp13 += tmp14 + MULTIPLY(z4, FIX(2.205608352)); /* c3+c5+c9-c7 */ - tmp14 = MULTIPLY(z3 + z4, - FIX(0.657217813)); /* -c9 */ - tmp12 += tmp14; - tmp13 += tmp14; - tmp15 = MULTIPLY(tmp15, FIX(0.338443458)); /* c11 */ - tmp14 = tmp15 + MULTIPLY(z1, FIX(0.318774355)) - /* c9-c11 */ - MULTIPLY(z2, FIX(0.466105296)); /* c1-c7 */ - z1 = MULTIPLY(z3 - z2, FIX(0.937797057)); /* c7 */ - tmp14 += z1; - tmp15 += z1 + MULTIPLY(z3, FIX(0.384515595)) - /* c3-c7 */ - MULTIPLY(z4, FIX(1.742345811)); /* c1+c11 */ - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[12] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[11] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[10] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23 + tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[9] = range_limit[(int) RIGHT_SHIFT(tmp23 - tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp24 + tmp14, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp24 - tmp14, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp25 + tmp15, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp25 - tmp15, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp26, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += 8; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a 14x14 output block. - * - * Optimized algorithm with 20 multiplications in the 1-D kernel. - * cK represents sqrt(2) * cos(K*pi/28). - */ - -GLOBAL(void) -jpeg_idct_14x14 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15, tmp16; - INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25, tmp26; - INT32 z1, z2, z3, z4; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[8*14]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - z1 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - z1 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - z1 += ONE << (CONST_BITS-PASS1_BITS-1); - z4 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - z2 = MULTIPLY(z4, FIX(1.274162392)); /* c4 */ - z3 = MULTIPLY(z4, FIX(0.314692123)); /* c12 */ - z4 = MULTIPLY(z4, FIX(0.881747734)); /* c8 */ - - tmp10 = z1 + z2; - tmp11 = z1 + z3; - tmp12 = z1 - z4; - - tmp23 = RIGHT_SHIFT(z1 - ((z2 + z3 - z4) << 1), /* c0 = (c4+c12-c8)*2 */ - CONST_BITS-PASS1_BITS); - - z1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - z2 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); - - z3 = MULTIPLY(z1 + z2, FIX(1.105676686)); /* c6 */ - - tmp13 = z3 + MULTIPLY(z1, FIX(0.273079590)); /* c2-c6 */ - tmp14 = z3 - MULTIPLY(z2, FIX(1.719280954)); /* c6+c10 */ - tmp15 = MULTIPLY(z1, FIX(0.613604268)) - /* c10 */ - MULTIPLY(z2, FIX(1.378756276)); /* c2 */ - - tmp20 = tmp10 + tmp13; - tmp26 = tmp10 - tmp13; - tmp21 = tmp11 + tmp14; - tmp25 = tmp11 - tmp14; - tmp22 = tmp12 + tmp15; - tmp24 = tmp12 - tmp15; - - /* Odd part */ - - z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); - z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); - tmp13 = z4 << CONST_BITS; - - tmp14 = z1 + z3; - tmp11 = MULTIPLY(z1 + z2, FIX(1.334852607)); /* c3 */ - tmp12 = MULTIPLY(tmp14, FIX(1.197448846)); /* c5 */ - tmp10 = tmp11 + tmp12 + tmp13 - MULTIPLY(z1, FIX(1.126980169)); /* c3+c5-c1 */ - tmp14 = MULTIPLY(tmp14, FIX(0.752406978)); /* c9 */ - tmp16 = tmp14 - MULTIPLY(z1, FIX(1.061150426)); /* c9+c11-c13 */ - z1 -= z2; - tmp15 = MULTIPLY(z1, FIX(0.467085129)) - tmp13; /* c11 */ - tmp16 += tmp15; - z1 += z4; - z4 = MULTIPLY(z2 + z3, - FIX(0.158341681)) - tmp13; /* -c13 */ - tmp11 += z4 - MULTIPLY(z2, FIX(0.424103948)); /* c3-c9-c13 */ - tmp12 += z4 - MULTIPLY(z3, FIX(2.373959773)); /* c3+c5-c13 */ - z4 = MULTIPLY(z3 - z2, FIX(1.405321284)); /* c1 */ - tmp14 += z4 + tmp13 - MULTIPLY(z3, FIX(1.6906431334)); /* c1+c9-c11 */ - tmp15 += z4 + MULTIPLY(z2, FIX(0.674957567)); /* c1+c11-c5 */ - - tmp13 = (z1 - z3) << PASS1_BITS; - - /* Final output stage */ - - wsptr[8*0] = (int) RIGHT_SHIFT(tmp20 + tmp10, CONST_BITS-PASS1_BITS); - wsptr[8*13] = (int) RIGHT_SHIFT(tmp20 - tmp10, CONST_BITS-PASS1_BITS); - wsptr[8*1] = (int) RIGHT_SHIFT(tmp21 + tmp11, CONST_BITS-PASS1_BITS); - wsptr[8*12] = (int) RIGHT_SHIFT(tmp21 - tmp11, CONST_BITS-PASS1_BITS); - wsptr[8*2] = (int) RIGHT_SHIFT(tmp22 + tmp12, CONST_BITS-PASS1_BITS); - wsptr[8*11] = (int) RIGHT_SHIFT(tmp22 - tmp12, CONST_BITS-PASS1_BITS); - wsptr[8*3] = (int) (tmp23 + tmp13); - wsptr[8*10] = (int) (tmp23 - tmp13); - wsptr[8*4] = (int) RIGHT_SHIFT(tmp24 + tmp14, CONST_BITS-PASS1_BITS); - wsptr[8*9] = (int) RIGHT_SHIFT(tmp24 - tmp14, CONST_BITS-PASS1_BITS); - wsptr[8*5] = (int) RIGHT_SHIFT(tmp25 + tmp15, CONST_BITS-PASS1_BITS); - wsptr[8*8] = (int) RIGHT_SHIFT(tmp25 - tmp15, CONST_BITS-PASS1_BITS); - wsptr[8*6] = (int) RIGHT_SHIFT(tmp26 + tmp16, CONST_BITS-PASS1_BITS); - wsptr[8*7] = (int) RIGHT_SHIFT(tmp26 - tmp16, CONST_BITS-PASS1_BITS); - } - - /* Pass 2: process 14 rows from work array, store into output array. */ - - wsptr = workspace; - for (ctr = 0; ctr < 14; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - z1 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - z1 <<= CONST_BITS; - z4 = (INT32) wsptr[4]; - z2 = MULTIPLY(z4, FIX(1.274162392)); /* c4 */ - z3 = MULTIPLY(z4, FIX(0.314692123)); /* c12 */ - z4 = MULTIPLY(z4, FIX(0.881747734)); /* c8 */ - - tmp10 = z1 + z2; - tmp11 = z1 + z3; - tmp12 = z1 - z4; - - tmp23 = z1 - ((z2 + z3 - z4) << 1); /* c0 = (c4+c12-c8)*2 */ - - z1 = (INT32) wsptr[2]; - z2 = (INT32) wsptr[6]; - - z3 = MULTIPLY(z1 + z2, FIX(1.105676686)); /* c6 */ - - tmp13 = z3 + MULTIPLY(z1, FIX(0.273079590)); /* c2-c6 */ - tmp14 = z3 - MULTIPLY(z2, FIX(1.719280954)); /* c6+c10 */ - tmp15 = MULTIPLY(z1, FIX(0.613604268)) - /* c10 */ - MULTIPLY(z2, FIX(1.378756276)); /* c2 */ - - tmp20 = tmp10 + tmp13; - tmp26 = tmp10 - tmp13; - tmp21 = tmp11 + tmp14; - tmp25 = tmp11 - tmp14; - tmp22 = tmp12 + tmp15; - tmp24 = tmp12 - tmp15; - - /* Odd part */ - - z1 = (INT32) wsptr[1]; - z2 = (INT32) wsptr[3]; - z3 = (INT32) wsptr[5]; - z4 = (INT32) wsptr[7]; - z4 <<= CONST_BITS; - - tmp14 = z1 + z3; - tmp11 = MULTIPLY(z1 + z2, FIX(1.334852607)); /* c3 */ - tmp12 = MULTIPLY(tmp14, FIX(1.197448846)); /* c5 */ - tmp10 = tmp11 + tmp12 + z4 - MULTIPLY(z1, FIX(1.126980169)); /* c3+c5-c1 */ - tmp14 = MULTIPLY(tmp14, FIX(0.752406978)); /* c9 */ - tmp16 = tmp14 - MULTIPLY(z1, FIX(1.061150426)); /* c9+c11-c13 */ - z1 -= z2; - tmp15 = MULTIPLY(z1, FIX(0.467085129)) - z4; /* c11 */ - tmp16 += tmp15; - tmp13 = MULTIPLY(z2 + z3, - FIX(0.158341681)) - z4; /* -c13 */ - tmp11 += tmp13 - MULTIPLY(z2, FIX(0.424103948)); /* c3-c9-c13 */ - tmp12 += tmp13 - MULTIPLY(z3, FIX(2.373959773)); /* c3+c5-c13 */ - tmp13 = MULTIPLY(z3 - z2, FIX(1.405321284)); /* c1 */ - tmp14 += tmp13 + z4 - MULTIPLY(z3, FIX(1.6906431334)); /* c1+c9-c11 */ - tmp15 += tmp13 + MULTIPLY(z2, FIX(0.674957567)); /* c1+c11-c5 */ - - tmp13 = ((z1 - z3) << CONST_BITS) + z4; - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[13] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[12] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[11] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23 + tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[10] = range_limit[(int) RIGHT_SHIFT(tmp23 - tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp24 + tmp14, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[9] = range_limit[(int) RIGHT_SHIFT(tmp24 - tmp14, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp25 + tmp15, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp25 - tmp15, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp26 + tmp16, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp26 - tmp16, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += 8; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a 15x15 output block. - * - * Optimized algorithm with 22 multiplications in the 1-D kernel. - * cK represents sqrt(2) * cos(K*pi/30). - */ - -GLOBAL(void) -jpeg_idct_15x15 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15, tmp16; - INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25, tmp26, tmp27; - INT32 z1, z2, z3, z4; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[8*15]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - z1 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - z1 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - z1 += ONE << (CONST_BITS-PASS1_BITS-1); - - z2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - z3 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - z4 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); - - tmp10 = MULTIPLY(z4, FIX(0.437016024)); /* c12 */ - tmp11 = MULTIPLY(z4, FIX(1.144122806)); /* c6 */ - - tmp12 = z1 - tmp10; - tmp13 = z1 + tmp11; - z1 -= (tmp11 - tmp10) << 1; /* c0 = (c6-c12)*2 */ - - z4 = z2 - z3; - z3 += z2; - tmp10 = MULTIPLY(z3, FIX(1.337628990)); /* (c2+c4)/2 */ - tmp11 = MULTIPLY(z4, FIX(0.045680613)); /* (c2-c4)/2 */ - z2 = MULTIPLY(z2, FIX(1.439773946)); /* c4+c14 */ - - tmp20 = tmp13 + tmp10 + tmp11; - tmp23 = tmp12 - tmp10 + tmp11 + z2; - - tmp10 = MULTIPLY(z3, FIX(0.547059574)); /* (c8+c14)/2 */ - tmp11 = MULTIPLY(z4, FIX(0.399234004)); /* (c8-c14)/2 */ - - tmp25 = tmp13 - tmp10 - tmp11; - tmp26 = tmp12 + tmp10 - tmp11 - z2; - - tmp10 = MULTIPLY(z3, FIX(0.790569415)); /* (c6+c12)/2 */ - tmp11 = MULTIPLY(z4, FIX(0.353553391)); /* (c6-c12)/2 */ - - tmp21 = tmp12 + tmp10 + tmp11; - tmp24 = tmp13 - tmp10 + tmp11; - tmp11 += tmp11; - tmp22 = z1 + tmp11; /* c10 = c6-c12 */ - tmp27 = z1 - tmp11 - tmp11; /* c0 = (c6-c12)*2 */ - - /* Odd part */ - - z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - z4 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); - z3 = MULTIPLY(z4, FIX(1.224744871)); /* c5 */ - z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); - - tmp13 = z2 - z4; - tmp15 = MULTIPLY(z1 + tmp13, FIX(0.831253876)); /* c9 */ - tmp11 = tmp15 + MULTIPLY(z1, FIX(0.513743148)); /* c3-c9 */ - tmp14 = tmp15 - MULTIPLY(tmp13, FIX(2.176250899)); /* c3+c9 */ - - tmp13 = MULTIPLY(z2, - FIX(0.831253876)); /* -c9 */ - tmp15 = MULTIPLY(z2, - FIX(1.344997024)); /* -c3 */ - z2 = z1 - z4; - tmp12 = z3 + MULTIPLY(z2, FIX(1.406466353)); /* c1 */ - - tmp10 = tmp12 + MULTIPLY(z4, FIX(2.457431844)) - tmp15; /* c1+c7 */ - tmp16 = tmp12 - MULTIPLY(z1, FIX(1.112434820)) + tmp13; /* c1-c13 */ - tmp12 = MULTIPLY(z2, FIX(1.224744871)) - z3; /* c5 */ - z2 = MULTIPLY(z1 + z4, FIX(0.575212477)); /* c11 */ - tmp13 += z2 + MULTIPLY(z1, FIX(0.475753014)) - z3; /* c7-c11 */ - tmp15 += z2 - MULTIPLY(z4, FIX(0.869244010)) + z3; /* c11+c13 */ - - /* Final output stage */ - - wsptr[8*0] = (int) RIGHT_SHIFT(tmp20 + tmp10, CONST_BITS-PASS1_BITS); - wsptr[8*14] = (int) RIGHT_SHIFT(tmp20 - tmp10, CONST_BITS-PASS1_BITS); - wsptr[8*1] = (int) RIGHT_SHIFT(tmp21 + tmp11, CONST_BITS-PASS1_BITS); - wsptr[8*13] = (int) RIGHT_SHIFT(tmp21 - tmp11, CONST_BITS-PASS1_BITS); - wsptr[8*2] = (int) RIGHT_SHIFT(tmp22 + tmp12, CONST_BITS-PASS1_BITS); - wsptr[8*12] = (int) RIGHT_SHIFT(tmp22 - tmp12, CONST_BITS-PASS1_BITS); - wsptr[8*3] = (int) RIGHT_SHIFT(tmp23 + tmp13, CONST_BITS-PASS1_BITS); - wsptr[8*11] = (int) RIGHT_SHIFT(tmp23 - tmp13, CONST_BITS-PASS1_BITS); - wsptr[8*4] = (int) RIGHT_SHIFT(tmp24 + tmp14, CONST_BITS-PASS1_BITS); - wsptr[8*10] = (int) RIGHT_SHIFT(tmp24 - tmp14, CONST_BITS-PASS1_BITS); - wsptr[8*5] = (int) RIGHT_SHIFT(tmp25 + tmp15, CONST_BITS-PASS1_BITS); - wsptr[8*9] = (int) RIGHT_SHIFT(tmp25 - tmp15, CONST_BITS-PASS1_BITS); - wsptr[8*6] = (int) RIGHT_SHIFT(tmp26 + tmp16, CONST_BITS-PASS1_BITS); - wsptr[8*8] = (int) RIGHT_SHIFT(tmp26 - tmp16, CONST_BITS-PASS1_BITS); - wsptr[8*7] = (int) RIGHT_SHIFT(tmp27, CONST_BITS-PASS1_BITS); - } - - /* Pass 2: process 15 rows from work array, store into output array. */ - - wsptr = workspace; - for (ctr = 0; ctr < 15; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - z1 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - z1 <<= CONST_BITS; - - z2 = (INT32) wsptr[2]; - z3 = (INT32) wsptr[4]; - z4 = (INT32) wsptr[6]; - - tmp10 = MULTIPLY(z4, FIX(0.437016024)); /* c12 */ - tmp11 = MULTIPLY(z4, FIX(1.144122806)); /* c6 */ - - tmp12 = z1 - tmp10; - tmp13 = z1 + tmp11; - z1 -= (tmp11 - tmp10) << 1; /* c0 = (c6-c12)*2 */ - - z4 = z2 - z3; - z3 += z2; - tmp10 = MULTIPLY(z3, FIX(1.337628990)); /* (c2+c4)/2 */ - tmp11 = MULTIPLY(z4, FIX(0.045680613)); /* (c2-c4)/2 */ - z2 = MULTIPLY(z2, FIX(1.439773946)); /* c4+c14 */ - - tmp20 = tmp13 + tmp10 + tmp11; - tmp23 = tmp12 - tmp10 + tmp11 + z2; - - tmp10 = MULTIPLY(z3, FIX(0.547059574)); /* (c8+c14)/2 */ - tmp11 = MULTIPLY(z4, FIX(0.399234004)); /* (c8-c14)/2 */ - - tmp25 = tmp13 - tmp10 - tmp11; - tmp26 = tmp12 + tmp10 - tmp11 - z2; - - tmp10 = MULTIPLY(z3, FIX(0.790569415)); /* (c6+c12)/2 */ - tmp11 = MULTIPLY(z4, FIX(0.353553391)); /* (c6-c12)/2 */ - - tmp21 = tmp12 + tmp10 + tmp11; - tmp24 = tmp13 - tmp10 + tmp11; - tmp11 += tmp11; - tmp22 = z1 + tmp11; /* c10 = c6-c12 */ - tmp27 = z1 - tmp11 - tmp11; /* c0 = (c6-c12)*2 */ - - /* Odd part */ - - z1 = (INT32) wsptr[1]; - z2 = (INT32) wsptr[3]; - z4 = (INT32) wsptr[5]; - z3 = MULTIPLY(z4, FIX(1.224744871)); /* c5 */ - z4 = (INT32) wsptr[7]; - - tmp13 = z2 - z4; - tmp15 = MULTIPLY(z1 + tmp13, FIX(0.831253876)); /* c9 */ - tmp11 = tmp15 + MULTIPLY(z1, FIX(0.513743148)); /* c3-c9 */ - tmp14 = tmp15 - MULTIPLY(tmp13, FIX(2.176250899)); /* c3+c9 */ - - tmp13 = MULTIPLY(z2, - FIX(0.831253876)); /* -c9 */ - tmp15 = MULTIPLY(z2, - FIX(1.344997024)); /* -c3 */ - z2 = z1 - z4; - tmp12 = z3 + MULTIPLY(z2, FIX(1.406466353)); /* c1 */ - - tmp10 = tmp12 + MULTIPLY(z4, FIX(2.457431844)) - tmp15; /* c1+c7 */ - tmp16 = tmp12 - MULTIPLY(z1, FIX(1.112434820)) + tmp13; /* c1-c13 */ - tmp12 = MULTIPLY(z2, FIX(1.224744871)) - z3; /* c5 */ - z2 = MULTIPLY(z1 + z4, FIX(0.575212477)); /* c11 */ - tmp13 += z2 + MULTIPLY(z1, FIX(0.475753014)) - z3; /* c7-c11 */ - tmp15 += z2 - MULTIPLY(z4, FIX(0.869244010)) + z3; /* c11+c13 */ - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[14] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[13] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[12] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23 + tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[11] = range_limit[(int) RIGHT_SHIFT(tmp23 - tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp24 + tmp14, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[10] = range_limit[(int) RIGHT_SHIFT(tmp24 - tmp14, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp25 + tmp15, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[9] = range_limit[(int) RIGHT_SHIFT(tmp25 - tmp15, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp26 + tmp16, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp26 - tmp16, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp27, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += 8; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a 16x16 output block. - * - * Optimized algorithm with 28 multiplications in the 1-D kernel. - * cK represents sqrt(2) * cos(K*pi/32). - */ - -GLOBAL(void) -jpeg_idct_16x16 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3, tmp10, tmp11, tmp12, tmp13; - INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25, tmp26, tmp27; - INT32 z1, z2, z3, z4; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[8*16]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - tmp0 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - tmp0 += ONE << (CONST_BITS-PASS1_BITS-1); - - z1 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - tmp1 = MULTIPLY(z1, FIX(1.306562965)); /* c4[16] = c2[8] */ - tmp2 = MULTIPLY(z1, FIX_0_541196100); /* c12[16] = c6[8] */ - - tmp10 = tmp0 + tmp1; - tmp11 = tmp0 - tmp1; - tmp12 = tmp0 + tmp2; - tmp13 = tmp0 - tmp2; - - z1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - z2 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); - z3 = z1 - z2; - z4 = MULTIPLY(z3, FIX(0.275899379)); /* c14[16] = c7[8] */ - z3 = MULTIPLY(z3, FIX(1.387039845)); /* c2[16] = c1[8] */ - - tmp0 = z3 + MULTIPLY(z2, FIX_2_562915447); /* (c6+c2)[16] = (c3+c1)[8] */ - tmp1 = z4 + MULTIPLY(z1, FIX_0_899976223); /* (c6-c14)[16] = (c3-c7)[8] */ - tmp2 = z3 - MULTIPLY(z1, FIX(0.601344887)); /* (c2-c10)[16] = (c1-c5)[8] */ - tmp3 = z4 - MULTIPLY(z2, FIX(0.509795579)); /* (c10-c14)[16] = (c5-c7)[8] */ - - tmp20 = tmp10 + tmp0; - tmp27 = tmp10 - tmp0; - tmp21 = tmp12 + tmp1; - tmp26 = tmp12 - tmp1; - tmp22 = tmp13 + tmp2; - tmp25 = tmp13 - tmp2; - tmp23 = tmp11 + tmp3; - tmp24 = tmp11 - tmp3; - - /* Odd part */ - - z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); - z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); - - tmp11 = z1 + z3; - - tmp1 = MULTIPLY(z1 + z2, FIX(1.353318001)); /* c3 */ - tmp2 = MULTIPLY(tmp11, FIX(1.247225013)); /* c5 */ - tmp3 = MULTIPLY(z1 + z4, FIX(1.093201867)); /* c7 */ - tmp10 = MULTIPLY(z1 - z4, FIX(0.897167586)); /* c9 */ - tmp11 = MULTIPLY(tmp11, FIX(0.666655658)); /* c11 */ - tmp12 = MULTIPLY(z1 - z2, FIX(0.410524528)); /* c13 */ - tmp0 = tmp1 + tmp2 + tmp3 - - MULTIPLY(z1, FIX(2.286341144)); /* c7+c5+c3-c1 */ - tmp13 = tmp10 + tmp11 + tmp12 - - MULTIPLY(z1, FIX(1.835730603)); /* c9+c11+c13-c15 */ - z1 = MULTIPLY(z2 + z3, FIX(0.138617169)); /* c15 */ - tmp1 += z1 + MULTIPLY(z2, FIX(0.071888074)); /* c9+c11-c3-c15 */ - tmp2 += z1 - MULTIPLY(z3, FIX(1.125726048)); /* c5+c7+c15-c3 */ - z1 = MULTIPLY(z3 - z2, FIX(1.407403738)); /* c1 */ - tmp11 += z1 - MULTIPLY(z3, FIX(0.766367282)); /* c1+c11-c9-c13 */ - tmp12 += z1 + MULTIPLY(z2, FIX(1.971951411)); /* c1+c5+c13-c7 */ - z2 += z4; - z1 = MULTIPLY(z2, - FIX(0.666655658)); /* -c11 */ - tmp1 += z1; - tmp3 += z1 + MULTIPLY(z4, FIX(1.065388962)); /* c3+c11+c15-c7 */ - z2 = MULTIPLY(z2, - FIX(1.247225013)); /* -c5 */ - tmp10 += z2 + MULTIPLY(z4, FIX(3.141271809)); /* c1+c5+c9-c13 */ - tmp12 += z2; - z2 = MULTIPLY(z3 + z4, - FIX(1.353318001)); /* -c3 */ - tmp2 += z2; - tmp3 += z2; - z2 = MULTIPLY(z4 - z3, FIX(0.410524528)); /* c13 */ - tmp10 += z2; - tmp11 += z2; - - /* Final output stage */ - - wsptr[8*0] = (int) RIGHT_SHIFT(tmp20 + tmp0, CONST_BITS-PASS1_BITS); - wsptr[8*15] = (int) RIGHT_SHIFT(tmp20 - tmp0, CONST_BITS-PASS1_BITS); - wsptr[8*1] = (int) RIGHT_SHIFT(tmp21 + tmp1, CONST_BITS-PASS1_BITS); - wsptr[8*14] = (int) RIGHT_SHIFT(tmp21 - tmp1, CONST_BITS-PASS1_BITS); - wsptr[8*2] = (int) RIGHT_SHIFT(tmp22 + tmp2, CONST_BITS-PASS1_BITS); - wsptr[8*13] = (int) RIGHT_SHIFT(tmp22 - tmp2, CONST_BITS-PASS1_BITS); - wsptr[8*3] = (int) RIGHT_SHIFT(tmp23 + tmp3, CONST_BITS-PASS1_BITS); - wsptr[8*12] = (int) RIGHT_SHIFT(tmp23 - tmp3, CONST_BITS-PASS1_BITS); - wsptr[8*4] = (int) RIGHT_SHIFT(tmp24 + tmp10, CONST_BITS-PASS1_BITS); - wsptr[8*11] = (int) RIGHT_SHIFT(tmp24 - tmp10, CONST_BITS-PASS1_BITS); - wsptr[8*5] = (int) RIGHT_SHIFT(tmp25 + tmp11, CONST_BITS-PASS1_BITS); - wsptr[8*10] = (int) RIGHT_SHIFT(tmp25 - tmp11, CONST_BITS-PASS1_BITS); - wsptr[8*6] = (int) RIGHT_SHIFT(tmp26 + tmp12, CONST_BITS-PASS1_BITS); - wsptr[8*9] = (int) RIGHT_SHIFT(tmp26 - tmp12, CONST_BITS-PASS1_BITS); - wsptr[8*7] = (int) RIGHT_SHIFT(tmp27 + tmp13, CONST_BITS-PASS1_BITS); - wsptr[8*8] = (int) RIGHT_SHIFT(tmp27 - tmp13, CONST_BITS-PASS1_BITS); - } - - /* Pass 2: process 16 rows from work array, store into output array. */ - - wsptr = workspace; - for (ctr = 0; ctr < 16; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - tmp0 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - tmp0 <<= CONST_BITS; - - z1 = (INT32) wsptr[4]; - tmp1 = MULTIPLY(z1, FIX(1.306562965)); /* c4[16] = c2[8] */ - tmp2 = MULTIPLY(z1, FIX_0_541196100); /* c12[16] = c6[8] */ - - tmp10 = tmp0 + tmp1; - tmp11 = tmp0 - tmp1; - tmp12 = tmp0 + tmp2; - tmp13 = tmp0 - tmp2; - - z1 = (INT32) wsptr[2]; - z2 = (INT32) wsptr[6]; - z3 = z1 - z2; - z4 = MULTIPLY(z3, FIX(0.275899379)); /* c14[16] = c7[8] */ - z3 = MULTIPLY(z3, FIX(1.387039845)); /* c2[16] = c1[8] */ - - tmp0 = z3 + MULTIPLY(z2, FIX_2_562915447); /* (c6+c2)[16] = (c3+c1)[8] */ - tmp1 = z4 + MULTIPLY(z1, FIX_0_899976223); /* (c6-c14)[16] = (c3-c7)[8] */ - tmp2 = z3 - MULTIPLY(z1, FIX(0.601344887)); /* (c2-c10)[16] = (c1-c5)[8] */ - tmp3 = z4 - MULTIPLY(z2, FIX(0.509795579)); /* (c10-c14)[16] = (c5-c7)[8] */ - - tmp20 = tmp10 + tmp0; - tmp27 = tmp10 - tmp0; - tmp21 = tmp12 + tmp1; - tmp26 = tmp12 - tmp1; - tmp22 = tmp13 + tmp2; - tmp25 = tmp13 - tmp2; - tmp23 = tmp11 + tmp3; - tmp24 = tmp11 - tmp3; - - /* Odd part */ - - z1 = (INT32) wsptr[1]; - z2 = (INT32) wsptr[3]; - z3 = (INT32) wsptr[5]; - z4 = (INT32) wsptr[7]; - - tmp11 = z1 + z3; - - tmp1 = MULTIPLY(z1 + z2, FIX(1.353318001)); /* c3 */ - tmp2 = MULTIPLY(tmp11, FIX(1.247225013)); /* c5 */ - tmp3 = MULTIPLY(z1 + z4, FIX(1.093201867)); /* c7 */ - tmp10 = MULTIPLY(z1 - z4, FIX(0.897167586)); /* c9 */ - tmp11 = MULTIPLY(tmp11, FIX(0.666655658)); /* c11 */ - tmp12 = MULTIPLY(z1 - z2, FIX(0.410524528)); /* c13 */ - tmp0 = tmp1 + tmp2 + tmp3 - - MULTIPLY(z1, FIX(2.286341144)); /* c7+c5+c3-c1 */ - tmp13 = tmp10 + tmp11 + tmp12 - - MULTIPLY(z1, FIX(1.835730603)); /* c9+c11+c13-c15 */ - z1 = MULTIPLY(z2 + z3, FIX(0.138617169)); /* c15 */ - tmp1 += z1 + MULTIPLY(z2, FIX(0.071888074)); /* c9+c11-c3-c15 */ - tmp2 += z1 - MULTIPLY(z3, FIX(1.125726048)); /* c5+c7+c15-c3 */ - z1 = MULTIPLY(z3 - z2, FIX(1.407403738)); /* c1 */ - tmp11 += z1 - MULTIPLY(z3, FIX(0.766367282)); /* c1+c11-c9-c13 */ - tmp12 += z1 + MULTIPLY(z2, FIX(1.971951411)); /* c1+c5+c13-c7 */ - z2 += z4; - z1 = MULTIPLY(z2, - FIX(0.666655658)); /* -c11 */ - tmp1 += z1; - tmp3 += z1 + MULTIPLY(z4, FIX(1.065388962)); /* c3+c11+c15-c7 */ - z2 = MULTIPLY(z2, - FIX(1.247225013)); /* -c5 */ - tmp10 += z2 + MULTIPLY(z4, FIX(3.141271809)); /* c1+c5+c9-c13 */ - tmp12 += z2; - z2 = MULTIPLY(z3 + z4, - FIX(1.353318001)); /* -c3 */ - tmp2 += z2; - tmp3 += z2; - z2 = MULTIPLY(z4 - z3, FIX(0.410524528)); /* c13 */ - tmp10 += z2; - tmp11 += z2; - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[15] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp1, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[14] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp1, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[13] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23 + tmp3, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[12] = range_limit[(int) RIGHT_SHIFT(tmp23 - tmp3, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp24 + tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[11] = range_limit[(int) RIGHT_SHIFT(tmp24 - tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp25 + tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[10] = range_limit[(int) RIGHT_SHIFT(tmp25 - tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp26 + tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[9] = range_limit[(int) RIGHT_SHIFT(tmp26 - tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp27 + tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp27 - tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += 8; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a 16x8 output block. - * - * 8-point IDCT in pass 1 (columns), 16-point in pass 2 (rows). - */ - -GLOBAL(void) -jpeg_idct_16x8 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3, tmp10, tmp11, tmp12, tmp13; - INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25, tmp26, tmp27; - INT32 z1, z2, z3, z4; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[8*8]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. - * Note results are scaled up by sqrt(8) compared to a true IDCT; - * furthermore, we scale the results by 2**PASS1_BITS. - * 8-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/16). - */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = DCTSIZE; ctr > 0; ctr--) { - /* Due to quantization, we will usually find that many of the input - * coefficients are zero, especially the AC terms. We can exploit this - * by short-circuiting the IDCT calculation for any column in which all - * the AC terms are zero. In that case each output is equal to the - * DC coefficient (with scale factor as needed). - * With typical images and quantization tables, half or more of the - * column DCT calculations can be simplified this way. - */ - - if (inptr[DCTSIZE*1] == 0 && inptr[DCTSIZE*2] == 0 && - inptr[DCTSIZE*3] == 0 && inptr[DCTSIZE*4] == 0 && - inptr[DCTSIZE*5] == 0 && inptr[DCTSIZE*6] == 0 && - inptr[DCTSIZE*7] == 0) { - /* AC terms all zero */ - int dcval = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]) << PASS1_BITS; - - wsptr[DCTSIZE*0] = dcval; - wsptr[DCTSIZE*1] = dcval; - wsptr[DCTSIZE*2] = dcval; - wsptr[DCTSIZE*3] = dcval; - wsptr[DCTSIZE*4] = dcval; - wsptr[DCTSIZE*5] = dcval; - wsptr[DCTSIZE*6] = dcval; - wsptr[DCTSIZE*7] = dcval; - - inptr++; /* advance pointers to next column */ - quantptr++; - wsptr++; - continue; - } - - /* Even part: reverse the even part of the forward DCT. - * The rotator is c(-6). - */ - - z2 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - z3 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - z2 <<= CONST_BITS; - z3 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - z2 += ONE << (CONST_BITS-PASS1_BITS-1); - - tmp0 = z2 + z3; - tmp1 = z2 - z3; - - z2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - z3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); - - z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ - tmp2 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ - tmp3 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ - - tmp10 = tmp0 + tmp2; - tmp13 = tmp0 - tmp2; - tmp11 = tmp1 + tmp3; - tmp12 = tmp1 - tmp3; - - /* Odd part per figure 8; the matrix is unitary and hence its - * transpose is its inverse. i0..i3 are y7,y5,y3,y1 respectively. - */ - - tmp0 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); - tmp1 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); - tmp2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - tmp3 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - - z2 = tmp0 + tmp2; - z3 = tmp1 + tmp3; - - z1 = MULTIPLY(z2 + z3, FIX_1_175875602); /* c3 */ - z2 = MULTIPLY(z2, - FIX_1_961570560); /* -c3-c5 */ - z3 = MULTIPLY(z3, - FIX_0_390180644); /* -c3+c5 */ - z2 += z1; - z3 += z1; - - z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ - tmp0 = MULTIPLY(tmp0, FIX_0_298631336); /* -c1+c3+c5-c7 */ - tmp3 = MULTIPLY(tmp3, FIX_1_501321110); /* c1+c3-c5-c7 */ - tmp0 += z1 + z2; - tmp3 += z1 + z3; - - z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ - tmp1 = MULTIPLY(tmp1, FIX_2_053119869); /* c1+c3-c5+c7 */ - tmp2 = MULTIPLY(tmp2, FIX_3_072711026); /* c1+c3+c5-c7 */ - tmp1 += z1 + z3; - tmp2 += z1 + z2; - - /* Final output stage: inputs are tmp10..tmp13, tmp0..tmp3 */ - - wsptr[DCTSIZE*0] = (int) RIGHT_SHIFT(tmp10 + tmp3, CONST_BITS-PASS1_BITS); - wsptr[DCTSIZE*7] = (int) RIGHT_SHIFT(tmp10 - tmp3, CONST_BITS-PASS1_BITS); - wsptr[DCTSIZE*1] = (int) RIGHT_SHIFT(tmp11 + tmp2, CONST_BITS-PASS1_BITS); - wsptr[DCTSIZE*6] = (int) RIGHT_SHIFT(tmp11 - tmp2, CONST_BITS-PASS1_BITS); - wsptr[DCTSIZE*2] = (int) RIGHT_SHIFT(tmp12 + tmp1, CONST_BITS-PASS1_BITS); - wsptr[DCTSIZE*5] = (int) RIGHT_SHIFT(tmp12 - tmp1, CONST_BITS-PASS1_BITS); - wsptr[DCTSIZE*3] = (int) RIGHT_SHIFT(tmp13 + tmp0, CONST_BITS-PASS1_BITS); - wsptr[DCTSIZE*4] = (int) RIGHT_SHIFT(tmp13 - tmp0, CONST_BITS-PASS1_BITS); - - inptr++; /* advance pointers to next column */ - quantptr++; - wsptr++; - } - - /* Pass 2: process 8 rows from work array, store into output array. - * 16-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/32). - */ - - wsptr = workspace; - for (ctr = 0; ctr < 8; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - tmp0 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - tmp0 <<= CONST_BITS; - - z1 = (INT32) wsptr[4]; - tmp1 = MULTIPLY(z1, FIX(1.306562965)); /* c4[16] = c2[8] */ - tmp2 = MULTIPLY(z1, FIX_0_541196100); /* c12[16] = c6[8] */ - - tmp10 = tmp0 + tmp1; - tmp11 = tmp0 - tmp1; - tmp12 = tmp0 + tmp2; - tmp13 = tmp0 - tmp2; - - z1 = (INT32) wsptr[2]; - z2 = (INT32) wsptr[6]; - z3 = z1 - z2; - z4 = MULTIPLY(z3, FIX(0.275899379)); /* c14[16] = c7[8] */ - z3 = MULTIPLY(z3, FIX(1.387039845)); /* c2[16] = c1[8] */ - - tmp0 = z3 + MULTIPLY(z2, FIX_2_562915447); /* (c6+c2)[16] = (c3+c1)[8] */ - tmp1 = z4 + MULTIPLY(z1, FIX_0_899976223); /* (c6-c14)[16] = (c3-c7)[8] */ - tmp2 = z3 - MULTIPLY(z1, FIX(0.601344887)); /* (c2-c10)[16] = (c1-c5)[8] */ - tmp3 = z4 - MULTIPLY(z2, FIX(0.509795579)); /* (c10-c14)[16] = (c5-c7)[8] */ - - tmp20 = tmp10 + tmp0; - tmp27 = tmp10 - tmp0; - tmp21 = tmp12 + tmp1; - tmp26 = tmp12 - tmp1; - tmp22 = tmp13 + tmp2; - tmp25 = tmp13 - tmp2; - tmp23 = tmp11 + tmp3; - tmp24 = tmp11 - tmp3; - - /* Odd part */ - - z1 = (INT32) wsptr[1]; - z2 = (INT32) wsptr[3]; - z3 = (INT32) wsptr[5]; - z4 = (INT32) wsptr[7]; - - tmp11 = z1 + z3; - - tmp1 = MULTIPLY(z1 + z2, FIX(1.353318001)); /* c3 */ - tmp2 = MULTIPLY(tmp11, FIX(1.247225013)); /* c5 */ - tmp3 = MULTIPLY(z1 + z4, FIX(1.093201867)); /* c7 */ - tmp10 = MULTIPLY(z1 - z4, FIX(0.897167586)); /* c9 */ - tmp11 = MULTIPLY(tmp11, FIX(0.666655658)); /* c11 */ - tmp12 = MULTIPLY(z1 - z2, FIX(0.410524528)); /* c13 */ - tmp0 = tmp1 + tmp2 + tmp3 - - MULTIPLY(z1, FIX(2.286341144)); /* c7+c5+c3-c1 */ - tmp13 = tmp10 + tmp11 + tmp12 - - MULTIPLY(z1, FIX(1.835730603)); /* c9+c11+c13-c15 */ - z1 = MULTIPLY(z2 + z3, FIX(0.138617169)); /* c15 */ - tmp1 += z1 + MULTIPLY(z2, FIX(0.071888074)); /* c9+c11-c3-c15 */ - tmp2 += z1 - MULTIPLY(z3, FIX(1.125726048)); /* c5+c7+c15-c3 */ - z1 = MULTIPLY(z3 - z2, FIX(1.407403738)); /* c1 */ - tmp11 += z1 - MULTIPLY(z3, FIX(0.766367282)); /* c1+c11-c9-c13 */ - tmp12 += z1 + MULTIPLY(z2, FIX(1.971951411)); /* c1+c5+c13-c7 */ - z2 += z4; - z1 = MULTIPLY(z2, - FIX(0.666655658)); /* -c11 */ - tmp1 += z1; - tmp3 += z1 + MULTIPLY(z4, FIX(1.065388962)); /* c3+c11+c15-c7 */ - z2 = MULTIPLY(z2, - FIX(1.247225013)); /* -c5 */ - tmp10 += z2 + MULTIPLY(z4, FIX(3.141271809)); /* c1+c5+c9-c13 */ - tmp12 += z2; - z2 = MULTIPLY(z3 + z4, - FIX(1.353318001)); /* -c3 */ - tmp2 += z2; - tmp3 += z2; - z2 = MULTIPLY(z4 - z3, FIX(0.410524528)); /* c13 */ - tmp10 += z2; - tmp11 += z2; - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[15] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp1, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[14] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp1, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[13] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23 + tmp3, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[12] = range_limit[(int) RIGHT_SHIFT(tmp23 - tmp3, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp24 + tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[11] = range_limit[(int) RIGHT_SHIFT(tmp24 - tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp25 + tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[10] = range_limit[(int) RIGHT_SHIFT(tmp25 - tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp26 + tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[9] = range_limit[(int) RIGHT_SHIFT(tmp26 - tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp27 + tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp27 - tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += 8; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a 14x7 output block. - * - * 7-point IDCT in pass 1 (columns), 14-point in pass 2 (rows). - */ - -GLOBAL(void) -jpeg_idct_14x7 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15, tmp16; - INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25, tmp26; - INT32 z1, z2, z3, z4; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[8*7]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. - * 7-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/14). - */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - tmp23 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - tmp23 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - tmp23 += ONE << (CONST_BITS-PASS1_BITS-1); - - z1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - z2 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - z3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); - - tmp20 = MULTIPLY(z2 - z3, FIX(0.881747734)); /* c4 */ - tmp22 = MULTIPLY(z1 - z2, FIX(0.314692123)); /* c6 */ - tmp21 = tmp20 + tmp22 + tmp23 - MULTIPLY(z2, FIX(1.841218003)); /* c2+c4-c6 */ - tmp10 = z1 + z3; - z2 -= tmp10; - tmp10 = MULTIPLY(tmp10, FIX(1.274162392)) + tmp23; /* c2 */ - tmp20 += tmp10 - MULTIPLY(z3, FIX(0.077722536)); /* c2-c4-c6 */ - tmp22 += tmp10 - MULTIPLY(z1, FIX(2.470602249)); /* c2+c4+c6 */ - tmp23 += MULTIPLY(z2, FIX(1.414213562)); /* c0 */ - - /* Odd part */ - - z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); - - tmp11 = MULTIPLY(z1 + z2, FIX(0.935414347)); /* (c3+c1-c5)/2 */ - tmp12 = MULTIPLY(z1 - z2, FIX(0.170262339)); /* (c3+c5-c1)/2 */ - tmp10 = tmp11 - tmp12; - tmp11 += tmp12; - tmp12 = MULTIPLY(z2 + z3, - FIX(1.378756276)); /* -c1 */ - tmp11 += tmp12; - z2 = MULTIPLY(z1 + z3, FIX(0.613604268)); /* c5 */ - tmp10 += z2; - tmp12 += z2 + MULTIPLY(z3, FIX(1.870828693)); /* c3+c1-c5 */ - - /* Final output stage */ - - wsptr[8*0] = (int) RIGHT_SHIFT(tmp20 + tmp10, CONST_BITS-PASS1_BITS); - wsptr[8*6] = (int) RIGHT_SHIFT(tmp20 - tmp10, CONST_BITS-PASS1_BITS); - wsptr[8*1] = (int) RIGHT_SHIFT(tmp21 + tmp11, CONST_BITS-PASS1_BITS); - wsptr[8*5] = (int) RIGHT_SHIFT(tmp21 - tmp11, CONST_BITS-PASS1_BITS); - wsptr[8*2] = (int) RIGHT_SHIFT(tmp22 + tmp12, CONST_BITS-PASS1_BITS); - wsptr[8*4] = (int) RIGHT_SHIFT(tmp22 - tmp12, CONST_BITS-PASS1_BITS); - wsptr[8*3] = (int) RIGHT_SHIFT(tmp23, CONST_BITS-PASS1_BITS); - } - - /* Pass 2: process 7 rows from work array, store into output array. - * 14-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/28). - */ - - wsptr = workspace; - for (ctr = 0; ctr < 7; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - z1 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - z1 <<= CONST_BITS; - z4 = (INT32) wsptr[4]; - z2 = MULTIPLY(z4, FIX(1.274162392)); /* c4 */ - z3 = MULTIPLY(z4, FIX(0.314692123)); /* c12 */ - z4 = MULTIPLY(z4, FIX(0.881747734)); /* c8 */ - - tmp10 = z1 + z2; - tmp11 = z1 + z3; - tmp12 = z1 - z4; - - tmp23 = z1 - ((z2 + z3 - z4) << 1); /* c0 = (c4+c12-c8)*2 */ - - z1 = (INT32) wsptr[2]; - z2 = (INT32) wsptr[6]; - - z3 = MULTIPLY(z1 + z2, FIX(1.105676686)); /* c6 */ - - tmp13 = z3 + MULTIPLY(z1, FIX(0.273079590)); /* c2-c6 */ - tmp14 = z3 - MULTIPLY(z2, FIX(1.719280954)); /* c6+c10 */ - tmp15 = MULTIPLY(z1, FIX(0.613604268)) - /* c10 */ - MULTIPLY(z2, FIX(1.378756276)); /* c2 */ - - tmp20 = tmp10 + tmp13; - tmp26 = tmp10 - tmp13; - tmp21 = tmp11 + tmp14; - tmp25 = tmp11 - tmp14; - tmp22 = tmp12 + tmp15; - tmp24 = tmp12 - tmp15; - - /* Odd part */ - - z1 = (INT32) wsptr[1]; - z2 = (INT32) wsptr[3]; - z3 = (INT32) wsptr[5]; - z4 = (INT32) wsptr[7]; - z4 <<= CONST_BITS; - - tmp14 = z1 + z3; - tmp11 = MULTIPLY(z1 + z2, FIX(1.334852607)); /* c3 */ - tmp12 = MULTIPLY(tmp14, FIX(1.197448846)); /* c5 */ - tmp10 = tmp11 + tmp12 + z4 - MULTIPLY(z1, FIX(1.126980169)); /* c3+c5-c1 */ - tmp14 = MULTIPLY(tmp14, FIX(0.752406978)); /* c9 */ - tmp16 = tmp14 - MULTIPLY(z1, FIX(1.061150426)); /* c9+c11-c13 */ - z1 -= z2; - tmp15 = MULTIPLY(z1, FIX(0.467085129)) - z4; /* c11 */ - tmp16 += tmp15; - tmp13 = MULTIPLY(z2 + z3, - FIX(0.158341681)) - z4; /* -c13 */ - tmp11 += tmp13 - MULTIPLY(z2, FIX(0.424103948)); /* c3-c9-c13 */ - tmp12 += tmp13 - MULTIPLY(z3, FIX(2.373959773)); /* c3+c5-c13 */ - tmp13 = MULTIPLY(z3 - z2, FIX(1.405321284)); /* c1 */ - tmp14 += tmp13 + z4 - MULTIPLY(z3, FIX(1.6906431334)); /* c1+c9-c11 */ - tmp15 += tmp13 + MULTIPLY(z2, FIX(0.674957567)); /* c1+c11-c5 */ - - tmp13 = ((z1 - z3) << CONST_BITS) + z4; - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[13] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[12] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[11] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23 + tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[10] = range_limit[(int) RIGHT_SHIFT(tmp23 - tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp24 + tmp14, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[9] = range_limit[(int) RIGHT_SHIFT(tmp24 - tmp14, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp25 + tmp15, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp25 - tmp15, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp26 + tmp16, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp26 - tmp16, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += 8; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a 12x6 output block. - * - * 6-point IDCT in pass 1 (columns), 12-point in pass 2 (rows). - */ - -GLOBAL(void) -jpeg_idct_12x6 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15; - INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25; - INT32 z1, z2, z3, z4; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[8*6]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. - * 6-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/12). - */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - tmp10 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - tmp10 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - tmp10 += ONE << (CONST_BITS-PASS1_BITS-1); - tmp12 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - tmp20 = MULTIPLY(tmp12, FIX(0.707106781)); /* c4 */ - tmp11 = tmp10 + tmp20; - tmp21 = RIGHT_SHIFT(tmp10 - tmp20 - tmp20, CONST_BITS-PASS1_BITS); - tmp20 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - tmp10 = MULTIPLY(tmp20, FIX(1.224744871)); /* c2 */ - tmp20 = tmp11 + tmp10; - tmp22 = tmp11 - tmp10; - - /* Odd part */ - - z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); - tmp11 = MULTIPLY(z1 + z3, FIX(0.366025404)); /* c5 */ - tmp10 = tmp11 + ((z1 + z2) << CONST_BITS); - tmp12 = tmp11 + ((z3 - z2) << CONST_BITS); - tmp11 = (z1 - z2 - z3) << PASS1_BITS; - - /* Final output stage */ - - wsptr[8*0] = (int) RIGHT_SHIFT(tmp20 + tmp10, CONST_BITS-PASS1_BITS); - wsptr[8*5] = (int) RIGHT_SHIFT(tmp20 - tmp10, CONST_BITS-PASS1_BITS); - wsptr[8*1] = (int) (tmp21 + tmp11); - wsptr[8*4] = (int) (tmp21 - tmp11); - wsptr[8*2] = (int) RIGHT_SHIFT(tmp22 + tmp12, CONST_BITS-PASS1_BITS); - wsptr[8*3] = (int) RIGHT_SHIFT(tmp22 - tmp12, CONST_BITS-PASS1_BITS); - } - - /* Pass 2: process 6 rows from work array, store into output array. - * 12-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/24). - */ - - wsptr = workspace; - for (ctr = 0; ctr < 6; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - z3 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - z3 <<= CONST_BITS; - - z4 = (INT32) wsptr[4]; - z4 = MULTIPLY(z4, FIX(1.224744871)); /* c4 */ - - tmp10 = z3 + z4; - tmp11 = z3 - z4; - - z1 = (INT32) wsptr[2]; - z4 = MULTIPLY(z1, FIX(1.366025404)); /* c2 */ - z1 <<= CONST_BITS; - z2 = (INT32) wsptr[6]; - z2 <<= CONST_BITS; - - tmp12 = z1 - z2; - - tmp21 = z3 + tmp12; - tmp24 = z3 - tmp12; - - tmp12 = z4 + z2; - - tmp20 = tmp10 + tmp12; - tmp25 = tmp10 - tmp12; - - tmp12 = z4 - z1 - z2; - - tmp22 = tmp11 + tmp12; - tmp23 = tmp11 - tmp12; - - /* Odd part */ - - z1 = (INT32) wsptr[1]; - z2 = (INT32) wsptr[3]; - z3 = (INT32) wsptr[5]; - z4 = (INT32) wsptr[7]; - - tmp11 = MULTIPLY(z2, FIX(1.306562965)); /* c3 */ - tmp14 = MULTIPLY(z2, - FIX_0_541196100); /* -c9 */ - - tmp10 = z1 + z3; - tmp15 = MULTIPLY(tmp10 + z4, FIX(0.860918669)); /* c7 */ - tmp12 = tmp15 + MULTIPLY(tmp10, FIX(0.261052384)); /* c5-c7 */ - tmp10 = tmp12 + tmp11 + MULTIPLY(z1, FIX(0.280143716)); /* c1-c5 */ - tmp13 = MULTIPLY(z3 + z4, - FIX(1.045510580)); /* -(c7+c11) */ - tmp12 += tmp13 + tmp14 - MULTIPLY(z3, FIX(1.478575242)); /* c1+c5-c7-c11 */ - tmp13 += tmp15 - tmp11 + MULTIPLY(z4, FIX(1.586706681)); /* c1+c11 */ - tmp15 += tmp14 - MULTIPLY(z1, FIX(0.676326758)) - /* c7-c11 */ - MULTIPLY(z4, FIX(1.982889723)); /* c5+c7 */ - - z1 -= z4; - z2 -= z3; - z3 = MULTIPLY(z1 + z2, FIX_0_541196100); /* c9 */ - tmp11 = z3 + MULTIPLY(z1, FIX_0_765366865); /* c3-c9 */ - tmp14 = z3 - MULTIPLY(z2, FIX_1_847759065); /* c3+c9 */ - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[11] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[10] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[9] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23 + tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp23 - tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp24 + tmp14, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp24 - tmp14, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp25 + tmp15, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp25 - tmp15, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += 8; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a 10x5 output block. - * - * 5-point IDCT in pass 1 (columns), 10-point in pass 2 (rows). - */ - -GLOBAL(void) -jpeg_idct_10x5 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp10, tmp11, tmp12, tmp13, tmp14; - INT32 tmp20, tmp21, tmp22, tmp23, tmp24; - INT32 z1, z2, z3, z4; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[8*5]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. - * 5-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/10). - */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - tmp12 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - tmp12 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - tmp12 += ONE << (CONST_BITS-PASS1_BITS-1); - tmp13 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - tmp14 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - z1 = MULTIPLY(tmp13 + tmp14, FIX(0.790569415)); /* (c2+c4)/2 */ - z2 = MULTIPLY(tmp13 - tmp14, FIX(0.353553391)); /* (c2-c4)/2 */ - z3 = tmp12 + z2; - tmp10 = z3 + z1; - tmp11 = z3 - z1; - tmp12 -= z2 << 2; - - /* Odd part */ - - z2 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - z3 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - - z1 = MULTIPLY(z2 + z3, FIX(0.831253876)); /* c3 */ - tmp13 = z1 + MULTIPLY(z2, FIX(0.513743148)); /* c1-c3 */ - tmp14 = z1 - MULTIPLY(z3, FIX(2.176250899)); /* c1+c3 */ - - /* Final output stage */ - - wsptr[8*0] = (int) RIGHT_SHIFT(tmp10 + tmp13, CONST_BITS-PASS1_BITS); - wsptr[8*4] = (int) RIGHT_SHIFT(tmp10 - tmp13, CONST_BITS-PASS1_BITS); - wsptr[8*1] = (int) RIGHT_SHIFT(tmp11 + tmp14, CONST_BITS-PASS1_BITS); - wsptr[8*3] = (int) RIGHT_SHIFT(tmp11 - tmp14, CONST_BITS-PASS1_BITS); - wsptr[8*2] = (int) RIGHT_SHIFT(tmp12, CONST_BITS-PASS1_BITS); - } - - /* Pass 2: process 5 rows from work array, store into output array. - * 10-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/20). - */ - - wsptr = workspace; - for (ctr = 0; ctr < 5; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - z3 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - z3 <<= CONST_BITS; - z4 = (INT32) wsptr[4]; - z1 = MULTIPLY(z4, FIX(1.144122806)); /* c4 */ - z2 = MULTIPLY(z4, FIX(0.437016024)); /* c8 */ - tmp10 = z3 + z1; - tmp11 = z3 - z2; - - tmp22 = z3 - ((z1 - z2) << 1); /* c0 = (c4-c8)*2 */ - - z2 = (INT32) wsptr[2]; - z3 = (INT32) wsptr[6]; - - z1 = MULTIPLY(z2 + z3, FIX(0.831253876)); /* c6 */ - tmp12 = z1 + MULTIPLY(z2, FIX(0.513743148)); /* c2-c6 */ - tmp13 = z1 - MULTIPLY(z3, FIX(2.176250899)); /* c2+c6 */ - - tmp20 = tmp10 + tmp12; - tmp24 = tmp10 - tmp12; - tmp21 = tmp11 + tmp13; - tmp23 = tmp11 - tmp13; - - /* Odd part */ - - z1 = (INT32) wsptr[1]; - z2 = (INT32) wsptr[3]; - z3 = (INT32) wsptr[5]; - z3 <<= CONST_BITS; - z4 = (INT32) wsptr[7]; - - tmp11 = z2 + z4; - tmp13 = z2 - z4; - - tmp12 = MULTIPLY(tmp13, FIX(0.309016994)); /* (c3-c7)/2 */ - - z2 = MULTIPLY(tmp11, FIX(0.951056516)); /* (c3+c7)/2 */ - z4 = z3 + tmp12; - - tmp10 = MULTIPLY(z1, FIX(1.396802247)) + z2 + z4; /* c1 */ - tmp14 = MULTIPLY(z1, FIX(0.221231742)) - z2 + z4; /* c9 */ - - z2 = MULTIPLY(tmp11, FIX(0.587785252)); /* (c1-c9)/2 */ - z4 = z3 - tmp12 - (tmp13 << (CONST_BITS - 1)); - - tmp12 = ((z1 - tmp13) << CONST_BITS) - z3; - - tmp11 = MULTIPLY(z1, FIX(1.260073511)) - z2 - z4; /* c3 */ - tmp13 = MULTIPLY(z1, FIX(0.642039522)) - z2 + z4; /* c7 */ - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[9] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[8] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23 + tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp23 - tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp24 + tmp14, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp24 - tmp14, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += 8; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing an 8x4 output block. - * - * 4-point IDCT in pass 1 (columns), 8-point in pass 2 (rows). - */ - -GLOBAL(void) -jpeg_idct_8x4 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3; - INT32 tmp10, tmp11, tmp12, tmp13; - INT32 z1, z2, z3; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[8*4]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. - * 4-point IDCT kernel, - * cK represents sqrt(2) * cos(K*pi/16) [refers to 8-point IDCT]. - */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - tmp2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - - tmp10 = (tmp0 + tmp2) << PASS1_BITS; - tmp12 = (tmp0 - tmp2) << PASS1_BITS; - - /* Odd part */ - /* Same rotation as in the even part of the 8x8 LL&M IDCT */ - - z2 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - z3 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - - z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ - /* Add fudge factor here for final descale. */ - z1 += ONE << (CONST_BITS-PASS1_BITS-1); - tmp0 = RIGHT_SHIFT(z1 + MULTIPLY(z2, FIX_0_765366865), /* c2-c6 */ - CONST_BITS-PASS1_BITS); - tmp2 = RIGHT_SHIFT(z1 - MULTIPLY(z3, FIX_1_847759065), /* c2+c6 */ - CONST_BITS-PASS1_BITS); - - /* Final output stage */ - - wsptr[8*0] = (int) (tmp10 + tmp0); - wsptr[8*3] = (int) (tmp10 - tmp0); - wsptr[8*1] = (int) (tmp12 + tmp2); - wsptr[8*2] = (int) (tmp12 - tmp2); - } - - /* Pass 2: process rows from work array, store into output array. - * Note that we must descale the results by a factor of 8 == 2**3, - * and also undo the PASS1_BITS scaling. - * 8-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/16). - */ - - wsptr = workspace; - for (ctr = 0; ctr < 4; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part: reverse the even part of the forward DCT. - * The rotator is c(-6). - */ - - /* Add range center and fudge factor for final descale and range-limit. */ - z2 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - z3 = (INT32) wsptr[4]; - - tmp0 = (z2 + z3) << CONST_BITS; - tmp1 = (z2 - z3) << CONST_BITS; - - z2 = (INT32) wsptr[2]; - z3 = (INT32) wsptr[6]; - - z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ - tmp2 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ - tmp3 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ - - tmp10 = tmp0 + tmp2; - tmp13 = tmp0 - tmp2; - tmp11 = tmp1 + tmp3; - tmp12 = tmp1 - tmp3; - - /* Odd part per figure 8; the matrix is unitary and hence its - * transpose is its inverse. i0..i3 are y7,y5,y3,y1 respectively. - */ - - tmp0 = (INT32) wsptr[7]; - tmp1 = (INT32) wsptr[5]; - tmp2 = (INT32) wsptr[3]; - tmp3 = (INT32) wsptr[1]; - - z2 = tmp0 + tmp2; - z3 = tmp1 + tmp3; - - z1 = MULTIPLY(z2 + z3, FIX_1_175875602); /* c3 */ - z2 = MULTIPLY(z2, - FIX_1_961570560); /* -c3-c5 */ - z3 = MULTIPLY(z3, - FIX_0_390180644); /* -c3+c5 */ - z2 += z1; - z3 += z1; - - z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ - tmp0 = MULTIPLY(tmp0, FIX_0_298631336); /* -c1+c3+c5-c7 */ - tmp3 = MULTIPLY(tmp3, FIX_1_501321110); /* c1+c3-c5-c7 */ - tmp0 += z1 + z2; - tmp3 += z1 + z3; - - z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ - tmp1 = MULTIPLY(tmp1, FIX_2_053119869); /* c1+c3-c5+c7 */ - tmp2 = MULTIPLY(tmp2, FIX_3_072711026); /* c1+c3+c5-c7 */ - tmp1 += z1 + z3; - tmp2 += z1 + z2; - - /* Final output stage: inputs are tmp10..tmp13, tmp0..tmp3 */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp3, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp3, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp11 + tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp11 - tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12 + tmp1, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp12 - tmp1, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp13 + tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp13 - tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += DCTSIZE; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a 6x3 output block. - * - * 3-point IDCT in pass 1 (columns), 6-point in pass 2 (rows). - */ - -GLOBAL(void) -jpeg_idct_6x3 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp0, tmp1, tmp2, tmp10, tmp11, tmp12; - INT32 z1, z2, z3; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[6*3]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. - * 3-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/6). - */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 6; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - tmp0 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - tmp0 += ONE << (CONST_BITS-PASS1_BITS-1); - tmp2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - tmp12 = MULTIPLY(tmp2, FIX(0.707106781)); /* c2 */ - tmp10 = tmp0 + tmp12; - tmp2 = tmp0 - tmp12 - tmp12; - - /* Odd part */ - - tmp12 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - tmp0 = MULTIPLY(tmp12, FIX(1.224744871)); /* c1 */ - - /* Final output stage */ - - wsptr[6*0] = (int) RIGHT_SHIFT(tmp10 + tmp0, CONST_BITS-PASS1_BITS); - wsptr[6*2] = (int) RIGHT_SHIFT(tmp10 - tmp0, CONST_BITS-PASS1_BITS); - wsptr[6*1] = (int) RIGHT_SHIFT(tmp2, CONST_BITS-PASS1_BITS); - } - - /* Pass 2: process 3 rows from work array, store into output array. - * 6-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/12). - */ - - wsptr = workspace; - for (ctr = 0; ctr < 3; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - tmp0 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - tmp0 <<= CONST_BITS; - tmp2 = (INT32) wsptr[4]; - tmp10 = MULTIPLY(tmp2, FIX(0.707106781)); /* c4 */ - tmp1 = tmp0 + tmp10; - tmp11 = tmp0 - tmp10 - tmp10; - tmp10 = (INT32) wsptr[2]; - tmp0 = MULTIPLY(tmp10, FIX(1.224744871)); /* c2 */ - tmp10 = tmp1 + tmp0; - tmp12 = tmp1 - tmp0; - - /* Odd part */ - - z1 = (INT32) wsptr[1]; - z2 = (INT32) wsptr[3]; - z3 = (INT32) wsptr[5]; - tmp1 = MULTIPLY(z1 + z3, FIX(0.366025404)); /* c5 */ - tmp0 = tmp1 + ((z1 + z2) << CONST_BITS); - tmp2 = tmp1 + ((z3 - z2) << CONST_BITS); - tmp1 = (z1 - z2 - z3) << CONST_BITS; - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp11 + tmp1, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp11 - tmp1, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12 + tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp12 - tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += 6; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a 4x2 output block. - * - * 2-point IDCT in pass 1 (columns), 4-point in pass 2 (rows). - */ - -GLOBAL(void) -jpeg_idct_4x2 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp0, tmp2, tmp10, tmp12; - INT32 z1, z2, z3; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - INT32 * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - INT32 workspace[4*2]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 4; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - tmp10 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - - /* Odd part */ - - tmp0 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - - /* Final output stage */ - - wsptr[4*0] = tmp10 + tmp0; - wsptr[4*1] = tmp10 - tmp0; - } - - /* Pass 2: process 2 rows from work array, store into output array. - * 4-point IDCT kernel, - * cK represents sqrt(2) * cos(K*pi/16) [refers to 8-point IDCT]. - */ - - wsptr = workspace; - for (ctr = 0; ctr < 2; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - tmp0 = wsptr[0] + ((((INT32) RANGE_CENTER) << 3) + (ONE << 2)); - tmp2 = wsptr[2]; - - tmp10 = (tmp0 + tmp2) << CONST_BITS; - tmp12 = (tmp0 - tmp2) << CONST_BITS; - - /* Odd part */ - /* Same rotation as in the even part of the 8x8 LL&M IDCT */ - - z2 = wsptr[1]; - z3 = wsptr[3]; - - z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ - tmp0 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ - tmp2 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp0, - CONST_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp0, - CONST_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp12 + tmp2, - CONST_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12 - tmp2, - CONST_BITS+3) - & RANGE_MASK]; - - wsptr += 4; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a 2x1 output block. - * - * 1-point IDCT in pass 1 (columns), 2-point in pass 2 (rows). - */ - -GLOBAL(void) -jpeg_idct_2x1 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - DCTELEM tmp0, tmp1; - ISLOW_MULT_TYPE * quantptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - ISHIFT_TEMPS - - /* Pass 1: empty. */ - - /* Pass 2: process 1 row from input, store into output array. */ - - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - outptr = output_buf[0] + output_col; - - /* Even part */ - - tmp0 = DEQUANTIZE(coef_block[0], quantptr[0]); - /* Add range center and fudge factor for final descale and range-limit. */ - tmp0 += (((DCTELEM) RANGE_CENTER) << 3) + (1 << 2); - - /* Odd part */ - - tmp1 = DEQUANTIZE(coef_block[1], quantptr[1]); - - /* Final output stage */ - - outptr[0] = range_limit[(int) IRIGHT_SHIFT(tmp0 + tmp1, 3) & RANGE_MASK]; - outptr[1] = range_limit[(int) IRIGHT_SHIFT(tmp0 - tmp1, 3) & RANGE_MASK]; -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing an 8x16 output block. - * - * 16-point IDCT in pass 1 (columns), 8-point in pass 2 (rows). - */ - -GLOBAL(void) -jpeg_idct_8x16 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3, tmp10, tmp11, tmp12, tmp13; - INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25, tmp26, tmp27; - INT32 z1, z2, z3, z4; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[8*16]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. - * 16-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/32). - */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 8; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - tmp0 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - tmp0 += ONE << (CONST_BITS-PASS1_BITS-1); - - z1 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - tmp1 = MULTIPLY(z1, FIX(1.306562965)); /* c4[16] = c2[8] */ - tmp2 = MULTIPLY(z1, FIX_0_541196100); /* c12[16] = c6[8] */ - - tmp10 = tmp0 + tmp1; - tmp11 = tmp0 - tmp1; - tmp12 = tmp0 + tmp2; - tmp13 = tmp0 - tmp2; - - z1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - z2 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); - z3 = z1 - z2; - z4 = MULTIPLY(z3, FIX(0.275899379)); /* c14[16] = c7[8] */ - z3 = MULTIPLY(z3, FIX(1.387039845)); /* c2[16] = c1[8] */ - - tmp0 = z3 + MULTIPLY(z2, FIX_2_562915447); /* (c6+c2)[16] = (c3+c1)[8] */ - tmp1 = z4 + MULTIPLY(z1, FIX_0_899976223); /* (c6-c14)[16] = (c3-c7)[8] */ - tmp2 = z3 - MULTIPLY(z1, FIX(0.601344887)); /* (c2-c10)[16] = (c1-c5)[8] */ - tmp3 = z4 - MULTIPLY(z2, FIX(0.509795579)); /* (c10-c14)[16] = (c5-c7)[8] */ - - tmp20 = tmp10 + tmp0; - tmp27 = tmp10 - tmp0; - tmp21 = tmp12 + tmp1; - tmp26 = tmp12 - tmp1; - tmp22 = tmp13 + tmp2; - tmp25 = tmp13 - tmp2; - tmp23 = tmp11 + tmp3; - tmp24 = tmp11 - tmp3; - - /* Odd part */ - - z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); - z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); - - tmp11 = z1 + z3; - - tmp1 = MULTIPLY(z1 + z2, FIX(1.353318001)); /* c3 */ - tmp2 = MULTIPLY(tmp11, FIX(1.247225013)); /* c5 */ - tmp3 = MULTIPLY(z1 + z4, FIX(1.093201867)); /* c7 */ - tmp10 = MULTIPLY(z1 - z4, FIX(0.897167586)); /* c9 */ - tmp11 = MULTIPLY(tmp11, FIX(0.666655658)); /* c11 */ - tmp12 = MULTIPLY(z1 - z2, FIX(0.410524528)); /* c13 */ - tmp0 = tmp1 + tmp2 + tmp3 - - MULTIPLY(z1, FIX(2.286341144)); /* c7+c5+c3-c1 */ - tmp13 = tmp10 + tmp11 + tmp12 - - MULTIPLY(z1, FIX(1.835730603)); /* c9+c11+c13-c15 */ - z1 = MULTIPLY(z2 + z3, FIX(0.138617169)); /* c15 */ - tmp1 += z1 + MULTIPLY(z2, FIX(0.071888074)); /* c9+c11-c3-c15 */ - tmp2 += z1 - MULTIPLY(z3, FIX(1.125726048)); /* c5+c7+c15-c3 */ - z1 = MULTIPLY(z3 - z2, FIX(1.407403738)); /* c1 */ - tmp11 += z1 - MULTIPLY(z3, FIX(0.766367282)); /* c1+c11-c9-c13 */ - tmp12 += z1 + MULTIPLY(z2, FIX(1.971951411)); /* c1+c5+c13-c7 */ - z2 += z4; - z1 = MULTIPLY(z2, - FIX(0.666655658)); /* -c11 */ - tmp1 += z1; - tmp3 += z1 + MULTIPLY(z4, FIX(1.065388962)); /* c3+c11+c15-c7 */ - z2 = MULTIPLY(z2, - FIX(1.247225013)); /* -c5 */ - tmp10 += z2 + MULTIPLY(z4, FIX(3.141271809)); /* c1+c5+c9-c13 */ - tmp12 += z2; - z2 = MULTIPLY(z3 + z4, - FIX(1.353318001)); /* -c3 */ - tmp2 += z2; - tmp3 += z2; - z2 = MULTIPLY(z4 - z3, FIX(0.410524528)); /* c13 */ - tmp10 += z2; - tmp11 += z2; - - /* Final output stage */ - - wsptr[8*0] = (int) RIGHT_SHIFT(tmp20 + tmp0, CONST_BITS-PASS1_BITS); - wsptr[8*15] = (int) RIGHT_SHIFT(tmp20 - tmp0, CONST_BITS-PASS1_BITS); - wsptr[8*1] = (int) RIGHT_SHIFT(tmp21 + tmp1, CONST_BITS-PASS1_BITS); - wsptr[8*14] = (int) RIGHT_SHIFT(tmp21 - tmp1, CONST_BITS-PASS1_BITS); - wsptr[8*2] = (int) RIGHT_SHIFT(tmp22 + tmp2, CONST_BITS-PASS1_BITS); - wsptr[8*13] = (int) RIGHT_SHIFT(tmp22 - tmp2, CONST_BITS-PASS1_BITS); - wsptr[8*3] = (int) RIGHT_SHIFT(tmp23 + tmp3, CONST_BITS-PASS1_BITS); - wsptr[8*12] = (int) RIGHT_SHIFT(tmp23 - tmp3, CONST_BITS-PASS1_BITS); - wsptr[8*4] = (int) RIGHT_SHIFT(tmp24 + tmp10, CONST_BITS-PASS1_BITS); - wsptr[8*11] = (int) RIGHT_SHIFT(tmp24 - tmp10, CONST_BITS-PASS1_BITS); - wsptr[8*5] = (int) RIGHT_SHIFT(tmp25 + tmp11, CONST_BITS-PASS1_BITS); - wsptr[8*10] = (int) RIGHT_SHIFT(tmp25 - tmp11, CONST_BITS-PASS1_BITS); - wsptr[8*6] = (int) RIGHT_SHIFT(tmp26 + tmp12, CONST_BITS-PASS1_BITS); - wsptr[8*9] = (int) RIGHT_SHIFT(tmp26 - tmp12, CONST_BITS-PASS1_BITS); - wsptr[8*7] = (int) RIGHT_SHIFT(tmp27 + tmp13, CONST_BITS-PASS1_BITS); - wsptr[8*8] = (int) RIGHT_SHIFT(tmp27 - tmp13, CONST_BITS-PASS1_BITS); - } - - /* Pass 2: process rows from work array, store into output array. - * Note that we must descale the results by a factor of 8 == 2**3, - * and also undo the PASS1_BITS scaling. - * 8-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/16). - */ - - wsptr = workspace; - for (ctr = 0; ctr < 16; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part: reverse the even part of the forward DCT. - * The rotator is c(-6). - */ - - /* Add range center and fudge factor for final descale and range-limit. */ - z2 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - z3 = (INT32) wsptr[4]; - - tmp0 = (z2 + z3) << CONST_BITS; - tmp1 = (z2 - z3) << CONST_BITS; - - z2 = (INT32) wsptr[2]; - z3 = (INT32) wsptr[6]; - - z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ - tmp2 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ - tmp3 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ - - tmp10 = tmp0 + tmp2; - tmp13 = tmp0 - tmp2; - tmp11 = tmp1 + tmp3; - tmp12 = tmp1 - tmp3; - - /* Odd part per figure 8; the matrix is unitary and hence its - * transpose is its inverse. i0..i3 are y7,y5,y3,y1 respectively. - */ - - tmp0 = (INT32) wsptr[7]; - tmp1 = (INT32) wsptr[5]; - tmp2 = (INT32) wsptr[3]; - tmp3 = (INT32) wsptr[1]; - - z2 = tmp0 + tmp2; - z3 = tmp1 + tmp3; - - z1 = MULTIPLY(z2 + z3, FIX_1_175875602); /* c3 */ - z2 = MULTIPLY(z2, - FIX_1_961570560); /* -c3-c5 */ - z3 = MULTIPLY(z3, - FIX_0_390180644); /* -c3+c5 */ - z2 += z1; - z3 += z1; - - z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ - tmp0 = MULTIPLY(tmp0, FIX_0_298631336); /* -c1+c3+c5-c7 */ - tmp3 = MULTIPLY(tmp3, FIX_1_501321110); /* c1+c3-c5-c7 */ - tmp0 += z1 + z2; - tmp3 += z1 + z3; - - z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ - tmp1 = MULTIPLY(tmp1, FIX_2_053119869); /* c1+c3-c5+c7 */ - tmp2 = MULTIPLY(tmp2, FIX_3_072711026); /* c1+c3+c5-c7 */ - tmp1 += z1 + z3; - tmp2 += z1 + z2; - - /* Final output stage: inputs are tmp10..tmp13, tmp0..tmp3 */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp3, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[7] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp3, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp11 + tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp11 - tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12 + tmp1, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp12 - tmp1, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp13 + tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp13 - tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += DCTSIZE; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a 7x14 output block. - * - * 14-point IDCT in pass 1 (columns), 7-point in pass 2 (rows). - */ - -GLOBAL(void) -jpeg_idct_7x14 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15, tmp16; - INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25, tmp26; - INT32 z1, z2, z3, z4; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[7*14]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. - * 14-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/28). - */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 7; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - z1 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - z1 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - z1 += ONE << (CONST_BITS-PASS1_BITS-1); - z4 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - z2 = MULTIPLY(z4, FIX(1.274162392)); /* c4 */ - z3 = MULTIPLY(z4, FIX(0.314692123)); /* c12 */ - z4 = MULTIPLY(z4, FIX(0.881747734)); /* c8 */ - - tmp10 = z1 + z2; - tmp11 = z1 + z3; - tmp12 = z1 - z4; - - tmp23 = RIGHT_SHIFT(z1 - ((z2 + z3 - z4) << 1), /* c0 = (c4+c12-c8)*2 */ - CONST_BITS-PASS1_BITS); - - z1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - z2 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); - - z3 = MULTIPLY(z1 + z2, FIX(1.105676686)); /* c6 */ - - tmp13 = z3 + MULTIPLY(z1, FIX(0.273079590)); /* c2-c6 */ - tmp14 = z3 - MULTIPLY(z2, FIX(1.719280954)); /* c6+c10 */ - tmp15 = MULTIPLY(z1, FIX(0.613604268)) - /* c10 */ - MULTIPLY(z2, FIX(1.378756276)); /* c2 */ - - tmp20 = tmp10 + tmp13; - tmp26 = tmp10 - tmp13; - tmp21 = tmp11 + tmp14; - tmp25 = tmp11 - tmp14; - tmp22 = tmp12 + tmp15; - tmp24 = tmp12 - tmp15; - - /* Odd part */ - - z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); - z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); - tmp13 = z4 << CONST_BITS; - - tmp14 = z1 + z3; - tmp11 = MULTIPLY(z1 + z2, FIX(1.334852607)); /* c3 */ - tmp12 = MULTIPLY(tmp14, FIX(1.197448846)); /* c5 */ - tmp10 = tmp11 + tmp12 + tmp13 - MULTIPLY(z1, FIX(1.126980169)); /* c3+c5-c1 */ - tmp14 = MULTIPLY(tmp14, FIX(0.752406978)); /* c9 */ - tmp16 = tmp14 - MULTIPLY(z1, FIX(1.061150426)); /* c9+c11-c13 */ - z1 -= z2; - tmp15 = MULTIPLY(z1, FIX(0.467085129)) - tmp13; /* c11 */ - tmp16 += tmp15; - z1 += z4; - z4 = MULTIPLY(z2 + z3, - FIX(0.158341681)) - tmp13; /* -c13 */ - tmp11 += z4 - MULTIPLY(z2, FIX(0.424103948)); /* c3-c9-c13 */ - tmp12 += z4 - MULTIPLY(z3, FIX(2.373959773)); /* c3+c5-c13 */ - z4 = MULTIPLY(z3 - z2, FIX(1.405321284)); /* c1 */ - tmp14 += z4 + tmp13 - MULTIPLY(z3, FIX(1.6906431334)); /* c1+c9-c11 */ - tmp15 += z4 + MULTIPLY(z2, FIX(0.674957567)); /* c1+c11-c5 */ - - tmp13 = (z1 - z3) << PASS1_BITS; - - /* Final output stage */ - - wsptr[7*0] = (int) RIGHT_SHIFT(tmp20 + tmp10, CONST_BITS-PASS1_BITS); - wsptr[7*13] = (int) RIGHT_SHIFT(tmp20 - tmp10, CONST_BITS-PASS1_BITS); - wsptr[7*1] = (int) RIGHT_SHIFT(tmp21 + tmp11, CONST_BITS-PASS1_BITS); - wsptr[7*12] = (int) RIGHT_SHIFT(tmp21 - tmp11, CONST_BITS-PASS1_BITS); - wsptr[7*2] = (int) RIGHT_SHIFT(tmp22 + tmp12, CONST_BITS-PASS1_BITS); - wsptr[7*11] = (int) RIGHT_SHIFT(tmp22 - tmp12, CONST_BITS-PASS1_BITS); - wsptr[7*3] = (int) (tmp23 + tmp13); - wsptr[7*10] = (int) (tmp23 - tmp13); - wsptr[7*4] = (int) RIGHT_SHIFT(tmp24 + tmp14, CONST_BITS-PASS1_BITS); - wsptr[7*9] = (int) RIGHT_SHIFT(tmp24 - tmp14, CONST_BITS-PASS1_BITS); - wsptr[7*5] = (int) RIGHT_SHIFT(tmp25 + tmp15, CONST_BITS-PASS1_BITS); - wsptr[7*8] = (int) RIGHT_SHIFT(tmp25 - tmp15, CONST_BITS-PASS1_BITS); - wsptr[7*6] = (int) RIGHT_SHIFT(tmp26 + tmp16, CONST_BITS-PASS1_BITS); - wsptr[7*7] = (int) RIGHT_SHIFT(tmp26 - tmp16, CONST_BITS-PASS1_BITS); - } - - /* Pass 2: process 14 rows from work array, store into output array. - * 7-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/14). - */ - - wsptr = workspace; - for (ctr = 0; ctr < 14; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - tmp23 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - tmp23 <<= CONST_BITS; - - z1 = (INT32) wsptr[2]; - z2 = (INT32) wsptr[4]; - z3 = (INT32) wsptr[6]; - - tmp20 = MULTIPLY(z2 - z3, FIX(0.881747734)); /* c4 */ - tmp22 = MULTIPLY(z1 - z2, FIX(0.314692123)); /* c6 */ - tmp21 = tmp20 + tmp22 + tmp23 - MULTIPLY(z2, FIX(1.841218003)); /* c2+c4-c6 */ - tmp10 = z1 + z3; - z2 -= tmp10; - tmp10 = MULTIPLY(tmp10, FIX(1.274162392)) + tmp23; /* c2 */ - tmp20 += tmp10 - MULTIPLY(z3, FIX(0.077722536)); /* c2-c4-c6 */ - tmp22 += tmp10 - MULTIPLY(z1, FIX(2.470602249)); /* c2+c4+c6 */ - tmp23 += MULTIPLY(z2, FIX(1.414213562)); /* c0 */ - - /* Odd part */ - - z1 = (INT32) wsptr[1]; - z2 = (INT32) wsptr[3]; - z3 = (INT32) wsptr[5]; - - tmp11 = MULTIPLY(z1 + z2, FIX(0.935414347)); /* (c3+c1-c5)/2 */ - tmp12 = MULTIPLY(z1 - z2, FIX(0.170262339)); /* (c3+c5-c1)/2 */ - tmp10 = tmp11 - tmp12; - tmp11 += tmp12; - tmp12 = MULTIPLY(z2 + z3, - FIX(1.378756276)); /* -c1 */ - tmp11 += tmp12; - z2 = MULTIPLY(z1 + z3, FIX(0.613604268)); /* c5 */ - tmp10 += z2; - tmp12 += z2 + MULTIPLY(z3, FIX(1.870828693)); /* c3+c1-c5 */ - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[6] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp23, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += 7; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a 6x12 output block. - * - * 12-point IDCT in pass 1 (columns), 6-point in pass 2 (rows). - */ - -GLOBAL(void) -jpeg_idct_6x12 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp10, tmp11, tmp12, tmp13, tmp14, tmp15; - INT32 tmp20, tmp21, tmp22, tmp23, tmp24, tmp25; - INT32 z1, z2, z3, z4; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[6*12]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. - * 12-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/24). - */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 6; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - z3 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - z3 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - z3 += ONE << (CONST_BITS-PASS1_BITS-1); - - z4 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - z4 = MULTIPLY(z4, FIX(1.224744871)); /* c4 */ - - tmp10 = z3 + z4; - tmp11 = z3 - z4; - - z1 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - z4 = MULTIPLY(z1, FIX(1.366025404)); /* c2 */ - z1 <<= CONST_BITS; - z2 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); - z2 <<= CONST_BITS; - - tmp12 = z1 - z2; - - tmp21 = z3 + tmp12; - tmp24 = z3 - tmp12; - - tmp12 = z4 + z2; - - tmp20 = tmp10 + tmp12; - tmp25 = tmp10 - tmp12; - - tmp12 = z4 - z1 - z2; - - tmp22 = tmp11 + tmp12; - tmp23 = tmp11 - tmp12; - - /* Odd part */ - - z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); - z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); - - tmp11 = MULTIPLY(z2, FIX(1.306562965)); /* c3 */ - tmp14 = MULTIPLY(z2, - FIX_0_541196100); /* -c9 */ - - tmp10 = z1 + z3; - tmp15 = MULTIPLY(tmp10 + z4, FIX(0.860918669)); /* c7 */ - tmp12 = tmp15 + MULTIPLY(tmp10, FIX(0.261052384)); /* c5-c7 */ - tmp10 = tmp12 + tmp11 + MULTIPLY(z1, FIX(0.280143716)); /* c1-c5 */ - tmp13 = MULTIPLY(z3 + z4, - FIX(1.045510580)); /* -(c7+c11) */ - tmp12 += tmp13 + tmp14 - MULTIPLY(z3, FIX(1.478575242)); /* c1+c5-c7-c11 */ - tmp13 += tmp15 - tmp11 + MULTIPLY(z4, FIX(1.586706681)); /* c1+c11 */ - tmp15 += tmp14 - MULTIPLY(z1, FIX(0.676326758)) - /* c7-c11 */ - MULTIPLY(z4, FIX(1.982889723)); /* c5+c7 */ - - z1 -= z4; - z2 -= z3; - z3 = MULTIPLY(z1 + z2, FIX_0_541196100); /* c9 */ - tmp11 = z3 + MULTIPLY(z1, FIX_0_765366865); /* c3-c9 */ - tmp14 = z3 - MULTIPLY(z2, FIX_1_847759065); /* c3+c9 */ - - /* Final output stage */ - - wsptr[6*0] = (int) RIGHT_SHIFT(tmp20 + tmp10, CONST_BITS-PASS1_BITS); - wsptr[6*11] = (int) RIGHT_SHIFT(tmp20 - tmp10, CONST_BITS-PASS1_BITS); - wsptr[6*1] = (int) RIGHT_SHIFT(tmp21 + tmp11, CONST_BITS-PASS1_BITS); - wsptr[6*10] = (int) RIGHT_SHIFT(tmp21 - tmp11, CONST_BITS-PASS1_BITS); - wsptr[6*2] = (int) RIGHT_SHIFT(tmp22 + tmp12, CONST_BITS-PASS1_BITS); - wsptr[6*9] = (int) RIGHT_SHIFT(tmp22 - tmp12, CONST_BITS-PASS1_BITS); - wsptr[6*3] = (int) RIGHT_SHIFT(tmp23 + tmp13, CONST_BITS-PASS1_BITS); - wsptr[6*8] = (int) RIGHT_SHIFT(tmp23 - tmp13, CONST_BITS-PASS1_BITS); - wsptr[6*4] = (int) RIGHT_SHIFT(tmp24 + tmp14, CONST_BITS-PASS1_BITS); - wsptr[6*7] = (int) RIGHT_SHIFT(tmp24 - tmp14, CONST_BITS-PASS1_BITS); - wsptr[6*5] = (int) RIGHT_SHIFT(tmp25 + tmp15, CONST_BITS-PASS1_BITS); - wsptr[6*6] = (int) RIGHT_SHIFT(tmp25 - tmp15, CONST_BITS-PASS1_BITS); - } - - /* Pass 2: process 12 rows from work array, store into output array. - * 6-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/12). - */ - - wsptr = workspace; - for (ctr = 0; ctr < 12; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - tmp10 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - tmp10 <<= CONST_BITS; - tmp12 = (INT32) wsptr[4]; - tmp20 = MULTIPLY(tmp12, FIX(0.707106781)); /* c4 */ - tmp11 = tmp10 + tmp20; - tmp21 = tmp10 - tmp20 - tmp20; - tmp20 = (INT32) wsptr[2]; - tmp10 = MULTIPLY(tmp20, FIX(1.224744871)); /* c2 */ - tmp20 = tmp11 + tmp10; - tmp22 = tmp11 - tmp10; - - /* Odd part */ - - z1 = (INT32) wsptr[1]; - z2 = (INT32) wsptr[3]; - z3 = (INT32) wsptr[5]; - tmp11 = MULTIPLY(z1 + z3, FIX(0.366025404)); /* c5 */ - tmp10 = tmp11 + ((z1 + z2) << CONST_BITS); - tmp12 = tmp11 + ((z3 - z2) << CONST_BITS); - tmp11 = (z1 - z2 - z3) << CONST_BITS; - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp20 + tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[5] = range_limit[(int) RIGHT_SHIFT(tmp20 - tmp10, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp21 + tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp21 - tmp11, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp22 + tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp22 - tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += 6; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a 5x10 output block. - * - * 10-point IDCT in pass 1 (columns), 5-point in pass 2 (rows). - */ - -GLOBAL(void) -jpeg_idct_5x10 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp10, tmp11, tmp12, tmp13, tmp14; - INT32 tmp20, tmp21, tmp22, tmp23, tmp24; - INT32 z1, z2, z3, z4, z5; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[5*10]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. - * 10-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/20). - */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 5; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - z3 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - z3 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - z3 += ONE << (CONST_BITS-PASS1_BITS-1); - z4 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - z1 = MULTIPLY(z4, FIX(1.144122806)); /* c4 */ - z2 = MULTIPLY(z4, FIX(0.437016024)); /* c8 */ - tmp10 = z3 + z1; - tmp11 = z3 - z2; - - tmp22 = RIGHT_SHIFT(z3 - ((z1 - z2) << 1), /* c0 = (c4-c8)*2 */ - CONST_BITS-PASS1_BITS); - - z2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - z3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); - - z1 = MULTIPLY(z2 + z3, FIX(0.831253876)); /* c6 */ - tmp12 = z1 + MULTIPLY(z2, FIX(0.513743148)); /* c2-c6 */ - tmp13 = z1 - MULTIPLY(z3, FIX(2.176250899)); /* c2+c6 */ - - tmp20 = tmp10 + tmp12; - tmp24 = tmp10 - tmp12; - tmp21 = tmp11 + tmp13; - tmp23 = tmp11 - tmp13; - - /* Odd part */ - - z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); - z4 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); - - tmp11 = z2 + z4; - tmp13 = z2 - z4; - - tmp12 = MULTIPLY(tmp13, FIX(0.309016994)); /* (c3-c7)/2 */ - z5 = z3 << CONST_BITS; - - z2 = MULTIPLY(tmp11, FIX(0.951056516)); /* (c3+c7)/2 */ - z4 = z5 + tmp12; - - tmp10 = MULTIPLY(z1, FIX(1.396802247)) + z2 + z4; /* c1 */ - tmp14 = MULTIPLY(z1, FIX(0.221231742)) - z2 + z4; /* c9 */ - - z2 = MULTIPLY(tmp11, FIX(0.587785252)); /* (c1-c9)/2 */ - z4 = z5 - tmp12 - (tmp13 << (CONST_BITS - 1)); - - tmp12 = (z1 - tmp13 - z3) << PASS1_BITS; - - tmp11 = MULTIPLY(z1, FIX(1.260073511)) - z2 - z4; /* c3 */ - tmp13 = MULTIPLY(z1, FIX(0.642039522)) - z2 + z4; /* c7 */ - - /* Final output stage */ - - wsptr[5*0] = (int) RIGHT_SHIFT(tmp20 + tmp10, CONST_BITS-PASS1_BITS); - wsptr[5*9] = (int) RIGHT_SHIFT(tmp20 - tmp10, CONST_BITS-PASS1_BITS); - wsptr[5*1] = (int) RIGHT_SHIFT(tmp21 + tmp11, CONST_BITS-PASS1_BITS); - wsptr[5*8] = (int) RIGHT_SHIFT(tmp21 - tmp11, CONST_BITS-PASS1_BITS); - wsptr[5*2] = (int) (tmp22 + tmp12); - wsptr[5*7] = (int) (tmp22 - tmp12); - wsptr[5*3] = (int) RIGHT_SHIFT(tmp23 + tmp13, CONST_BITS-PASS1_BITS); - wsptr[5*6] = (int) RIGHT_SHIFT(tmp23 - tmp13, CONST_BITS-PASS1_BITS); - wsptr[5*4] = (int) RIGHT_SHIFT(tmp24 + tmp14, CONST_BITS-PASS1_BITS); - wsptr[5*5] = (int) RIGHT_SHIFT(tmp24 - tmp14, CONST_BITS-PASS1_BITS); - } - - /* Pass 2: process 10 rows from work array, store into output array. - * 5-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/10). - */ - - wsptr = workspace; - for (ctr = 0; ctr < 10; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - tmp12 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - tmp12 <<= CONST_BITS; - tmp13 = (INT32) wsptr[2]; - tmp14 = (INT32) wsptr[4]; - z1 = MULTIPLY(tmp13 + tmp14, FIX(0.790569415)); /* (c2+c4)/2 */ - z2 = MULTIPLY(tmp13 - tmp14, FIX(0.353553391)); /* (c2-c4)/2 */ - z3 = tmp12 + z2; - tmp10 = z3 + z1; - tmp11 = z3 - z1; - tmp12 -= z2 << 2; - - /* Odd part */ - - z2 = (INT32) wsptr[1]; - z3 = (INT32) wsptr[3]; - - z1 = MULTIPLY(z2 + z3, FIX(0.831253876)); /* c3 */ - tmp13 = z1 + MULTIPLY(z2, FIX(0.513743148)); /* c1-c3 */ - tmp14 = z1 - MULTIPLY(z3, FIX(2.176250899)); /* c1+c3 */ - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[4] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp13, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp11 + tmp14, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp11 - tmp14, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += 5; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a 4x8 output block. - * - * 8-point IDCT in pass 1 (columns), 4-point in pass 2 (rows). - */ - -GLOBAL(void) -jpeg_idct_4x8 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp0, tmp1, tmp2, tmp3; - INT32 tmp10, tmp11, tmp12, tmp13; - INT32 z1, z2, z3; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[4*8]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. - * Note results are scaled up by sqrt(8) compared to a true IDCT; - * furthermore, we scale the results by 2**PASS1_BITS. - * 8-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/16). - */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 4; ctr > 0; ctr--) { - /* Due to quantization, we will usually find that many of the input - * coefficients are zero, especially the AC terms. We can exploit this - * by short-circuiting the IDCT calculation for any column in which all - * the AC terms are zero. In that case each output is equal to the - * DC coefficient (with scale factor as needed). - * With typical images and quantization tables, half or more of the - * column DCT calculations can be simplified this way. - */ - - if (inptr[DCTSIZE*1] == 0 && inptr[DCTSIZE*2] == 0 && - inptr[DCTSIZE*3] == 0 && inptr[DCTSIZE*4] == 0 && - inptr[DCTSIZE*5] == 0 && inptr[DCTSIZE*6] == 0 && - inptr[DCTSIZE*7] == 0) { - /* AC terms all zero */ - int dcval = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]) << PASS1_BITS; - - wsptr[4*0] = dcval; - wsptr[4*1] = dcval; - wsptr[4*2] = dcval; - wsptr[4*3] = dcval; - wsptr[4*4] = dcval; - wsptr[4*5] = dcval; - wsptr[4*6] = dcval; - wsptr[4*7] = dcval; - - inptr++; /* advance pointers to next column */ - quantptr++; - wsptr++; - continue; - } - - /* Even part: reverse the even part of the forward DCT. - * The rotator is c(-6). - */ - - z2 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - z3 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - z2 <<= CONST_BITS; - z3 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - z2 += ONE << (CONST_BITS-PASS1_BITS-1); - - tmp0 = z2 + z3; - tmp1 = z2 - z3; - - z2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - z3 = DEQUANTIZE(inptr[DCTSIZE*6], quantptr[DCTSIZE*6]); - - z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ - tmp2 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ - tmp3 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ - - tmp10 = tmp0 + tmp2; - tmp13 = tmp0 - tmp2; - tmp11 = tmp1 + tmp3; - tmp12 = tmp1 - tmp3; - - /* Odd part per figure 8; the matrix is unitary and hence its - * transpose is its inverse. i0..i3 are y7,y5,y3,y1 respectively. - */ - - tmp0 = DEQUANTIZE(inptr[DCTSIZE*7], quantptr[DCTSIZE*7]); - tmp1 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); - tmp2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - tmp3 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - - z2 = tmp0 + tmp2; - z3 = tmp1 + tmp3; - - z1 = MULTIPLY(z2 + z3, FIX_1_175875602); /* c3 */ - z2 = MULTIPLY(z2, - FIX_1_961570560); /* -c3-c5 */ - z3 = MULTIPLY(z3, - FIX_0_390180644); /* -c3+c5 */ - z2 += z1; - z3 += z1; - - z1 = MULTIPLY(tmp0 + tmp3, - FIX_0_899976223); /* -c3+c7 */ - tmp0 = MULTIPLY(tmp0, FIX_0_298631336); /* -c1+c3+c5-c7 */ - tmp3 = MULTIPLY(tmp3, FIX_1_501321110); /* c1+c3-c5-c7 */ - tmp0 += z1 + z2; - tmp3 += z1 + z3; - - z1 = MULTIPLY(tmp1 + tmp2, - FIX_2_562915447); /* -c1-c3 */ - tmp1 = MULTIPLY(tmp1, FIX_2_053119869); /* c1+c3-c5+c7 */ - tmp2 = MULTIPLY(tmp2, FIX_3_072711026); /* c1+c3+c5-c7 */ - tmp1 += z1 + z3; - tmp2 += z1 + z2; - - /* Final output stage: inputs are tmp10..tmp13, tmp0..tmp3 */ - - wsptr[4*0] = (int) RIGHT_SHIFT(tmp10 + tmp3, CONST_BITS-PASS1_BITS); - wsptr[4*7] = (int) RIGHT_SHIFT(tmp10 - tmp3, CONST_BITS-PASS1_BITS); - wsptr[4*1] = (int) RIGHT_SHIFT(tmp11 + tmp2, CONST_BITS-PASS1_BITS); - wsptr[4*6] = (int) RIGHT_SHIFT(tmp11 - tmp2, CONST_BITS-PASS1_BITS); - wsptr[4*2] = (int) RIGHT_SHIFT(tmp12 + tmp1, CONST_BITS-PASS1_BITS); - wsptr[4*5] = (int) RIGHT_SHIFT(tmp12 - tmp1, CONST_BITS-PASS1_BITS); - wsptr[4*3] = (int) RIGHT_SHIFT(tmp13 + tmp0, CONST_BITS-PASS1_BITS); - wsptr[4*4] = (int) RIGHT_SHIFT(tmp13 - tmp0, CONST_BITS-PASS1_BITS); - - inptr++; /* advance pointers to next column */ - quantptr++; - wsptr++; - } - - /* Pass 2: process 8 rows from work array, store into output array. - * 4-point IDCT kernel, - * cK represents sqrt(2) * cos(K*pi/16) [refers to 8-point IDCT]. - */ - - wsptr = workspace; - for (ctr = 0; ctr < 8; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - tmp0 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - tmp2 = (INT32) wsptr[2]; - - tmp10 = (tmp0 + tmp2) << CONST_BITS; - tmp12 = (tmp0 - tmp2) << CONST_BITS; - - /* Odd part */ - /* Same rotation as in the even part of the 8x8 LL&M IDCT */ - - z2 = (INT32) wsptr[1]; - z3 = (INT32) wsptr[3]; - - z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ - tmp0 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ - tmp2 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[3] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp12 + tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp12 - tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += 4; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a 3x6 output block. - * - * 6-point IDCT in pass 1 (columns), 3-point in pass 2 (rows). - */ - -GLOBAL(void) -jpeg_idct_3x6 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp0, tmp1, tmp2, tmp10, tmp11, tmp12; - INT32 z1, z2, z3; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - int * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - int workspace[3*6]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. - * 6-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/12). - */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 3; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - tmp0 <<= CONST_BITS; - /* Add fudge factor here for final descale. */ - tmp0 += ONE << (CONST_BITS-PASS1_BITS-1); - tmp2 = DEQUANTIZE(inptr[DCTSIZE*4], quantptr[DCTSIZE*4]); - tmp10 = MULTIPLY(tmp2, FIX(0.707106781)); /* c4 */ - tmp1 = tmp0 + tmp10; - tmp11 = RIGHT_SHIFT(tmp0 - tmp10 - tmp10, CONST_BITS-PASS1_BITS); - tmp10 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - tmp0 = MULTIPLY(tmp10, FIX(1.224744871)); /* c2 */ - tmp10 = tmp1 + tmp0; - tmp12 = tmp1 - tmp0; - - /* Odd part */ - - z1 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - z2 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - z3 = DEQUANTIZE(inptr[DCTSIZE*5], quantptr[DCTSIZE*5]); - tmp1 = MULTIPLY(z1 + z3, FIX(0.366025404)); /* c5 */ - tmp0 = tmp1 + ((z1 + z2) << CONST_BITS); - tmp2 = tmp1 + ((z3 - z2) << CONST_BITS); - tmp1 = (z1 - z2 - z3) << PASS1_BITS; - - /* Final output stage */ - - wsptr[3*0] = (int) RIGHT_SHIFT(tmp10 + tmp0, CONST_BITS-PASS1_BITS); - wsptr[3*5] = (int) RIGHT_SHIFT(tmp10 - tmp0, CONST_BITS-PASS1_BITS); - wsptr[3*1] = (int) (tmp11 + tmp1); - wsptr[3*4] = (int) (tmp11 - tmp1); - wsptr[3*2] = (int) RIGHT_SHIFT(tmp12 + tmp2, CONST_BITS-PASS1_BITS); - wsptr[3*3] = (int) RIGHT_SHIFT(tmp12 - tmp2, CONST_BITS-PASS1_BITS); - } - - /* Pass 2: process 6 rows from work array, store into output array. - * 3-point IDCT kernel, cK represents sqrt(2) * cos(K*pi/6). - */ - - wsptr = workspace; - for (ctr = 0; ctr < 6; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - tmp0 = (INT32) wsptr[0] + - ((((INT32) RANGE_CENTER) << (PASS1_BITS+3)) + - (ONE << (PASS1_BITS+2))); - tmp0 <<= CONST_BITS; - tmp2 = (INT32) wsptr[2]; - tmp12 = MULTIPLY(tmp2, FIX(0.707106781)); /* c2 */ - tmp10 = tmp0 + tmp12; - tmp2 = tmp0 - tmp12 - tmp12; - - /* Odd part */ - - tmp12 = (INT32) wsptr[1]; - tmp0 = MULTIPLY(tmp12, FIX(1.224744871)); /* c1 */ - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[2] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp0, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp2, - CONST_BITS+PASS1_BITS+3) - & RANGE_MASK]; - - wsptr += 3; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a 2x4 output block. - * - * 4-point IDCT in pass 1 (columns), 2-point in pass 2 (rows). - */ - -GLOBAL(void) -jpeg_idct_2x4 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - INT32 tmp0, tmp2, tmp10, tmp12; - INT32 z1, z2, z3; - JCOEFPTR inptr; - ISLOW_MULT_TYPE * quantptr; - INT32 * wsptr; - JSAMPROW outptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - int ctr; - INT32 workspace[2*4]; /* buffers data between passes */ - SHIFT_TEMPS - - /* Pass 1: process columns from input, store into work array. - * 4-point IDCT kernel, - * cK represents sqrt(2) * cos(K*pi/16) [refers to 8-point IDCT]. - */ - - inptr = coef_block; - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - wsptr = workspace; - for (ctr = 0; ctr < 2; ctr++, inptr++, quantptr++, wsptr++) { - /* Even part */ - - tmp0 = DEQUANTIZE(inptr[DCTSIZE*0], quantptr[DCTSIZE*0]); - tmp2 = DEQUANTIZE(inptr[DCTSIZE*2], quantptr[DCTSIZE*2]); - - tmp10 = (tmp0 + tmp2) << CONST_BITS; - tmp12 = (tmp0 - tmp2) << CONST_BITS; - - /* Odd part */ - /* Same rotation as in the even part of the 8x8 LL&M IDCT */ - - z2 = DEQUANTIZE(inptr[DCTSIZE*1], quantptr[DCTSIZE*1]); - z3 = DEQUANTIZE(inptr[DCTSIZE*3], quantptr[DCTSIZE*3]); - - z1 = MULTIPLY(z2 + z3, FIX_0_541196100); /* c6 */ - tmp0 = z1 + MULTIPLY(z2, FIX_0_765366865); /* c2-c6 */ - tmp2 = z1 - MULTIPLY(z3, FIX_1_847759065); /* c2+c6 */ - - /* Final output stage */ - - wsptr[2*0] = tmp10 + tmp0; - wsptr[2*3] = tmp10 - tmp0; - wsptr[2*1] = tmp12 + tmp2; - wsptr[2*2] = tmp12 - tmp2; - } - - /* Pass 2: process 4 rows from work array, store into output array. */ - - wsptr = workspace; - for (ctr = 0; ctr < 4; ctr++) { - outptr = output_buf[ctr] + output_col; - - /* Even part */ - - /* Add range center and fudge factor for final descale and range-limit. */ - tmp10 = wsptr[0] + - ((((INT32) RANGE_CENTER) << (CONST_BITS+3)) + - (ONE << (CONST_BITS+2))); - - /* Odd part */ - - tmp0 = wsptr[1]; - - /* Final output stage */ - - outptr[0] = range_limit[(int) RIGHT_SHIFT(tmp10 + tmp0, CONST_BITS+3) - & RANGE_MASK]; - outptr[1] = range_limit[(int) RIGHT_SHIFT(tmp10 - tmp0, CONST_BITS+3) - & RANGE_MASK]; - - wsptr += 2; /* advance pointer to next row */ - } -} - - -/* - * Perform dequantization and inverse DCT on one block of coefficients, - * producing a 1x2 output block. - * - * 2-point IDCT in pass 1 (columns), 1-point in pass 2 (rows). - */ - -GLOBAL(void) -jpeg_idct_1x2 (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col) -{ - DCTELEM tmp0, tmp1; - ISLOW_MULT_TYPE * quantptr; - JSAMPLE *range_limit = IDCT_range_limit(cinfo); - ISHIFT_TEMPS - - /* Process 1 column from input, store into output array. */ - - quantptr = (ISLOW_MULT_TYPE *) compptr->dct_table; - - /* Even part */ - - tmp0 = DEQUANTIZE(coef_block[DCTSIZE*0], quantptr[DCTSIZE*0]); - /* Add range center and fudge factor for final descale and range-limit. */ - tmp0 += (((DCTELEM) RANGE_CENTER) << 3) + (1 << 2); - - /* Odd part */ - - tmp1 = DEQUANTIZE(coef_block[DCTSIZE*1], quantptr[DCTSIZE*1]); - - /* Final output stage */ - - output_buf[0][output_col] = - range_limit[(int) IRIGHT_SHIFT(tmp0 + tmp1, 3) & RANGE_MASK]; - output_buf[1][output_col] = - range_limit[(int) IRIGHT_SHIFT(tmp0 - tmp1, 3) & RANGE_MASK]; -} - -#endif /* IDCT_SCALING_SUPPORTED */ -#endif /* DCT_ISLOW_SUPPORTED */ diff --git a/cmake/externals/libjpeg/jinclude.h b/cmake/externals/libjpeg/jinclude.h deleted file mode 100644 index 12ea8cd2f..000000000 --- a/cmake/externals/libjpeg/jinclude.h +++ /dev/null @@ -1,157 +0,0 @@ -/* - * jinclude.h - * - * Copyright (C) 1991-1994, Thomas G. Lane. - * Modified 2017-2022 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file exists to provide a single place to fix any problems with - * including the wrong system include files. (Common problems are taken - * care of by the standard jconfig symbols, but on really weird systems - * you may have to edit this file.) - * - * NOTE: this file is NOT intended to be included by applications using - * the JPEG library. Most applications need only include jpeglib.h. - */ - - -/* Include auto-config file to find out which system include files we need. */ - -#include "jconfig.h" /* auto configuration options */ -#define JCONFIG_INCLUDED /* so that jpeglib.h doesn't do it again */ - -/* - * We need the NULL macro and size_t typedef. - * On an ANSI-conforming system it is sufficient to include . - * Otherwise, we get them from or ; we may have to - * pull in as well. - * Note that the core JPEG library does not require ; - * only the default error handler and data source/destination modules do. - * But we must pull it in because of the references to FILE in jpeglib.h. - * You can remove those references if you want to compile without . - */ - -#ifdef HAVE_STDDEF_H -#include -#endif - -#ifdef HAVE_STDLIB_H -#include -#endif - -#ifdef NEED_SYS_TYPES_H -#include -#endif - -#include - -/* - * We need memory copying and zeroing functions, plus strncpy(). - * ANSI and System V implementations declare these in . - * BSD doesn't have the mem() functions, but it does have bcopy()/bzero(). - * Some systems may declare memset and memcpy in . - * - * NOTE: we assume the size parameters to these functions are of type size_t. - * Change the casts in these macros if not! - */ - -#ifdef NEED_BSD_STRINGS - -#include -#define MEMZERO(target,size) bzero((void *)(target), (size_t)(size)) -#define MEMCOPY(dest,src,size) bcopy((const void *)(src), (void *)(dest), (size_t)(size)) - -#else /* not BSD, assume ANSI/SysV string lib */ - -#include -#define MEMZERO(target,size) memset((void *)(target), 0, (size_t)(size)) -#define MEMCOPY(dest,src,size) memcpy((void *)(dest), (const void *)(src), (size_t)(size)) - -#endif - -/* - * In ANSI C, and indeed any rational implementation, size_t is also the - * type returned by sizeof(). However, it seems there are some irrational - * implementations out there, in which sizeof() returns an int even though - * size_t is defined as long or unsigned long. To ensure consistent results - * we always use this SIZEOF() macro in place of using sizeof() directly. - */ - -#define SIZEOF(object) ((size_t) sizeof(object)) - -/* - * The modules that use fread() and fwrite() always invoke them through - * these macros. On some systems you may need to twiddle the argument casts. - * CAUTION: argument order is different from underlying functions! - * - * Furthermore, macros are provided for fflush() and ferror() in order - * to facilitate adaption by applications using an own FILE class. - * - * You can define your own custom file I/O functions in jconfig.h and - * #define JPEG_HAVE_FILE_IO_CUSTOM there to prevent redefinition here. - * - * You can #define JPEG_USE_FILE_IO_CUSTOM in jconfig.h to use custom file - * I/O functions implemented in Delphi VCL (Visual Component Library) - * in Vcl.Imaging.jpeg.pas for the TJPEGImage component utilizing - * the Delphi RTL (Run-Time Library) TMemoryStream component: - * - * procedure jpeg_stdio_src(var cinfo: jpeg_decompress_struct; - * input_file: TStream); external; - * - * procedure jpeg_stdio_dest(var cinfo: jpeg_compress_struct; - * output_file: TStream); external; - * - * function jfread(var buf; recsize, reccount: Integer; S: TStream): Integer; - * begin - * Result := S.Read(buf, recsize * reccount); - * end; - * - * function jfwrite(const buf; recsize, reccount: Integer; S: TStream): Integer; - * begin - * Result := S.Write(buf, recsize * reccount); - * end; - * - * function jfflush(S: TStream): Integer; - * begin - * Result := 0; - * end; - * - * function jferror(S: TStream): Integer; - * begin - * Result := 0; - * end; - * - * TMemoryStream of Delphi RTL has the distinctive feature to provide dynamic - * memory buffer management with a file/stream-based interface, particularly for - * the write (output) operation, which is easier to apply compared with direct - * implementations as given in jdatadst.c for memory destination. Those direct - * implementations of dynamic memory write tend to be more difficult to use, - * so providing an option like TMemoryStream may be a useful alternative. - * - * The CFile/CMemFile classes of the Microsoft Foundation Class (MFC) Library - * may be used in a similar fashion. - */ - -#ifndef JPEG_HAVE_FILE_IO_CUSTOM -#ifdef JPEG_USE_FILE_IO_CUSTOM -extern size_t jfread(void * __ptr, size_t __size, size_t __n, FILE * __stream); -extern size_t jfwrite(const void * __ptr, size_t __size, size_t __n, FILE * __stream); -extern int jfflush(FILE * __stream); -extern int jferror(FILE * __fp); - -#define JFREAD(file,buf,sizeofbuf) \ - ((size_t) jfread((void *) (buf), (size_t) 1, (size_t) (sizeofbuf), (file))) -#define JFWRITE(file,buf,sizeofbuf) \ - ((size_t) jfwrite((const void *) (buf), (size_t) 1, (size_t) (sizeofbuf), (file))) -#define JFFLUSH(file) jfflush(file) -#define JFERROR(file) jferror(file) -#else -#define JFREAD(file,buf,sizeofbuf) \ - ((size_t) fread((void *) (buf), (size_t) 1, (size_t) (sizeofbuf), (file))) -#define JFWRITE(file,buf,sizeofbuf) \ - ((size_t) fwrite((const void *) (buf), (size_t) 1, (size_t) (sizeofbuf), (file))) -#define JFFLUSH(file) fflush(file) -#define JFERROR(file) ferror(file) -#endif -#endif diff --git a/cmake/externals/libjpeg/jmemansi.c b/cmake/externals/libjpeg/jmemansi.c deleted file mode 100644 index 2d93e4962..000000000 --- a/cmake/externals/libjpeg/jmemansi.c +++ /dev/null @@ -1,167 +0,0 @@ -/* - * jmemansi.c - * - * Copyright (C) 1992-1996, Thomas G. Lane. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file provides a simple generic implementation of the system- - * dependent portion of the JPEG memory manager. This implementation - * assumes that you have the ANSI-standard library routine tmpfile(). - * Also, the problem of determining the amount of memory available - * is shoved onto the user. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" -#include "jmemsys.h" /* import the system-dependent declarations */ - -#ifndef HAVE_STDLIB_H /* should declare malloc(),free() */ -extern void * malloc JPP((size_t size)); -extern void free JPP((void *ptr)); -#endif - -#ifndef SEEK_SET /* pre-ANSI systems may not define this; */ -#define SEEK_SET 0 /* if not, assume 0 is correct */ -#endif - - -/* - * Memory allocation and freeing are controlled by the regular library - * routines malloc() and free(). - */ - -GLOBAL(void *) -jpeg_get_small (j_common_ptr cinfo, size_t sizeofobject) -{ - return (void *) malloc(sizeofobject); -} - -GLOBAL(void) -jpeg_free_small (j_common_ptr cinfo, void * object, size_t sizeofobject) -{ - free(object); -} - - -/* - * "Large" objects are treated the same as "small" ones. - * NB: although we include FAR keywords in the routine declarations, - * this file won't actually work in 80x86 small/medium model; at least, - * you probably won't be able to process useful-size images in only 64KB. - */ - -GLOBAL(void FAR *) -jpeg_get_large (j_common_ptr cinfo, size_t sizeofobject) -{ - return (void FAR *) malloc(sizeofobject); -} - -GLOBAL(void) -jpeg_free_large (j_common_ptr cinfo, void FAR * object, size_t sizeofobject) -{ - free(object); -} - - -/* - * This routine computes the total memory space available for allocation. - * It's impossible to do this in a portable way; our current solution is - * to make the user tell us (with a default value set at compile time). - * If you can actually get the available space, it's a good idea to subtract - * a slop factor of 5% or so. - */ - -#ifndef DEFAULT_MAX_MEM /* so can override from makefile */ -#define DEFAULT_MAX_MEM 1000000L /* default: one megabyte */ -#endif - -GLOBAL(long) -jpeg_mem_available (j_common_ptr cinfo, long min_bytes_needed, - long max_bytes_needed, long already_allocated) -{ - return cinfo->mem->max_memory_to_use - already_allocated; -} - - -/* - * Backing store (temporary file) management. - * Backing store objects are only used when the value returned by - * jpeg_mem_available is less than the total space needed. You can dispense - * with these routines if you have plenty of virtual memory; see jmemnobs.c. - */ - - -METHODDEF(void) -read_backing_store (j_common_ptr cinfo, backing_store_ptr info, - void FAR * buffer_address, - long file_offset, long byte_count) -{ - if (fseek(info->temp_file, file_offset, SEEK_SET)) - ERREXIT(cinfo, JERR_TFILE_SEEK); - if (JFREAD(info->temp_file, buffer_address, byte_count) - != (size_t) byte_count) - ERREXIT(cinfo, JERR_TFILE_READ); -} - - -METHODDEF(void) -write_backing_store (j_common_ptr cinfo, backing_store_ptr info, - void FAR * buffer_address, - long file_offset, long byte_count) -{ - if (fseek(info->temp_file, file_offset, SEEK_SET)) - ERREXIT(cinfo, JERR_TFILE_SEEK); - if (JFWRITE(info->temp_file, buffer_address, byte_count) - != (size_t) byte_count) - ERREXIT(cinfo, JERR_TFILE_WRITE); -} - - -METHODDEF(void) -close_backing_store (j_common_ptr cinfo, backing_store_ptr info) -{ - fclose(info->temp_file); - /* Since this implementation uses tmpfile() to create the file, - * no explicit file deletion is needed. - */ -} - - -/* - * Initial opening of a backing-store object. - * - * This version uses tmpfile(), which constructs a suitable file name - * behind the scenes. We don't have to use info->temp_name[] at all; - * indeed, we can't even find out the actual name of the temp file. - */ - -GLOBAL(void) -jpeg_open_backing_store (j_common_ptr cinfo, backing_store_ptr info, - long total_bytes_needed) -{ - if ((info->temp_file = tmpfile()) == NULL) - ERREXITS(cinfo, JERR_TFILE_CREATE, ""); - info->read_backing_store = read_backing_store; - info->write_backing_store = write_backing_store; - info->close_backing_store = close_backing_store; -} - - -/* - * These routines take care of any system-dependent initialization and - * cleanup required. - */ - -GLOBAL(long) -jpeg_mem_init (j_common_ptr cinfo) -{ - return DEFAULT_MAX_MEM; /* default for max_memory_to_use */ -} - -GLOBAL(void) -jpeg_mem_term (j_common_ptr cinfo) -{ - /* no work */ -} diff --git a/cmake/externals/libjpeg/jmemmgr.c b/cmake/externals/libjpeg/jmemmgr.c deleted file mode 100644 index 40377de47..000000000 --- a/cmake/externals/libjpeg/jmemmgr.c +++ /dev/null @@ -1,1115 +0,0 @@ -/* - * jmemmgr.c - * - * Copyright (C) 1991-1997, Thomas G. Lane. - * Modified 2011-2019 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains the JPEG system-independent memory management - * routines. This code is usable across a wide variety of machines; most - * of the system dependencies have been isolated in a separate file. - * The major functions provided here are: - * * pool-based allocation and freeing of memory; - * * policy decisions about how to divide available memory among the - * virtual arrays; - * * control logic for swapping virtual arrays between main memory and - * backing storage. - * The separate system-dependent file provides the actual backing-storage - * access code, and it contains the policy decision about how much total - * main memory to use. - * This file is system-dependent in the sense that some of its functions - * are unnecessary in some systems. For example, if there is enough virtual - * memory so that backing storage will never be used, much of the virtual - * array control logic could be removed. (Of course, if you have that much - * memory then you shouldn't care about a little bit of unused code...) - */ - -#define JPEG_INTERNALS -#define AM_MEMORY_MANAGER /* we define jvirt_Xarray_control structs */ -#include "jinclude.h" -#include "jpeglib.h" -#include "jmemsys.h" /* import the system-dependent declarations */ - -#ifndef NO_GETENV -#ifndef HAVE_STDLIB_H /* should declare getenv() */ -extern char * getenv JPP((const char * name)); -#endif -#endif - - -/* - * Some important notes: - * The allocation routines provided here must never return NULL. - * They should exit to error_exit if unsuccessful. - * - * It's not a good idea to try to merge the sarray and barray routines, - * even though they are textually almost the same, because samples are - * usually stored as bytes while coefficients are shorts or ints. Thus, - * in machines where byte pointers have a different representation from - * word pointers, the resulting machine code could not be the same. - */ - - -/* - * Many machines require storage alignment: longs must start on 4-byte - * boundaries, doubles on 8-byte boundaries, etc. On such machines, malloc() - * always returns pointers that are multiples of the worst-case alignment - * requirement, and we had better do so too. - * There isn't any really portable way to determine the worst-case alignment - * requirement. This module assumes that the alignment requirement is - * multiples of sizeof(ALIGN_TYPE). - * By default, we define ALIGN_TYPE as double. This is necessary on some - * workstations (where doubles really do need 8-byte alignment) and will work - * fine on nearly everything. If your machine has lesser alignment needs, - * you can save a few bytes by making ALIGN_TYPE smaller. - * The only place I know of where this will NOT work is certain Macintosh - * 680x0 compilers that define double as a 10-byte IEEE extended float. - * Doing 10-byte alignment is counterproductive because longwords won't be - * aligned well. Put "#define ALIGN_TYPE long" in jconfig.h if you have - * such a compiler. - */ - -#ifndef ALIGN_TYPE /* so can override from jconfig.h */ -#define ALIGN_TYPE double -#endif - - -/* - * We allocate objects from "pools", where each pool is gotten with a single - * request to jpeg_get_small() or jpeg_get_large(). There is no per-object - * overhead within a pool, except for alignment padding. Each pool has a - * header with a link to the next pool of the same class. - * Small and large pool headers are identical except that the latter's - * link pointer must be FAR on 80x86 machines. - * Notice that the "real" header fields are union'ed with a dummy ALIGN_TYPE - * field. This forces the compiler to make SIZEOF(small_pool_hdr) a multiple - * of the alignment requirement of ALIGN_TYPE. - */ - -typedef union small_pool_struct * small_pool_ptr; - -typedef union small_pool_struct { - struct { - small_pool_ptr next; /* next in list of pools */ - size_t bytes_used; /* how many bytes already used within pool */ - size_t bytes_left; /* bytes still available in this pool */ - } hdr; - ALIGN_TYPE dummy; /* included in union to ensure alignment */ -} small_pool_hdr; - -typedef union large_pool_struct FAR * large_pool_ptr; - -typedef union large_pool_struct { - struct { - large_pool_ptr next; /* next in list of pools */ - size_t bytes_used; /* how many bytes already used within pool */ - size_t bytes_left; /* bytes still available in this pool */ - } hdr; - ALIGN_TYPE dummy; /* included in union to ensure alignment */ -} large_pool_hdr; - - -/* - * Here is the full definition of a memory manager object. - */ - -typedef struct { - struct jpeg_memory_mgr pub; /* public fields */ - - /* Each pool identifier (lifetime class) names a linked list of pools. */ - small_pool_ptr small_list[JPOOL_NUMPOOLS]; - large_pool_ptr large_list[JPOOL_NUMPOOLS]; - - /* Since we only have one lifetime class of virtual arrays, only one - * linked list is necessary (for each datatype). Note that the virtual - * array control blocks being linked together are actually stored somewhere - * in the small-pool list. - */ - jvirt_sarray_ptr virt_sarray_list; - jvirt_barray_ptr virt_barray_list; - - /* This counts total space obtained from jpeg_get_small/large */ - size_t total_space_allocated; - - /* alloc_sarray and alloc_barray set this value for use by virtual - * array routines. - */ - JDIMENSION last_rowsperchunk; /* from most recent alloc_sarray/barray */ -} my_memory_mgr; - -typedef my_memory_mgr * my_mem_ptr; - - -/* - * The control blocks for virtual arrays. - * Note that these blocks are allocated in the "small" pool area. - * System-dependent info for the associated backing store (if any) is hidden - * inside the backing_store_info struct. - */ - -struct jvirt_sarray_control { - JSAMPARRAY mem_buffer; /* => the in-memory buffer */ - JDIMENSION rows_in_array; /* total virtual array height */ - JDIMENSION samplesperrow; /* width of array (and of memory buffer) */ - JDIMENSION maxaccess; /* max rows accessed by access_virt_sarray */ - JDIMENSION rows_in_mem; /* height of memory buffer */ - JDIMENSION rowsperchunk; /* allocation chunk size in mem_buffer */ - JDIMENSION cur_start_row; /* first logical row # in the buffer */ - JDIMENSION first_undef_row; /* row # of first uninitialized row */ - boolean pre_zero; /* pre-zero mode requested? */ - boolean dirty; /* do current buffer contents need written? */ - boolean b_s_open; /* is backing-store data valid? */ - jvirt_sarray_ptr next; /* link to next virtual sarray control block */ - backing_store_info b_s_info; /* System-dependent control info */ -}; - -struct jvirt_barray_control { - JBLOCKARRAY mem_buffer; /* => the in-memory buffer */ - JDIMENSION rows_in_array; /* total virtual array height */ - JDIMENSION blocksperrow; /* width of array (and of memory buffer) */ - JDIMENSION maxaccess; /* max rows accessed by access_virt_barray */ - JDIMENSION rows_in_mem; /* height of memory buffer */ - JDIMENSION rowsperchunk; /* allocation chunk size in mem_buffer */ - JDIMENSION cur_start_row; /* first logical row # in the buffer */ - JDIMENSION first_undef_row; /* row # of first uninitialized row */ - boolean pre_zero; /* pre-zero mode requested? */ - boolean dirty; /* do current buffer contents need written? */ - boolean b_s_open; /* is backing-store data valid? */ - jvirt_barray_ptr next; /* link to next virtual barray control block */ - backing_store_info b_s_info; /* System-dependent control info */ -}; - - -#ifdef MEM_STATS /* optional extra stuff for statistics */ - -LOCAL(void) -print_mem_stats (j_common_ptr cinfo, int pool_id) -{ - my_mem_ptr mem = (my_mem_ptr) cinfo->mem; - small_pool_ptr shdr_ptr; - large_pool_ptr lhdr_ptr; - - /* Since this is only a debugging stub, we can cheat a little by using - * fprintf directly rather than going through the trace message code. - * This is helpful because message parm array can't handle longs. - */ - fprintf(stderr, "Freeing pool %d, total space = %ld\n", - pool_id, (long) mem->total_space_allocated); - - for (lhdr_ptr = mem->large_list[pool_id]; lhdr_ptr != NULL; - lhdr_ptr = lhdr_ptr->hdr.next) { - fprintf(stderr, " Large chunk used %ld\n", - (long) lhdr_ptr->hdr.bytes_used); - } - - for (shdr_ptr = mem->small_list[pool_id]; shdr_ptr != NULL; - shdr_ptr = shdr_ptr->hdr.next) { - fprintf(stderr, " Small chunk used %ld free %ld\n", - (long) shdr_ptr->hdr.bytes_used, - (long) shdr_ptr->hdr.bytes_left); - } -} - -#endif /* MEM_STATS */ - - -LOCAL(noreturn_t) -out_of_memory (j_common_ptr cinfo, int which) -/* Report an out-of-memory error and stop execution */ -/* If we compiled MEM_STATS support, report alloc requests before dying */ -{ -#ifdef MEM_STATS - cinfo->err->trace_level = 2; /* force self_destruct to report stats */ -#endif - ERREXIT1(cinfo, JERR_OUT_OF_MEMORY, which); -} - - -/* - * Allocation of "small" objects. - * - * For these, we use pooled storage. When a new pool must be created, - * we try to get enough space for the current request plus a "slop" factor, - * where the slop will be the amount of leftover space in the new pool. - * The speed vs. space tradeoff is largely determined by the slop values. - * A different slop value is provided for each pool class (lifetime), - * and we also distinguish the first pool of a class from later ones. - * NOTE: the values given work fairly well on both 16- and 32-bit-int - * machines, but may be too small if longs are 64 bits or more. - */ - -static const size_t first_pool_slop[JPOOL_NUMPOOLS] = -{ - 1600, /* first PERMANENT pool */ - 16000 /* first IMAGE pool */ -}; - -static const size_t extra_pool_slop[JPOOL_NUMPOOLS] = -{ - 0, /* additional PERMANENT pools */ - 5000 /* additional IMAGE pools */ -}; - -#define MIN_SLOP 50 /* greater than 0 to avoid futile looping */ - - -METHODDEF(void *) -alloc_small (j_common_ptr cinfo, int pool_id, size_t sizeofobject) -/* Allocate a "small" object */ -{ - my_mem_ptr mem = (my_mem_ptr) cinfo->mem; - small_pool_ptr hdr_ptr, prev_hdr_ptr; - size_t odd_bytes, min_request, slop; - char * data_ptr; - - /* Check for unsatisfiable request (do now to ensure no overflow below) */ - if (sizeofobject > (size_t) MAX_ALLOC_CHUNK - SIZEOF(small_pool_hdr)) - out_of_memory(cinfo, 1); /* request exceeds malloc's ability */ - - /* Round up the requested size to a multiple of SIZEOF(ALIGN_TYPE) */ - odd_bytes = sizeofobject % SIZEOF(ALIGN_TYPE); - if (odd_bytes > 0) - sizeofobject += SIZEOF(ALIGN_TYPE) - odd_bytes; - - /* See if space is available in any existing pool */ - if (pool_id < 0 || pool_id >= JPOOL_NUMPOOLS) - ERREXIT1(cinfo, JERR_BAD_POOL_ID, pool_id); /* safety check */ - prev_hdr_ptr = NULL; - hdr_ptr = mem->small_list[pool_id]; - while (hdr_ptr != NULL) { - if (hdr_ptr->hdr.bytes_left >= sizeofobject) - break; /* found pool with enough space */ - prev_hdr_ptr = hdr_ptr; - hdr_ptr = hdr_ptr->hdr.next; - } - - /* Time to make a new pool? */ - if (hdr_ptr == NULL) { - /* min_request is what we need now, slop is what will be leftover */ - min_request = sizeofobject + SIZEOF(small_pool_hdr); - if (prev_hdr_ptr == NULL) /* first pool in class? */ - slop = first_pool_slop[pool_id]; - else - slop = extra_pool_slop[pool_id]; - /* Don't ask for more than MAX_ALLOC_CHUNK */ - if (slop > (size_t) MAX_ALLOC_CHUNK - min_request) - slop = (size_t) MAX_ALLOC_CHUNK - min_request; - /* Try to get space, if fail reduce slop and try again */ - for (;;) { - hdr_ptr = (small_pool_ptr) jpeg_get_small(cinfo, min_request + slop); - if (hdr_ptr != NULL) - break; - slop /= 2; - if (slop < MIN_SLOP) /* give up when it gets real small */ - out_of_memory(cinfo, 2); /* jpeg_get_small failed */ - } - mem->total_space_allocated += min_request + slop; - /* Success, initialize the new pool header and add to end of list */ - hdr_ptr->hdr.next = NULL; - hdr_ptr->hdr.bytes_used = 0; - hdr_ptr->hdr.bytes_left = sizeofobject + slop; - if (prev_hdr_ptr == NULL) /* first pool in class? */ - mem->small_list[pool_id] = hdr_ptr; - else - prev_hdr_ptr->hdr.next = hdr_ptr; - } - - /* OK, allocate the object from the current pool */ - data_ptr = (char *) (hdr_ptr + 1); /* point to first data byte in pool */ - data_ptr += hdr_ptr->hdr.bytes_used; /* point to place for object */ - hdr_ptr->hdr.bytes_used += sizeofobject; - hdr_ptr->hdr.bytes_left -= sizeofobject; - - return (void *) data_ptr; -} - - -/* - * Allocation of "large" objects. - * - * The external semantics of these are the same as "small" objects, - * except that FAR pointers are used on 80x86. However the pool - * management heuristics are quite different. We assume that each - * request is large enough that it may as well be passed directly to - * jpeg_get_large; the pool management just links everything together - * so that we can free it all on demand. - * Note: the major use of "large" objects is in JSAMPARRAY and JBLOCKARRAY - * structures. The routines that create these structures (see below) - * deliberately bunch rows together to ensure a large request size. - */ - -METHODDEF(void FAR *) -alloc_large (j_common_ptr cinfo, int pool_id, size_t sizeofobject) -/* Allocate a "large" object */ -{ - my_mem_ptr mem = (my_mem_ptr) cinfo->mem; - large_pool_ptr hdr_ptr; - size_t odd_bytes; - - /* Check for unsatisfiable request (do now to ensure no overflow below) */ - if (sizeofobject > (size_t) MAX_ALLOC_CHUNK - SIZEOF(large_pool_hdr)) - out_of_memory(cinfo, 3); /* request exceeds malloc's ability */ - - /* Round up the requested size to a multiple of SIZEOF(ALIGN_TYPE) */ - odd_bytes = sizeofobject % SIZEOF(ALIGN_TYPE); - if (odd_bytes > 0) - sizeofobject += SIZEOF(ALIGN_TYPE) - odd_bytes; - - /* Always make a new pool */ - if (pool_id < 0 || pool_id >= JPOOL_NUMPOOLS) - ERREXIT1(cinfo, JERR_BAD_POOL_ID, pool_id); /* safety check */ - - hdr_ptr = (large_pool_ptr) jpeg_get_large(cinfo, sizeofobject + - SIZEOF(large_pool_hdr)); - if (hdr_ptr == NULL) - out_of_memory(cinfo, 4); /* jpeg_get_large failed */ - mem->total_space_allocated += sizeofobject + SIZEOF(large_pool_hdr); - - /* Success, initialize the new pool header and add to list */ - hdr_ptr->hdr.next = mem->large_list[pool_id]; - /* We maintain space counts in each pool header for statistical purposes, - * even though they are not needed for allocation. - */ - hdr_ptr->hdr.bytes_used = sizeofobject; - hdr_ptr->hdr.bytes_left = 0; - mem->large_list[pool_id] = hdr_ptr; - - return (void FAR *) (hdr_ptr + 1); /* point to first data byte in pool */ -} - - -/* - * Creation of 2-D sample arrays. - * The pointers are in near heap, the samples themselves in FAR heap. - * - * To minimize allocation overhead and to allow I/O of large contiguous - * blocks, we allocate the sample rows in groups of as many rows as possible - * without exceeding MAX_ALLOC_CHUNK total bytes per allocation request. - * NB: the virtual array control routines, later in this file, know about - * this chunking of rows. The rowsperchunk value is left in the mem manager - * object so that it can be saved away if this sarray is the workspace for - * a virtual array. - */ - -METHODDEF(JSAMPARRAY) -alloc_sarray (j_common_ptr cinfo, int pool_id, - JDIMENSION samplesperrow, JDIMENSION numrows) -/* Allocate a 2-D sample array */ -{ - my_mem_ptr mem = (my_mem_ptr) cinfo->mem; - JSAMPARRAY result; - JSAMPROW workspace; - JDIMENSION rowsperchunk, currow, i; - long ltemp; - - /* Calculate max # of rows allowed in one allocation chunk */ - ltemp = (MAX_ALLOC_CHUNK - SIZEOF(large_pool_hdr)) / - ((long) samplesperrow * SIZEOF(JSAMPLE)); - if (ltemp <= 0) - ERREXIT(cinfo, JERR_WIDTH_OVERFLOW); - if (ltemp < (long) numrows) - rowsperchunk = (JDIMENSION) ltemp; - else - rowsperchunk = numrows; - mem->last_rowsperchunk = rowsperchunk; - - /* Get space for row pointers (small object) */ - result = (JSAMPARRAY) alloc_small(cinfo, pool_id, - (size_t) numrows * SIZEOF(JSAMPROW)); - - /* Get the rows themselves (large objects) */ - currow = 0; - while (currow < numrows) { - rowsperchunk = MIN(rowsperchunk, numrows - currow); - workspace = (JSAMPROW) alloc_large(cinfo, pool_id, - (size_t) rowsperchunk * (size_t) samplesperrow * SIZEOF(JSAMPLE)); - for (i = rowsperchunk; i > 0; i--) { - result[currow++] = workspace; - workspace += samplesperrow; - } - } - - return result; -} - - -/* - * Creation of 2-D coefficient-block arrays. - * This is essentially the same as the code for sample arrays, above. - */ - -METHODDEF(JBLOCKARRAY) -alloc_barray (j_common_ptr cinfo, int pool_id, - JDIMENSION blocksperrow, JDIMENSION numrows) -/* Allocate a 2-D coefficient-block array */ -{ - my_mem_ptr mem = (my_mem_ptr) cinfo->mem; - JBLOCKARRAY result; - JBLOCKROW workspace; - JDIMENSION rowsperchunk, currow, i; - long ltemp; - - /* Calculate max # of rows allowed in one allocation chunk */ - ltemp = (MAX_ALLOC_CHUNK - SIZEOF(large_pool_hdr)) / - ((long) blocksperrow * SIZEOF(JBLOCK)); - if (ltemp <= 0) - ERREXIT(cinfo, JERR_WIDTH_OVERFLOW); - if (ltemp < (long) numrows) - rowsperchunk = (JDIMENSION) ltemp; - else - rowsperchunk = numrows; - mem->last_rowsperchunk = rowsperchunk; - - /* Get space for row pointers (small object) */ - result = (JBLOCKARRAY) alloc_small(cinfo, pool_id, - (size_t) numrows * SIZEOF(JBLOCKROW)); - - /* Get the rows themselves (large objects) */ - currow = 0; - while (currow < numrows) { - rowsperchunk = MIN(rowsperchunk, numrows - currow); - workspace = (JBLOCKROW) alloc_large(cinfo, pool_id, - (size_t) rowsperchunk * (size_t) blocksperrow * SIZEOF(JBLOCK)); - for (i = rowsperchunk; i > 0; i--) { - result[currow++] = workspace; - workspace += blocksperrow; - } - } - - return result; -} - - -/* - * About virtual array management: - * - * The above "normal" array routines are only used to allocate strip buffers - * (as wide as the image, but just a few rows high). Full-image-sized buffers - * are handled as "virtual" arrays. The array is still accessed a strip at a - * time, but the memory manager must save the whole array for repeated - * accesses. The intended implementation is that there is a strip buffer in - * memory (as high as is possible given the desired memory limit), plus a - * backing file that holds the rest of the array. - * - * The request_virt_array routines are told the total size of the image and - * the maximum number of rows that will be accessed at once. The in-memory - * buffer must be at least as large as the maxaccess value. - * - * The request routines create control blocks but not the in-memory buffers. - * That is postponed until realize_virt_arrays is called. At that time the - * total amount of space needed is known (approximately, anyway), so free - * memory can be divided up fairly. - * - * The access_virt_array routines are responsible for making a specific strip - * area accessible (after reading or writing the backing file, if necessary). - * Note that the access routines are told whether the caller intends to modify - * the accessed strip; during a read-only pass this saves having to rewrite - * data to disk. The access routines are also responsible for pre-zeroing - * any newly accessed rows, if pre-zeroing was requested. - * - * In current usage, the access requests are usually for nonoverlapping - * strips; that is, successive access start_row numbers differ by exactly - * num_rows = maxaccess. This means we can get good performance with simple - * buffer dump/reload logic, by making the in-memory buffer be a multiple - * of the access height; then there will never be accesses across bufferload - * boundaries. The code will still work with overlapping access requests, - * but it doesn't handle bufferload overlaps very efficiently. - */ - - -METHODDEF(jvirt_sarray_ptr) -request_virt_sarray (j_common_ptr cinfo, int pool_id, boolean pre_zero, - JDIMENSION samplesperrow, JDIMENSION numrows, - JDIMENSION maxaccess) -/* Request a virtual 2-D sample array */ -{ - my_mem_ptr mem = (my_mem_ptr) cinfo->mem; - jvirt_sarray_ptr result; - - /* Only IMAGE-lifetime virtual arrays are currently supported */ - if (pool_id != JPOOL_IMAGE) - ERREXIT1(cinfo, JERR_BAD_POOL_ID, pool_id); /* safety check */ - - /* get control block */ - result = (jvirt_sarray_ptr) alloc_small(cinfo, pool_id, - SIZEOF(struct jvirt_sarray_control)); - - result->mem_buffer = NULL; /* marks array not yet realized */ - result->rows_in_array = numrows; - result->samplesperrow = samplesperrow; - result->maxaccess = maxaccess; - result->pre_zero = pre_zero; - result->b_s_open = FALSE; /* no associated backing-store object */ - result->next = mem->virt_sarray_list; /* add to list of virtual arrays */ - mem->virt_sarray_list = result; - - return result; -} - - -METHODDEF(jvirt_barray_ptr) -request_virt_barray (j_common_ptr cinfo, int pool_id, boolean pre_zero, - JDIMENSION blocksperrow, JDIMENSION numrows, - JDIMENSION maxaccess) -/* Request a virtual 2-D coefficient-block array */ -{ - my_mem_ptr mem = (my_mem_ptr) cinfo->mem; - jvirt_barray_ptr result; - - /* Only IMAGE-lifetime virtual arrays are currently supported */ - if (pool_id != JPOOL_IMAGE) - ERREXIT1(cinfo, JERR_BAD_POOL_ID, pool_id); /* safety check */ - - /* get control block */ - result = (jvirt_barray_ptr) alloc_small(cinfo, pool_id, - SIZEOF(struct jvirt_barray_control)); - - result->mem_buffer = NULL; /* marks array not yet realized */ - result->rows_in_array = numrows; - result->blocksperrow = blocksperrow; - result->maxaccess = maxaccess; - result->pre_zero = pre_zero; - result->b_s_open = FALSE; /* no associated backing-store object */ - result->next = mem->virt_barray_list; /* add to list of virtual arrays */ - mem->virt_barray_list = result; - - return result; -} - - -METHODDEF(void) -realize_virt_arrays (j_common_ptr cinfo) -/* Allocate the in-memory buffers for any unrealized virtual arrays */ -{ - my_mem_ptr mem = (my_mem_ptr) cinfo->mem; - long bytesperrow, space_per_minheight, maximum_space; - long avail_mem, minheights, max_minheights; - jvirt_sarray_ptr sptr; - jvirt_barray_ptr bptr; - - /* Compute the minimum space needed (maxaccess rows in each buffer) - * and the maximum space needed (full image height in each buffer). - * These may be of use to the system-dependent jpeg_mem_available routine. - */ - space_per_minheight = 0; - maximum_space = 0; - for (sptr = mem->virt_sarray_list; sptr != NULL; sptr = sptr->next) { - if (sptr->mem_buffer == NULL) { /* if not realized yet */ - bytesperrow = (long) sptr->samplesperrow * SIZEOF(JSAMPLE); - space_per_minheight += (long) sptr->maxaccess * bytesperrow; - maximum_space += (long) sptr->rows_in_array * bytesperrow; - } - } - for (bptr = mem->virt_barray_list; bptr != NULL; bptr = bptr->next) { - if (bptr->mem_buffer == NULL) { /* if not realized yet */ - bytesperrow = (long) bptr->blocksperrow * SIZEOF(JBLOCK); - space_per_minheight += (long) bptr->maxaccess * bytesperrow; - maximum_space += (long) bptr->rows_in_array * bytesperrow; - } - } - - if (space_per_minheight <= 0) - return; /* no unrealized arrays, no work */ - - /* Determine amount of memory to actually use; this is system-dependent. */ - avail_mem = jpeg_mem_available(cinfo, space_per_minheight, maximum_space, - (long) mem->total_space_allocated); - - /* If the maximum space needed is available, make all the buffers full - * height; otherwise parcel it out with the same number of minheights - * in each buffer. - */ - if (avail_mem >= maximum_space) - max_minheights = 1000000000L; - else { - max_minheights = avail_mem / space_per_minheight; - /* If there doesn't seem to be enough space, try to get the minimum - * anyway. This allows a "stub" implementation of jpeg_mem_available(). - */ - if (max_minheights <= 0) - max_minheights = 1; - } - - /* Allocate the in-memory buffers and initialize backing store as needed. */ - - for (sptr = mem->virt_sarray_list; sptr != NULL; sptr = sptr->next) { - if (sptr->mem_buffer == NULL) { /* if not realized yet */ - minheights = ((long) sptr->rows_in_array - 1L) / sptr->maxaccess + 1L; - if (minheights <= max_minheights) { - /* This buffer fits in memory */ - sptr->rows_in_mem = sptr->rows_in_array; - } else { - /* It doesn't fit in memory, create backing store. */ - sptr->rows_in_mem = (JDIMENSION) (max_minheights * sptr->maxaccess); - jpeg_open_backing_store(cinfo, & sptr->b_s_info, - (long) sptr->rows_in_array * - (long) sptr->samplesperrow * - (long) SIZEOF(JSAMPLE)); - sptr->b_s_open = TRUE; - } - sptr->mem_buffer = alloc_sarray(cinfo, JPOOL_IMAGE, - sptr->samplesperrow, sptr->rows_in_mem); - sptr->rowsperchunk = mem->last_rowsperchunk; - sptr->cur_start_row = 0; - sptr->first_undef_row = 0; - sptr->dirty = FALSE; - } - } - - for (bptr = mem->virt_barray_list; bptr != NULL; bptr = bptr->next) { - if (bptr->mem_buffer == NULL) { /* if not realized yet */ - minheights = ((long) bptr->rows_in_array - 1L) / bptr->maxaccess + 1L; - if (minheights <= max_minheights) { - /* This buffer fits in memory */ - bptr->rows_in_mem = bptr->rows_in_array; - } else { - /* It doesn't fit in memory, create backing store. */ - bptr->rows_in_mem = (JDIMENSION) (max_minheights * bptr->maxaccess); - jpeg_open_backing_store(cinfo, & bptr->b_s_info, - (long) bptr->rows_in_array * - (long) bptr->blocksperrow * - (long) SIZEOF(JBLOCK)); - bptr->b_s_open = TRUE; - } - bptr->mem_buffer = alloc_barray(cinfo, JPOOL_IMAGE, - bptr->blocksperrow, bptr->rows_in_mem); - bptr->rowsperchunk = mem->last_rowsperchunk; - bptr->cur_start_row = 0; - bptr->first_undef_row = 0; - bptr->dirty = FALSE; - } - } -} - - -LOCAL(void) -do_sarray_io (j_common_ptr cinfo, jvirt_sarray_ptr ptr, boolean writing) -/* Do backing store read or write of a virtual sample array */ -{ - long bytesperrow, file_offset, byte_count, rows, thisrow, i; - - bytesperrow = (long) ptr->samplesperrow * SIZEOF(JSAMPLE); - file_offset = (long) ptr->cur_start_row * bytesperrow; - /* Loop to read or write each allocation chunk in mem_buffer */ - for (i = 0; i < (long) ptr->rows_in_mem; i += ptr->rowsperchunk) { - /* One chunk, but check for short chunk at end of buffer */ - rows = MIN((long) ptr->rowsperchunk, (long) ptr->rows_in_mem - i); - /* Transfer no more than is currently defined */ - thisrow = (long) ptr->cur_start_row + i; - rows = MIN(rows, (long) ptr->first_undef_row - thisrow); - /* Transfer no more than fits in file */ - rows = MIN(rows, (long) ptr->rows_in_array - thisrow); - if (rows <= 0) /* this chunk might be past end of file! */ - break; - byte_count = rows * bytesperrow; - if (writing) - (*ptr->b_s_info.write_backing_store) (cinfo, & ptr->b_s_info, - (void FAR *) ptr->mem_buffer[i], - file_offset, byte_count); - else - (*ptr->b_s_info.read_backing_store) (cinfo, & ptr->b_s_info, - (void FAR *) ptr->mem_buffer[i], - file_offset, byte_count); - file_offset += byte_count; - } -} - - -LOCAL(void) -do_barray_io (j_common_ptr cinfo, jvirt_barray_ptr ptr, boolean writing) -/* Do backing store read or write of a virtual coefficient-block array */ -{ - long bytesperrow, file_offset, byte_count, rows, thisrow, i; - - bytesperrow = (long) ptr->blocksperrow * SIZEOF(JBLOCK); - file_offset = (long) ptr->cur_start_row * bytesperrow; - /* Loop to read or write each allocation chunk in mem_buffer */ - for (i = 0; i < (long) ptr->rows_in_mem; i += ptr->rowsperchunk) { - /* One chunk, but check for short chunk at end of buffer */ - rows = MIN((long) ptr->rowsperchunk, (long) ptr->rows_in_mem - i); - /* Transfer no more than is currently defined */ - thisrow = (long) ptr->cur_start_row + i; - rows = MIN(rows, (long) ptr->first_undef_row - thisrow); - /* Transfer no more than fits in file */ - rows = MIN(rows, (long) ptr->rows_in_array - thisrow); - if (rows <= 0) /* this chunk might be past end of file! */ - break; - byte_count = rows * bytesperrow; - if (writing) - (*ptr->b_s_info.write_backing_store) (cinfo, & ptr->b_s_info, - (void FAR *) ptr->mem_buffer[i], - file_offset, byte_count); - else - (*ptr->b_s_info.read_backing_store) (cinfo, & ptr->b_s_info, - (void FAR *) ptr->mem_buffer[i], - file_offset, byte_count); - file_offset += byte_count; - } -} - - -METHODDEF(JSAMPARRAY) -access_virt_sarray (j_common_ptr cinfo, jvirt_sarray_ptr ptr, - JDIMENSION start_row, JDIMENSION num_rows, - boolean writable) -/* Access the part of a virtual sample array starting at start_row */ -/* and extending for num_rows rows. writable is true if */ -/* caller intends to modify the accessed area. */ -{ - JDIMENSION end_row = start_row + num_rows; - JDIMENSION undef_row; - - /* debugging check */ - if (end_row > ptr->rows_in_array || num_rows > ptr->maxaccess || - ptr->mem_buffer == NULL) - ERREXIT(cinfo, JERR_BAD_VIRTUAL_ACCESS); - - /* Make the desired part of the virtual array accessible */ - if (start_row < ptr->cur_start_row || - end_row > ptr->cur_start_row + ptr->rows_in_mem) { - if (! ptr->b_s_open) - ERREXIT(cinfo, JERR_VIRTUAL_BUG); - /* Flush old buffer contents if necessary */ - if (ptr->dirty) { - do_sarray_io(cinfo, ptr, TRUE); - ptr->dirty = FALSE; - } - /* Decide what part of virtual array to access. - * Algorithm: if target address > current window, assume forward scan, - * load starting at target address. If target address < current window, - * assume backward scan, load so that target area is top of window. - * Note that when switching from forward write to forward read, will have - * start_row = 0, so the limiting case applies and we load from 0 anyway. - */ - if (start_row > ptr->cur_start_row) { - ptr->cur_start_row = start_row; - } else { - /* use long arithmetic here to avoid overflow & unsigned problems */ - long ltemp; - - ltemp = (long) end_row - (long) ptr->rows_in_mem; - if (ltemp < 0) - ltemp = 0; /* don't fall off front end of file */ - ptr->cur_start_row = (JDIMENSION) ltemp; - } - /* Read in the selected part of the array. - * During the initial write pass, we will do no actual read - * because the selected part is all undefined. - */ - do_sarray_io(cinfo, ptr, FALSE); - } - /* Ensure the accessed part of the array is defined; prezero if needed. - * To improve locality of access, we only prezero the part of the array - * that the caller is about to access, not the entire in-memory array. - */ - if (ptr->first_undef_row < end_row) { - if (ptr->first_undef_row < start_row) { - if (writable) /* writer skipped over a section of array */ - ERREXIT(cinfo, JERR_BAD_VIRTUAL_ACCESS); - undef_row = start_row; /* but reader is allowed to read ahead */ - } else { - undef_row = ptr->first_undef_row; - } - if (writable) - ptr->first_undef_row = end_row; - if (ptr->pre_zero) { - size_t bytesperrow = (size_t) ptr->samplesperrow * SIZEOF(JSAMPLE); - undef_row -= ptr->cur_start_row; /* make indexes relative to buffer */ - end_row -= ptr->cur_start_row; - while (undef_row < end_row) { - FMEMZERO((void FAR *) ptr->mem_buffer[undef_row], bytesperrow); - undef_row++; - } - } else { - if (! writable) /* reader looking at undefined data */ - ERREXIT(cinfo, JERR_BAD_VIRTUAL_ACCESS); - } - } - /* Flag the buffer dirty if caller will write in it */ - if (writable) - ptr->dirty = TRUE; - /* Return address of proper part of the buffer */ - return ptr->mem_buffer + (start_row - ptr->cur_start_row); -} - - -METHODDEF(JBLOCKARRAY) -access_virt_barray (j_common_ptr cinfo, jvirt_barray_ptr ptr, - JDIMENSION start_row, JDIMENSION num_rows, - boolean writable) -/* Access the part of a virtual block array starting at start_row */ -/* and extending for num_rows rows. writable is true if */ -/* caller intends to modify the accessed area. */ -{ - JDIMENSION end_row = start_row + num_rows; - JDIMENSION undef_row; - - /* debugging check */ - if (end_row > ptr->rows_in_array || num_rows > ptr->maxaccess || - ptr->mem_buffer == NULL) - ERREXIT(cinfo, JERR_BAD_VIRTUAL_ACCESS); - - /* Make the desired part of the virtual array accessible */ - if (start_row < ptr->cur_start_row || - end_row > ptr->cur_start_row + ptr->rows_in_mem) { - if (! ptr->b_s_open) - ERREXIT(cinfo, JERR_VIRTUAL_BUG); - /* Flush old buffer contents if necessary */ - if (ptr->dirty) { - do_barray_io(cinfo, ptr, TRUE); - ptr->dirty = FALSE; - } - /* Decide what part of virtual array to access. - * Algorithm: if target address > current window, assume forward scan, - * load starting at target address. If target address < current window, - * assume backward scan, load so that target area is top of window. - * Note that when switching from forward write to forward read, will have - * start_row = 0, so the limiting case applies and we load from 0 anyway. - */ - if (start_row > ptr->cur_start_row) { - ptr->cur_start_row = start_row; - } else { - /* use long arithmetic here to avoid overflow & unsigned problems */ - long ltemp; - - ltemp = (long) end_row - (long) ptr->rows_in_mem; - if (ltemp < 0) - ltemp = 0; /* don't fall off front end of file */ - ptr->cur_start_row = (JDIMENSION) ltemp; - } - /* Read in the selected part of the array. - * During the initial write pass, we will do no actual read - * because the selected part is all undefined. - */ - do_barray_io(cinfo, ptr, FALSE); - } - /* Ensure the accessed part of the array is defined; prezero if needed. - * To improve locality of access, we only prezero the part of the array - * that the caller is about to access, not the entire in-memory array. - */ - if (ptr->first_undef_row < end_row) { - if (ptr->first_undef_row < start_row) { - if (writable) /* writer skipped over a section of array */ - ERREXIT(cinfo, JERR_BAD_VIRTUAL_ACCESS); - undef_row = start_row; /* but reader is allowed to read ahead */ - } else { - undef_row = ptr->first_undef_row; - } - if (writable) - ptr->first_undef_row = end_row; - if (ptr->pre_zero) { - size_t bytesperrow = (size_t) ptr->blocksperrow * SIZEOF(JBLOCK); - undef_row -= ptr->cur_start_row; /* make indexes relative to buffer */ - end_row -= ptr->cur_start_row; - while (undef_row < end_row) { - FMEMZERO((void FAR *) ptr->mem_buffer[undef_row], bytesperrow); - undef_row++; - } - } else { - if (! writable) /* reader looking at undefined data */ - ERREXIT(cinfo, JERR_BAD_VIRTUAL_ACCESS); - } - } - /* Flag the buffer dirty if caller will write in it */ - if (writable) - ptr->dirty = TRUE; - /* Return address of proper part of the buffer */ - return ptr->mem_buffer + (start_row - ptr->cur_start_row); -} - - -/* - * Release all objects belonging to a specified pool. - */ - -METHODDEF(void) -free_pool (j_common_ptr cinfo, int pool_id) -{ - my_mem_ptr mem = (my_mem_ptr) cinfo->mem; - small_pool_ptr shdr_ptr; - large_pool_ptr lhdr_ptr; - size_t space_freed; - - if (pool_id < 0 || pool_id >= JPOOL_NUMPOOLS) - ERREXIT1(cinfo, JERR_BAD_POOL_ID, pool_id); /* safety check */ - -#ifdef MEM_STATS - if (cinfo->err->trace_level > 1) - print_mem_stats(cinfo, pool_id); /* print pool's memory usage statistics */ -#endif - - /* If freeing IMAGE pool, close any virtual arrays first */ - if (pool_id == JPOOL_IMAGE) { - jvirt_sarray_ptr sptr; - jvirt_barray_ptr bptr; - - for (sptr = mem->virt_sarray_list; sptr != NULL; sptr = sptr->next) { - if (sptr->b_s_open) { /* there may be no backing store */ - sptr->b_s_open = FALSE; /* prevent recursive close if error */ - (*sptr->b_s_info.close_backing_store) (cinfo, & sptr->b_s_info); - } - } - mem->virt_sarray_list = NULL; - for (bptr = mem->virt_barray_list; bptr != NULL; bptr = bptr->next) { - if (bptr->b_s_open) { /* there may be no backing store */ - bptr->b_s_open = FALSE; /* prevent recursive close if error */ - (*bptr->b_s_info.close_backing_store) (cinfo, & bptr->b_s_info); - } - } - mem->virt_barray_list = NULL; - } - - /* Release large objects */ - lhdr_ptr = mem->large_list[pool_id]; - mem->large_list[pool_id] = NULL; - - while (lhdr_ptr != NULL) { - large_pool_ptr next_lhdr_ptr = lhdr_ptr->hdr.next; - space_freed = lhdr_ptr->hdr.bytes_used + - lhdr_ptr->hdr.bytes_left + - SIZEOF(large_pool_hdr); - jpeg_free_large(cinfo, (void FAR *) lhdr_ptr, space_freed); - mem->total_space_allocated -= space_freed; - lhdr_ptr = next_lhdr_ptr; - } - - /* Release small objects */ - shdr_ptr = mem->small_list[pool_id]; - mem->small_list[pool_id] = NULL; - - while (shdr_ptr != NULL) { - small_pool_ptr next_shdr_ptr = shdr_ptr->hdr.next; - space_freed = shdr_ptr->hdr.bytes_used + - shdr_ptr->hdr.bytes_left + - SIZEOF(small_pool_hdr); - jpeg_free_small(cinfo, (void *) shdr_ptr, space_freed); - mem->total_space_allocated -= space_freed; - shdr_ptr = next_shdr_ptr; - } -} - - -/* - * Close up shop entirely. - * Note that this cannot be called unless cinfo->mem is non-NULL. - */ - -METHODDEF(void) -self_destruct (j_common_ptr cinfo) -{ - int pool; - - /* Close all backing store, release all memory. - * Releasing pools in reverse order might help avoid fragmentation - * with some (brain-damaged) malloc libraries. - */ - for (pool = JPOOL_NUMPOOLS-1; pool >= JPOOL_PERMANENT; pool--) { - free_pool(cinfo, pool); - } - - /* Release the memory manager control block too. */ - jpeg_free_small(cinfo, (void *) cinfo->mem, SIZEOF(my_memory_mgr)); - cinfo->mem = NULL; /* ensures I will be called only once */ - - jpeg_mem_term(cinfo); /* system-dependent cleanup */ -} - - -/* - * Memory manager initialization. - * When this is called, only the error manager pointer is valid in cinfo! - */ - -GLOBAL(void) -jinit_memory_mgr (j_common_ptr cinfo) -{ - my_mem_ptr mem; - long max_to_use; - int pool; - size_t test_mac; - - cinfo->mem = NULL; /* for safety if init fails */ - - /* Check for configuration errors. - * SIZEOF(ALIGN_TYPE) should be a power of 2; otherwise, it probably - * doesn't reflect any real hardware alignment requirement. - * The test is a little tricky: for X>0, X and X-1 have no one-bits - * in common if and only if X is a power of 2, ie has only one one-bit. - * Some compilers may give an "unreachable code" warning here; ignore it. - */ - if ((SIZEOF(ALIGN_TYPE) & (SIZEOF(ALIGN_TYPE)-1)) != 0) - ERREXIT(cinfo, JERR_BAD_ALIGN_TYPE); - /* MAX_ALLOC_CHUNK must be representable as type size_t, and must be - * a multiple of SIZEOF(ALIGN_TYPE). - * Again, an "unreachable code" warning may be ignored here. - * But a "constant too large" warning means you need to fix MAX_ALLOC_CHUNK. - */ - test_mac = (size_t) MAX_ALLOC_CHUNK; - if ((long) test_mac != MAX_ALLOC_CHUNK || - (MAX_ALLOC_CHUNK % SIZEOF(ALIGN_TYPE)) != 0) - ERREXIT(cinfo, JERR_BAD_ALLOC_CHUNK); - - max_to_use = jpeg_mem_init(cinfo); /* system-dependent initialization */ - - /* Attempt to allocate memory manager's control block */ - mem = (my_mem_ptr) jpeg_get_small(cinfo, SIZEOF(my_memory_mgr)); - - if (mem == NULL) { - jpeg_mem_term(cinfo); /* system-dependent cleanup */ - ERREXIT1(cinfo, JERR_OUT_OF_MEMORY, 0); - } - - /* OK, fill in the method pointers */ - mem->pub.alloc_small = alloc_small; - mem->pub.alloc_large = alloc_large; - mem->pub.alloc_sarray = alloc_sarray; - mem->pub.alloc_barray = alloc_barray; - mem->pub.request_virt_sarray = request_virt_sarray; - mem->pub.request_virt_barray = request_virt_barray; - mem->pub.realize_virt_arrays = realize_virt_arrays; - mem->pub.access_virt_sarray = access_virt_sarray; - mem->pub.access_virt_barray = access_virt_barray; - mem->pub.free_pool = free_pool; - mem->pub.self_destruct = self_destruct; - - /* Make MAX_ALLOC_CHUNK accessible to other modules */ - mem->pub.max_alloc_chunk = MAX_ALLOC_CHUNK; - - /* Initialize working state */ - mem->pub.max_memory_to_use = max_to_use; - - for (pool = JPOOL_NUMPOOLS-1; pool >= JPOOL_PERMANENT; pool--) { - mem->small_list[pool] = NULL; - mem->large_list[pool] = NULL; - } - mem->virt_sarray_list = NULL; - mem->virt_barray_list = NULL; - - mem->total_space_allocated = SIZEOF(my_memory_mgr); - - /* Declare ourselves open for business */ - cinfo->mem = &mem->pub; - - /* Check for an environment variable JPEGMEM; if found, override the - * default max_memory setting from jpeg_mem_init. Note that the - * surrounding application may again override this value. - * If your system doesn't support getenv(), define NO_GETENV to disable - * this feature. - */ -#ifndef NO_GETENV - { char * memenv; - - if ((memenv = getenv("JPEGMEM")) != NULL) { - char ch = 'x'; - - if (sscanf(memenv, "%ld%c", &max_to_use, &ch) > 0) { - if (ch == 'm' || ch == 'M') - max_to_use *= 1000L; - mem->pub.max_memory_to_use = max_to_use * 1000L; - } - } - } -#endif - -} diff --git a/cmake/externals/libjpeg/jmemnobs.c b/cmake/externals/libjpeg/jmemnobs.c deleted file mode 100644 index a364cd822..000000000 --- a/cmake/externals/libjpeg/jmemnobs.c +++ /dev/null @@ -1,113 +0,0 @@ -/* - * jmemnobs.c - * - * Copyright (C) 1992-1996, Thomas G. Lane. - * Modified 2019 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file provides a really simple implementation of the system- - * dependent portion of the JPEG memory manager. This implementation - * assumes that no backing-store files are needed: all required space - * can be obtained from malloc(). - * This is very portable in the sense that it'll compile on almost anything, - * but you'd better have lots of main memory (or virtual memory) if you want - * to process big images. - * Note that the max_memory_to_use option is respected by this implementation. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" -#include "jmemsys.h" /* import the system-dependent declarations */ - -#ifndef HAVE_STDLIB_H /* should declare malloc(),free() */ -extern void * malloc JPP((size_t size)); -extern void free JPP((void *ptr)); -#endif - - -/* - * Memory allocation and freeing are controlled by the regular library - * routines malloc() and free(). - */ - -GLOBAL(void *) -jpeg_get_small (j_common_ptr cinfo, size_t sizeofobject) -{ - return (void *) malloc(sizeofobject); -} - -GLOBAL(void) -jpeg_free_small (j_common_ptr cinfo, void * object, size_t sizeofobject) -{ - free(object); -} - - -/* - * "Large" objects are treated the same as "small" ones. - * NB: although we include FAR keywords in the routine declarations, - * this file won't actually work in 80x86 small/medium model; at least, - * you probably won't be able to process useful-size images in only 64KB. - */ - -GLOBAL(void FAR *) -jpeg_get_large (j_common_ptr cinfo, size_t sizeofobject) -{ - return (void FAR *) malloc(sizeofobject); -} - -GLOBAL(void) -jpeg_free_large (j_common_ptr cinfo, void FAR * object, size_t sizeofobject) -{ - free(object); -} - - -/* - * This routine computes the total memory space available for allocation. - */ - -GLOBAL(long) -jpeg_mem_available (j_common_ptr cinfo, long min_bytes_needed, - long max_bytes_needed, long already_allocated) -{ - if (cinfo->mem->max_memory_to_use) - return cinfo->mem->max_memory_to_use - already_allocated; - - /* Here we say, "we got all you want bud!" */ - return max_bytes_needed; -} - - -/* - * Backing store (temporary file) management. - * Since jpeg_mem_available always promised the moon, - * this should never be called and we can just error out. - */ - -GLOBAL(void) -jpeg_open_backing_store (j_common_ptr cinfo, backing_store_ptr info, - long total_bytes_needed) -{ - ERREXIT(cinfo, JERR_NO_BACKING_STORE); -} - - -/* - * These routines take care of any system-dependent initialization and - * cleanup required. Here, there isn't any. - */ - -GLOBAL(long) -jpeg_mem_init (j_common_ptr cinfo) -{ - return 0; /* just set max_memory_to_use to 0 */ -} - -GLOBAL(void) -jpeg_mem_term (j_common_ptr cinfo) -{ - /* no work */ -} diff --git a/cmake/externals/libjpeg/jmemsys.h b/cmake/externals/libjpeg/jmemsys.h deleted file mode 100644 index 6c3c6d348..000000000 --- a/cmake/externals/libjpeg/jmemsys.h +++ /dev/null @@ -1,198 +0,0 @@ -/* - * jmemsys.h - * - * Copyright (C) 1992-1997, Thomas G. Lane. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This include file defines the interface between the system-independent - * and system-dependent portions of the JPEG memory manager. No other - * modules need include it. (The system-independent portion is jmemmgr.c; - * there are several different versions of the system-dependent portion.) - * - * This file works as-is for the system-dependent memory managers supplied - * in the IJG distribution. You may need to modify it if you write a - * custom memory manager. If system-dependent changes are needed in - * this file, the best method is to #ifdef them based on a configuration - * symbol supplied in jconfig.h, as we have done with USE_MSDOS_MEMMGR - * and USE_MAC_MEMMGR. - */ - - -/* Short forms of external names for systems with brain-damaged linkers. */ - -#ifdef NEED_SHORT_EXTERNAL_NAMES -#define jpeg_get_small jGetSmall -#define jpeg_free_small jFreeSmall -#define jpeg_get_large jGetLarge -#define jpeg_free_large jFreeLarge -#define jpeg_mem_available jMemAvail -#define jpeg_open_backing_store jOpenBackStore -#define jpeg_mem_init jMemInit -#define jpeg_mem_term jMemTerm -#endif /* NEED_SHORT_EXTERNAL_NAMES */ - - -/* - * These two functions are used to allocate and release small chunks of - * memory. (Typically the total amount requested through jpeg_get_small is - * no more than 20K or so; this will be requested in chunks of a few K each.) - * Behavior should be the same as for the standard library functions malloc - * and free; in particular, jpeg_get_small must return NULL on failure. - * On most systems, these ARE malloc and free. jpeg_free_small is passed the - * size of the object being freed, just in case it's needed. - * On an 80x86 machine using small-data memory model, these manage near heap. - */ - -EXTERN(void *) jpeg_get_small JPP((j_common_ptr cinfo, size_t sizeofobject)); -EXTERN(void) jpeg_free_small JPP((j_common_ptr cinfo, void * object, - size_t sizeofobject)); - -/* - * These two functions are used to allocate and release large chunks of - * memory (up to the total free space designated by jpeg_mem_available). - * The interface is the same as above, except that on an 80x86 machine, - * far pointers are used. On most other machines these are identical to - * the jpeg_get/free_small routines; but we keep them separate anyway, - * in case a different allocation strategy is desirable for large chunks. - */ - -EXTERN(void FAR *) jpeg_get_large JPP((j_common_ptr cinfo, - size_t sizeofobject)); -EXTERN(void) jpeg_free_large JPP((j_common_ptr cinfo, void FAR * object, - size_t sizeofobject)); - -/* - * The macro MAX_ALLOC_CHUNK designates the maximum number of bytes that may - * be requested in a single call to jpeg_get_large (and jpeg_get_small for that - * matter, but that case should never come into play). This macro is needed - * to model the 64Kb-segment-size limit of far addressing on 80x86 machines. - * On those machines, we expect that jconfig.h will provide a proper value. - * On machines with 32-bit flat address spaces, any large constant may be used. - * - * NB: jmemmgr.c expects that MAX_ALLOC_CHUNK will be representable as type - * size_t and will be a multiple of sizeof(align_type). - */ - -#ifndef MAX_ALLOC_CHUNK /* may be overridden in jconfig.h */ -#define MAX_ALLOC_CHUNK 1000000000L -#endif - -/* - * This routine computes the total space still available for allocation by - * jpeg_get_large. If more space than this is needed, backing store will be - * used. NOTE: any memory already allocated must not be counted. - * - * There is a minimum space requirement, corresponding to the minimum - * feasible buffer sizes; jmemmgr.c will request that much space even if - * jpeg_mem_available returns zero. The maximum space needed, enough to hold - * all working storage in memory, is also passed in case it is useful. - * Finally, the total space already allocated is passed. If no better - * method is available, cinfo->mem->max_memory_to_use - already_allocated - * is often a suitable calculation. - * - * It is OK for jpeg_mem_available to underestimate the space available - * (that'll just lead to more backing-store access than is really necessary). - * However, an overestimate will lead to failure. Hence it's wise to subtract - * a slop factor from the true available space. 5% should be enough. - * - * On machines with lots of virtual memory, any large constant may be returned. - * Conversely, zero may be returned to always use the minimum amount of memory. - */ - -EXTERN(long) jpeg_mem_available JPP((j_common_ptr cinfo, - long min_bytes_needed, - long max_bytes_needed, - long already_allocated)); - - -/* - * This structure holds whatever state is needed to access a single - * backing-store object. The read/write/close method pointers are called - * by jmemmgr.c to manipulate the backing-store object; all other fields - * are private to the system-dependent backing store routines. - */ - -#define TEMP_NAME_LENGTH 64 /* max length of a temporary file's name */ - - -#ifdef USE_MSDOS_MEMMGR /* DOS-specific junk */ - -typedef unsigned short XMSH; /* type of extended-memory handles */ -typedef unsigned short EMSH; /* type of expanded-memory handles */ - -typedef union { - short file_handle; /* DOS file handle if it's a temp file */ - XMSH xms_handle; /* handle if it's a chunk of XMS */ - EMSH ems_handle; /* handle if it's a chunk of EMS */ -} handle_union; - -#endif /* USE_MSDOS_MEMMGR */ - -#ifdef USE_MAC_MEMMGR /* Mac-specific junk */ -#include -#endif /* USE_MAC_MEMMGR */ - - -typedef struct backing_store_struct * backing_store_ptr; - -typedef struct backing_store_struct { - /* Methods for reading/writing/closing this backing-store object */ - JMETHOD(void, read_backing_store, (j_common_ptr cinfo, - backing_store_ptr info, - void FAR * buffer_address, - long file_offset, long byte_count)); - JMETHOD(void, write_backing_store, (j_common_ptr cinfo, - backing_store_ptr info, - void FAR * buffer_address, - long file_offset, long byte_count)); - JMETHOD(void, close_backing_store, (j_common_ptr cinfo, - backing_store_ptr info)); - - /* Private fields for system-dependent backing-store management */ -#ifdef USE_MSDOS_MEMMGR - /* For the MS-DOS manager (jmemdos.c), we need: */ - handle_union handle; /* reference to backing-store storage object */ - char temp_name[TEMP_NAME_LENGTH]; /* name if it's a file */ -#else -#ifdef USE_MAC_MEMMGR - /* For the Mac manager (jmemmac.c), we need: */ - short temp_file; /* file reference number to temp file */ - FSSpec tempSpec; /* the FSSpec for the temp file */ - char temp_name[TEMP_NAME_LENGTH]; /* name if it's a file */ -#else - /* For a typical implementation with temp files, we need: */ - FILE * temp_file; /* stdio reference to temp file */ - char temp_name[TEMP_NAME_LENGTH]; /* name of temp file */ -#endif -#endif -} backing_store_info; - - -/* - * Initial opening of a backing-store object. This must fill in the - * read/write/close pointers in the object. The read/write routines - * may take an error exit if the specified maximum file size is exceeded. - * (If jpeg_mem_available always returns a large value, this routine can - * just take an error exit.) - */ - -EXTERN(void) jpeg_open_backing_store JPP((j_common_ptr cinfo, - backing_store_ptr info, - long total_bytes_needed)); - - -/* - * These routines take care of any system-dependent initialization and - * cleanup required. jpeg_mem_init will be called before anything is - * allocated (and, therefore, nothing in cinfo is of use except the error - * manager pointer). It should return a suitable default value for - * max_memory_to_use; this may subsequently be overridden by the surrounding - * application. (Note that max_memory_to_use is only important if - * jpeg_mem_available chooses to consult it ... no one else will.) - * jpeg_mem_term may assume that all requested memory has been freed and that - * all opened backing-store objects have been closed. - */ - -EXTERN(long) jpeg_mem_init JPP((j_common_ptr cinfo)); -EXTERN(void) jpeg_mem_term JPP((j_common_ptr cinfo)); diff --git a/cmake/externals/libjpeg/jmorecfg.h b/cmake/externals/libjpeg/jmorecfg.h deleted file mode 100644 index 4638d6af2..000000000 --- a/cmake/externals/libjpeg/jmorecfg.h +++ /dev/null @@ -1,457 +0,0 @@ -/* - * jmorecfg.h - * - * Copyright (C) 1991-1997, Thomas G. Lane. - * Modified 1997-2022 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains additional configuration options that customize the - * JPEG software for special applications or support machine-dependent - * optimizations. Most users will not need to touch this file. - */ - - -/* - * Define BITS_IN_JSAMPLE as either - * 8 for 8-bit sample values (the usual setting) - * 9 for 9-bit sample values - * 10 for 10-bit sample values - * 11 for 11-bit sample values - * 12 for 12-bit sample values - * Only 8, 9, 10, 11, and 12 bits sample data precision are supported for - * full-feature DCT processing. Further depths up to 16-bit may be added - * later for the lossless modes of operation. - * Run-time selection and conversion of data precision will be added later - * and are currently not supported, sorry. - * Exception: The transcoding part (jpegtran) supports all settings in a - * single instance, since it operates on the level of DCT coefficients and - * not sample values. The DCT coefficients are of the same type (16 bits) - * in all cases (see below). - */ - -#define BITS_IN_JSAMPLE 8 /* use 8, 9, 10, 11, or 12 */ - - -/* - * Maximum number of components (color channels) allowed in JPEG image. - * To meet the letter of the JPEG spec, set this to 255. However, darn - * few applications need more than 4 channels (maybe 5 for CMYK + alpha - * mask). We recommend 10 as a reasonable compromise; use 4 if you are - * really short on memory. (Each allowed component costs a hundred or so - * bytes of storage, whether actually used in an image or not.) - */ - -#define MAX_COMPONENTS 10 /* maximum number of image components */ - - -/* - * Basic data types. - * You may need to change these if you have a machine with unusual data - * type sizes; for example, "char" not 8 bits, "short" not 16 bits, - * or "long" not 32 bits. We don't care whether "int" is 16 or 32 bits, - * but it had better be at least 16. - */ - -/* Representation of a single sample (pixel element value). - * We frequently allocate large arrays of these, so it's important to keep - * them small. But if you have memory to burn and access to char or short - * arrays is very slow on your hardware, you might want to change these. - */ - -#if BITS_IN_JSAMPLE == 8 -/* JSAMPLE should be the smallest type that will hold the values 0..255. - * You can use a signed char by having GETJSAMPLE mask it with 0xFF. - */ - -#ifdef HAVE_UNSIGNED_CHAR - -typedef unsigned char JSAMPLE; -#define GETJSAMPLE(value) ((int) (value)) - -#else /* not HAVE_UNSIGNED_CHAR */ - -typedef char JSAMPLE; -#ifdef CHAR_IS_UNSIGNED -#define GETJSAMPLE(value) ((int) (value)) -#else -#define GETJSAMPLE(value) ((int) (value) & 0xFF) -#endif /* CHAR_IS_UNSIGNED */ - -#endif /* HAVE_UNSIGNED_CHAR */ - -#define MAXJSAMPLE 255 -#define CENTERJSAMPLE 128 - -#endif /* BITS_IN_JSAMPLE == 8 */ - - -#if BITS_IN_JSAMPLE == 9 -/* JSAMPLE should be the smallest type that will hold the values 0..511. - * On nearly all machines "short" will do nicely. - */ - -typedef short JSAMPLE; -#define GETJSAMPLE(value) ((int) (value)) - -#define MAXJSAMPLE 511 -#define CENTERJSAMPLE 256 - -#endif /* BITS_IN_JSAMPLE == 9 */ - - -#if BITS_IN_JSAMPLE == 10 -/* JSAMPLE should be the smallest type that will hold the values 0..1023. - * On nearly all machines "short" will do nicely. - */ - -typedef short JSAMPLE; -#define GETJSAMPLE(value) ((int) (value)) - -#define MAXJSAMPLE 1023 -#define CENTERJSAMPLE 512 - -#endif /* BITS_IN_JSAMPLE == 10 */ - - -#if BITS_IN_JSAMPLE == 11 -/* JSAMPLE should be the smallest type that will hold the values 0..2047. - * On nearly all machines "short" will do nicely. - */ - -typedef short JSAMPLE; -#define GETJSAMPLE(value) ((int) (value)) - -#define MAXJSAMPLE 2047 -#define CENTERJSAMPLE 1024 - -#endif /* BITS_IN_JSAMPLE == 11 */ - - -#if BITS_IN_JSAMPLE == 12 -/* JSAMPLE should be the smallest type that will hold the values 0..4095. - * On nearly all machines "short" will do nicely. - */ - -typedef short JSAMPLE; -#define GETJSAMPLE(value) ((int) (value)) - -#define MAXJSAMPLE 4095 -#define CENTERJSAMPLE 2048 - -#endif /* BITS_IN_JSAMPLE == 12 */ - - -/* Representation of a DCT frequency coefficient. - * This should be a signed value of at least 16 bits; "short" is usually OK. - * Again, we allocate large arrays of these, but you can change to int - * if you have memory to burn and "short" is really slow. - */ - -typedef short JCOEF; - - -/* Compressed datastreams are represented as arrays of JOCTET. - * These must be EXACTLY 8 bits wide, at least once they are written to - * external storage. Note that when using the stdio data source/destination - * managers, this is also the data type passed to fread/fwrite. - */ - -#ifdef HAVE_UNSIGNED_CHAR - -typedef unsigned char JOCTET; -#define GETJOCTET(value) (value) - -#else /* not HAVE_UNSIGNED_CHAR */ - -typedef char JOCTET; -#ifdef CHAR_IS_UNSIGNED -#define GETJOCTET(value) (value) -#else -#define GETJOCTET(value) ((value) & 0xFF) -#endif /* CHAR_IS_UNSIGNED */ - -#endif /* HAVE_UNSIGNED_CHAR */ - - -/* These typedefs are used for various table entries and so forth. - * They must be at least as wide as specified; but making them too big - * won't cost a huge amount of memory, so we don't provide special - * extraction code like we did for JSAMPLE. (In other words, these - * typedefs live at a different point on the speed/space tradeoff curve.) - */ - -/* UINT8 must hold at least the values 0..255. */ - -#ifdef HAVE_UNSIGNED_CHAR -typedef unsigned char UINT8; -#else /* not HAVE_UNSIGNED_CHAR */ -#ifdef CHAR_IS_UNSIGNED -typedef char UINT8; -#else /* not CHAR_IS_UNSIGNED */ -typedef short UINT8; -#endif /* CHAR_IS_UNSIGNED */ -#endif /* HAVE_UNSIGNED_CHAR */ - -/* UINT16 must hold at least the values 0..65535. */ - -#ifdef HAVE_UNSIGNED_SHORT -typedef unsigned short UINT16; -#else /* not HAVE_UNSIGNED_SHORT */ -typedef unsigned int UINT16; -#endif /* HAVE_UNSIGNED_SHORT */ - -/* INT16 must hold at least the values -32768..32767. */ - -#ifndef XMD_H /* X11/xmd.h correctly defines INT16 */ -typedef short INT16; -#endif - -/* INT32 must hold at least signed 32-bit values. */ - -#ifndef XMD_H /* X11/xmd.h correctly defines INT32 */ -#ifndef _BASETSD_H_ /* Microsoft defines it in basetsd.h */ -#ifndef _BASETSD_H /* MinGW is slightly different */ -#ifndef QGLOBAL_H /* Qt defines it in qglobal.h */ -typedef long INT32; -#endif -#endif -#endif -#endif - -/* Datatype used for image dimensions. The JPEG standard only supports - * images up to 64K*64K due to 16-bit fields in SOF markers. Therefore - * "unsigned int" is sufficient on all machines. However, if you need to - * handle larger images and you don't mind deviating from the spec, you - * can change this datatype. - */ - -typedef unsigned int JDIMENSION; - -#define JPEG_MAX_DIMENSION 65500L /* a tad under 64K to prevent overflows */ - - -/* These macros are used in all function definitions and extern declarations. - * You could modify them if you need to change function linkage conventions; - * in particular, you'll need to do that to make the library a Windows DLL. - * Another application is to make all functions global for use with debuggers - * or code profilers that require it. - */ - -/* a function called through method pointers: */ -#define METHODDEF(type) static type -/* a function used only in its module: */ -#define LOCAL(type) static type -/* a function referenced thru EXTERNs: */ -#define GLOBAL(type) type -/* a reference to a GLOBAL function: */ -#define EXTERN(type) extern type - - -/* This macro is used to declare a "method", that is, a function pointer. - * We want to supply prototype parameters if the compiler can cope. - * Note that the arglist parameter must be parenthesized! - * Again, you can customize this if you need special linkage keywords. - */ - -#ifdef HAVE_PROTOTYPES -#define JMETHOD(type,methodname,arglist) type (*methodname) arglist -#else -#define JMETHOD(type,methodname,arglist) type (*methodname) () -#endif - - -/* The noreturn type identifier is used to declare functions - * which cannot return. - * Compilers can thus create more optimized code and perform - * better checks for warnings and errors. - * Static analyzer tools can make improved inferences about - * execution paths and are prevented from giving false alerts. - * - * Unfortunately, the proposed specifications of corresponding - * extensions in the Dec 2011 ISO C standard revision (C11), - * GCC, MSVC, etc. are not viable. - * Thus we introduce a user defined type to declare noreturn - * functions at least for clarity. A proper compiler would - * have a suitable noreturn type to match in place of void. - */ - -#ifndef HAVE_NORETURN_T -typedef void noreturn_t; -#endif - - -/* Here is the pseudo-keyword for declaring pointers that must be "far" - * on 80x86 machines. Most of the specialized coding for 80x86 is handled - * by just saying "FAR *" where such a pointer is needed. In a few places - * explicit coding is needed; see uses of the NEED_FAR_POINTERS symbol. - */ - -#ifndef FAR -#ifdef NEED_FAR_POINTERS -#define FAR far -#else -#define FAR -#endif -#endif - - -/* - * On a few systems, type boolean and/or its values FALSE, TRUE may appear - * in standard header files. Or you may have conflicts with application- - * specific header files that you want to include together with these files. - * Defining HAVE_BOOLEAN before including jpeglib.h should make it work. - */ - -#ifndef HAVE_BOOLEAN -#if defined FALSE || defined TRUE || defined QGLOBAL_H -/* Qt3 defines FALSE and TRUE as "const" variables in qglobal.h */ -typedef int boolean; -#ifndef FALSE /* in case these macros already exist */ -#define FALSE 0 /* values of boolean */ -#endif -#ifndef TRUE -#define TRUE 1 -#endif -#else -typedef enum { FALSE = 0, TRUE = 1 } boolean; -#endif -#endif - - -/* - * The remaining options affect code selection within the JPEG library, - * but they don't need to be visible to most applications using the library. - * To minimize application namespace pollution, the symbols won't be - * defined unless JPEG_INTERNALS or JPEG_INTERNAL_OPTIONS has been defined. - */ - -#ifdef JPEG_INTERNALS -#define JPEG_INTERNAL_OPTIONS -#endif - -#ifdef JPEG_INTERNAL_OPTIONS - - -/* - * These defines indicate whether to include various optional functions. - * Undefining some of these symbols will produce a smaller but less capable - * library. Note that you can leave certain source files out of the - * compilation/linking process if you've #undef'd the corresponding symbols. - * (You may HAVE to do that if your compiler doesn't like null source files.) - */ - -/* Capability options common to encoder and decoder: */ - -#define DCT_ISLOW_SUPPORTED /* slow but accurate integer algorithm */ -#define DCT_IFAST_SUPPORTED /* faster, less accurate integer method */ -#define DCT_FLOAT_SUPPORTED /* floating-point: accurate, fast on fast HW */ - -/* Encoder capability options: */ - -#define C_ARITH_CODING_SUPPORTED /* Arithmetic coding back end? */ -#define C_MULTISCAN_FILES_SUPPORTED /* Multiple-scan JPEG files? */ -#define C_PROGRESSIVE_SUPPORTED /* Progressive JPEG? (Requires MULTISCAN) */ -#define DCT_SCALING_SUPPORTED /* Input rescaling via DCT? (Requires DCT_ISLOW) */ -#define ENTROPY_OPT_SUPPORTED /* Optimization of entropy coding parms? */ -/* Note: if you selected more than 8-bit data precision, it is dangerous to - * turn off ENTROPY_OPT_SUPPORTED. The standard Huffman tables are only - * good for 8-bit precision, so arithmetic coding is recommended for higher - * precision. The Huffman encoder normally uses entropy optimization to - * compute usable tables for higher precision. Otherwise, you'll have to - * supply different default Huffman tables. - * The exact same statements apply for progressive JPEG: the default tables - * don't work for progressive mode. (This may get fixed, however.) - */ -#define INPUT_SMOOTHING_SUPPORTED /* Input image smoothing option? */ - -/* Decoder capability options: */ - -#define D_ARITH_CODING_SUPPORTED /* Arithmetic coding back end? */ -#define D_MULTISCAN_FILES_SUPPORTED /* Multiple-scan JPEG files? */ -#define D_PROGRESSIVE_SUPPORTED /* Progressive JPEG? (Requires MULTISCAN) */ -#define IDCT_SCALING_SUPPORTED /* Output rescaling via IDCT? (Requires DCT_ISLOW) */ -#define SAVE_MARKERS_SUPPORTED /* jpeg_save_markers() needed? */ -#define BLOCK_SMOOTHING_SUPPORTED /* Block smoothing? (Progressive only) */ -#undef UPSAMPLE_SCALING_SUPPORTED /* Output rescaling at upsample stage? */ -#define UPSAMPLE_MERGING_SUPPORTED /* Fast path for sloppy upsampling? */ -#define QUANT_1PASS_SUPPORTED /* 1-pass color quantization? */ -#define QUANT_2PASS_SUPPORTED /* 2-pass color quantization? */ - -/* more capability options later, no doubt */ - - -/* - * Ordering of RGB data in scanlines passed to or from the application. - * If your application wants to deal with data in the order B,G,R, just - * #define JPEG_USE_RGB_CUSTOM in jconfig.h, or define your own custom - * order in jconfig.h and #define JPEG_HAVE_RGB_CUSTOM. - * You can also deal with formats such as R,G,B,X (one extra byte per pixel) - * by changing RGB_PIXELSIZE. - * Note that changing the offsets will also change - * the order in which colormap data is organized. - * RESTRICTIONS: - * 1. The sample applications cjpeg,djpeg do NOT support modified RGB formats. - * 2. The color quantizer modules will not behave desirably if RGB_PIXELSIZE - * is not 3 (they don't understand about dummy color components!). - * So you can't use color quantization if you change that value. - */ - -#ifndef JPEG_HAVE_RGB_CUSTOM -#ifdef JPEG_USE_RGB_CUSTOM -#define RGB_RED 2 /* Offset of Red in an RGB scanline element */ -#define RGB_GREEN 1 /* Offset of Green */ -#define RGB_BLUE 0 /* Offset of Blue */ -#else -#define RGB_RED 0 /* Offset of Red in an RGB scanline element */ -#define RGB_GREEN 1 /* Offset of Green */ -#define RGB_BLUE 2 /* Offset of Blue */ -#endif -#define RGB_PIXELSIZE 3 /* JSAMPLEs per RGB scanline element */ -#endif - - -/* Definitions for speed-related optimizations. */ - - -/* If your compiler supports inline functions, define INLINE - * as the inline keyword; otherwise define it as empty. - */ - -#ifndef INLINE -#ifdef __GNUC__ /* for instance, GNU C knows about inline */ -#define INLINE __inline__ -#endif -#ifndef INLINE -#define INLINE /* default is to define it as empty */ -#endif -#endif - - -/* On some machines (notably 68000 series) "int" is 32 bits, but multiplying - * two 16-bit shorts is faster than multiplying two ints. Define MULTIPLIER - * as short on such a machine. MULTIPLIER must be at least 16 bits wide. - */ - -#ifndef MULTIPLIER -#define MULTIPLIER int /* type for fastest integer multiply */ -#endif - - -/* FAST_FLOAT should be either float or double, whichever is done faster - * by your compiler. (Note that this type is only used in the floating point - * DCT routines, so it only matters if you've defined DCT_FLOAT_SUPPORTED.) - * Typically, float is faster in ANSI C compilers, while double is faster in - * pre-ANSI compilers (because they insist on converting to double anyway). - * The code below therefore chooses float if we have ANSI-style prototypes. - */ - -#ifndef FAST_FLOAT -#ifdef HAVE_PROTOTYPES -#define FAST_FLOAT float -#else -#define FAST_FLOAT double -#endif -#endif - -#endif /* JPEG_INTERNAL_OPTIONS */ diff --git a/cmake/externals/libjpeg/jpegint.h b/cmake/externals/libjpeg/jpegint.h deleted file mode 100644 index 3528bff5b..000000000 --- a/cmake/externals/libjpeg/jpegint.h +++ /dev/null @@ -1,445 +0,0 @@ -/* - * jpegint.h - * - * Copyright (C) 1991-1997, Thomas G. Lane. - * Modified 1997-2020 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file provides common declarations for the various JPEG modules. - * These declarations are considered internal to the JPEG library; most - * applications using the library shouldn't need to include this file. - */ - - -/* Declarations for both compression & decompression */ - -typedef enum { /* Operating modes for buffer controllers */ - JBUF_PASS_THRU, /* Plain stripwise operation */ - /* Remaining modes require a full-image buffer to have been created */ - JBUF_SAVE_SOURCE, /* Run source subobject only, save output */ - JBUF_CRANK_DEST, /* Run dest subobject only, using saved data */ - JBUF_SAVE_AND_PASS /* Run both subobjects, save output */ -} J_BUF_MODE; - -/* Values of global_state field (jdapi.c has some dependencies on ordering!) */ -#define CSTATE_START 100 /* after create_compress */ -#define CSTATE_SCANNING 101 /* start_compress done, write_scanlines OK */ -#define CSTATE_RAW_OK 102 /* start_compress done, write_raw_data OK */ -#define CSTATE_WRCOEFS 103 /* jpeg_write_coefficients done */ -#define DSTATE_START 200 /* after create_decompress */ -#define DSTATE_INHEADER 201 /* reading header markers, no SOS yet */ -#define DSTATE_READY 202 /* found SOS, ready for start_decompress */ -#define DSTATE_PRELOAD 203 /* reading multiscan file in start_decompress*/ -#define DSTATE_PRESCAN 204 /* performing dummy pass for 2-pass quant */ -#define DSTATE_SCANNING 205 /* start_decompress done, read_scanlines OK */ -#define DSTATE_RAW_OK 206 /* start_decompress done, read_raw_data OK */ -#define DSTATE_BUFIMAGE 207 /* expecting jpeg_start_output */ -#define DSTATE_BUFPOST 208 /* looking for SOS/EOI in jpeg_finish_output */ -#define DSTATE_RDCOEFS 209 /* reading file in jpeg_read_coefficients */ -#define DSTATE_STOPPING 210 /* looking for EOI in jpeg_finish_decompress */ - - -/* Declarations for compression modules */ - -/* Master control module */ -struct jpeg_comp_master { - JMETHOD(void, prepare_for_pass, (j_compress_ptr cinfo)); - JMETHOD(void, pass_startup, (j_compress_ptr cinfo)); - JMETHOD(void, finish_pass, (j_compress_ptr cinfo)); - - /* State variables made visible to other modules */ - boolean call_pass_startup; /* True if pass_startup must be called */ - boolean is_last_pass; /* True during last pass */ -}; - -/* Main buffer control (downsampled-data buffer) */ -struct jpeg_c_main_controller { - JMETHOD(void, start_pass, (j_compress_ptr cinfo, J_BUF_MODE pass_mode)); - JMETHOD(void, process_data, (j_compress_ptr cinfo, - JSAMPARRAY input_buf, JDIMENSION *in_row_ctr, - JDIMENSION in_rows_avail)); -}; - -/* Compression preprocessing (downsampling input buffer control) */ -struct jpeg_c_prep_controller { - JMETHOD(void, start_pass, (j_compress_ptr cinfo, J_BUF_MODE pass_mode)); - JMETHOD(void, pre_process_data, (j_compress_ptr cinfo, - JSAMPARRAY input_buf, - JDIMENSION *in_row_ctr, - JDIMENSION in_rows_avail, - JSAMPIMAGE output_buf, - JDIMENSION *out_row_group_ctr, - JDIMENSION out_row_groups_avail)); -}; - -/* Coefficient buffer control */ -struct jpeg_c_coef_controller { - JMETHOD(void, start_pass, (j_compress_ptr cinfo, J_BUF_MODE pass_mode)); - JMETHOD(boolean, compress_data, (j_compress_ptr cinfo, - JSAMPIMAGE input_buf)); -}; - -/* Colorspace conversion */ -struct jpeg_color_converter { - JMETHOD(void, start_pass, (j_compress_ptr cinfo)); - JMETHOD(void, color_convert, (j_compress_ptr cinfo, - JSAMPARRAY input_buf, JSAMPIMAGE output_buf, - JDIMENSION output_row, int num_rows)); -}; - -/* Downsampling */ -struct jpeg_downsampler { - JMETHOD(void, start_pass, (j_compress_ptr cinfo)); - JMETHOD(void, downsample, (j_compress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION in_row_index, - JSAMPIMAGE output_buf, - JDIMENSION out_row_group_index)); - - boolean need_context_rows; /* TRUE if need rows above & below */ -}; - -/* Forward DCT (also controls coefficient quantization) */ -typedef JMETHOD(void, forward_DCT_ptr, - (j_compress_ptr cinfo, jpeg_component_info * compptr, - JSAMPARRAY sample_data, JBLOCKROW coef_blocks, - JDIMENSION start_col, JDIMENSION num_blocks)); - -struct jpeg_forward_dct { - JMETHOD(void, start_pass, (j_compress_ptr cinfo)); - /* It is useful to allow each component to have a separate FDCT method. */ - forward_DCT_ptr forward_DCT[MAX_COMPONENTS]; -}; - -/* Entropy encoding */ -struct jpeg_entropy_encoder { - JMETHOD(void, start_pass, (j_compress_ptr cinfo, boolean gather_statistics)); - JMETHOD(boolean, encode_mcu, (j_compress_ptr cinfo, JBLOCKARRAY MCU_data)); - JMETHOD(void, finish_pass, (j_compress_ptr cinfo)); -}; - -/* Marker writing */ -struct jpeg_marker_writer { - JMETHOD(void, write_file_header, (j_compress_ptr cinfo)); - JMETHOD(void, write_frame_header, (j_compress_ptr cinfo)); - JMETHOD(void, write_scan_header, (j_compress_ptr cinfo)); - JMETHOD(void, write_file_trailer, (j_compress_ptr cinfo)); - JMETHOD(void, write_tables_only, (j_compress_ptr cinfo)); - /* These routines are exported to allow insertion of extra markers */ - /* Probably only COM and APPn markers should be written this way */ - JMETHOD(void, write_marker_header, (j_compress_ptr cinfo, int marker, - unsigned int datalen)); - JMETHOD(void, write_marker_byte, (j_compress_ptr cinfo, int val)); -}; - - -/* Declarations for decompression modules */ - -/* Master control module */ -struct jpeg_decomp_master { - JMETHOD(void, prepare_for_output_pass, (j_decompress_ptr cinfo)); - JMETHOD(void, finish_output_pass, (j_decompress_ptr cinfo)); - - /* State variables made visible to other modules */ - boolean is_dummy_pass; /* True during 1st pass for 2-pass quant */ -}; - -/* Input control module */ -struct jpeg_input_controller { - JMETHOD(int, consume_input, (j_decompress_ptr cinfo)); - JMETHOD(void, reset_input_controller, (j_decompress_ptr cinfo)); - JMETHOD(void, start_input_pass, (j_decompress_ptr cinfo)); - JMETHOD(void, finish_input_pass, (j_decompress_ptr cinfo)); - - /* State variables made visible to other modules */ - boolean has_multiple_scans; /* True if file has multiple scans */ - boolean eoi_reached; /* True when EOI has been consumed */ -}; - -/* Main buffer control (downsampled-data buffer) */ -struct jpeg_d_main_controller { - JMETHOD(void, start_pass, (j_decompress_ptr cinfo, J_BUF_MODE pass_mode)); - JMETHOD(void, process_data, (j_decompress_ptr cinfo, - JSAMPARRAY output_buf, JDIMENSION *out_row_ctr, - JDIMENSION out_rows_avail)); -}; - -/* Coefficient buffer control */ -struct jpeg_d_coef_controller { - JMETHOD(void, start_input_pass, (j_decompress_ptr cinfo)); - JMETHOD(int, consume_data, (j_decompress_ptr cinfo)); - JMETHOD(void, start_output_pass, (j_decompress_ptr cinfo)); - JMETHOD(int, decompress_data, (j_decompress_ptr cinfo, - JSAMPIMAGE output_buf)); - /* Pointer to array of coefficient virtual arrays, or NULL if none */ - jvirt_barray_ptr *coef_arrays; -}; - -/* Decompression postprocessing (color quantization buffer control) */ -struct jpeg_d_post_controller { - JMETHOD(void, start_pass, (j_decompress_ptr cinfo, J_BUF_MODE pass_mode)); - JMETHOD(void, post_process_data, (j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, - JDIMENSION *in_row_group_ctr, - JDIMENSION in_row_groups_avail, - JSAMPARRAY output_buf, - JDIMENSION *out_row_ctr, - JDIMENSION out_rows_avail)); -}; - -/* Marker reading & parsing */ -struct jpeg_marker_reader { - JMETHOD(void, reset_marker_reader, (j_decompress_ptr cinfo)); - /* Read markers until SOS or EOI. - * Returns same codes as are defined for jpeg_consume_input: - * JPEG_SUSPENDED, JPEG_REACHED_SOS, or JPEG_REACHED_EOI. - */ - JMETHOD(int, read_markers, (j_decompress_ptr cinfo)); - /* Read a restart marker --- exported for use by entropy decoder only */ - jpeg_marker_parser_method read_restart_marker; - - /* State of marker reader --- nominally internal, but applications - * supplying COM or APPn handlers might like to know the state. - */ - boolean saw_SOI; /* found SOI? */ - boolean saw_SOF; /* found SOF? */ - int next_restart_num; /* next restart number expected (0-7) */ - unsigned int discarded_bytes; /* # of bytes skipped looking for a marker */ -}; - -/* Entropy decoding */ -struct jpeg_entropy_decoder { - JMETHOD(void, start_pass, (j_decompress_ptr cinfo)); - JMETHOD(boolean, decode_mcu, (j_decompress_ptr cinfo, JBLOCKARRAY MCU_data)); - JMETHOD(void, finish_pass, (j_decompress_ptr cinfo)); -}; - -/* Inverse DCT (also performs dequantization) */ -typedef JMETHOD(void, inverse_DCT_method_ptr, - (j_decompress_ptr cinfo, jpeg_component_info * compptr, - JCOEFPTR coef_block, - JSAMPARRAY output_buf, JDIMENSION output_col)); - -struct jpeg_inverse_dct { - JMETHOD(void, start_pass, (j_decompress_ptr cinfo)); - /* It is useful to allow each component to have a separate IDCT method. */ - inverse_DCT_method_ptr inverse_DCT[MAX_COMPONENTS]; -}; - -/* Upsampling (note that upsampler must also call color converter) */ -struct jpeg_upsampler { - JMETHOD(void, start_pass, (j_decompress_ptr cinfo)); - JMETHOD(void, upsample, (j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, - JDIMENSION *in_row_group_ctr, - JDIMENSION in_row_groups_avail, - JSAMPARRAY output_buf, - JDIMENSION *out_row_ctr, - JDIMENSION out_rows_avail)); - - boolean need_context_rows; /* TRUE if need rows above & below */ -}; - -/* Colorspace conversion */ -struct jpeg_color_deconverter { - JMETHOD(void, start_pass, (j_decompress_ptr cinfo)); - JMETHOD(void, color_convert, (j_decompress_ptr cinfo, - JSAMPIMAGE input_buf, JDIMENSION input_row, - JSAMPARRAY output_buf, int num_rows)); -}; - -/* Color quantization or color precision reduction */ -struct jpeg_color_quantizer { - JMETHOD(void, start_pass, (j_decompress_ptr cinfo, boolean is_pre_scan)); - JMETHOD(void, color_quantize, (j_decompress_ptr cinfo, - JSAMPARRAY input_buf, JSAMPARRAY output_buf, - int num_rows)); - JMETHOD(void, finish_pass, (j_decompress_ptr cinfo)); - JMETHOD(void, new_color_map, (j_decompress_ptr cinfo)); -}; - - -/* Definition of range extension bits for decompression processes. - * See the comments with prepare_range_limit_table (in jdmaster.c) - * for more info. - * The recommended default value for normal applications is 2. - * Applications with special requirements may use a different value. - * For example, Ghostscript wants to use 3 for proper handling of - * wacky images with oversize coefficient values. - */ - -#define RANGE_BITS 2 -#define RANGE_CENTER (CENTERJSAMPLE << RANGE_BITS) - - -/* Miscellaneous useful macros */ - -#undef MAX -#define MAX(a,b) ((a) > (b) ? (a) : (b)) -#undef MIN -#define MIN(a,b) ((a) < (b) ? (a) : (b)) - - -/* We assume that right shift corresponds to signed division by 2 with - * rounding towards minus infinity. This is correct for typical "arithmetic - * shift" instructions that shift in copies of the sign bit. But some - * C compilers implement >> with an unsigned shift. For these machines you - * must define RIGHT_SHIFT_IS_UNSIGNED. - * RIGHT_SHIFT provides a proper signed right shift of an INT32 quantity. - * It is only applied with constant shift counts. SHIFT_TEMPS must be - * included in the variables of any routine using RIGHT_SHIFT. - */ - -#ifdef RIGHT_SHIFT_IS_UNSIGNED -#define SHIFT_TEMPS INT32 shift_temp; -#define RIGHT_SHIFT(x,shft) \ - ((shift_temp = (x)) < 0 ? \ - (shift_temp >> (shft)) | ((~((INT32) 0)) << (32-(shft))) : \ - (shift_temp >> (shft))) -#else -#define SHIFT_TEMPS -#define RIGHT_SHIFT(x,shft) ((x) >> (shft)) -#endif - -/* Descale and correctly round an INT32 value that's scaled by N bits. - * We assume RIGHT_SHIFT rounds towards minus infinity, so adding - * the fudge factor is correct for either sign of X. - */ - -#define DESCALE(x,n) RIGHT_SHIFT((x) + ((INT32) 1 << ((n)-1)), n) - - -/* Short forms of external names for systems with brain-damaged linkers. */ - -#ifdef NEED_SHORT_EXTERNAL_NAMES -#define jinit_compress_master jICompress -#define jinit_c_master_control jICMaster -#define jinit_c_main_controller jICMainC -#define jinit_c_prep_controller jICPrepC -#define jinit_c_coef_controller jICCoefC -#define jinit_color_converter jICColor -#define jinit_downsampler jIDownsampler -#define jinit_forward_dct jIFDCT -#define jinit_huff_encoder jIHEncoder -#define jinit_arith_encoder jIAEncoder -#define jinit_marker_writer jIMWriter -#define jinit_master_decompress jIDMaster -#define jinit_d_main_controller jIDMainC -#define jinit_d_coef_controller jIDCoefC -#define jinit_d_post_controller jIDPostC -#define jinit_input_controller jIInCtlr -#define jinit_marker_reader jIMReader -#define jinit_huff_decoder jIHDecoder -#define jinit_arith_decoder jIADecoder -#define jinit_inverse_dct jIIDCT -#define jinit_upsampler jIUpsampler -#define jinit_color_deconverter jIDColor -#define jinit_1pass_quantizer jI1Quant -#define jinit_2pass_quantizer jI2Quant -#define jinit_merged_upsampler jIMUpsampler -#define jinit_memory_mgr jIMemMgr -#define jdiv_round_up jDivRound -#define jround_up jRound -#define jzero_far jZeroFar -#define jcopy_sample_rows jCopySamples -#define jcopy_block_row jCopyBlocks -#define jpeg_zigzag_order jZIGTable -#define jpeg_natural_order jZAGTable -#define jpeg_natural_order7 jZAG7Table -#define jpeg_natural_order6 jZAG6Table -#define jpeg_natural_order5 jZAG5Table -#define jpeg_natural_order4 jZAG4Table -#define jpeg_natural_order3 jZAG3Table -#define jpeg_natural_order2 jZAG2Table -#define jpeg_aritab jAriTab -#endif /* NEED_SHORT_EXTERNAL_NAMES */ - - -/* On normal machines we can apply MEMCOPY() and MEMZERO() to sample arrays - * and coefficient-block arrays. This won't work on 80x86 because the arrays - * are FAR and we're assuming a small-pointer memory model. However, some - * DOS compilers provide far-pointer versions of memcpy() and memset() even - * in the small-model libraries. These will be used if USE_FMEM is defined. - * Otherwise, the routines in jutils.c do it the hard way. - */ - -#ifndef NEED_FAR_POINTERS /* normal case, same as regular macro */ -#define FMEMZERO(target,size) MEMZERO(target,size) -#else /* 80x86 case */ -#ifdef USE_FMEM -#define FMEMZERO(target,size) _fmemset((void FAR *)(target), 0, (size_t)(size)) -#else -EXTERN(void) jzero_far JPP((void FAR * target, size_t bytestozero)); -#define FMEMZERO(target,size) jzero_far(target, size) -#endif -#endif - - -/* Compression module initialization routines */ -EXTERN(void) jinit_compress_master JPP((j_compress_ptr cinfo)); -EXTERN(void) jinit_c_master_control JPP((j_compress_ptr cinfo, - boolean transcode_only)); -EXTERN(void) jinit_c_main_controller JPP((j_compress_ptr cinfo, - boolean need_full_buffer)); -EXTERN(void) jinit_c_prep_controller JPP((j_compress_ptr cinfo, - boolean need_full_buffer)); -EXTERN(void) jinit_c_coef_controller JPP((j_compress_ptr cinfo, - boolean need_full_buffer)); -EXTERN(void) jinit_color_converter JPP((j_compress_ptr cinfo)); -EXTERN(void) jinit_downsampler JPP((j_compress_ptr cinfo)); -EXTERN(void) jinit_forward_dct JPP((j_compress_ptr cinfo)); -EXTERN(void) jinit_huff_encoder JPP((j_compress_ptr cinfo)); -EXTERN(void) jinit_arith_encoder JPP((j_compress_ptr cinfo)); -EXTERN(void) jinit_marker_writer JPP((j_compress_ptr cinfo)); -/* Decompression module initialization routines */ -EXTERN(void) jinit_master_decompress JPP((j_decompress_ptr cinfo)); -EXTERN(void) jinit_d_main_controller JPP((j_decompress_ptr cinfo, - boolean need_full_buffer)); -EXTERN(void) jinit_d_coef_controller JPP((j_decompress_ptr cinfo, - boolean need_full_buffer)); -EXTERN(void) jinit_d_post_controller JPP((j_decompress_ptr cinfo, - boolean need_full_buffer)); -EXTERN(void) jinit_input_controller JPP((j_decompress_ptr cinfo)); -EXTERN(void) jinit_marker_reader JPP((j_decompress_ptr cinfo)); -EXTERN(void) jinit_huff_decoder JPP((j_decompress_ptr cinfo)); -EXTERN(void) jinit_arith_decoder JPP((j_decompress_ptr cinfo)); -EXTERN(void) jinit_inverse_dct JPP((j_decompress_ptr cinfo)); -EXTERN(void) jinit_upsampler JPP((j_decompress_ptr cinfo)); -EXTERN(void) jinit_color_deconverter JPP((j_decompress_ptr cinfo)); -EXTERN(void) jinit_1pass_quantizer JPP((j_decompress_ptr cinfo)); -EXTERN(void) jinit_2pass_quantizer JPP((j_decompress_ptr cinfo)); -EXTERN(void) jinit_merged_upsampler JPP((j_decompress_ptr cinfo)); -/* Memory manager initialization */ -EXTERN(void) jinit_memory_mgr JPP((j_common_ptr cinfo)); - -/* Utility routines in jutils.c */ -EXTERN(long) jdiv_round_up JPP((long a, long b)); -EXTERN(long) jround_up JPP((long a, long b)); -EXTERN(void) jcopy_sample_rows JPP((JSAMPARRAY input_array, - JSAMPARRAY output_array, - int num_rows, JDIMENSION num_cols)); -EXTERN(void) jcopy_block_row JPP((JBLOCKROW input_row, JBLOCKROW output_row, - JDIMENSION num_blocks)); -/* Constant tables in jutils.c */ -#if 0 /* This table is not actually needed in v6a */ -extern const int jpeg_zigzag_order[]; /* natural coef order to zigzag order */ -#endif -extern const int jpeg_natural_order[]; /* zigzag coef order to natural order */ -extern const int jpeg_natural_order7[]; /* zz to natural order for 7x7 block */ -extern const int jpeg_natural_order6[]; /* zz to natural order for 6x6 block */ -extern const int jpeg_natural_order5[]; /* zz to natural order for 5x5 block */ -extern const int jpeg_natural_order4[]; /* zz to natural order for 4x4 block */ -extern const int jpeg_natural_order3[]; /* zz to natural order for 3x3 block */ -extern const int jpeg_natural_order2[]; /* zz to natural order for 2x2 block */ - -/* Arithmetic coding probability estimation tables in jaricom.c */ -extern const INT32 jpeg_aritab[]; - -/* Suppress undefined-structure complaints if necessary. */ - -#ifdef INCOMPLETE_TYPES_BROKEN -#ifndef AM_MEMORY_MANAGER /* only jmemmgr.c defines these */ -struct jvirt_sarray_control { long dummy; }; -struct jvirt_barray_control { long dummy; }; -#endif -#endif /* INCOMPLETE_TYPES_BROKEN */ diff --git a/cmake/externals/libjpeg/jpeglib.h b/cmake/externals/libjpeg/jpeglib.h deleted file mode 100644 index e7e15ab2c..000000000 --- a/cmake/externals/libjpeg/jpeglib.h +++ /dev/null @@ -1,1183 +0,0 @@ -/* - * jpeglib.h - * - * Copyright (C) 1991-1998, Thomas G. Lane. - * Modified 2002-2022 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file defines the application interface for the JPEG library. - * Most applications using the library need only include this file, - * and perhaps jerror.h if they want to know the exact error codes. - */ - -#ifndef JPEGLIB_H -#define JPEGLIB_H - -/* - * First we include the configuration files that record how this - * installation of the JPEG library is set up. jconfig.h can be - * generated automatically for many systems. jmorecfg.h contains - * manual configuration options that most people need not worry about. - */ - -#ifndef JCONFIG_INCLUDED /* in case jinclude.h already did */ -#include "jconfig.h" /* widely used configuration options */ -#endif -#include "jmorecfg.h" /* seldom changed options */ - - -#ifdef __cplusplus -#ifndef DONT_USE_EXTERN_C -extern "C" { -#endif -#endif - -/* Version IDs for the JPEG library. - * Might be useful for tests like "#if JPEG_LIB_VERSION >= 90". - */ - -#define JPEG_LIB_VERSION 90 /* Compatibility version 9.0 */ -#define JPEG_LIB_VERSION_MAJOR 9 -#define JPEG_LIB_VERSION_MINOR 6 - - -/* Various constants determining the sizes of things. - * All of these are specified by the JPEG standard, - * so don't change them if you want to be compatible. - */ - -#define DCTSIZE 8 /* The basic DCT block is 8x8 coefficients */ -#define DCTSIZE2 64 /* DCTSIZE squared; # of elements in a block */ -#define NUM_QUANT_TBLS 4 /* Quantization tables are numbered 0..3 */ -#define NUM_HUFF_TBLS 4 /* Huffman tables are numbered 0..3 */ -#define NUM_ARITH_TBLS 16 /* Arith-coding tables are numbered 0..15 */ -#define MAX_COMPS_IN_SCAN 4 /* JPEG limit on # of components in one scan */ -#define MAX_SAMP_FACTOR 4 /* JPEG limit on sampling factors */ -/* Unfortunately, some bozo at Adobe saw no reason to be bound by the standard; - * the PostScript DCT filter can emit files with many more than 10 blocks/MCU. - * If you happen to run across such a file, you can up D_MAX_BLOCKS_IN_MCU - * to handle it. We even let you do this from the jconfig.h file. However, - * we strongly discourage changing C_MAX_BLOCKS_IN_MCU; just because Adobe - * sometimes emits noncompliant files doesn't mean you should too. - */ -#define C_MAX_BLOCKS_IN_MCU 10 /* compressor's limit on blocks per MCU */ -#ifndef D_MAX_BLOCKS_IN_MCU -#define D_MAX_BLOCKS_IN_MCU 10 /* decompressor's limit on blocks per MCU */ -#endif - - -/* Data structures for images (arrays of samples and of DCT coefficients). - * On 80x86 machines, the image arrays are too big for near pointers, - * but the pointer arrays can fit in near memory. - */ - -typedef JSAMPLE FAR *JSAMPROW; /* ptr to one image row of pixel samples. */ -typedef JSAMPROW *JSAMPARRAY; /* ptr to some rows (a 2-D sample array) */ -typedef JSAMPARRAY *JSAMPIMAGE; /* a 3-D sample array: top index is color */ - -typedef JCOEF JBLOCK[DCTSIZE2]; /* one block of coefficients */ -typedef JBLOCK FAR *JBLOCKROW; /* pointer to one row of coefficient blocks */ -typedef JBLOCKROW *JBLOCKARRAY; /* a 2-D array of coefficient blocks */ -typedef JBLOCKARRAY *JBLOCKIMAGE; /* a 3-D array of coefficient blocks */ - -typedef JCOEF FAR *JCOEFPTR; /* useful in a couple of places */ - - -/* Types for JPEG compression parameters and working tables. */ - - -/* DCT coefficient quantization tables. */ - -typedef struct { - /* This array gives the coefficient quantizers in natural array order - * (not the zigzag order in which they are stored in a JPEG DQT marker). - * CAUTION: IJG versions prior to v6a kept this array in zigzag order. - */ - UINT16 quantval[DCTSIZE2]; /* quantization step for each coefficient */ - /* This field is used only during compression. It's initialized FALSE when - * the table is created, and set TRUE when it's been output to the file. - * You could suppress output of a table by setting this to TRUE. - * (See jpeg_suppress_tables for an example.) - */ - boolean sent_table; /* TRUE when table has been output */ -} JQUANT_TBL; - - -/* Huffman coding tables. */ - -typedef struct { - /* These two fields directly represent the contents of a JPEG DHT marker */ - UINT8 bits[17]; /* bits[k] = # of symbols with codes of */ - /* length k bits; bits[0] is unused */ - UINT8 huffval[256]; /* The symbols, in order of incr code length */ - /* This field is used only during compression. It's initialized FALSE when - * the table is created, and set TRUE when it's been output to the file. - * You could suppress output of a table by setting this to TRUE. - * (See jpeg_suppress_tables for an example.) - */ - boolean sent_table; /* TRUE when table has been output */ -} JHUFF_TBL; - - -/* Basic info about one component (color channel). */ - -typedef struct { - /* These values are fixed over the whole image. */ - /* For compression, they must be supplied by parameter setup; */ - /* for decompression, they are read from the SOF marker. */ - int component_id; /* identifier for this component (0..255) */ - int component_index; /* its index in SOF or cinfo->comp_info[] */ - int h_samp_factor; /* horizontal sampling factor (1..4) */ - int v_samp_factor; /* vertical sampling factor (1..4) */ - int quant_tbl_no; /* quantization table selector (0..3) */ - /* These values may vary between scans. */ - /* For compression, they must be supplied by parameter setup; */ - /* for decompression, they are read from the SOS marker. */ - /* The decompressor output side may not use these variables. */ - int dc_tbl_no; /* DC entropy table selector (0..3) */ - int ac_tbl_no; /* AC entropy table selector (0..3) */ - - /* Remaining fields should be treated as private by applications. */ - - /* These values are computed during compression or decompression startup: */ - /* Component's size in DCT blocks. - * Any dummy blocks added to complete an MCU are not counted; therefore - * these values do not depend on whether a scan is interleaved or not. - */ - JDIMENSION width_in_blocks; - JDIMENSION height_in_blocks; - /* Size of a DCT block in samples, - * reflecting any scaling we choose to apply during the DCT step. - * Values from 1 to 16 are supported. - * Note that different components may receive different DCT scalings. - */ - int DCT_h_scaled_size; - int DCT_v_scaled_size; - /* The downsampled dimensions are the component's actual, unpadded number - * of samples at the main buffer (preprocessing/compression interface); - * DCT scaling is included, so - * downsampled_width = - * ceil(image_width * Hi/Hmax * DCT_h_scaled_size/block_size) - * and similarly for height. - */ - JDIMENSION downsampled_width; /* actual width in samples */ - JDIMENSION downsampled_height; /* actual height in samples */ - /* For decompression, in cases where some of the components will be - * ignored (eg grayscale output from YCbCr image), we can skip most - * computations for the unused components. - * For compression, some of the components will need further quantization - * scale by factor of 2 after DCT (eg BG_YCC output from normal RGB input). - * The field is first set TRUE for decompression, FALSE for compression - * in initial_setup, and then adapted in color conversion setup. - */ - boolean component_needed; - - /* These values are computed before starting a scan of the component. */ - /* The decompressor output side may not use these variables. */ - int MCU_width; /* number of blocks per MCU, horizontally */ - int MCU_height; /* number of blocks per MCU, vertically */ - int MCU_blocks; /* MCU_width * MCU_height */ - int MCU_sample_width; /* MCU width in samples: MCU_width * DCT_h_scaled_size */ - int last_col_width; /* # of non-dummy blocks across in last MCU */ - int last_row_height; /* # of non-dummy blocks down in last MCU */ - - /* Saved quantization table for component; NULL if none yet saved. - * See jdinput.c comments about the need for this information. - * This field is currently used only for decompression. - */ - JQUANT_TBL * quant_table; - - /* Private per-component storage for DCT or IDCT subsystem. */ - void * dct_table; -} jpeg_component_info; - - -/* The script for encoding a multiple-scan file is an array of these: */ - -typedef struct { - int comps_in_scan; /* number of components encoded in this scan */ - int component_index[MAX_COMPS_IN_SCAN]; /* their SOF/comp_info[] indexes */ - int Ss, Se; /* progressive JPEG spectral selection parms */ - int Ah, Al; /* progressive JPEG successive approx. parms */ -} jpeg_scan_info; - -/* The decompressor can save APPn and COM markers in a list of these: */ - -typedef struct jpeg_marker_struct FAR * jpeg_saved_marker_ptr; - -struct jpeg_marker_struct { - jpeg_saved_marker_ptr next; /* next in list, or NULL */ - UINT8 marker; /* marker code: JPEG_COM, or JPEG_APP0+n */ - unsigned int original_length; /* # bytes of data in the file */ - unsigned int data_length; /* # bytes of data saved at data[] */ - JOCTET FAR * data; /* the data contained in the marker */ - /* the marker length word is not counted in data_length or original_length */ -}; - -/* Known color spaces. */ - -typedef enum { - JCS_UNKNOWN, /* error/unspecified */ - JCS_GRAYSCALE, /* monochrome */ - JCS_RGB, /* red/green/blue, standard RGB (sRGB) */ - JCS_YCbCr, /* Y/Cb/Cr (also known as YUV), standard YCC */ - JCS_CMYK, /* C/M/Y/K */ - JCS_YCCK, /* Y/Cb/Cr/K */ - JCS_BG_RGB, /* big gamut red/green/blue, bg-sRGB */ - JCS_BG_YCC /* big gamut Y/Cb/Cr, bg-sYCC */ -} J_COLOR_SPACE; - -/* Supported color transforms. */ - -typedef enum { - JCT_NONE = 0, - JCT_SUBTRACT_GREEN = 1 -} J_COLOR_TRANSFORM; - -/* DCT/IDCT algorithm options. */ - -typedef enum { - JDCT_ISLOW, /* slow but accurate integer algorithm */ - JDCT_IFAST, /* faster, less accurate integer method */ - JDCT_FLOAT /* floating-point: accurate, fast on fast HW */ -} J_DCT_METHOD; - -#ifndef JDCT_DEFAULT /* may be overridden in jconfig.h */ -#define JDCT_DEFAULT JDCT_ISLOW -#endif -#ifndef JDCT_FASTEST /* may be overridden in jconfig.h */ -#define JDCT_FASTEST JDCT_IFAST -#endif - -/* Dithering options for decompression. */ - -typedef enum { - JDITHER_NONE, /* no dithering */ - JDITHER_ORDERED, /* simple ordered dither */ - JDITHER_FS /* Floyd-Steinberg error diffusion dither */ -} J_DITHER_MODE; - - -/* Common fields between JPEG compression and decompression master structs. */ - -#define jpeg_common_fields \ - struct jpeg_error_mgr * err; /* Error handler module */\ - struct jpeg_memory_mgr * mem; /* Memory manager module */\ - struct jpeg_progress_mgr * progress; /* Progress monitor, or NULL if none */\ - void * client_data; /* Available for use by application */\ - boolean is_decompressor; /* So common code can tell which is which */\ - int global_state /* For checking call sequence validity */ - -/* Routines that are to be used by both halves of the library are declared - * to receive a pointer to this structure. There are no actual instances of - * jpeg_common_struct, only of jpeg_compress_struct and jpeg_decompress_struct. - */ -struct jpeg_common_struct { - jpeg_common_fields; /* Fields common to both master struct types */ - /* Additional fields follow in an actual jpeg_compress_struct or - * jpeg_decompress_struct. All three structs must agree on these - * initial fields! (This would be a lot cleaner in C++.) - */ -}; - -typedef struct jpeg_common_struct * j_common_ptr; -typedef struct jpeg_compress_struct * j_compress_ptr; -typedef struct jpeg_decompress_struct * j_decompress_ptr; - - -/* Master record for a compression instance */ - -struct jpeg_compress_struct { - jpeg_common_fields; /* Fields shared with jpeg_decompress_struct */ - - /* Destination for compressed data */ - struct jpeg_destination_mgr * dest; - - /* Description of source image --- these fields must be filled in by - * outer application before starting compression. in_color_space must - * be correct before you can even call jpeg_set_defaults(). - */ - - JDIMENSION image_width; /* input image width */ - JDIMENSION image_height; /* input image height */ - int input_components; /* # of color components in input image */ - J_COLOR_SPACE in_color_space; /* colorspace of input image */ - - double input_gamma; /* image gamma of input image */ - - /* Compression parameters --- these fields must be set before calling - * jpeg_start_compress(). We recommend calling jpeg_set_defaults() to - * initialize everything to reasonable defaults, then changing anything - * the application specifically wants to change. That way you won't get - * burnt when new parameters are added. Also note that there are several - * helper routines to simplify changing parameters. - */ - - unsigned int scale_num, scale_denom; /* fraction by which to scale image */ - - JDIMENSION jpeg_width; /* scaled JPEG image width */ - JDIMENSION jpeg_height; /* scaled JPEG image height */ - /* Dimensions of actual JPEG image that will be written to file, - * derived from input dimensions by scaling factors above. - * These fields are computed by jpeg_start_compress(). - * You can also use jpeg_calc_jpeg_dimensions() to determine these values - * in advance of calling jpeg_start_compress(). - */ - - int data_precision; /* bits of precision in image data */ - - int num_components; /* # of color components in JPEG image */ - J_COLOR_SPACE jpeg_color_space; /* colorspace of JPEG image */ - - jpeg_component_info * comp_info; - /* comp_info[i] describes component that appears i'th in SOF */ - - JQUANT_TBL * quant_tbl_ptrs[NUM_QUANT_TBLS]; - int q_scale_factor[NUM_QUANT_TBLS]; - /* ptrs to coefficient quantization tables, or NULL if not defined, - * and corresponding scale factors (percentage, initialized 100). - */ - - JHUFF_TBL * dc_huff_tbl_ptrs[NUM_HUFF_TBLS]; - JHUFF_TBL * ac_huff_tbl_ptrs[NUM_HUFF_TBLS]; - /* ptrs to Huffman coding tables, or NULL if not defined */ - - UINT8 arith_dc_L[NUM_ARITH_TBLS]; /* L values for DC arith-coding tables */ - UINT8 arith_dc_U[NUM_ARITH_TBLS]; /* U values for DC arith-coding tables */ - UINT8 arith_ac_K[NUM_ARITH_TBLS]; /* Kx values for AC arith-coding tables */ - - int num_scans; /* # of entries in scan_info array */ - const jpeg_scan_info * scan_info; /* script for multi-scan file, or NULL */ - /* The default value of scan_info is NULL, which causes a single-scan - * sequential JPEG file to be emitted. To create a multi-scan file, - * set num_scans and scan_info to point to an array of scan definitions. - */ - - boolean raw_data_in; /* TRUE=caller supplies downsampled data */ - boolean arith_code; /* TRUE=arithmetic coding, FALSE=Huffman */ - boolean optimize_coding; /* TRUE=optimize entropy encoding parms */ - boolean CCIR601_sampling; /* TRUE=first samples are cosited */ - boolean do_fancy_downsampling; /* TRUE=apply fancy downsampling */ - int smoothing_factor; /* 1..100, or 0 for no input smoothing */ - J_DCT_METHOD dct_method; /* DCT algorithm selector */ - - /* The restart interval can be specified in absolute MCUs by setting - * restart_interval, or in MCU rows by setting restart_in_rows - * (in which case the correct restart_interval will be figured - * for each scan). - */ - unsigned int restart_interval; /* MCUs per restart, or 0 for no restart */ - int restart_in_rows; /* if > 0, MCU rows per restart interval */ - - /* Parameters controlling emission of special markers. */ - - boolean write_JFIF_header; /* should a JFIF marker be written? */ - UINT8 JFIF_major_version; /* What to write for the JFIF version number */ - UINT8 JFIF_minor_version; - /* These three values are not used by the JPEG code, merely copied */ - /* into the JFIF APP0 marker. density_unit can be 0 for unknown, */ - /* 1 for dots/inch, or 2 for dots/cm. Note that the pixel aspect */ - /* ratio is defined by X_density/Y_density even when density_unit=0. */ - UINT8 density_unit; /* JFIF code for pixel size units */ - UINT16 X_density; /* Horizontal pixel density */ - UINT16 Y_density; /* Vertical pixel density */ - boolean write_Adobe_marker; /* should an Adobe marker be written? */ - - J_COLOR_TRANSFORM color_transform; - /* Color transform identifier, writes LSE marker if nonzero */ - - /* State variable: index of next scanline to be written to - * jpeg_write_scanlines(). Application may use this to control its - * processing loop, e.g., "while (next_scanline < image_height)". - */ - - JDIMENSION next_scanline; /* 0 .. image_height-1 */ - - /* Remaining fields are known throughout compressor, but generally - * should not be touched by a surrounding application. - */ - - /* - * These fields are computed during compression startup - */ - boolean progressive_mode; /* TRUE if scan script uses progressive mode */ - int max_h_samp_factor; /* largest h_samp_factor */ - int max_v_samp_factor; /* largest v_samp_factor */ - - int min_DCT_h_scaled_size; /* smallest DCT_h_scaled_size of any component */ - int min_DCT_v_scaled_size; /* smallest DCT_v_scaled_size of any component */ - - JDIMENSION total_iMCU_rows; /* # of iMCU rows to be input to coef ctlr */ - /* The coefficient controller receives data in units of MCU rows as defined - * for fully interleaved scans (whether the JPEG file is interleaved or not). - * There are v_samp_factor * DCT_v_scaled_size sample rows of each component - * in an "iMCU" (interleaved MCU) row. - */ - - /* - * These fields are valid during any one scan. - * They describe the components and MCUs actually appearing in the scan. - */ - int comps_in_scan; /* # of JPEG components in this scan */ - jpeg_component_info * cur_comp_info[MAX_COMPS_IN_SCAN]; - /* *cur_comp_info[i] describes component that appears i'th in SOS */ - - JDIMENSION MCUs_per_row; /* # of MCUs across the image */ - JDIMENSION MCU_rows_in_scan; /* # of MCU rows in the image */ - - int blocks_in_MCU; /* # of DCT blocks per MCU */ - int MCU_membership[C_MAX_BLOCKS_IN_MCU]; - /* MCU_membership[i] is index in cur_comp_info of component owning */ - /* i'th block in an MCU */ - - int Ss, Se, Ah, Al; /* progressive JPEG parameters for scan */ - - int block_size; /* the basic DCT block size: 1..16 */ - const int * natural_order; /* natural-order position array */ - int lim_Se; /* min( Se, DCTSIZE2-1 ) */ - - /* - * Links to compression subobjects (methods and private variables of modules) - */ - struct jpeg_comp_master * master; - struct jpeg_c_main_controller * main; - struct jpeg_c_prep_controller * prep; - struct jpeg_c_coef_controller * coef; - struct jpeg_marker_writer * marker; - struct jpeg_color_converter * cconvert; - struct jpeg_downsampler * downsample; - struct jpeg_forward_dct * fdct; - struct jpeg_entropy_encoder * entropy; - jpeg_scan_info * script_space; /* workspace for jpeg_simple_progression */ - int script_space_size; -}; - - -/* Master record for a decompression instance */ - -struct jpeg_decompress_struct { - jpeg_common_fields; /* Fields shared with jpeg_compress_struct */ - - /* Source of compressed data */ - struct jpeg_source_mgr * src; - - /* Basic description of image --- filled in by jpeg_read_header(). */ - /* Application may inspect these values to decide how to process image. */ - - JDIMENSION image_width; /* nominal image width (from SOF marker) */ - JDIMENSION image_height; /* nominal image height */ - int num_components; /* # of color components in JPEG image */ - J_COLOR_SPACE jpeg_color_space; /* colorspace of JPEG image */ - - /* Decompression processing parameters --- these fields must be set before - * calling jpeg_start_decompress(). Note that jpeg_read_header() initializes - * them to default values. - */ - - J_COLOR_SPACE out_color_space; /* colorspace for output */ - - unsigned int scale_num, scale_denom; /* fraction by which to scale image */ - - double output_gamma; /* image gamma wanted in output */ - - boolean buffered_image; /* TRUE=multiple output passes */ - boolean raw_data_out; /* TRUE=downsampled data wanted */ - - J_DCT_METHOD dct_method; /* IDCT algorithm selector */ - boolean do_fancy_upsampling; /* TRUE=apply fancy upsampling */ - boolean do_block_smoothing; /* TRUE=apply interblock smoothing */ - - boolean quantize_colors; /* TRUE=colormapped output wanted */ - /* the following are ignored if not quantize_colors: */ - J_DITHER_MODE dither_mode; /* type of color dithering to use */ - boolean two_pass_quantize; /* TRUE=use two-pass color quantization */ - int desired_number_of_colors; /* max # colors to use in created colormap */ - /* these are significant only in buffered-image mode: */ - boolean enable_1pass_quant; /* enable future use of 1-pass quantizer */ - boolean enable_external_quant;/* enable future use of external colormap */ - boolean enable_2pass_quant; /* enable future use of 2-pass quantizer */ - - /* Description of actual output image that will be returned to application. - * These fields are computed by jpeg_start_decompress(). - * You can also use jpeg_calc_output_dimensions() to determine these values - * in advance of calling jpeg_start_decompress(). - */ - - JDIMENSION output_width; /* scaled image width */ - JDIMENSION output_height; /* scaled image height */ - int out_color_components; /* # of color components in out_color_space */ - int output_components; /* # of color components returned */ - /* output_components is 1 (a colormap index) when quantizing colors; - * otherwise it equals out_color_components. - */ - int rec_outbuf_height; /* min recommended height of scanline buffer */ - /* If the buffer passed to jpeg_read_scanlines() is less than this many rows - * high, space and time will be wasted due to unnecessary data copying. - * Usually rec_outbuf_height will be 1 or 2, at most 4. - */ - - /* When quantizing colors, the output colormap is described by these fields. - * The application can supply a colormap by setting colormap non-NULL before - * calling jpeg_start_decompress; otherwise a colormap is created during - * jpeg_start_decompress or jpeg_start_output. - * The map has out_color_components rows and actual_number_of_colors columns. - */ - int actual_number_of_colors; /* number of entries in use */ - JSAMPARRAY colormap; /* The color map as a 2-D pixel array */ - - /* State variables: these variables indicate the progress of decompression. - * The application may examine these but must not modify them. - */ - - /* Row index of next scanline to be read from jpeg_read_scanlines(). - * Application may use this to control its processing loop, e.g., - * "while (output_scanline < output_height)". - */ - JDIMENSION output_scanline; /* 0 .. output_height-1 */ - - /* Current input scan number and number of iMCU rows completed in scan. - * These indicate the progress of the decompressor input side. - */ - int input_scan_number; /* Number of SOS markers seen so far */ - JDIMENSION input_iMCU_row; /* Number of iMCU rows completed */ - - /* The "output scan number" is the notional scan being displayed by the - * output side. The decompressor will not allow output scan/row number - * to get ahead of input scan/row, but it can fall arbitrarily far behind. - */ - int output_scan_number; /* Nominal scan number being displayed */ - JDIMENSION output_iMCU_row; /* Number of iMCU rows read */ - - /* Current progression status. coef_bits[c][i] indicates the precision - * with which component c's DCT coefficient i (in zigzag order) is known. - * It is -1 when no data has yet been received, otherwise it is the point - * transform (shift) value for the most recent scan of the coefficient - * (thus, 0 at completion of the progression). - * This pointer is NULL when reading a non-progressive file. - */ - int (*coef_bits)[DCTSIZE2]; /* -1 or current Al value for each coef */ - - /* Internal JPEG parameters --- the application usually need not look at - * these fields. Note that the decompressor output side may not use - * any parameters that can change between scans. - */ - - /* Quantization and Huffman tables are carried forward across input - * datastreams when processing abbreviated JPEG datastreams. - */ - - JQUANT_TBL * quant_tbl_ptrs[NUM_QUANT_TBLS]; - /* ptrs to coefficient quantization tables, or NULL if not defined */ - - JHUFF_TBL * dc_huff_tbl_ptrs[NUM_HUFF_TBLS]; - JHUFF_TBL * ac_huff_tbl_ptrs[NUM_HUFF_TBLS]; - /* ptrs to Huffman coding tables, or NULL if not defined */ - - /* These parameters are never carried across datastreams, since they - * are given in SOF/SOS markers or defined to be reset by SOI. - */ - - int data_precision; /* bits of precision in image data */ - - jpeg_component_info * comp_info; - /* comp_info[i] describes component that appears i'th in SOF */ - - boolean is_baseline; /* TRUE if Baseline SOF0 encountered */ - boolean progressive_mode; /* TRUE if SOFn specifies progressive mode */ - boolean arith_code; /* TRUE=arithmetic coding, FALSE=Huffman */ - - UINT8 arith_dc_L[NUM_ARITH_TBLS]; /* L values for DC arith-coding tables */ - UINT8 arith_dc_U[NUM_ARITH_TBLS]; /* U values for DC arith-coding tables */ - UINT8 arith_ac_K[NUM_ARITH_TBLS]; /* Kx values for AC arith-coding tables */ - - unsigned int restart_interval; /* MCUs per restart interval, or 0 for no restart */ - - /* These fields record data obtained from optional markers recognized by - * the JPEG library. - */ - boolean saw_JFIF_marker; /* TRUE iff a JFIF APP0 marker was found */ - /* Data copied from JFIF marker; only valid if saw_JFIF_marker is TRUE: */ - UINT8 JFIF_major_version; /* JFIF version number */ - UINT8 JFIF_minor_version; - UINT8 density_unit; /* JFIF code for pixel size units */ - UINT16 X_density; /* Horizontal pixel density */ - UINT16 Y_density; /* Vertical pixel density */ - boolean saw_Adobe_marker; /* TRUE iff an Adobe APP14 marker was found */ - UINT8 Adobe_transform; /* Color transform code from Adobe marker */ - - J_COLOR_TRANSFORM color_transform; - /* Color transform identifier derived from LSE marker, otherwise zero */ - - boolean CCIR601_sampling; /* TRUE=first samples are cosited */ - - /* Aside from the specific data retained from APPn markers known to the - * library, the uninterpreted contents of any or all APPn and COM markers - * can be saved in a list for examination by the application. - */ - jpeg_saved_marker_ptr marker_list; /* Head of list of saved markers */ - - /* Remaining fields are known throughout decompressor, but generally - * should not be touched by a surrounding application. - */ - - /* - * These fields are computed during decompression startup - */ - int max_h_samp_factor; /* largest h_samp_factor */ - int max_v_samp_factor; /* largest v_samp_factor */ - - int min_DCT_h_scaled_size; /* smallest DCT_h_scaled_size of any component */ - int min_DCT_v_scaled_size; /* smallest DCT_v_scaled_size of any component */ - - JDIMENSION total_iMCU_rows; /* # of iMCU rows in image */ - /* The coefficient controller's input and output progress is measured in - * units of "iMCU" (interleaved MCU) rows. These are the same as MCU rows - * in fully interleaved JPEG scans, but are used whether the scan is - * interleaved or not. We define an iMCU row as v_samp_factor DCT block - * rows of each component. Therefore, the IDCT output contains - * v_samp_factor * DCT_v_scaled_size sample rows of a component per iMCU row. - */ - - JSAMPLE * sample_range_limit; /* table for fast range-limiting */ - - /* - * These fields are valid during any one scan. - * They describe the components and MCUs actually appearing in the scan. - * Note that the decompressor output side must not use these fields. - */ - int comps_in_scan; /* # of JPEG components in this scan */ - jpeg_component_info * cur_comp_info[MAX_COMPS_IN_SCAN]; - /* *cur_comp_info[i] describes component that appears i'th in SOS */ - - JDIMENSION MCUs_per_row; /* # of MCUs across the image */ - JDIMENSION MCU_rows_in_scan; /* # of MCU rows in the image */ - - int blocks_in_MCU; /* # of DCT blocks per MCU */ - int MCU_membership[D_MAX_BLOCKS_IN_MCU]; - /* MCU_membership[i] is index in cur_comp_info of component owning */ - /* i'th block in an MCU */ - - int Ss, Se, Ah, Al; /* progressive JPEG parameters for scan */ - - /* These fields are derived from Se of first SOS marker. - */ - int block_size; /* the basic DCT block size: 1..16 */ - const int * natural_order; /* natural-order position array for entropy decode */ - int lim_Se; /* min( Se, DCTSIZE2-1 ) for entropy decode */ - - /* This field is shared between entropy decoder and marker parser. - * It is either zero or the code of a JPEG marker that has been - * read from the data source, but has not yet been processed. - */ - int unread_marker; - - /* - * Links to decompression subobjects (methods, private variables of modules) - */ - struct jpeg_decomp_master * master; - struct jpeg_d_main_controller * main; - struct jpeg_d_coef_controller * coef; - struct jpeg_d_post_controller * post; - struct jpeg_input_controller * inputctl; - struct jpeg_marker_reader * marker; - struct jpeg_entropy_decoder * entropy; - struct jpeg_inverse_dct * idct; - struct jpeg_upsampler * upsample; - struct jpeg_color_deconverter * cconvert; - struct jpeg_color_quantizer * cquantize; -}; - - -/* "Object" declarations for JPEG modules that may be supplied or called - * directly by the surrounding application. - * As with all objects in the JPEG library, these structs only define the - * publicly visible methods and state variables of a module. Additional - * private fields may exist after the public ones. - */ - - -/* Error handler object */ - -struct jpeg_error_mgr { - /* Error exit handler: does not return to caller */ - JMETHOD(noreturn_t, error_exit, (j_common_ptr cinfo)); - /* Conditionally emit a trace or warning message */ - JMETHOD(void, emit_message, (j_common_ptr cinfo, int msg_level)); - /* Routine that actually outputs a trace or error message */ - JMETHOD(void, output_message, (j_common_ptr cinfo)); - /* Format a message string for the most recent JPEG error or message */ - JMETHOD(void, format_message, (j_common_ptr cinfo, char * buffer)); -#define JMSG_LENGTH_MAX 200 /* recommended size of format_message buffer */ - /* Reset error state variables at start of a new image */ - JMETHOD(void, reset_error_mgr, (j_common_ptr cinfo)); - - /* The message ID code and any parameters are saved here. - * A message can have one string parameter or up to 8 int parameters. - */ - int msg_code; -#define JMSG_STR_PARM_MAX 80 - union { - int i[8]; - char s[JMSG_STR_PARM_MAX]; - } msg_parm; - - /* Standard state variables for error facility */ - - int trace_level; /* max msg_level that will be displayed */ - - /* For recoverable corrupt-data errors, we emit a warning message, - * but keep going unless emit_message chooses to abort. emit_message - * should count warnings in num_warnings. The surrounding application - * can check for bad data by seeing if num_warnings is nonzero at the - * end of processing. - */ - long num_warnings; /* number of corrupt-data warnings */ - - /* These fields point to the table(s) of error message strings. - * An application can change the table pointer to switch to a different - * message list (typically, to change the language in which errors are - * reported). Some applications may wish to add additional error codes - * that will be handled by the JPEG library error mechanism; the second - * table pointer is used for this purpose. - * - * First table includes all errors generated by JPEG library itself. - * Error code 0 is reserved for a "no such error string" message. - */ - const char * const * jpeg_message_table; /* Library errors */ - int last_jpeg_message; /* Table contains strings 0..last_jpeg_message */ - /* Second table can be added by application (see cjpeg/djpeg for example). - * It contains strings numbered first_addon_message..last_addon_message. - */ - const char * const * addon_message_table; /* Non-library errors */ - int first_addon_message; /* code for first string in addon table */ - int last_addon_message; /* code for last string in addon table */ -}; - - -/* Progress monitor object */ - -struct jpeg_progress_mgr { - JMETHOD(void, progress_monitor, (j_common_ptr cinfo)); - - long pass_counter; /* work units completed in this pass */ - long pass_limit; /* total number of work units in this pass */ - int completed_passes; /* passes completed so far */ - int total_passes; /* total number of passes expected */ -}; - - -/* Data destination object for compression */ - -struct jpeg_destination_mgr { - JOCTET * next_output_byte; /* => next byte to write in buffer */ - size_t free_in_buffer; /* # of byte spaces remaining in buffer */ - - JMETHOD(void, init_destination, (j_compress_ptr cinfo)); - JMETHOD(boolean, empty_output_buffer, (j_compress_ptr cinfo)); - JMETHOD(void, term_destination, (j_compress_ptr cinfo)); -}; - - -/* Data source object for decompression */ - -struct jpeg_source_mgr { - const JOCTET * next_input_byte; /* => next byte to read from buffer */ - size_t bytes_in_buffer; /* # of bytes remaining in buffer */ - - JMETHOD(void, init_source, (j_decompress_ptr cinfo)); - JMETHOD(boolean, fill_input_buffer, (j_decompress_ptr cinfo)); - JMETHOD(void, skip_input_data, (j_decompress_ptr cinfo, long num_bytes)); - JMETHOD(boolean, resync_to_restart, (j_decompress_ptr cinfo, int desired)); - JMETHOD(void, term_source, (j_decompress_ptr cinfo)); -}; - - -/* Memory manager object. - * Allocates "small" objects (a few K total), "large" objects (tens of K), - * and "really big" objects (virtual arrays with backing store if needed). - * The memory manager does not allow individual objects to be freed; rather, - * each created object is assigned to a pool, and whole pools can be freed - * at once. This is faster and more convenient than remembering exactly what - * to free, especially where malloc()/free() are not too speedy. - * NB: alloc routines never return NULL. They exit to error_exit if not - * successful. - */ - -#define JPOOL_PERMANENT 0 /* lasts until master record is destroyed */ -#define JPOOL_IMAGE 1 /* lasts until done with image/datastream */ -#define JPOOL_NUMPOOLS 2 - -typedef struct jvirt_sarray_control * jvirt_sarray_ptr; -typedef struct jvirt_barray_control * jvirt_barray_ptr; - - -struct jpeg_memory_mgr { - /* Method pointers */ - JMETHOD(void *, alloc_small, (j_common_ptr cinfo, int pool_id, - size_t sizeofobject)); - JMETHOD(void FAR *, alloc_large, (j_common_ptr cinfo, int pool_id, - size_t sizeofobject)); - JMETHOD(JSAMPARRAY, alloc_sarray, (j_common_ptr cinfo, int pool_id, - JDIMENSION samplesperrow, - JDIMENSION numrows)); - JMETHOD(JBLOCKARRAY, alloc_barray, (j_common_ptr cinfo, int pool_id, - JDIMENSION blocksperrow, - JDIMENSION numrows)); - JMETHOD(jvirt_sarray_ptr, request_virt_sarray, (j_common_ptr cinfo, - int pool_id, - boolean pre_zero, - JDIMENSION samplesperrow, - JDIMENSION numrows, - JDIMENSION maxaccess)); - JMETHOD(jvirt_barray_ptr, request_virt_barray, (j_common_ptr cinfo, - int pool_id, - boolean pre_zero, - JDIMENSION blocksperrow, - JDIMENSION numrows, - JDIMENSION maxaccess)); - JMETHOD(void, realize_virt_arrays, (j_common_ptr cinfo)); - JMETHOD(JSAMPARRAY, access_virt_sarray, (j_common_ptr cinfo, - jvirt_sarray_ptr ptr, - JDIMENSION start_row, - JDIMENSION num_rows, - boolean writable)); - JMETHOD(JBLOCKARRAY, access_virt_barray, (j_common_ptr cinfo, - jvirt_barray_ptr ptr, - JDIMENSION start_row, - JDIMENSION num_rows, - boolean writable)); - JMETHOD(void, free_pool, (j_common_ptr cinfo, int pool_id)); - JMETHOD(void, self_destruct, (j_common_ptr cinfo)); - - /* Limit on memory allocation for this JPEG object. (Note that this is - * merely advisory, not a guaranteed maximum; it only affects the space - * used for virtual-array buffers.) May be changed by outer application - * after creating the JPEG object. - */ - long max_memory_to_use; - - /* Maximum allocation request accepted by alloc_large. */ - long max_alloc_chunk; -}; - - -/* Routine signature for application-supplied marker processing methods. - * Need not pass marker code since it is stored in cinfo->unread_marker. - */ -typedef JMETHOD(boolean, jpeg_marker_parser_method, (j_decompress_ptr cinfo)); - - -/* Declarations for routines called by application. - * The JPP macro hides prototype parameters from compilers that can't cope. - * Note JPP requires double parentheses. - */ - -#ifdef HAVE_PROTOTYPES -#define JPP(arglist) arglist -#else -#define JPP(arglist) () -#endif - - -/* Short forms of external names for systems with brain-damaged linkers. - * We shorten external names to be unique in the first six letters, which - * is good enough for all known systems. - * (If your compiler itself needs names to be unique in less than 15 - * characters, you are out of luck. Get a better compiler.) - */ - -#ifdef NEED_SHORT_EXTERNAL_NAMES -#define jpeg_std_error jStdError -#define jpeg_CreateCompress jCreaCompress -#define jpeg_CreateDecompress jCreaDecompress -#define jpeg_destroy_compress jDestCompress -#define jpeg_destroy_decompress jDestDecompress -#define jpeg_stdio_dest jStdDest -#define jpeg_stdio_src jStdSrc -#define jpeg_mem_dest jMemDest -#define jpeg_mem_src jMemSrc -#define jpeg_set_defaults jSetDefaults -#define jpeg_set_colorspace jSetColorspace -#define jpeg_default_colorspace jDefColorspace -#define jpeg_set_quality jSetQuality -#define jpeg_set_linear_quality jSetLQuality -#define jpeg_default_qtables jDefQTables -#define jpeg_add_quant_table jAddQuantTable -#define jpeg_quality_scaling jQualityScaling -#define jpeg_simple_progression jSimProgress -#define jpeg_suppress_tables jSuppressTables -#define jpeg_alloc_quant_table jAlcQTable -#define jpeg_alloc_huff_table jAlcHTable -#define jpeg_std_huff_table jStdHTable -#define jpeg_start_compress jStrtCompress -#define jpeg_write_scanlines jWrtScanlines -#define jpeg_finish_compress jFinCompress -#define jpeg_calc_jpeg_dimensions jCjpegDimensions -#define jpeg_write_raw_data jWrtRawData -#define jpeg_write_marker jWrtMarker -#define jpeg_write_m_header jWrtMHeader -#define jpeg_write_m_byte jWrtMByte -#define jpeg_write_tables jWrtTables -#define jpeg_read_header jReadHeader -#define jpeg_start_decompress jStrtDecompress -#define jpeg_read_scanlines jReadScanlines -#define jpeg_finish_decompress jFinDecompress -#define jpeg_read_raw_data jReadRawData -#define jpeg_has_multiple_scans jHasMultScn -#define jpeg_start_output jStrtOutput -#define jpeg_finish_output jFinOutput -#define jpeg_input_complete jInComplete -#define jpeg_new_colormap jNewCMap -#define jpeg_consume_input jConsumeInput -#define jpeg_core_output_dimensions jCoreDimensions -#define jpeg_calc_output_dimensions jCalcDimensions -#define jpeg_save_markers jSaveMarkers -#define jpeg_set_marker_processor jSetMarker -#define jpeg_read_coefficients jReadCoefs -#define jpeg_write_coefficients jWrtCoefs -#define jpeg_copy_critical_parameters jCopyCrit -#define jpeg_abort_compress jAbrtCompress -#define jpeg_abort_decompress jAbrtDecompress -#define jpeg_abort jAbort -#define jpeg_destroy jDestroy -#define jpeg_resync_to_restart jResyncRestart -#endif /* NEED_SHORT_EXTERNAL_NAMES */ - - -/* Default error-management setup */ -EXTERN(struct jpeg_error_mgr *) jpeg_std_error - JPP((struct jpeg_error_mgr * err)); - -/* Initialization of JPEG compression objects. - * jpeg_create_compress() and jpeg_create_decompress() are the exported - * names that applications should call. These expand to calls on - * jpeg_CreateCompress and jpeg_CreateDecompress with additional information - * passed for version mismatch checking. - * NB: you must set up the error-manager BEFORE calling jpeg_create_xxx. - */ -#define jpeg_create_compress(cinfo) \ - jpeg_CreateCompress((cinfo), JPEG_LIB_VERSION, \ - (size_t) sizeof(struct jpeg_compress_struct)) -#define jpeg_create_decompress(cinfo) \ - jpeg_CreateDecompress((cinfo), JPEG_LIB_VERSION, \ - (size_t) sizeof(struct jpeg_decompress_struct)) -EXTERN(void) jpeg_CreateCompress JPP((j_compress_ptr cinfo, - int version, size_t structsize)); -EXTERN(void) jpeg_CreateDecompress JPP((j_decompress_ptr cinfo, - int version, size_t structsize)); -/* Destruction of JPEG compression objects */ -EXTERN(void) jpeg_destroy_compress JPP((j_compress_ptr cinfo)); -EXTERN(void) jpeg_destroy_decompress JPP((j_decompress_ptr cinfo)); - -/* Standard data source and destination managers: stdio streams. */ -/* Caller is responsible for opening the file before and closing after. */ -EXTERN(void) jpeg_stdio_dest JPP((j_compress_ptr cinfo, FILE * outfile)); -EXTERN(void) jpeg_stdio_src JPP((j_decompress_ptr cinfo, FILE * infile)); - -/* Data source and destination managers: memory buffers. */ -EXTERN(void) jpeg_mem_dest JPP((j_compress_ptr cinfo, - unsigned char ** outbuffer, - size_t * outsize)); -EXTERN(void) jpeg_mem_src JPP((j_decompress_ptr cinfo, - const unsigned char * inbuffer, - size_t insize)); - -/* Default parameter setup for compression */ -EXTERN(void) jpeg_set_defaults JPP((j_compress_ptr cinfo)); -/* Compression parameter setup aids */ -EXTERN(void) jpeg_set_colorspace JPP((j_compress_ptr cinfo, - J_COLOR_SPACE colorspace)); -EXTERN(void) jpeg_default_colorspace JPP((j_compress_ptr cinfo)); -EXTERN(void) jpeg_set_quality JPP((j_compress_ptr cinfo, int quality, - boolean force_baseline)); -EXTERN(void) jpeg_set_linear_quality JPP((j_compress_ptr cinfo, - int scale_factor, - boolean force_baseline)); -EXTERN(void) jpeg_default_qtables JPP((j_compress_ptr cinfo, - boolean force_baseline)); -EXTERN(void) jpeg_add_quant_table JPP((j_compress_ptr cinfo, int which_tbl, - const unsigned int *basic_table, - int scale_factor, - boolean force_baseline)); -EXTERN(int) jpeg_quality_scaling JPP((int quality)); -EXTERN(void) jpeg_simple_progression JPP((j_compress_ptr cinfo)); -EXTERN(void) jpeg_suppress_tables JPP((j_compress_ptr cinfo, - boolean suppress)); -EXTERN(JQUANT_TBL *) jpeg_alloc_quant_table JPP((j_common_ptr cinfo)); -EXTERN(JHUFF_TBL *) jpeg_alloc_huff_table JPP((j_common_ptr cinfo)); -EXTERN(JHUFF_TBL *) jpeg_std_huff_table JPP((j_common_ptr cinfo, - boolean isDC, int tblno)); - -/* Main entry points for compression */ -EXTERN(void) jpeg_start_compress JPP((j_compress_ptr cinfo, - boolean write_all_tables)); -EXTERN(JDIMENSION) jpeg_write_scanlines JPP((j_compress_ptr cinfo, - JSAMPARRAY scanlines, - JDIMENSION num_lines)); -EXTERN(void) jpeg_finish_compress JPP((j_compress_ptr cinfo)); - -/* Precalculate JPEG dimensions for current compression parameters. */ -EXTERN(void) jpeg_calc_jpeg_dimensions JPP((j_compress_ptr cinfo)); - -/* Replaces jpeg_write_scanlines when writing raw downsampled data. */ -EXTERN(JDIMENSION) jpeg_write_raw_data JPP((j_compress_ptr cinfo, - JSAMPIMAGE data, - JDIMENSION num_lines)); - -/* Write a special marker. See libjpeg.txt concerning safe usage. */ -EXTERN(void) jpeg_write_marker - JPP((j_compress_ptr cinfo, int marker, - const JOCTET * dataptr, unsigned int datalen)); -/* Same, but piecemeal. */ -EXTERN(void) jpeg_write_m_header - JPP((j_compress_ptr cinfo, int marker, unsigned int datalen)); -EXTERN(void) jpeg_write_m_byte - JPP((j_compress_ptr cinfo, int val)); - -/* Alternate compression function: just write an abbreviated table file */ -EXTERN(void) jpeg_write_tables JPP((j_compress_ptr cinfo)); - -/* Decompression startup: read start of JPEG datastream to see what's there */ -EXTERN(int) jpeg_read_header JPP((j_decompress_ptr cinfo, - boolean require_image)); -/* Return value is one of: */ -#define JPEG_SUSPENDED 0 /* Suspended due to lack of input data */ -#define JPEG_HEADER_OK 1 /* Found valid image datastream */ -#define JPEG_HEADER_TABLES_ONLY 2 /* Found valid table-specs-only datastream */ -/* If you pass require_image = TRUE (normal case), you need not check for - * a TABLES_ONLY return code; an abbreviated file will cause an error exit. - * JPEG_SUSPENDED is only possible if you use a data source module that can - * give a suspension return (the stdio source module doesn't). - */ - -/* Main entry points for decompression */ -EXTERN(boolean) jpeg_start_decompress JPP((j_decompress_ptr cinfo)); -EXTERN(JDIMENSION) jpeg_read_scanlines JPP((j_decompress_ptr cinfo, - JSAMPARRAY scanlines, - JDIMENSION max_lines)); -EXTERN(boolean) jpeg_finish_decompress JPP((j_decompress_ptr cinfo)); - -/* Replaces jpeg_read_scanlines when reading raw downsampled data. */ -EXTERN(JDIMENSION) jpeg_read_raw_data JPP((j_decompress_ptr cinfo, - JSAMPIMAGE data, - JDIMENSION max_lines)); - -/* Additional entry points for buffered-image mode. */ -EXTERN(boolean) jpeg_has_multiple_scans JPP((j_decompress_ptr cinfo)); -EXTERN(boolean) jpeg_start_output JPP((j_decompress_ptr cinfo, - int scan_number)); -EXTERN(boolean) jpeg_finish_output JPP((j_decompress_ptr cinfo)); -EXTERN(boolean) jpeg_input_complete JPP((j_decompress_ptr cinfo)); -EXTERN(void) jpeg_new_colormap JPP((j_decompress_ptr cinfo)); -EXTERN(int) jpeg_consume_input JPP((j_decompress_ptr cinfo)); -/* Return value is one of: */ -/* #define JPEG_SUSPENDED 0 Suspended due to lack of input data */ -#define JPEG_REACHED_SOS 1 /* Reached start of new scan */ -#define JPEG_REACHED_EOI 2 /* Reached end of image */ -#define JPEG_ROW_COMPLETED 3 /* Completed one iMCU row */ -#define JPEG_SCAN_COMPLETED 4 /* Completed last iMCU row of a scan */ - -/* Precalculate output dimensions for current decompression parameters. */ -EXTERN(void) jpeg_core_output_dimensions JPP((j_decompress_ptr cinfo)); -EXTERN(void) jpeg_calc_output_dimensions JPP((j_decompress_ptr cinfo)); - -/* Control saving of COM and APPn markers into marker_list. */ -EXTERN(void) jpeg_save_markers - JPP((j_decompress_ptr cinfo, int marker_code, - unsigned int length_limit)); - -/* Install a special processing method for COM or APPn markers. */ -EXTERN(void) jpeg_set_marker_processor - JPP((j_decompress_ptr cinfo, int marker_code, - jpeg_marker_parser_method routine)); - -/* Read or write raw DCT coefficients --- useful for lossless transcoding. */ -EXTERN(jvirt_barray_ptr *) jpeg_read_coefficients JPP((j_decompress_ptr cinfo)); -EXTERN(void) jpeg_write_coefficients JPP((j_compress_ptr cinfo, - jvirt_barray_ptr * coef_arrays)); -EXTERN(void) jpeg_copy_critical_parameters JPP((j_decompress_ptr srcinfo, - j_compress_ptr dstinfo)); - -/* If you choose to abort compression or decompression before completing - * jpeg_finish_(de)compress, then you need to clean up to release memory, - * temporary files, etc. You can just call jpeg_destroy_(de)compress - * if you're done with the JPEG object, but if you want to clean it up and - * reuse it, call this: - */ -EXTERN(void) jpeg_abort_compress JPP((j_compress_ptr cinfo)); -EXTERN(void) jpeg_abort_decompress JPP((j_decompress_ptr cinfo)); - -/* Generic versions of jpeg_abort and jpeg_destroy that work on either - * flavor of JPEG object. These may be more convenient in some places. - */ -EXTERN(void) jpeg_abort JPP((j_common_ptr cinfo)); -EXTERN(void) jpeg_destroy JPP((j_common_ptr cinfo)); - -/* Default restart-marker-resync procedure for use by data source modules */ -EXTERN(boolean) jpeg_resync_to_restart JPP((j_decompress_ptr cinfo, - int desired)); - - -/* These marker codes are exported since applications and data source modules - * are likely to want to use them. - */ - -#define JPEG_RST0 0xD0 /* RST0 marker code */ -#define JPEG_EOI 0xD9 /* EOI marker code */ -#define JPEG_APP0 0xE0 /* APP0 marker code */ -#define JPEG_COM 0xFE /* COM marker code */ - - -/* If we have a brain-damaged compiler that emits warnings (or worse, errors) - * for structure definitions that are never filled in, keep it quiet by - * supplying dummy definitions for the various substructures. - */ - -#ifdef INCOMPLETE_TYPES_BROKEN -#ifndef JPEG_INTERNALS /* will be defined in jpegint.h */ -struct jvirt_sarray_control { long dummy; }; -struct jvirt_barray_control { long dummy; }; -struct jpeg_comp_master { long dummy; }; -struct jpeg_c_main_controller { long dummy; }; -struct jpeg_c_prep_controller { long dummy; }; -struct jpeg_c_coef_controller { long dummy; }; -struct jpeg_marker_writer { long dummy; }; -struct jpeg_color_converter { long dummy; }; -struct jpeg_downsampler { long dummy; }; -struct jpeg_forward_dct { long dummy; }; -struct jpeg_entropy_encoder { long dummy; }; -struct jpeg_decomp_master { long dummy; }; -struct jpeg_d_main_controller { long dummy; }; -struct jpeg_d_coef_controller { long dummy; }; -struct jpeg_d_post_controller { long dummy; }; -struct jpeg_input_controller { long dummy; }; -struct jpeg_marker_reader { long dummy; }; -struct jpeg_entropy_decoder { long dummy; }; -struct jpeg_inverse_dct { long dummy; }; -struct jpeg_upsampler { long dummy; }; -struct jpeg_color_deconverter { long dummy; }; -struct jpeg_color_quantizer { long dummy; }; -#endif /* JPEG_INTERNALS */ -#endif /* INCOMPLETE_TYPES_BROKEN */ - - -/* - * The JPEG library modules define JPEG_INTERNALS before including this file. - * The internal structure declarations are read only when that is true. - * Applications using the library should not include jpegint.h, but may wish - * to include jerror.h. - */ - -#ifdef JPEG_INTERNALS -#include "jpegint.h" /* fetch private declarations */ -#include "jerror.h" /* fetch error codes too */ -#endif - -#ifdef __cplusplus -#ifndef DONT_USE_EXTERN_C -} -#endif -#endif - -#endif /* JPEGLIB_H */ diff --git a/cmake/externals/libjpeg/jquant1.c b/cmake/externals/libjpeg/jquant1.c deleted file mode 100644 index 60b1843e7..000000000 --- a/cmake/externals/libjpeg/jquant1.c +++ /dev/null @@ -1,851 +0,0 @@ -/* - * jquant1.c - * - * Copyright (C) 1991-1996, Thomas G. Lane. - * Modified 2011-2020 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains 1-pass color quantization (color mapping) routines. - * These routines provide mapping to a fixed color map using equally spaced - * color values. Optional Floyd-Steinberg or ordered dithering is available. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - -#ifdef QUANT_1PASS_SUPPORTED - - -/* - * The main purpose of 1-pass quantization is to provide a fast, if not very - * high quality, colormapped output capability. A 2-pass quantizer usually - * gives better visual quality; however, for quantized grayscale output this - * quantizer is perfectly adequate. Dithering is highly recommended with this - * quantizer, though you can turn it off if you really want to. - * - * In 1-pass quantization the colormap must be chosen in advance of seeing the - * image. We use a map consisting of all combinations of Ncolors[i] color - * values for the i'th component. The Ncolors[] values are chosen so that - * their product, the total number of colors, is no more than that requested. - * (In most cases, the product will be somewhat less.) - * - * Since the colormap is orthogonal, the representative value for each color - * component can be determined without considering the other components; - * then these indexes can be combined into a colormap index by a standard - * N-dimensional-array-subscript calculation. Most of the arithmetic involved - * can be precalculated and stored in the lookup table colorindex[]. - * colorindex[i][j] maps pixel value j in component i to the nearest - * representative value (grid plane) for that component; this index is - * multiplied by the array stride for component i, so that the - * index of the colormap entry closest to a given pixel value is just - * sum( colorindex[component-number][pixel-component-value] ) - * Aside from being fast, this scheme allows for variable spacing between - * representative values with no additional lookup cost. - * - * If gamma correction has been applied in color conversion, it might be wise - * to adjust the color grid spacing so that the representative colors are - * equidistant in linear space. At this writing, gamma correction is not - * implemented by jdcolor, so nothing is done here. - */ - - -/* Declarations for ordered dithering. - * - * We use a standard 16x16 ordered dither array. The basic concept of ordered - * dithering is described in many references, for instance Dale Schumacher's - * chapter II.2 of Graphics Gems II (James Arvo, ed. Academic Press, 1991). - * In place of Schumacher's comparisons against a "threshold" value, we add a - * "dither" value to the input pixel and then round the result to the nearest - * output value. The dither value is equivalent to (0.5 - threshold) times - * the distance between output values. For ordered dithering, we assume that - * the output colors are equally spaced; if not, results will probably be - * worse, since the dither may be too much or too little at a given point. - * - * The normal calculation would be to form pixel value + dither, range-limit - * this to 0..MAXJSAMPLE, and then index into the colorindex table as usual. - * We can skip the separate range-limiting step by extending the colorindex - * table in both directions. - */ - -#define ODITHER_SIZE 16 /* dimension of dither matrix */ -/* NB: if ODITHER_SIZE is not a power of 2, ODITHER_MASK uses will break */ -#define ODITHER_CELLS (ODITHER_SIZE*ODITHER_SIZE) /* # cells in matrix */ -#define ODITHER_MASK (ODITHER_SIZE-1) /* mask for wrapping around counters */ - -typedef int ODITHER_MATRIX[ODITHER_SIZE][ODITHER_SIZE]; -typedef int (*ODITHER_MATRIX_PTR)[ODITHER_SIZE]; - -static const UINT8 base_dither_matrix[ODITHER_SIZE][ODITHER_SIZE] = { - /* Bayer's order-4 dither array. Generated by the code given in - * Stephen Hawley's article "Ordered Dithering" in Graphics Gems I. - * The values in this array must range from 0 to ODITHER_CELLS-1. - */ - { 0,192, 48,240, 12,204, 60,252, 3,195, 51,243, 15,207, 63,255 }, - { 128, 64,176,112,140, 76,188,124,131, 67,179,115,143, 79,191,127 }, - { 32,224, 16,208, 44,236, 28,220, 35,227, 19,211, 47,239, 31,223 }, - { 160, 96,144, 80,172,108,156, 92,163, 99,147, 83,175,111,159, 95 }, - { 8,200, 56,248, 4,196, 52,244, 11,203, 59,251, 7,199, 55,247 }, - { 136, 72,184,120,132, 68,180,116,139, 75,187,123,135, 71,183,119 }, - { 40,232, 24,216, 36,228, 20,212, 43,235, 27,219, 39,231, 23,215 }, - { 168,104,152, 88,164,100,148, 84,171,107,155, 91,167,103,151, 87 }, - { 2,194, 50,242, 14,206, 62,254, 1,193, 49,241, 13,205, 61,253 }, - { 130, 66,178,114,142, 78,190,126,129, 65,177,113,141, 77,189,125 }, - { 34,226, 18,210, 46,238, 30,222, 33,225, 17,209, 45,237, 29,221 }, - { 162, 98,146, 82,174,110,158, 94,161, 97,145, 81,173,109,157, 93 }, - { 10,202, 58,250, 6,198, 54,246, 9,201, 57,249, 5,197, 53,245 }, - { 138, 74,186,122,134, 70,182,118,137, 73,185,121,133, 69,181,117 }, - { 42,234, 26,218, 38,230, 22,214, 41,233, 25,217, 37,229, 21,213 }, - { 170,106,154, 90,166,102,150, 86,169,105,153, 89,165,101,149, 85 } -}; - - -/* Declarations for Floyd-Steinberg dithering. - * - * Errors are accumulated into the array fserrors[], at a resolution of - * 1/16th of a pixel count. The error at a given pixel is propagated - * to its not-yet-processed neighbors using the standard F-S fractions, - * ... (here) 7/16 - * 3/16 5/16 1/16 - * We work left-to-right on even rows, right-to-left on odd rows. - * - * We can get away with a single array (holding one row's worth of errors) - * by using it to store the current row's errors at pixel columns not yet - * processed, but the next row's errors at columns already processed. We - * need only a few extra variables to hold the errors immediately around the - * current column. (If we are lucky, those variables are in registers, but - * even if not, they're probably cheaper to access than array elements are.) - * - * The fserrors[] array is indexed [component#][position]. - * We provide (#columns + 2) entries per component; the extra entry at each - * end saves us from special-casing the first and last pixels. - * - * Note: on a wide image, we might not have enough room in a PC's near data - * segment to hold the error array; so it is allocated with alloc_large. - */ - -#if BITS_IN_JSAMPLE == 8 -typedef INT16 FSERROR; /* 16 bits should be enough */ -typedef int LOCFSERROR; /* use 'int' for calculation temps */ -#else -typedef INT32 FSERROR; /* may need more than 16 bits */ -typedef INT32 LOCFSERROR; /* be sure calculation temps are big enough */ -#endif - -typedef FSERROR FAR *FSERRPTR; /* pointer to error array (in FAR storage!) */ - - -/* Private subobject */ - -#define MAX_Q_COMPS 4 /* max components I can handle */ - -typedef struct { - struct jpeg_color_quantizer pub; /* public fields */ - - /* Initially allocated colormap is saved here */ - JSAMPARRAY sv_colormap; /* The color map as a 2-D pixel array */ - int sv_actual; /* number of entries in use */ - - JSAMPARRAY colorindex; /* Precomputed mapping for speed */ - /* colorindex[i][j] = index of color closest to pixel value j in component i, - * premultiplied as described above. Since colormap indexes must fit into - * JSAMPLEs, the entries of this array will too. - */ - boolean is_padded; /* is the colorindex padded for odither? */ - - int Ncolors[MAX_Q_COMPS]; /* # of values alloced to each component */ - - /* Variables for ordered dithering */ - int row_index; /* cur row's vertical index in dither matrix */ - ODITHER_MATRIX_PTR odither[MAX_Q_COMPS]; /* one dither array per component */ - - /* Variables for Floyd-Steinberg dithering */ - FSERRPTR fserrors[MAX_Q_COMPS]; /* accumulated errors */ - boolean on_odd_row; /* flag to remember which row we are on */ -} my_cquantizer; - -typedef my_cquantizer * my_cquantize_ptr; - - -/* - * Policy-making subroutines for create_colormap and create_colorindex. - * These routines determine the colormap to be used. The rest of the module - * only assumes that the colormap is orthogonal. - * - * * select_ncolors decides how to divvy up the available colors - * among the components. - * * output_value defines the set of representative values for a component. - * * largest_input_value defines the mapping from input values to - * representative values for a component. - * Note that the latter two routines may impose different policies for - * different components, though this is not currently done. - */ - - -LOCAL(int) -select_ncolors (j_decompress_ptr cinfo, int Ncolors[]) -/* Determine allocation of desired colors to components, */ -/* and fill in Ncolors[] array to indicate choice. */ -/* Return value is total number of colors (product of Ncolors[] values). */ -{ - int nc = cinfo->out_color_components; /* number of color components */ - int max_colors = cinfo->desired_number_of_colors; - int total_colors, iroot, i, j; - boolean changed; - long temp; - static const int RGB_order[3] = { RGB_GREEN, RGB_RED, RGB_BLUE }; - - /* We can allocate at least the nc'th root of max_colors per component. */ - /* Compute floor(nc'th root of max_colors). */ - iroot = 1; - do { - iroot++; - temp = iroot; /* set temp = iroot ** nc */ - for (i = 1; i < nc; i++) - temp *= iroot; - } while (temp <= (long) max_colors); /* repeat till iroot exceeds root */ - iroot--; /* now iroot = floor(root) */ - - /* Must have at least 2 color values per component */ - if (iroot < 2) - ERREXIT1(cinfo, JERR_QUANT_FEW_COLORS, (int) temp); - - /* Initialize to iroot color values for each component */ - total_colors = 1; - for (i = 0; i < nc; i++) { - Ncolors[i] = iroot; - total_colors *= iroot; - } - /* We may be able to increment the count for one or more components without - * exceeding max_colors, though we know not all can be incremented. - * Sometimes, the first component can be incremented more than once! - * (Example: for 16 colors, we start at 2*2*2, go to 3*2*2, then 4*2*2.) - * In RGB colorspace, try to increment G first, then R, then B. - */ - do { - changed = FALSE; - for (i = 0; i < nc; i++) { - j = (cinfo->out_color_space == JCS_RGB ? RGB_order[i] : i); - /* calculate new total_colors if Ncolors[j] is incremented */ - temp = total_colors / Ncolors[j]; - temp *= Ncolors[j]+1; /* done in long arith to avoid oflo */ - if (temp > (long) max_colors) - break; /* won't fit, done with this pass */ - Ncolors[j]++; /* OK, apply the increment */ - total_colors = (int) temp; - changed = TRUE; - } - } while (changed); - - return total_colors; -} - - -LOCAL(int) -output_value (j_decompress_ptr cinfo, int ci, int j, int maxj) -/* Return j'th output value, where j will range from 0 to maxj */ -/* The output values must fall in 0..MAXJSAMPLE in increasing order */ -{ - /* We always provide values 0 and MAXJSAMPLE for each component; - * any additional values are equally spaced between these limits. - * (Forcing the upper and lower values to the limits ensures that - * dithering can't produce a color outside the selected gamut.) - */ - return (int) (((INT32) j * MAXJSAMPLE + maxj/2) / maxj); -} - - -LOCAL(int) -largest_input_value (j_decompress_ptr cinfo, int ci, int j, int maxj) -/* Return largest input value that should map to j'th output value */ -/* Must have largest(j=0) >= 0, and largest(j=maxj) >= MAXJSAMPLE */ -{ - /* Breakpoints are halfway between values returned by output_value */ - return (int) (((INT32) (2*j + 1) * MAXJSAMPLE + maxj) / (2*maxj)); -} - - -/* - * Create the colormap. - */ - -LOCAL(void) -create_colormap (j_decompress_ptr cinfo) -{ - my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; - JSAMPARRAY colormap; /* Created colormap */ - int total_colors; /* Number of distinct output colors */ - int i,j,k, nci, blksize, blkdist, ptr, val; - - /* Select number of colors for each component */ - total_colors = select_ncolors(cinfo, cquantize->Ncolors); - - /* Report selected color counts */ - if (cinfo->out_color_components == 3) - TRACEMS4(cinfo, 1, JTRC_QUANT_3_NCOLORS, - total_colors, cquantize->Ncolors[0], - cquantize->Ncolors[1], cquantize->Ncolors[2]); - else - TRACEMS1(cinfo, 1, JTRC_QUANT_NCOLORS, total_colors); - - /* Allocate and fill in the colormap. */ - /* The colors are ordered in the map in standard row-major order, */ - /* i.e. rightmost (highest-indexed) color changes most rapidly. */ - - colormap = (*cinfo->mem->alloc_sarray) ((j_common_ptr) cinfo, JPOOL_IMAGE, - (JDIMENSION) total_colors, (JDIMENSION) cinfo->out_color_components); - - /* blksize is number of adjacent repeated entries for a component */ - /* blkdist is distance between groups of identical entries for a component */ - blkdist = total_colors; - - for (i = 0; i < cinfo->out_color_components; i++) { - /* fill in colormap entries for i'th color component */ - nci = cquantize->Ncolors[i]; /* # of distinct values for this color */ - blksize = blkdist / nci; - for (j = 0; j < nci; j++) { - /* Compute j'th output value (out of nci) for component */ - val = output_value(cinfo, i, j, nci-1); - /* Fill in all colormap entries that have this value of this component */ - for (ptr = j * blksize; ptr < total_colors; ptr += blkdist) { - /* fill in blksize entries beginning at ptr */ - for (k = 0; k < blksize; k++) - colormap[i][ptr+k] = (JSAMPLE) val; - } - } - blkdist = blksize; /* blksize of this color is blkdist of next */ - } - - /* Save the colormap in private storage, - * where it will survive color quantization mode changes. - */ - cquantize->sv_colormap = colormap; - cquantize->sv_actual = total_colors; -} - - -/* - * Create the color index table. - */ - -LOCAL(void) -create_colorindex (j_decompress_ptr cinfo) -{ - my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; - JSAMPROW indexptr; - int i,j,k, nci, blksize, val, pad; - - /* For ordered dither, we pad the color index tables by MAXJSAMPLE in - * each direction (input index values can be -MAXJSAMPLE .. 2*MAXJSAMPLE). - * This is not necessary in the other dithering modes. However, we - * flag whether it was done in case user changes dithering mode. - */ - if (cinfo->dither_mode == JDITHER_ORDERED) { - pad = MAXJSAMPLE*2; - cquantize->is_padded = TRUE; - } else { - pad = 0; - cquantize->is_padded = FALSE; - } - - cquantize->colorindex = (*cinfo->mem->alloc_sarray) - ((j_common_ptr) cinfo, JPOOL_IMAGE, - (JDIMENSION) (MAXJSAMPLE+1 + pad), - (JDIMENSION) cinfo->out_color_components); - - /* blksize is number of adjacent repeated entries for a component */ - blksize = cquantize->sv_actual; - - for (i = 0; i < cinfo->out_color_components; i++) { - /* fill in colorindex entries for i'th color component */ - nci = cquantize->Ncolors[i]; /* # of distinct values for this color */ - blksize = blksize / nci; - - /* adjust colorindex pointers to provide padding at negative indexes. */ - if (pad) - cquantize->colorindex[i] += MAXJSAMPLE; - - /* in loop, val = index of current output value, */ - /* and k = largest j that maps to current val */ - indexptr = cquantize->colorindex[i]; - val = 0; - k = largest_input_value(cinfo, i, 0, nci-1); - for (j = 0; j <= MAXJSAMPLE; j++) { - while (j > k) /* advance val if past boundary */ - k = largest_input_value(cinfo, i, ++val, nci-1); - /* premultiply so that no multiplication needed in main processing */ - indexptr[j] = (JSAMPLE) (val * blksize); - } - /* Pad at both ends if necessary */ - if (pad) - for (j = 1; j <= MAXJSAMPLE; j++) { - indexptr[-j] = indexptr[0]; - indexptr[MAXJSAMPLE+j] = indexptr[MAXJSAMPLE]; - } - } -} - - -/* - * Create an ordered-dither array for a component having ncolors - * distinct output values. - */ - -LOCAL(ODITHER_MATRIX_PTR) -make_odither_array (j_decompress_ptr cinfo, int ncolors) -{ - ODITHER_MATRIX_PTR odither; - int j,k; - INT32 num,den; - - odither = (ODITHER_MATRIX_PTR) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(ODITHER_MATRIX)); - /* The inter-value distance for this color is MAXJSAMPLE/(ncolors-1). - * Hence the dither value for the matrix cell with fill order f - * (f=0..N-1) should be (N-1-2*f)/(2*N) * MAXJSAMPLE/(ncolors-1). - * On 16-bit-int machine, be careful to avoid overflow. - */ - den = 2 * ODITHER_CELLS * ((INT32) (ncolors - 1)); - for (j = 0; j < ODITHER_SIZE; j++) { - for (k = 0; k < ODITHER_SIZE; k++) { - num = ((INT32) (ODITHER_CELLS-1 - 2*((int)base_dither_matrix[j][k]))) - * MAXJSAMPLE; - /* Ensure round towards zero despite C's lack of consistency - * about rounding negative values in integer division... - */ - odither[j][k] = (int) (num<0 ? -((-num)/den) : num/den); - } - } - return odither; -} - - -/* - * Create the ordered-dither tables. - * Components having the same number of representative colors may - * share a dither table. - */ - -LOCAL(void) -create_odither_tables (j_decompress_ptr cinfo) -{ - my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; - ODITHER_MATRIX_PTR odither; - int i, j, nci; - - for (i = 0; i < cinfo->out_color_components; i++) { - nci = cquantize->Ncolors[i]; /* # of distinct values for this color */ - odither = NULL; /* search for matching prior component */ - for (j = 0; j < i; j++) { - if (nci == cquantize->Ncolors[j]) { - odither = cquantize->odither[j]; - break; - } - } - if (odither == NULL) /* need a new table? */ - odither = make_odither_array(cinfo, nci); - cquantize->odither[i] = odither; - } -} - - -/* - * Map some rows of pixels to the output colormapped representation. - */ - -METHODDEF(void) -color_quantize (j_decompress_ptr cinfo, JSAMPARRAY input_buf, - JSAMPARRAY output_buf, int num_rows) -/* General case, no dithering */ -{ - my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; - JSAMPARRAY colorindex = cquantize->colorindex; - register int pixcode, ci; - register JSAMPROW ptrin, ptrout; - int row; - JDIMENSION col; - JDIMENSION width = cinfo->output_width; - register int nc = cinfo->out_color_components; - - for (row = 0; row < num_rows; row++) { - ptrin = input_buf[row]; - ptrout = output_buf[row]; - for (col = width; col > 0; col--) { - pixcode = 0; - for (ci = 0; ci < nc; ci++) { - pixcode += GETJSAMPLE(colorindex[ci][GETJSAMPLE(*ptrin++)]); - } - *ptrout++ = (JSAMPLE) pixcode; - } - } -} - - -METHODDEF(void) -color_quantize3 (j_decompress_ptr cinfo, JSAMPARRAY input_buf, - JSAMPARRAY output_buf, int num_rows) -/* Fast path for out_color_components==3, no dithering */ -{ - my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; - register int pixcode; - register JSAMPROW ptrin, ptrout; - JSAMPROW colorindex0 = cquantize->colorindex[0]; - JSAMPROW colorindex1 = cquantize->colorindex[1]; - JSAMPROW colorindex2 = cquantize->colorindex[2]; - int row; - JDIMENSION col; - JDIMENSION width = cinfo->output_width; - - for (row = 0; row < num_rows; row++) { - ptrin = input_buf[row]; - ptrout = output_buf[row]; - for (col = width; col > 0; col--) { - pixcode = GETJSAMPLE(colorindex0[GETJSAMPLE(*ptrin++)]); - pixcode += GETJSAMPLE(colorindex1[GETJSAMPLE(*ptrin++)]); - pixcode += GETJSAMPLE(colorindex2[GETJSAMPLE(*ptrin++)]); - *ptrout++ = (JSAMPLE) pixcode; - } - } -} - - -METHODDEF(void) -quantize_ord_dither (j_decompress_ptr cinfo, JSAMPARRAY input_buf, - JSAMPARRAY output_buf, int num_rows) -/* General case, with ordered dithering */ -{ - my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; - register JSAMPROW input_ptr; - register JSAMPROW output_ptr; - JSAMPROW colorindex_ci; - int * dither; /* points to active row of dither matrix */ - int row_index, col_index; /* current indexes into dither matrix */ - int nc = cinfo->out_color_components; - int ci; - int row; - JDIMENSION col; - JDIMENSION width = cinfo->output_width; - - for (row = 0; row < num_rows; row++) { - /* Initialize output values to 0 so can process components separately */ - FMEMZERO((void FAR *) output_buf[row], (size_t) width * SIZEOF(JSAMPLE)); - row_index = cquantize->row_index; - for (ci = 0; ci < nc; ci++) { - input_ptr = input_buf[row] + ci; - output_ptr = output_buf[row]; - colorindex_ci = cquantize->colorindex[ci]; - dither = cquantize->odither[ci][row_index]; - col_index = 0; - - for (col = width; col > 0; col--) { - /* Form pixel value + dither, range-limit to 0..MAXJSAMPLE, - * select output value, accumulate into output code for this pixel. - * Range-limiting need not be done explicitly, as we have extended - * the colorindex table to produce the right answers for out-of-range - * inputs. The maximum dither is +- MAXJSAMPLE; this sets the - * required amount of padding. - */ - *output_ptr += colorindex_ci[GETJSAMPLE(*input_ptr)+dither[col_index]]; - input_ptr += nc; - output_ptr++; - col_index = (col_index + 1) & ODITHER_MASK; - } - } - /* Advance row index for next row */ - row_index = (row_index + 1) & ODITHER_MASK; - cquantize->row_index = row_index; - } -} - - -METHODDEF(void) -quantize3_ord_dither (j_decompress_ptr cinfo, JSAMPARRAY input_buf, - JSAMPARRAY output_buf, int num_rows) -/* Fast path for out_color_components==3, with ordered dithering */ -{ - my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; - register int pixcode; - register JSAMPROW input_ptr; - register JSAMPROW output_ptr; - JSAMPROW colorindex0 = cquantize->colorindex[0]; - JSAMPROW colorindex1 = cquantize->colorindex[1]; - JSAMPROW colorindex2 = cquantize->colorindex[2]; - int * dither0; /* points to active row of dither matrix */ - int * dither1; - int * dither2; - int row_index, col_index; /* current indexes into dither matrix */ - int row; - JDIMENSION col; - JDIMENSION width = cinfo->output_width; - - for (row = 0; row < num_rows; row++) { - row_index = cquantize->row_index; - input_ptr = input_buf[row]; - output_ptr = output_buf[row]; - dither0 = cquantize->odither[0][row_index]; - dither1 = cquantize->odither[1][row_index]; - dither2 = cquantize->odither[2][row_index]; - col_index = 0; - - for (col = width; col > 0; col--) { - pixcode = GETJSAMPLE(colorindex0[GETJSAMPLE(*input_ptr++) + - dither0[col_index]]); - pixcode += GETJSAMPLE(colorindex1[GETJSAMPLE(*input_ptr++) + - dither1[col_index]]); - pixcode += GETJSAMPLE(colorindex2[GETJSAMPLE(*input_ptr++) + - dither2[col_index]]); - *output_ptr++ = (JSAMPLE) pixcode; - col_index = (col_index + 1) & ODITHER_MASK; - } - row_index = (row_index + 1) & ODITHER_MASK; - cquantize->row_index = row_index; - } -} - - -METHODDEF(void) -quantize_fs_dither (j_decompress_ptr cinfo, JSAMPARRAY input_buf, - JSAMPARRAY output_buf, int num_rows) -/* General case, with Floyd-Steinberg dithering */ -{ - my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; - register LOCFSERROR cur; /* current error or pixel value */ - LOCFSERROR belowerr; /* error for pixel below cur */ - LOCFSERROR bpreverr; /* error for below/prev col */ - LOCFSERROR bnexterr; /* error for below/next col */ - LOCFSERROR delta; - register FSERRPTR errorptr; /* => fserrors[] at column before current */ - register JSAMPROW input_ptr; - register JSAMPROW output_ptr; - JSAMPROW colorindex_ci; - JSAMPROW colormap_ci; - int pixcode; - int nc = cinfo->out_color_components; - int dir; /* 1 for left-to-right, -1 for right-to-left */ - int dirnc; /* dir * nc */ - int ci; - int row; - JDIMENSION col; - JDIMENSION width = cinfo->output_width; - JSAMPLE *range_limit = cinfo->sample_range_limit; - SHIFT_TEMPS - - for (row = 0; row < num_rows; row++) { - /* Initialize output values to 0 so can process components separately */ - FMEMZERO((void FAR *) output_buf[row], (size_t) width * SIZEOF(JSAMPLE)); - for (ci = 0; ci < nc; ci++) { - input_ptr = input_buf[row] + ci; - output_ptr = output_buf[row]; - if (cquantize->on_odd_row) { - /* work right to left in this row */ - input_ptr += (width-1) * nc; /* so point to rightmost pixel */ - output_ptr += width-1; - dir = -1; - dirnc = -nc; - errorptr = cquantize->fserrors[ci] + (width+1); /* => entry after last column */ - } else { - /* work left to right in this row */ - dir = 1; - dirnc = nc; - errorptr = cquantize->fserrors[ci]; /* => entry before first column */ - } - colorindex_ci = cquantize->colorindex[ci]; - colormap_ci = cquantize->sv_colormap[ci]; - /* Preset error values: no error propagated to first pixel from left */ - cur = 0; - /* and no error propagated to row below yet */ - belowerr = bpreverr = 0; - - for (col = width; col > 0; col--) { - /* cur holds the error propagated from the previous pixel on the - * current line. Add the error propagated from the previous line - * to form the complete error correction term for this pixel, and - * round the error term (which is expressed * 16) to an integer. - * RIGHT_SHIFT rounds towards minus infinity, so adding 8 is correct - * for either sign of the error value. - * Note: errorptr points to *previous* column's array entry. - */ - cur = RIGHT_SHIFT(cur + errorptr[dir] + 8, 4); - /* Form pixel value + error, and range-limit to 0..MAXJSAMPLE. - * The maximum error is +- MAXJSAMPLE; this sets the required size - * of the range_limit array. - */ - cur += GETJSAMPLE(*input_ptr); - cur = GETJSAMPLE(range_limit[cur]); - /* Select output value, accumulate into output code for this pixel */ - pixcode = GETJSAMPLE(colorindex_ci[cur]); - *output_ptr += (JSAMPLE) pixcode; - /* Compute actual representation error at this pixel */ - /* Note: we can do this even though we don't have the final */ - /* pixel code, because the colormap is orthogonal. */ - cur -= GETJSAMPLE(colormap_ci[pixcode]); - /* Compute error fractions to be propagated to adjacent pixels. - * Add these into the running sums, and simultaneously shift the - * next-line error sums left by 1 column. - */ - bnexterr = cur; - delta = cur * 2; - cur += delta; /* form error * 3 */ - errorptr[0] = (FSERROR) (bpreverr + cur); - cur += delta; /* form error * 5 */ - bpreverr = belowerr + cur; - belowerr = bnexterr; - cur += delta; /* form error * 7 */ - /* At this point cur contains the 7/16 error value to be propagated - * to the next pixel on the current line, and all the errors for the - * next line have been shifted over. We are therefore ready to move on. - */ - input_ptr += dirnc; /* advance input ptr to next column */ - output_ptr += dir; /* advance output ptr to next column */ - errorptr += dir; /* advance errorptr to current column */ - } - /* Post-loop cleanup: we must unload the final error value into the - * final fserrors[] entry. Note we need not unload belowerr because - * it is for the dummy column before or after the actual array. - */ - errorptr[0] = (FSERROR) bpreverr; /* unload prev err into array */ - } - cquantize->on_odd_row = (cquantize->on_odd_row ? FALSE : TRUE); - } -} - - -/* - * Allocate workspace for Floyd-Steinberg errors. - */ - -LOCAL(void) -alloc_fs_workspace (j_decompress_ptr cinfo) -{ - my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; - size_t arraysize; - int i; - - arraysize = ((size_t) cinfo->output_width + (size_t) 2) * SIZEOF(FSERROR); - for (i = 0; i < cinfo->out_color_components; i++) { - cquantize->fserrors[i] = (FSERRPTR) (*cinfo->mem->alloc_large) - ((j_common_ptr) cinfo, JPOOL_IMAGE, arraysize); - } -} - - -/* - * Initialize for one-pass color quantization. - */ - -METHODDEF(void) -start_pass_1_quant (j_decompress_ptr cinfo, boolean is_pre_scan) -{ - my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; - size_t arraysize; - int i; - - /* Install my colormap. */ - cinfo->colormap = cquantize->sv_colormap; - cinfo->actual_number_of_colors = cquantize->sv_actual; - - /* Initialize for desired dithering mode. */ - switch (cinfo->dither_mode) { - case JDITHER_NONE: - if (cinfo->out_color_components == 3) - cquantize->pub.color_quantize = color_quantize3; - else - cquantize->pub.color_quantize = color_quantize; - break; - case JDITHER_ORDERED: - if (cinfo->out_color_components == 3) - cquantize->pub.color_quantize = quantize3_ord_dither; - else - cquantize->pub.color_quantize = quantize_ord_dither; - cquantize->row_index = 0; /* initialize state for ordered dither */ - /* If user changed to ordered dither from another mode, - * we must recreate the color index table with padding. - * This will cost extra space, but probably isn't very likely. - */ - if (! cquantize->is_padded) - create_colorindex(cinfo); - /* Create ordered-dither tables if we didn't already. */ - if (cquantize->odither[0] == NULL) - create_odither_tables(cinfo); - break; - case JDITHER_FS: - cquantize->pub.color_quantize = quantize_fs_dither; - cquantize->on_odd_row = FALSE; /* initialize state for F-S dither */ - /* Allocate Floyd-Steinberg workspace if didn't already. */ - if (cquantize->fserrors[0] == NULL) - alloc_fs_workspace(cinfo); - /* Initialize the propagated errors to zero. */ - arraysize = ((size_t) cinfo->output_width + (size_t) 2) * SIZEOF(FSERROR); - for (i = 0; i < cinfo->out_color_components; i++) - FMEMZERO((void FAR *) cquantize->fserrors[i], arraysize); - break; - default: - ERREXIT(cinfo, JERR_NOT_COMPILED); - } -} - - -/* - * Finish up at the end of the pass. - */ - -METHODDEF(void) -finish_pass_1_quant (j_decompress_ptr cinfo) -{ - /* no work in 1-pass case */ -} - - -/* - * Switch to a new external colormap between output passes. - * Shouldn't get to this module! - */ - -METHODDEF(void) -new_color_map_1_quant (j_decompress_ptr cinfo) -{ - ERREXIT(cinfo, JERR_MODE_CHANGE); -} - - -/* - * Module initialization routine for 1-pass color quantization. - */ - -GLOBAL(void) -jinit_1pass_quantizer (j_decompress_ptr cinfo) -{ - my_cquantize_ptr cquantize; - - cquantize = (my_cquantize_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_cquantizer)); - cinfo->cquantize = &cquantize->pub; - cquantize->pub.start_pass = start_pass_1_quant; - cquantize->pub.finish_pass = finish_pass_1_quant; - cquantize->pub.new_color_map = new_color_map_1_quant; - cquantize->fserrors[0] = NULL; /* Flag FS workspace not allocated */ - cquantize->odither[0] = NULL; /* Also flag odither arrays not allocated */ - - /* Make sure my internal arrays won't overflow */ - if (cinfo->out_color_components > MAX_Q_COMPS) - ERREXIT1(cinfo, JERR_QUANT_COMPONENTS, MAX_Q_COMPS); - /* Make sure colormap indexes can be represented by JSAMPLEs */ - if (cinfo->desired_number_of_colors > (MAXJSAMPLE+1)) - ERREXIT1(cinfo, JERR_QUANT_MANY_COLORS, MAXJSAMPLE+1); - - /* Create the colormap and color index table. */ - create_colormap(cinfo); - create_colorindex(cinfo); - - /* Allocate Floyd-Steinberg workspace now if requested. - * We do this now since it is FAR storage and may affect the memory - * manager's space calculations. If the user changes to FS dither - * mode in a later pass, we will allocate the space then, and will - * possibly overrun the max_memory_to_use setting. - */ - if (cinfo->dither_mode == JDITHER_FS) - alloc_fs_workspace(cinfo); -} - -#endif /* QUANT_1PASS_SUPPORTED */ diff --git a/cmake/externals/libjpeg/jquant2.c b/cmake/externals/libjpeg/jquant2.c deleted file mode 100644 index 662b9bcef..000000000 --- a/cmake/externals/libjpeg/jquant2.c +++ /dev/null @@ -1,1311 +0,0 @@ -/* - * jquant2.c - * - * Copyright (C) 1991-1996, Thomas G. Lane. - * Modified 2011-2020 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains 2-pass color quantization (color mapping) routines. - * These routines provide selection of a custom color map for an image, - * followed by mapping of the image to that color map, with optional - * Floyd-Steinberg dithering. - * It is also possible to use just the second pass to map to an arbitrary - * externally-given color map. - * - * Note: ordered dithering is not supported, since there isn't any fast - * way to compute intercolor distances; it's unclear that ordered dither's - * fundamental assumptions even hold with an irregularly spaced color map. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - -#ifdef QUANT_2PASS_SUPPORTED - - -/* - * This module implements the well-known Heckbert paradigm for color - * quantization. Most of the ideas used here can be traced back to - * Heckbert's seminal paper - * Heckbert, Paul. "Color Image Quantization for Frame Buffer Display", - * Proc. SIGGRAPH '82, Computer Graphics v.16 #3 (July 1982), pp 297-304. - * - * In the first pass over the image, we accumulate a histogram showing the - * usage count of each possible color. To keep the histogram to a reasonable - * size, we reduce the precision of the input; typical practice is to retain - * 5 or 6 bits per color, so that 8 or 4 different input values are counted - * in the same histogram cell. - * - * Next, the color-selection step begins with a box representing the whole - * color space, and repeatedly splits the "largest" remaining box until we - * have as many boxes as desired colors. Then the mean color in each - * remaining box becomes one of the possible output colors. - * - * The second pass over the image maps each input pixel to the closest output - * color (optionally after applying a Floyd-Steinberg dithering correction). - * This mapping is logically trivial, but making it go fast enough requires - * considerable care. - * - * Heckbert-style quantizers vary a good deal in their policies for choosing - * the "largest" box and deciding where to cut it. The particular policies - * used here have proved out well in experimental comparisons, but better ones - * may yet be found. - * - * In earlier versions of the IJG code, this module quantized in YCbCr color - * space, processing the raw upsampled data without a color conversion step. - * This allowed the color conversion math to be done only once per colormap - * entry, not once per pixel. However, that optimization precluded other - * useful optimizations (such as merging color conversion with upsampling) - * and it also interfered with desired capabilities such as quantizing to an - * externally-supplied colormap. We have therefore abandoned that approach. - * The present code works in the post-conversion color space, typically RGB. - * - * To improve the visual quality of the results, we actually work in scaled - * RGB space, giving G distances more weight than R, and R in turn more than - * B. To do everything in integer math, we must use integer scale factors. - * The 2/3/1 scale factors used here correspond loosely to the relative - * weights of the colors in the NTSC grayscale equation. - * If you want to use this code to quantize a non-RGB color space, you'll - * probably need to change these scale factors. - */ - -#define R_SCALE 2 /* scale R distances by this much */ -#define G_SCALE 3 /* scale G distances by this much */ -#define B_SCALE 1 /* and B by this much */ - -/* Relabel R/G/B as components 0/1/2, respecting the RGB ordering defined - * in jmorecfg.h. As the code stands, it will do the right thing for R,G,B - * and B,G,R orders. If you define some other weird order in jmorecfg.h, - * you'll get compile errors until you extend this logic. In that case - * you'll probably want to tweak the histogram sizes too. - */ - -#if RGB_RED == 0 -#define C0_SCALE R_SCALE -#endif -#if RGB_BLUE == 0 -#define C0_SCALE B_SCALE -#endif -#if RGB_GREEN == 1 -#define C1_SCALE G_SCALE -#endif -#if RGB_RED == 2 -#define C2_SCALE R_SCALE -#endif -#if RGB_BLUE == 2 -#define C2_SCALE B_SCALE -#endif - - -/* - * First we have the histogram data structure and routines for creating it. - * - * The number of bits of precision can be adjusted by changing these symbols. - * We recommend keeping 6 bits for G and 5 each for R and B. - * If you have plenty of memory and cycles, 6 bits all around gives marginally - * better results; if you are short of memory, 5 bits all around will save - * some space but degrade the results. - * To maintain a fully accurate histogram, we'd need to allocate a "long" - * (preferably unsigned long) for each cell. In practice this is overkill; - * we can get by with 16 bits per cell. Few of the cell counts will overflow, - * and clamping those that do overflow to the maximum value will give close- - * enough results. This reduces the recommended histogram size from 256Kb - * to 128Kb, which is a useful savings on PC-class machines. - * (In the second pass the histogram space is re-used for pixel mapping data; - * in that capacity, each cell must be able to store zero to the number of - * desired colors. 16 bits/cell is plenty for that too.) - * Since the JPEG code is intended to run in small memory model on 80x86 - * machines, we can't just allocate the histogram in one chunk. Instead - * of a true 3-D array, we use a row of pointers to 2-D arrays. Each - * pointer corresponds to a C0 value (typically 2^5 = 32 pointers) and - * each 2-D array has 2^6*2^5 = 2048 or 2^6*2^6 = 4096 entries. Note that - * on 80x86 machines, the pointer row is in near memory but the actual - * arrays are in far memory (same arrangement as we use for image arrays). - */ - -#define MAXNUMCOLORS (MAXJSAMPLE+1) /* maximum size of colormap */ - -/* These will do the right thing for either R,G,B or B,G,R color order, - * but you may not like the results for other color orders. - */ -#define HIST_C0_BITS 5 /* bits of precision in R/B histogram */ -#define HIST_C1_BITS 6 /* bits of precision in G histogram */ -#define HIST_C2_BITS 5 /* bits of precision in B/R histogram */ - -/* Number of elements along histogram axes. */ -#define HIST_C0_ELEMS (1<cquantize; - register JSAMPROW ptr; - register histptr histp; - register hist3d histogram = cquantize->histogram; - int row; - JDIMENSION col; - JDIMENSION width = cinfo->output_width; - - for (row = 0; row < num_rows; row++) { - ptr = input_buf[row]; - for (col = width; col > 0; col--) { - /* get pixel value and index into the histogram */ - histp = & histogram[GETJSAMPLE(ptr[0]) >> C0_SHIFT] - [GETJSAMPLE(ptr[1]) >> C1_SHIFT] - [GETJSAMPLE(ptr[2]) >> C2_SHIFT]; - /* increment, check for overflow and undo increment if so. */ - if (++(*histp) <= 0) - (*histp)--; - ptr += 3; - } - } -} - - -/* - * Next we have the really interesting routines: selection of a colormap - * given the completed histogram. - * These routines work with a list of "boxes", each representing a rectangular - * subset of the input color space (to histogram precision). - */ - -typedef struct { - /* The bounds of the box (inclusive); expressed as histogram indexes */ - int c0min, c0max; - int c1min, c1max; - int c2min, c2max; - /* The volume (actually 2-norm) of the box */ - INT32 volume; - /* The number of nonzero histogram cells within this box */ - long colorcount; -} box; - -typedef box * boxptr; - - -LOCAL(boxptr) -find_biggest_color_pop (boxptr boxlist, int numboxes) -/* Find the splittable box with the largest color population */ -/* Returns NULL if no splittable boxes remain */ -{ - register boxptr boxp; - register int i; - register long maxc = 0; - boxptr which = NULL; - - for (i = 0, boxp = boxlist; i < numboxes; i++, boxp++) { - if (boxp->colorcount > maxc && boxp->volume > 0) { - which = boxp; - maxc = boxp->colorcount; - } - } - return which; -} - - -LOCAL(boxptr) -find_biggest_volume (boxptr boxlist, int numboxes) -/* Find the splittable box with the largest (scaled) volume */ -/* Returns NULL if no splittable boxes remain */ -{ - register boxptr boxp; - register int i; - register INT32 maxv = 0; - boxptr which = NULL; - - for (i = 0, boxp = boxlist; i < numboxes; i++, boxp++) { - if (boxp->volume > maxv) { - which = boxp; - maxv = boxp->volume; - } - } - return which; -} - - -LOCAL(void) -update_box (j_decompress_ptr cinfo, boxptr boxp) -/* Shrink the min/max bounds of a box to enclose only nonzero elements, */ -/* and recompute its volume and population */ -{ - my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; - hist3d histogram = cquantize->histogram; - histptr histp; - int c0,c1,c2; - int c0min,c0max,c1min,c1max,c2min,c2max; - INT32 dist0,dist1,dist2; - long ccount; - - c0min = boxp->c0min; c0max = boxp->c0max; - c1min = boxp->c1min; c1max = boxp->c1max; - c2min = boxp->c2min; c2max = boxp->c2max; - - if (c0max > c0min) - for (c0 = c0min; c0 <= c0max; c0++) - for (c1 = c1min; c1 <= c1max; c1++) { - histp = & histogram[c0][c1][c2min]; - for (c2 = c2min; c2 <= c2max; c2++) - if (*histp++ != 0) { - boxp->c0min = c0min = c0; - goto have_c0min; - } - } - have_c0min: - if (c0max > c0min) - for (c0 = c0max; c0 >= c0min; c0--) - for (c1 = c1min; c1 <= c1max; c1++) { - histp = & histogram[c0][c1][c2min]; - for (c2 = c2min; c2 <= c2max; c2++) - if (*histp++ != 0) { - boxp->c0max = c0max = c0; - goto have_c0max; - } - } - have_c0max: - if (c1max > c1min) - for (c1 = c1min; c1 <= c1max; c1++) - for (c0 = c0min; c0 <= c0max; c0++) { - histp = & histogram[c0][c1][c2min]; - for (c2 = c2min; c2 <= c2max; c2++) - if (*histp++ != 0) { - boxp->c1min = c1min = c1; - goto have_c1min; - } - } - have_c1min: - if (c1max > c1min) - for (c1 = c1max; c1 >= c1min; c1--) - for (c0 = c0min; c0 <= c0max; c0++) { - histp = & histogram[c0][c1][c2min]; - for (c2 = c2min; c2 <= c2max; c2++) - if (*histp++ != 0) { - boxp->c1max = c1max = c1; - goto have_c1max; - } - } - have_c1max: - if (c2max > c2min) - for (c2 = c2min; c2 <= c2max; c2++) - for (c0 = c0min; c0 <= c0max; c0++) { - histp = & histogram[c0][c1min][c2]; - for (c1 = c1min; c1 <= c1max; c1++, histp += HIST_C2_ELEMS) - if (*histp != 0) { - boxp->c2min = c2min = c2; - goto have_c2min; - } - } - have_c2min: - if (c2max > c2min) - for (c2 = c2max; c2 >= c2min; c2--) - for (c0 = c0min; c0 <= c0max; c0++) { - histp = & histogram[c0][c1min][c2]; - for (c1 = c1min; c1 <= c1max; c1++, histp += HIST_C2_ELEMS) - if (*histp != 0) { - boxp->c2max = c2max = c2; - goto have_c2max; - } - } - have_c2max: - - /* Update box volume. - * We use 2-norm rather than real volume here; this biases the method - * against making long narrow boxes, and it has the side benefit that - * a box is splittable iff norm > 0. - * Since the differences are expressed in histogram-cell units, - * we have to shift back to JSAMPLE units to get consistent distances; - * after which, we scale according to the selected distance scale factors. - */ - dist0 = ((c0max - c0min) << C0_SHIFT) * C0_SCALE; - dist1 = ((c1max - c1min) << C1_SHIFT) * C1_SCALE; - dist2 = ((c2max - c2min) << C2_SHIFT) * C2_SCALE; - boxp->volume = dist0*dist0 + dist1*dist1 + dist2*dist2; - - /* Now scan remaining volume of box and compute population */ - ccount = 0; - for (c0 = c0min; c0 <= c0max; c0++) - for (c1 = c1min; c1 <= c1max; c1++) { - histp = & histogram[c0][c1][c2min]; - for (c2 = c2min; c2 <= c2max; c2++, histp++) - if (*histp != 0) { - ccount++; - } - } - boxp->colorcount = ccount; -} - - -LOCAL(int) -median_cut (j_decompress_ptr cinfo, boxptr boxlist, int numboxes, - int desired_colors) -/* Repeatedly select and split the largest box until we have enough boxes */ -{ - int n,lb; - int c0,c1,c2,cmax; - register boxptr b1,b2; - - while (numboxes < desired_colors) { - /* Select box to split. - * Current algorithm: by population for first half, then by volume. - */ - if (numboxes*2 <= desired_colors) { - b1 = find_biggest_color_pop(boxlist, numboxes); - } else { - b1 = find_biggest_volume(boxlist, numboxes); - } - if (b1 == NULL) /* no splittable boxes left! */ - break; - b2 = &boxlist[numboxes]; /* where new box will go */ - /* Copy the color bounds to the new box. */ - b2->c0max = b1->c0max; b2->c1max = b1->c1max; b2->c2max = b1->c2max; - b2->c0min = b1->c0min; b2->c1min = b1->c1min; b2->c2min = b1->c2min; - /* Choose which axis to split the box on. - * Current algorithm: longest scaled axis. - * See notes in update_box about scaling distances. - */ - c0 = ((b1->c0max - b1->c0min) << C0_SHIFT) * C0_SCALE; - c1 = ((b1->c1max - b1->c1min) << C1_SHIFT) * C1_SCALE; - c2 = ((b1->c2max - b1->c2min) << C2_SHIFT) * C2_SCALE; - /* We want to break any ties in favor of green, then red, blue last. - * This code does the right thing for R,G,B or B,G,R color orders only. - */ -#if RGB_RED == 0 - cmax = c1; n = 1; - if (c0 > cmax) { cmax = c0; n = 0; } - if (c2 > cmax) { n = 2; } -#else - cmax = c1; n = 1; - if (c2 > cmax) { cmax = c2; n = 2; } - if (c0 > cmax) { n = 0; } -#endif - /* Choose split point along selected axis, and update box bounds. - * Current algorithm: split at halfway point. - * (Since the box has been shrunk to minimum volume, - * any split will produce two nonempty subboxes.) - * Note that lb value is max for lower box, so must be < old max. - */ - switch (n) { - case 0: - lb = (b1->c0max + b1->c0min) / 2; - b1->c0max = lb; - b2->c0min = lb+1; - break; - case 1: - lb = (b1->c1max + b1->c1min) / 2; - b1->c1max = lb; - b2->c1min = lb+1; - break; - case 2: - lb = (b1->c2max + b1->c2min) / 2; - b1->c2max = lb; - b2->c2min = lb+1; - break; - } - /* Update stats for boxes */ - update_box(cinfo, b1); - update_box(cinfo, b2); - numboxes++; - } - return numboxes; -} - - -LOCAL(void) -compute_color (j_decompress_ptr cinfo, boxptr boxp, int icolor) -/* Compute representative color for a box, put it in colormap[icolor] */ -{ - /* Current algorithm: mean weighted by pixels (not colors) */ - /* Note it is important to get the rounding correct! */ - my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; - hist3d histogram = cquantize->histogram; - histptr histp; - int c0,c1,c2; - int c0min,c0max,c1min,c1max,c2min,c2max; - long count; - long total = 0; - long c0total = 0; - long c1total = 0; - long c2total = 0; - - c0min = boxp->c0min; c0max = boxp->c0max; - c1min = boxp->c1min; c1max = boxp->c1max; - c2min = boxp->c2min; c2max = boxp->c2max; - - for (c0 = c0min; c0 <= c0max; c0++) - for (c1 = c1min; c1 <= c1max; c1++) { - histp = & histogram[c0][c1][c2min]; - for (c2 = c2min; c2 <= c2max; c2++) { - if ((count = *histp++) != 0) { - total += count; - c0total += ((c0 << C0_SHIFT) + ((1<>1)) * count; - c1total += ((c1 << C1_SHIFT) + ((1<>1)) * count; - c2total += ((c2 << C2_SHIFT) + ((1<>1)) * count; - } - } - } - - cinfo->colormap[0][icolor] = (JSAMPLE) ((c0total + (total>>1)) / total); - cinfo->colormap[1][icolor] = (JSAMPLE) ((c1total + (total>>1)) / total); - cinfo->colormap[2][icolor] = (JSAMPLE) ((c2total + (total>>1)) / total); -} - - -LOCAL(void) -select_colors (j_decompress_ptr cinfo, int desired_colors) -/* Master routine for color selection */ -{ - boxptr boxlist; - int numboxes; - int i; - - /* Allocate workspace for box list */ - boxlist = (boxptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, desired_colors * SIZEOF(box)); - /* Initialize one box containing whole space */ - numboxes = 1; - boxlist[0].c0min = 0; - boxlist[0].c0max = MAXJSAMPLE >> C0_SHIFT; - boxlist[0].c1min = 0; - boxlist[0].c1max = MAXJSAMPLE >> C1_SHIFT; - boxlist[0].c2min = 0; - boxlist[0].c2max = MAXJSAMPLE >> C2_SHIFT; - /* Shrink it to actually-used volume and set its statistics */ - update_box(cinfo, & boxlist[0]); - /* Perform median-cut to produce final box list */ - numboxes = median_cut(cinfo, boxlist, numboxes, desired_colors); - /* Compute the representative color for each box, fill colormap */ - for (i = 0; i < numboxes; i++) - compute_color(cinfo, & boxlist[i], i); - cinfo->actual_number_of_colors = numboxes; - TRACEMS1(cinfo, 1, JTRC_QUANT_SELECTED, numboxes); -} - - -/* - * These routines are concerned with the time-critical task of mapping input - * colors to the nearest color in the selected colormap. - * - * We re-use the histogram space as an "inverse color map", essentially a - * cache for the results of nearest-color searches. All colors within a - * histogram cell will be mapped to the same colormap entry, namely the one - * closest to the cell's center. This may not be quite the closest entry to - * the actual input color, but it's almost as good. A zero in the cache - * indicates we haven't found the nearest color for that cell yet; the array - * is cleared to zeroes before starting the mapping pass. When we find the - * nearest color for a cell, its colormap index plus one is recorded in the - * cache for future use. The pass2 scanning routines call fill_inverse_cmap - * when they need to use an unfilled entry in the cache. - * - * Our method of efficiently finding nearest colors is based on the "locally - * sorted search" idea described by Heckbert and on the incremental distance - * calculation described by Spencer W. Thomas in chapter III.1 of Graphics - * Gems II (James Arvo, ed. Academic Press, 1991). Thomas points out that - * the distances from a given colormap entry to each cell of the histogram can - * be computed quickly using an incremental method: the differences between - * distances to adjacent cells themselves differ by a constant. This allows a - * fairly fast implementation of the "brute force" approach of computing the - * distance from every colormap entry to every histogram cell. Unfortunately, - * it needs a work array to hold the best-distance-so-far for each histogram - * cell (because the inner loop has to be over cells, not colormap entries). - * The work array elements have to be INT32s, so the work array would need - * 256Kb at our recommended precision. This is not feasible in DOS machines. - * - * To get around these problems, we apply Thomas' method to compute the - * nearest colors for only the cells within a small subbox of the histogram. - * The work array need be only as big as the subbox, so the memory usage - * problem is solved. Furthermore, we need not fill subboxes that are never - * referenced in pass2; many images use only part of the color gamut, so a - * fair amount of work is saved. An additional advantage of this - * approach is that we can apply Heckbert's locality criterion to quickly - * eliminate colormap entries that are far away from the subbox; typically - * three-fourths of the colormap entries are rejected by Heckbert's criterion, - * and we need not compute their distances to individual cells in the subbox. - * The speed of this approach is heavily influenced by the subbox size: too - * small means too much overhead, too big loses because Heckbert's criterion - * can't eliminate as many colormap entries. Empirically the best subbox - * size seems to be about 1/512th of the histogram (1/8th in each direction). - * - * Thomas' article also describes a refined method which is asymptotically - * faster than the brute-force method, but it is also far more complex and - * cannot efficiently be applied to small subboxes. It is therefore not - * useful for programs intended to be portable to DOS machines. On machines - * with plenty of memory, filling the whole histogram in one shot with Thomas' - * refined method might be faster than the present code --- but then again, - * it might not be any faster, and it's certainly more complicated. - */ - - -/* log2(histogram cells in update box) for each axis; this can be adjusted */ -#define BOX_C0_LOG (HIST_C0_BITS-3) -#define BOX_C1_LOG (HIST_C1_BITS-3) -#define BOX_C2_LOG (HIST_C2_BITS-3) - -#define BOX_C0_ELEMS (1<actual_number_of_colors; - int maxc0, maxc1, maxc2; - int centerc0, centerc1, centerc2; - int i, x, ncolors; - INT32 minmaxdist, min_dist, max_dist, tdist; - INT32 mindist[MAXNUMCOLORS]; /* min distance to colormap entry i */ - - /* Compute true coordinates of update box's upper corner and center. - * Actually we compute the coordinates of the center of the upper-corner - * histogram cell, which are the upper bounds of the volume we care about. - * Note that since ">>" rounds down, the "center" values may be closer to - * min than to max; hence comparisons to them must be "<=", not "<". - */ - maxc0 = minc0 + ((1 << BOX_C0_SHIFT) - (1 << C0_SHIFT)); - centerc0 = (minc0 + maxc0) >> 1; - maxc1 = minc1 + ((1 << BOX_C1_SHIFT) - (1 << C1_SHIFT)); - centerc1 = (minc1 + maxc1) >> 1; - maxc2 = minc2 + ((1 << BOX_C2_SHIFT) - (1 << C2_SHIFT)); - centerc2 = (minc2 + maxc2) >> 1; - - /* For each color in colormap, find: - * 1. its minimum squared-distance to any point in the update box - * (zero if color is within update box); - * 2. its maximum squared-distance to any point in the update box. - * Both of these can be found by considering only the corners of the box. - * We save the minimum distance for each color in mindist[]; - * only the smallest maximum distance is of interest. - */ - minmaxdist = 0x7FFFFFFFL; - - for (i = 0; i < numcolors; i++) { - /* We compute the squared-c0-distance term, then add in the other two. */ - x = GETJSAMPLE(cinfo->colormap[0][i]); - if (x < minc0) { - tdist = (x - minc0) * C0_SCALE; - min_dist = tdist*tdist; - tdist = (x - maxc0) * C0_SCALE; - max_dist = tdist*tdist; - } else if (x > maxc0) { - tdist = (x - maxc0) * C0_SCALE; - min_dist = tdist*tdist; - tdist = (x - minc0) * C0_SCALE; - max_dist = tdist*tdist; - } else { - /* within cell range so no contribution to min_dist */ - min_dist = 0; - if (x <= centerc0) { - tdist = (x - maxc0) * C0_SCALE; - max_dist = tdist*tdist; - } else { - tdist = (x - minc0) * C0_SCALE; - max_dist = tdist*tdist; - } - } - - x = GETJSAMPLE(cinfo->colormap[1][i]); - if (x < minc1) { - tdist = (x - minc1) * C1_SCALE; - min_dist += tdist*tdist; - tdist = (x - maxc1) * C1_SCALE; - max_dist += tdist*tdist; - } else if (x > maxc1) { - tdist = (x - maxc1) * C1_SCALE; - min_dist += tdist*tdist; - tdist = (x - minc1) * C1_SCALE; - max_dist += tdist*tdist; - } else { - /* within cell range so no contribution to min_dist */ - if (x <= centerc1) { - tdist = (x - maxc1) * C1_SCALE; - max_dist += tdist*tdist; - } else { - tdist = (x - minc1) * C1_SCALE; - max_dist += tdist*tdist; - } - } - - x = GETJSAMPLE(cinfo->colormap[2][i]); - if (x < minc2) { - tdist = (x - minc2) * C2_SCALE; - min_dist += tdist*tdist; - tdist = (x - maxc2) * C2_SCALE; - max_dist += tdist*tdist; - } else if (x > maxc2) { - tdist = (x - maxc2) * C2_SCALE; - min_dist += tdist*tdist; - tdist = (x - minc2) * C2_SCALE; - max_dist += tdist*tdist; - } else { - /* within cell range so no contribution to min_dist */ - if (x <= centerc2) { - tdist = (x - maxc2) * C2_SCALE; - max_dist += tdist*tdist; - } else { - tdist = (x - minc2) * C2_SCALE; - max_dist += tdist*tdist; - } - } - - mindist[i] = min_dist; /* save away the results */ - if (max_dist < minmaxdist) - minmaxdist = max_dist; - } - - /* Now we know that no cell in the update box is more than minmaxdist - * away from some colormap entry. Therefore, only colors that are - * within minmaxdist of some part of the box need be considered. - */ - ncolors = 0; - for (i = 0; i < numcolors; i++) { - if (mindist[i] <= minmaxdist) - colorlist[ncolors++] = (JSAMPLE) i; - } - return ncolors; -} - - -LOCAL(void) -find_best_colors (j_decompress_ptr cinfo, int minc0, int minc1, int minc2, - int numcolors, JSAMPLE colorlist[], JSAMPLE bestcolor[]) -/* Find the closest colormap entry for each cell in the update box, - * given the list of candidate colors prepared by find_nearby_colors. - * Return the indexes of the closest entries in the bestcolor[] array. - * This routine uses Thomas' incremental distance calculation method to - * find the distance from a colormap entry to successive cells in the box. - */ -{ - int ic0, ic1, ic2; - int i, icolor; - register INT32 * bptr; /* pointer into bestdist[] array */ - JSAMPLE * cptr; /* pointer into bestcolor[] array */ - INT32 dist0, dist1; /* initial distance values */ - register INT32 dist2; /* current distance in inner loop */ - INT32 xx0, xx1; /* distance increments */ - register INT32 xx2; - INT32 inc0, inc1, inc2; /* initial values for increments */ - /* This array holds the distance to the nearest-so-far color for each cell */ - INT32 bestdist[BOX_C0_ELEMS * BOX_C1_ELEMS * BOX_C2_ELEMS]; - - /* Initialize best-distance for each cell of the update box */ - bptr = bestdist; - for (i = BOX_C0_ELEMS*BOX_C1_ELEMS*BOX_C2_ELEMS-1; i >= 0; i--) - *bptr++ = 0x7FFFFFFFL; - - /* For each color selected by find_nearby_colors, - * compute its distance to the center of each cell in the box. - * If that's less than best-so-far, update best distance and color number. - */ - - /* Nominal steps between cell centers ("x" in Thomas article) */ -#define STEP_C0 ((1 << C0_SHIFT) * C0_SCALE) -#define STEP_C1 ((1 << C1_SHIFT) * C1_SCALE) -#define STEP_C2 ((1 << C2_SHIFT) * C2_SCALE) - - for (i = 0; i < numcolors; i++) { - icolor = GETJSAMPLE(colorlist[i]); - /* Compute (square of) distance from minc0/c1/c2 to this color */ - inc0 = (minc0 - GETJSAMPLE(cinfo->colormap[0][icolor])) * C0_SCALE; - dist0 = inc0*inc0; - inc1 = (minc1 - GETJSAMPLE(cinfo->colormap[1][icolor])) * C1_SCALE; - dist0 += inc1*inc1; - inc2 = (minc2 - GETJSAMPLE(cinfo->colormap[2][icolor])) * C2_SCALE; - dist0 += inc2*inc2; - /* Form the initial difference increments */ - inc0 = inc0 * (2 * STEP_C0) + STEP_C0 * STEP_C0; - inc1 = inc1 * (2 * STEP_C1) + STEP_C1 * STEP_C1; - inc2 = inc2 * (2 * STEP_C2) + STEP_C2 * STEP_C2; - /* Now loop over all cells in box, updating distance per Thomas method */ - bptr = bestdist; - cptr = bestcolor; - xx0 = inc0; - for (ic0 = BOX_C0_ELEMS-1; ic0 >= 0; ic0--) { - dist1 = dist0; - xx1 = inc1; - for (ic1 = BOX_C1_ELEMS-1; ic1 >= 0; ic1--) { - dist2 = dist1; - xx2 = inc2; - for (ic2 = BOX_C2_ELEMS-1; ic2 >= 0; ic2--) { - if (dist2 < *bptr) { - *bptr = dist2; - *cptr = (JSAMPLE) icolor; - } - dist2 += xx2; - xx2 += 2 * STEP_C2 * STEP_C2; - bptr++; - cptr++; - } - dist1 += xx1; - xx1 += 2 * STEP_C1 * STEP_C1; - } - dist0 += xx0; - xx0 += 2 * STEP_C0 * STEP_C0; - } - } -} - - -LOCAL(void) -fill_inverse_cmap (j_decompress_ptr cinfo, int c0, int c1, int c2) -/* Fill the inverse-colormap entries in the update box that contains */ -/* histogram cell c0/c1/c2. (Only that one cell MUST be filled, but */ -/* we can fill as many others as we wish.) */ -{ - my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; - hist3d histogram = cquantize->histogram; - int minc0, minc1, minc2; /* lower left corner of update box */ - int ic0, ic1, ic2; - register JSAMPLE * cptr; /* pointer into bestcolor[] array */ - register histptr cachep; /* pointer into main cache array */ - /* This array lists the candidate colormap indexes. */ - JSAMPLE colorlist[MAXNUMCOLORS]; - int numcolors; /* number of candidate colors */ - /* This array holds the actually closest colormap index for each cell. */ - JSAMPLE bestcolor[BOX_C0_ELEMS * BOX_C1_ELEMS * BOX_C2_ELEMS]; - - /* Convert cell coordinates to update box ID */ - c0 >>= BOX_C0_LOG; - c1 >>= BOX_C1_LOG; - c2 >>= BOX_C2_LOG; - - /* Compute true coordinates of update box's origin corner. - * Actually we compute the coordinates of the center of the corner - * histogram cell, which are the lower bounds of the volume we care about. - */ - minc0 = (c0 << BOX_C0_SHIFT) + ((1 << C0_SHIFT) >> 1); - minc1 = (c1 << BOX_C1_SHIFT) + ((1 << C1_SHIFT) >> 1); - minc2 = (c2 << BOX_C2_SHIFT) + ((1 << C2_SHIFT) >> 1); - - /* Determine which colormap entries are close enough to be candidates - * for the nearest entry to some cell in the update box. - */ - numcolors = find_nearby_colors(cinfo, minc0, minc1, minc2, colorlist); - - /* Determine the actually nearest colors. */ - find_best_colors(cinfo, minc0, minc1, minc2, numcolors, colorlist, - bestcolor); - - /* Save the best color numbers (plus 1) in the main cache array */ - c0 <<= BOX_C0_LOG; /* convert ID back to base cell indexes */ - c1 <<= BOX_C1_LOG; - c2 <<= BOX_C2_LOG; - cptr = bestcolor; - for (ic0 = 0; ic0 < BOX_C0_ELEMS; ic0++) { - for (ic1 = 0; ic1 < BOX_C1_ELEMS; ic1++) { - cachep = & histogram[c0+ic0][c1+ic1][c2]; - for (ic2 = 0; ic2 < BOX_C2_ELEMS; ic2++) { - *cachep++ = (histcell) (GETJSAMPLE(*cptr++) + 1); - } - } - } -} - - -/* - * Map some rows of pixels to the output colormapped representation. - */ - -METHODDEF(void) -pass2_no_dither (j_decompress_ptr cinfo, - JSAMPARRAY input_buf, JSAMPARRAY output_buf, int num_rows) -/* This version performs no dithering */ -{ - my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; - hist3d histogram = cquantize->histogram; - register JSAMPROW inptr, outptr; - register histptr cachep; - register int c0, c1, c2; - int row; - JDIMENSION col; - JDIMENSION width = cinfo->output_width; - - for (row = 0; row < num_rows; row++) { - inptr = input_buf[row]; - outptr = output_buf[row]; - for (col = width; col > 0; col--) { - /* get pixel value and index into the cache */ - c0 = GETJSAMPLE(*inptr++) >> C0_SHIFT; - c1 = GETJSAMPLE(*inptr++) >> C1_SHIFT; - c2 = GETJSAMPLE(*inptr++) >> C2_SHIFT; - cachep = & histogram[c0][c1][c2]; - /* If we have not seen this color before, find nearest colormap entry */ - /* and update the cache */ - if (*cachep == 0) - fill_inverse_cmap(cinfo, c0,c1,c2); - /* Now emit the colormap index for this cell */ - *outptr++ = (JSAMPLE) (*cachep - 1); - } - } -} - - -METHODDEF(void) -pass2_fs_dither (j_decompress_ptr cinfo, - JSAMPARRAY input_buf, JSAMPARRAY output_buf, int num_rows) -/* This version performs Floyd-Steinberg dithering */ -{ - my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; - hist3d histogram = cquantize->histogram; - register LOCFSERROR cur0, cur1, cur2; /* current error or pixel value */ - LOCFSERROR belowerr0, belowerr1, belowerr2; /* error for pixel below cur */ - LOCFSERROR bpreverr0, bpreverr1, bpreverr2; /* error for below/prev col */ - register FSERRPTR errorptr; /* => fserrors[] at column before current */ - JSAMPROW inptr; /* => current input pixel */ - JSAMPROW outptr; /* => current output pixel */ - histptr cachep; - int dir; /* +1 or -1 depending on direction */ - int dir3; /* 3*dir, for advancing inptr & errorptr */ - int row; - JDIMENSION col; - JDIMENSION width = cinfo->output_width; - JSAMPLE *range_limit = cinfo->sample_range_limit; - int *error_limit = cquantize->error_limiter; - JSAMPROW colormap0 = cinfo->colormap[0]; - JSAMPROW colormap1 = cinfo->colormap[1]; - JSAMPROW colormap2 = cinfo->colormap[2]; - SHIFT_TEMPS - - for (row = 0; row < num_rows; row++) { - inptr = input_buf[row]; - outptr = output_buf[row]; - if (cquantize->on_odd_row) { - /* work right to left in this row */ - inptr += (width-1) * 3; /* so point to rightmost pixel */ - outptr += width-1; - dir = -1; - dir3 = -3; - errorptr = cquantize->fserrors + (width+1)*3; /* => entry after last column */ - cquantize->on_odd_row = FALSE; /* flip for next time */ - } else { - /* work left to right in this row */ - dir = 1; - dir3 = 3; - errorptr = cquantize->fserrors; /* => entry before first real column */ - cquantize->on_odd_row = TRUE; /* flip for next time */ - } - /* Preset error values: no error propagated to first pixel from left */ - cur0 = cur1 = cur2 = 0; - /* and no error propagated to row below yet */ - belowerr0 = belowerr1 = belowerr2 = 0; - bpreverr0 = bpreverr1 = bpreverr2 = 0; - - for (col = width; col > 0; col--) { - /* curN holds the error propagated from the previous pixel on the - * current line. Add the error propagated from the previous line - * to form the complete error correction term for this pixel, and - * round the error term (which is expressed * 16) to an integer. - * RIGHT_SHIFT rounds towards minus infinity, so adding 8 is correct - * for either sign of the error value. - * Note: errorptr points to *previous* column's array entry. - */ - cur0 = RIGHT_SHIFT(cur0 + errorptr[dir3+0] + 8, 4); - cur1 = RIGHT_SHIFT(cur1 + errorptr[dir3+1] + 8, 4); - cur2 = RIGHT_SHIFT(cur2 + errorptr[dir3+2] + 8, 4); - /* Limit the error using transfer function set by init_error_limit. - * See comments with init_error_limit for rationale. - */ - cur0 = error_limit[cur0]; - cur1 = error_limit[cur1]; - cur2 = error_limit[cur2]; - /* Form pixel value + error, and range-limit to 0..MAXJSAMPLE. - * The maximum error is +- MAXJSAMPLE (or less with error limiting); - * this sets the required size of the range_limit array. - */ - cur0 += GETJSAMPLE(inptr[0]); - cur1 += GETJSAMPLE(inptr[1]); - cur2 += GETJSAMPLE(inptr[2]); - cur0 = GETJSAMPLE(range_limit[cur0]); - cur1 = GETJSAMPLE(range_limit[cur1]); - cur2 = GETJSAMPLE(range_limit[cur2]); - /* Index into the cache with adjusted pixel value */ - cachep = & histogram[cur0>>C0_SHIFT][cur1>>C1_SHIFT][cur2>>C2_SHIFT]; - /* If we have not seen this color before, find nearest colormap */ - /* entry and update the cache */ - if (*cachep == 0) - fill_inverse_cmap(cinfo, cur0>>C0_SHIFT,cur1>>C1_SHIFT,cur2>>C2_SHIFT); - /* Now emit the colormap index for this cell */ - { register int pixcode = *cachep - 1; - *outptr = (JSAMPLE) pixcode; - /* Compute representation error for this pixel */ - cur0 -= GETJSAMPLE(colormap0[pixcode]); - cur1 -= GETJSAMPLE(colormap1[pixcode]); - cur2 -= GETJSAMPLE(colormap2[pixcode]); - } - /* Compute error fractions to be propagated to adjacent pixels. - * Add these into the running sums, and simultaneously shift the - * next-line error sums left by 1 column. - */ - { register LOCFSERROR bnexterr, delta; - - bnexterr = cur0; /* Process component 0 */ - delta = cur0 * 2; - cur0 += delta; /* form error * 3 */ - errorptr[0] = (FSERROR) (bpreverr0 + cur0); - cur0 += delta; /* form error * 5 */ - bpreverr0 = belowerr0 + cur0; - belowerr0 = bnexterr; - cur0 += delta; /* form error * 7 */ - bnexterr = cur1; /* Process component 1 */ - delta = cur1 * 2; - cur1 += delta; /* form error * 3 */ - errorptr[1] = (FSERROR) (bpreverr1 + cur1); - cur1 += delta; /* form error * 5 */ - bpreverr1 = belowerr1 + cur1; - belowerr1 = bnexterr; - cur1 += delta; /* form error * 7 */ - bnexterr = cur2; /* Process component 2 */ - delta = cur2 * 2; - cur2 += delta; /* form error * 3 */ - errorptr[2] = (FSERROR) (bpreverr2 + cur2); - cur2 += delta; /* form error * 5 */ - bpreverr2 = belowerr2 + cur2; - belowerr2 = bnexterr; - cur2 += delta; /* form error * 7 */ - } - /* At this point curN contains the 7/16 error value to be propagated - * to the next pixel on the current line, and all the errors for the - * next line have been shifted over. We are therefore ready to move on. - */ - inptr += dir3; /* Advance pixel pointers to next column */ - outptr += dir; - errorptr += dir3; /* advance errorptr to current column */ - } - /* Post-loop cleanup: we must unload the final error values into the - * final fserrors[] entry. Note we need not unload belowerrN because - * it is for the dummy column before or after the actual array. - */ - errorptr[0] = (FSERROR) bpreverr0; /* unload prev errs into array */ - errorptr[1] = (FSERROR) bpreverr1; - errorptr[2] = (FSERROR) bpreverr2; - } -} - - -/* - * Initialize the error-limiting transfer function (lookup table). - * The raw F-S error computation can potentially compute error values of up to - * +- MAXJSAMPLE. But we want the maximum correction applied to a pixel to be - * much less, otherwise obviously wrong pixels will be created. (Typical - * effects include weird fringes at color-area boundaries, isolated bright - * pixels in a dark area, etc.) The standard advice for avoiding this problem - * is to ensure that the "corners" of the color cube are allocated as output - * colors; then repeated errors in the same direction cannot cause cascading - * error buildup. However, that only prevents the error from getting - * completely out of hand; Aaron Giles reports that error limiting improves - * the results even with corner colors allocated. - * A simple clamping of the error values to about +- MAXJSAMPLE/8 works pretty - * well, but the smoother transfer function used below is even better. Thanks - * to Aaron Giles for this idea. - */ - -LOCAL(void) -init_error_limit (j_decompress_ptr cinfo) -/* Allocate and fill in the error_limiter table */ -{ - my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; - int * table; - int in, out; - - table = (int *) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, (MAXJSAMPLE*2+1) * SIZEOF(int)); - table += MAXJSAMPLE; /* so can index -MAXJSAMPLE .. +MAXJSAMPLE */ - cquantize->error_limiter = table; - -#define STEPSIZE ((MAXJSAMPLE+1)/16) - /* Map errors 1:1 up to +- MAXJSAMPLE/16 */ - out = 0; - for (in = 0; in < STEPSIZE; in++, out++) { - table[in] = out; table[-in] = -out; - } - /* Map errors 1:2 up to +- 3*MAXJSAMPLE/16 */ - for (; in < STEPSIZE*3; in++, out += (in&1) ? 0 : 1) { - table[in] = out; table[-in] = -out; - } - /* Clamp the rest to final out value (which is (MAXJSAMPLE+1)/8) */ - for (; in <= MAXJSAMPLE; in++) { - table[in] = out; table[-in] = -out; - } -#undef STEPSIZE -} - - -/* - * Finish up at the end of each pass. - */ - -METHODDEF(void) -finish_pass1 (j_decompress_ptr cinfo) -{ - my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; - - /* Select the representative colors and fill in cinfo->colormap */ - cinfo->colormap = cquantize->sv_colormap; - select_colors(cinfo, cquantize->desired); - /* Force next pass to zero the color index table */ - cquantize->needs_zeroed = TRUE; -} - - -METHODDEF(void) -finish_pass2 (j_decompress_ptr cinfo) -{ - /* no work */ -} - - -/* - * Initialize for each processing pass. - */ - -METHODDEF(void) -start_pass_2_quant (j_decompress_ptr cinfo, boolean is_pre_scan) -{ - my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; - hist3d histogram = cquantize->histogram; - int i; - - /* Only F-S dithering or no dithering is supported. */ - /* If user asks for ordered dither, give him F-S. */ - if (cinfo->dither_mode != JDITHER_NONE) - cinfo->dither_mode = JDITHER_FS; - - if (is_pre_scan) { - /* Set up method pointers */ - cquantize->pub.color_quantize = prescan_quantize; - cquantize->pub.finish_pass = finish_pass1; - cquantize->needs_zeroed = TRUE; /* Always zero histogram */ - } else { - /* Set up method pointers */ - if (cinfo->dither_mode == JDITHER_FS) - cquantize->pub.color_quantize = pass2_fs_dither; - else - cquantize->pub.color_quantize = pass2_no_dither; - cquantize->pub.finish_pass = finish_pass2; - - /* Make sure color count is acceptable */ - i = cinfo->actual_number_of_colors; - if (i < 1) - ERREXIT1(cinfo, JERR_QUANT_FEW_COLORS, 1); - if (i > MAXNUMCOLORS) - ERREXIT1(cinfo, JERR_QUANT_MANY_COLORS, MAXNUMCOLORS); - - if (cinfo->dither_mode == JDITHER_FS) { - size_t arraysize = ((size_t) cinfo->output_width + (size_t) 2) - * (3 * SIZEOF(FSERROR)); - /* Allocate Floyd-Steinberg workspace if we didn't already. */ - if (cquantize->fserrors == NULL) - cquantize->fserrors = (FSERRPTR) (*cinfo->mem->alloc_large) - ((j_common_ptr) cinfo, JPOOL_IMAGE, arraysize); - /* Initialize the propagated errors to zero. */ - FMEMZERO((void FAR *) cquantize->fserrors, arraysize); - /* Make the error-limit table if we didn't already. */ - if (cquantize->error_limiter == NULL) - init_error_limit(cinfo); - cquantize->on_odd_row = FALSE; - } - - } - /* Zero the histogram or inverse color map, if necessary */ - if (cquantize->needs_zeroed) { - for (i = 0; i < HIST_C0_ELEMS; i++) { - FMEMZERO((void FAR *) histogram[i], - HIST_C1_ELEMS*HIST_C2_ELEMS * SIZEOF(histcell)); - } - cquantize->needs_zeroed = FALSE; - } -} - - -/* - * Switch to a new external colormap between output passes. - */ - -METHODDEF(void) -new_color_map_2_quant (j_decompress_ptr cinfo) -{ - my_cquantize_ptr cquantize = (my_cquantize_ptr) cinfo->cquantize; - - /* Reset the inverse color map */ - cquantize->needs_zeroed = TRUE; -} - - -/* - * Module initialization routine for 2-pass color quantization. - */ - -GLOBAL(void) -jinit_2pass_quantizer (j_decompress_ptr cinfo) -{ - my_cquantize_ptr cquantize; - int i; - - cquantize = (my_cquantize_ptr) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, SIZEOF(my_cquantizer)); - cinfo->cquantize = &cquantize->pub; - cquantize->pub.start_pass = start_pass_2_quant; - cquantize->pub.new_color_map = new_color_map_2_quant; - cquantize->fserrors = NULL; /* flag optional arrays not allocated */ - cquantize->error_limiter = NULL; - - /* Make sure jdmaster didn't give me a case I can't handle */ - if (cinfo->out_color_components != 3) - ERREXIT(cinfo, JERR_NOTIMPL); - - /* Allocate the histogram/inverse colormap storage */ - cquantize->histogram = (hist3d) (*cinfo->mem->alloc_small) - ((j_common_ptr) cinfo, JPOOL_IMAGE, HIST_C0_ELEMS * SIZEOF(hist2d)); - for (i = 0; i < HIST_C0_ELEMS; i++) { - cquantize->histogram[i] = (hist2d) (*cinfo->mem->alloc_large) - ((j_common_ptr) cinfo, JPOOL_IMAGE, - HIST_C1_ELEMS*HIST_C2_ELEMS * SIZEOF(histcell)); - } - cquantize->needs_zeroed = TRUE; /* histogram is garbage now */ - - /* Allocate storage for the completed colormap, if required. - * We do this now since it is FAR storage and may affect - * the memory manager's space calculations. - */ - if (cinfo->enable_2pass_quant) { - /* Make sure color count is acceptable */ - int desired = cinfo->desired_number_of_colors; - /* Lower bound on # of colors ... somewhat arbitrary as long as > 0 */ - if (desired < 8) - ERREXIT1(cinfo, JERR_QUANT_FEW_COLORS, 8); - /* Make sure colormap indexes can be represented by JSAMPLEs */ - if (desired > MAXNUMCOLORS) - ERREXIT1(cinfo, JERR_QUANT_MANY_COLORS, MAXNUMCOLORS); - cquantize->sv_colormap = (*cinfo->mem->alloc_sarray) - ((j_common_ptr) cinfo, JPOOL_IMAGE, - (JDIMENSION) desired, (JDIMENSION) 3); - cquantize->desired = desired; - } else - cquantize->sv_colormap = NULL; - - /* Only F-S dithering or no dithering is supported. */ - /* If user asks for ordered dither, give him F-S. */ - if (cinfo->dither_mode != JDITHER_NONE) - cinfo->dither_mode = JDITHER_FS; - - /* Allocate Floyd-Steinberg workspace if necessary. - * This isn't really needed until pass 2, but again it is FAR storage. - * Although we will cope with a later change in dither_mode, - * we do not promise to honor max_memory_to_use if dither_mode changes. - */ - if (cinfo->dither_mode == JDITHER_FS) { - cquantize->fserrors = (FSERRPTR) (*cinfo->mem->alloc_large) - ((j_common_ptr) cinfo, JPOOL_IMAGE, - ((size_t) cinfo->output_width + (size_t) 2) * (3 * SIZEOF(FSERROR))); - /* Might as well create the error-limiting table too. */ - init_error_limit(cinfo); - } -} - -#endif /* QUANT_2PASS_SUPPORTED */ diff --git a/cmake/externals/libjpeg/jutils.c b/cmake/externals/libjpeg/jutils.c deleted file mode 100644 index 31e16dfb5..000000000 --- a/cmake/externals/libjpeg/jutils.c +++ /dev/null @@ -1,224 +0,0 @@ -/* - * jutils.c - * - * Copyright (C) 1991-1996, Thomas G. Lane. - * Modified 2009-2020 by Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains tables and miscellaneous utility routines needed - * for both compression and decompression. - * Note we prefix all global names with "j" to minimize conflicts with - * a surrounding application. - */ - -#define JPEG_INTERNALS -#include "jinclude.h" -#include "jpeglib.h" - - -/* - * jpeg_zigzag_order[i] is the zigzag-order position of the i'th element - * of a DCT block read in natural order (left to right, top to bottom). - */ - -#if 0 /* This table is not actually needed in v6a */ - -const int jpeg_zigzag_order[DCTSIZE2] = { - 0, 1, 5, 6, 14, 15, 27, 28, - 2, 4, 7, 13, 16, 26, 29, 42, - 3, 8, 12, 17, 25, 30, 41, 43, - 9, 11, 18, 24, 31, 40, 44, 53, - 10, 19, 23, 32, 39, 45, 52, 54, - 20, 22, 33, 38, 46, 51, 55, 60, - 21, 34, 37, 47, 50, 56, 59, 61, - 35, 36, 48, 49, 57, 58, 62, 63 -}; - -#endif - -/* - * jpeg_natural_order[i] is the natural-order position of the i'th element - * of zigzag order. - * - * When reading corrupted data, the Huffman decoders could attempt - * to reference an entry beyond the end of this array (if the decoded - * zero run length reaches past the end of the block). To prevent - * wild stores without adding an inner-loop test, we put some extra - * "63"s after the real entries. This will cause the extra coefficient - * to be stored in location 63 of the block, not somewhere random. - * The worst case would be a run-length of 15, which means we need 16 - * fake entries. - */ - -const int jpeg_natural_order[DCTSIZE2+16] = { - 0, 1, 8, 16, 9, 2, 3, 10, - 17, 24, 32, 25, 18, 11, 4, 5, - 12, 19, 26, 33, 40, 48, 41, 34, - 27, 20, 13, 6, 7, 14, 21, 28, - 35, 42, 49, 56, 57, 50, 43, 36, - 29, 22, 15, 23, 30, 37, 44, 51, - 58, 59, 52, 45, 38, 31, 39, 46, - 53, 60, 61, 54, 47, 55, 62, 63, - 63, 63, 63, 63, 63, 63, 63, 63, /* extra entries for safety in decoder */ - 63, 63, 63, 63, 63, 63, 63, 63 -}; - -const int jpeg_natural_order7[7*7+16] = { - 0, 1, 8, 16, 9, 2, 3, 10, - 17, 24, 32, 25, 18, 11, 4, 5, - 12, 19, 26, 33, 40, 48, 41, 34, - 27, 20, 13, 6, 14, 21, 28, 35, - 42, 49, 50, 43, 36, 29, 22, 30, - 37, 44, 51, 52, 45, 38, 46, 53, - 54, - 63, 63, 63, 63, 63, 63, 63, 63, /* extra entries for safety in decoder */ - 63, 63, 63, 63, 63, 63, 63, 63 -}; - -const int jpeg_natural_order6[6*6+16] = { - 0, 1, 8, 16, 9, 2, 3, 10, - 17, 24, 32, 25, 18, 11, 4, 5, - 12, 19, 26, 33, 40, 41, 34, 27, - 20, 13, 21, 28, 35, 42, 43, 36, - 29, 37, 44, 45, - 63, 63, 63, 63, 63, 63, 63, 63, /* extra entries for safety in decoder */ - 63, 63, 63, 63, 63, 63, 63, 63 -}; - -const int jpeg_natural_order5[5*5+16] = { - 0, 1, 8, 16, 9, 2, 3, 10, - 17, 24, 32, 25, 18, 11, 4, 12, - 19, 26, 33, 34, 27, 20, 28, 35, - 36, - 63, 63, 63, 63, 63, 63, 63, 63, /* extra entries for safety in decoder */ - 63, 63, 63, 63, 63, 63, 63, 63 -}; - -const int jpeg_natural_order4[4*4+16] = { - 0, 1, 8, 16, 9, 2, 3, 10, - 17, 24, 25, 18, 11, 19, 26, 27, - 63, 63, 63, 63, 63, 63, 63, 63, /* extra entries for safety in decoder */ - 63, 63, 63, 63, 63, 63, 63, 63 -}; - -const int jpeg_natural_order3[3*3+16] = { - 0, 1, 8, 16, 9, 2, 10, 17, - 18, - 63, 63, 63, 63, 63, 63, 63, 63, /* extra entries for safety in decoder */ - 63, 63, 63, 63, 63, 63, 63, 63 -}; - -const int jpeg_natural_order2[2*2+16] = { - 0, 1, 8, 9, - 63, 63, 63, 63, 63, 63, 63, 63, /* extra entries for safety in decoder */ - 63, 63, 63, 63, 63, 63, 63, 63 -}; - - -/* - * Arithmetic utilities - */ - -GLOBAL(long) -jdiv_round_up (long a, long b) -/* Compute a/b rounded up to next integer, ie, ceil(a/b) */ -/* Assumes a >= 0, b > 0 */ -{ - return (a + b - 1L) / b; -} - - -GLOBAL(long) -jround_up (long a, long b) -/* Compute a rounded up to next multiple of b, ie, ceil(a/b)*b */ -/* Assumes a >= 0, b > 0 */ -{ - a += b - 1L; - return a - (a % b); -} - - -/* On normal machines we can apply MEMCOPY() and MEMZERO() to sample arrays - * and coefficient-block arrays. This won't work on 80x86 because the arrays - * are FAR and we're assuming a small-pointer memory model. However, some - * DOS compilers provide far-pointer versions of memcpy() and memset() even - * in the small-model libraries. These will be used if USE_FMEM is defined. - * Otherwise, the routines below do it the hard way. (The performance cost - * is not all that great, because these routines aren't very heavily used.) - */ - -#ifndef NEED_FAR_POINTERS /* normal case, same as regular macro */ -#define FMEMCOPY(dest,src,size) MEMCOPY(dest,src,size) -#else /* 80x86 case, define if we can */ -#ifdef USE_FMEM -#define FMEMCOPY(dest,src,size) _fmemcpy((void FAR *)(dest), (const void FAR *)(src), (size_t)(size)) -#else -/* This function is for use by the FMEMZERO macro defined in jpegint.h. - * Do not call this function directly, use the FMEMZERO macro instead. - */ -GLOBAL(void) -jzero_far (void FAR * target, size_t bytestozero) -/* Zero out a chunk of FAR memory. */ -/* This might be sample-array data, block-array data, or alloc_large data. */ -{ - register char FAR * ptr = (char FAR *) target; - register size_t count; - - for (count = bytestozero; count > 0; count--) { - *ptr++ = 0; - } -} -#endif -#endif - - -GLOBAL(void) -jcopy_sample_rows (JSAMPARRAY input_array, - JSAMPARRAY output_array, - int num_rows, JDIMENSION num_cols) -/* Copy some rows of samples from one place to another. - * num_rows rows are copied from *input_array++ to *output_array++; - * these areas may overlap for duplication. - * The source and destination arrays must be at least as wide as num_cols. - */ -{ - register JSAMPROW inptr, outptr; -#ifdef FMEMCOPY - register size_t count = (size_t) num_cols * SIZEOF(JSAMPLE); -#else - register JDIMENSION count; -#endif - register int row; - - for (row = num_rows; row > 0; row--) { - inptr = *input_array++; - outptr = *output_array++; -#ifdef FMEMCOPY - FMEMCOPY(outptr, inptr, count); -#else - for (count = num_cols; count > 0; count--) - *outptr++ = *inptr++; /* needn't bother with GETJSAMPLE() here */ -#endif - } -} - - -GLOBAL(void) -jcopy_block_row (JBLOCKROW input_row, JBLOCKROW output_row, - JDIMENSION num_blocks) -/* Copy a row of coefficient blocks from one place to another. */ -{ -#ifdef FMEMCOPY - FMEMCOPY(output_row, input_row, (size_t) num_blocks * (DCTSIZE2 * SIZEOF(JCOEF))); -#else - register JCOEFPTR inptr, outptr; - register long count; - - inptr = (JCOEFPTR) input_row; - outptr = (JCOEFPTR) output_row; - for (count = (long) num_blocks * DCTSIZE2; count > 0; count--) { - *outptr++ = *inptr++; - } -#endif -} diff --git a/cmake/externals/libjpeg/jversion.h b/cmake/externals/libjpeg/jversion.h deleted file mode 100644 index df53ef5e5..000000000 --- a/cmake/externals/libjpeg/jversion.h +++ /dev/null @@ -1,14 +0,0 @@ -/* - * jversion.h - * - * Copyright (C) 1991-2024, Thomas G. Lane, Guido Vollbeding. - * This file is part of the Independent JPEG Group's software. - * For conditions of distribution and use, see the accompanying README file. - * - * This file contains software version identification. - */ - - -#define JVERSION "9f 14-Jan-2024" - -#define JCOPYRIGHT "Copyright (C) 2024, Thomas G. Lane, Guido Vollbeding" diff --git a/cmake/externals/libspng/CMakeLists.txt b/cmake/externals/libspng/CMakeLists.txt deleted file mode 100644 index c277888e5..000000000 --- a/cmake/externals/libspng/CMakeLists.txt +++ /dev/null @@ -1,39 +0,0 @@ -# ---------------------------------------------------------------------------- -# CMake file for libspng. See root CMakeLists.txt -# -# ---------------------------------------------------------------------------- - -project(SPNG) - -set(CURR_INCLUDE_DIR "${CMAKE_CURRENT_LIST_DIR}") -set_property(GLOBAL PROPERTY SPNG_INCLUDE_DIR ${CURR_INCLUDE_DIR}) - -file(GLOB_RECURSE spng_headers RELATIVE "${CMAKE_CURRENT_LIST_DIR}" "*.h") -file(GLOB_RECURSE spng_sources RELATIVE "${CMAKE_CURRENT_LIST_DIR}" "*.c") - -message(STATUS "libspng will be used as PNG codec") - -# ---------------------------------------------------------------------------------- -# Define the library target: -# ---------------------------------------------------------------------------------- - -if(MSVC) - add_definitions(-D_CRT_SECURE_NO_DEPRECATE) -endif(MSVC) - -add_library(SPNG STATIC ${OPENCV_3RDPARTY_EXCLUDE_FROM_ALL} ${spng_headers} ${spng_sources}) -target_link_libraries(SPNG ${ZLIB_LIBRARIES}) - -set_target_properties(SPNG - PROPERTIES OUTPUT_NAME SPNG - DEBUG_POSTFIX "${OPENCV_DEBUG_POSTFIX}" - COMPILE_PDB_NAME SPNG - COMPILE_PDB_NAME_DEBUG "SPNG${OPENCV_DEBUG_POSTFIX}" - ARCHIVE_OUTPUT_DIRECTORY "${3P_LIBRARY_OUTPUT_PATH}" - ) - -target_compile_definitions(SPNG PUBLIC SPNG_STATIC) - -if(ENABLE_SOLUTION_FOLDERS) - set_target_properties(SPNG PROPERTIES FOLDER "3rdparty") -endif() \ No newline at end of file diff --git a/cmake/externals/libspng/LICENSE b/cmake/externals/libspng/LICENSE deleted file mode 100644 index a29d5a2fa..000000000 --- a/cmake/externals/libspng/LICENSE +++ /dev/null @@ -1,25 +0,0 @@ -BSD 2-Clause License - -Copyright (c) 2018-2023, Randy -All rights reserved. - -Redistribution and use in source and binary forms, with or without -modification, are permitted provided that the following conditions are met: - -* Redistributions of source code must retain the above copyright notice, this - list of conditions and the following disclaimer. - -* Redistributions in binary form must reproduce the above copyright notice, - this list of conditions and the following disclaimer in the documentation - and/or other materials provided with the distribution. - -THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" -AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE -IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE -DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE -FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL -DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR -SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER -CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, -OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE -OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. \ No newline at end of file diff --git a/cmake/externals/libspng/spng.c b/cmake/externals/libspng/spng.c deleted file mode 100644 index b22b7110a..000000000 --- a/cmake/externals/libspng/spng.c +++ /dev/null @@ -1,6980 +0,0 @@ -/* SPDX-License-Identifier: (BSD-2-Clause AND libpng-2.0) */ - -#define SPNG__BUILD - -#include "spng.h" - -#include -#include -#include -#include - -#define ZLIB_CONST - -#ifdef __FRAMAC__ - #define SPNG_DISABLE_OPT - #include "tests/framac_stubs.h" -#else - #ifdef SPNG_USE_MINIZ - #include - #else - #include - #endif -#endif - -#ifdef SPNG_MULTITHREADING - #include -#endif - -/* Not build options, edit at your own risk! */ -#define SPNG_READ_SIZE (8192) -#define SPNG_WRITE_SIZE SPNG_READ_SIZE -#define SPNG_MAX_CHUNK_COUNT (1000) - -#define SPNG_TARGET_CLONES(x) - -#ifndef SPNG_DISABLE_OPT - - #if defined(__i386__) || defined(__x86_64__) || defined(_M_IX86) || defined(_M_X64) - #define SPNG_X86 - - #if defined(__x86_64__) || defined(_M_X64) - #define SPNG_X86_64 - #endif - - #elif defined(__aarch64__) || defined(_M_ARM64) /* || defined(__ARM_NEON) */ - #define SPNG_ARM /* NOTE: only arm64 builds are tested! */ - #else - #pragma message "disabling SIMD optimizations for unknown target" - #define SPNG_DISABLE_OPT - #endif - - #if defined(SPNG_X86_64) && defined(SPNG_ENABLE_TARGET_CLONES) - #undef SPNG_TARGET_CLONES - #define SPNG_TARGET_CLONES(x) __attribute__((target_clones(x))) - #else - #define SPNG_TARGET_CLONES(x) - #endif - - #ifndef SPNG_DISABLE_OPT - static void defilter_sub3(size_t rowbytes, unsigned char *row); - static void defilter_sub4(size_t rowbytes, unsigned char *row); - static void defilter_avg3(size_t rowbytes, unsigned char *row, const unsigned char *prev); - static void defilter_avg4(size_t rowbytes, unsigned char *row, const unsigned char *prev); - static void defilter_paeth3(size_t rowbytes, unsigned char *row, const unsigned char *prev); - static void defilter_paeth4(size_t rowbytes, unsigned char *row, const unsigned char *prev); - - #if defined(SPNG_ARM) - static uint32_t expand_palette_rgba8_neon(unsigned char *row, const unsigned char *scanline, const unsigned char *plte, uint32_t width); - static uint32_t expand_palette_rgb8_neon(unsigned char *row, const unsigned char *scanline, const unsigned char *plte, uint32_t width); - #endif - #endif -#endif - -#if defined(_MSC_VER) - #pragma warning(push) - #pragma warning(disable: 4244) -#endif - -#if (defined(__BYTE_ORDER__) && __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__) || defined(__BIG_ENDIAN__) - #define SPNG_BIG_ENDIAN -#else - #define SPNG_LITTLE_ENDIAN -#endif - -enum spng_state -{ - SPNG_STATE_INVALID = 0, - SPNG_STATE_INIT = 1, /* No PNG buffer/stream is set */ - SPNG_STATE_INPUT, /* Decoder input PNG was set */ - SPNG_STATE_OUTPUT = SPNG_STATE_INPUT, /* Encoder output was set */ - SPNG_STATE_IHDR, /* IHDR was read/written */ - SPNG_STATE_FIRST_IDAT, /* Encoded up to / reached first IDAT */ - SPNG_STATE_DECODE_INIT, /* Decoder is ready for progressive reads */ - SPNG_STATE_ENCODE_INIT = SPNG_STATE_DECODE_INIT, - SPNG_STATE_EOI, /* Reached the last scanline/row */ - SPNG_STATE_LAST_IDAT, /* Reached last IDAT, set at end of decode_image() */ - SPNG_STATE_AFTER_IDAT, /* */ - SPNG_STATE_IEND, /* Reached IEND */ -}; - -enum spng__internal -{ - SPNG__IO_SIGNAL = 1 << 9, - SPNG__CTX_FLAGS_ALL = (SPNG_CTX_IGNORE_ADLER32 | SPNG_CTX_ENCODER) -}; - -#define SPNG_STR(x) _SPNG_STR(x) -#define _SPNG_STR(x) #x - -#define SPNG_VERSION_STRING SPNG_STR(SPNG_VERSION_MAJOR) "." \ - SPNG_STR(SPNG_VERSION_MINOR) "." \ - SPNG_STR(SPNG_VERSION_PATCH) - -#define SPNG_GET_CHUNK_BOILERPLATE(chunk) \ - if(ctx == NULL) return 1; \ - int ret = read_chunks(ctx, 0); \ - if(ret) return ret; \ - if(!ctx->stored.chunk) return SPNG_ECHUNKAVAIL; \ - if(chunk == NULL) return 1 - -#define SPNG_SET_CHUNK_BOILERPLATE(chunk) \ - if(ctx == NULL || chunk == NULL) return 1; \ - if(ctx->data == NULL && !ctx->encode_only) return SPNG_ENOSRC; \ - int ret = read_chunks(ctx, 0); \ - if(ret) return ret - -/* Determine if the spng_option can be overriden/optimized */ -#define spng__optimize(option) (ctx->optimize_option & (1 << option)) - -struct spng_subimage -{ - uint32_t width; - uint32_t height; - size_t out_width; /* byte width based on output format */ - size_t scanline_width; -}; - -struct spng_text2 -{ - int type; - char *keyword; - char *text; - - size_t text_length; - - uint8_t compression_flag; /* iTXt only */ - char *language_tag; /* iTXt only */ - char *translated_keyword; /* iTXt only */ - - size_t cache_usage; - char user_keyword_storage[80]; -}; - -struct decode_flags -{ - unsigned apply_trns: 1; - unsigned apply_gamma: 1; - unsigned use_sbit: 1; - unsigned indexed: 1; - unsigned do_scaling: 1; - unsigned interlaced: 1; - unsigned same_layout: 1; - unsigned zerocopy: 1; - unsigned unpack: 1; -}; - -struct encode_flags -{ - unsigned interlace: 1; - unsigned same_layout: 1; - unsigned to_bigendian: 1; - unsigned progressive: 1; - unsigned finalize: 1; - - enum spng_filter_choice filter_choice; -}; - -struct spng_chunk_bitfield -{ - unsigned ihdr: 1; - unsigned plte: 1; - unsigned chrm: 1; - unsigned iccp: 1; - unsigned gama: 1; - unsigned sbit: 1; - unsigned srgb: 1; - unsigned text: 1; - unsigned bkgd: 1; - unsigned hist: 1; - unsigned trns: 1; - unsigned phys: 1; - unsigned splt: 1; - unsigned time: 1; - unsigned offs: 1; - unsigned exif: 1; - unsigned unknown: 1; -}; - -/* Packed sample iterator */ -struct spng__iter -{ - const uint8_t mask; - unsigned shift_amount; - const unsigned initial_shift, bit_depth; - const unsigned char *samples; -}; - -union spng__decode_plte -{ - struct spng_plte_entry rgba[256]; - unsigned char rgb[256 * 3]; - unsigned char raw[256 * 4]; - uint32_t align_this; -}; - -struct spng__zlib_options -{ - int compression_level; - int window_bits; - int mem_level; - int strategy; - int data_type; -}; - -typedef void spng__undo(spng_ctx *ctx); - -struct spng_ctx -{ - size_t data_size; - size_t bytes_read; - size_t stream_buf_size; - unsigned char *stream_buf; - const unsigned char *data; - - /* User-defined pointers for streaming */ - spng_read_fn *read_fn; - spng_write_fn *write_fn; - void *stream_user_ptr; - - /* Used for buffer reads */ - const unsigned char *png_base; - size_t bytes_left; - size_t last_read_size; - - /* Used for encoding */ - int user_owns_out_png; - unsigned char *out_png; - unsigned char *write_ptr; - size_t out_png_size; - size_t bytes_encoded; - - /* These are updated by read/write_header()/read_chunk_bytes() */ - struct spng_chunk current_chunk; - uint32_t cur_chunk_bytes_left; - uint32_t cur_actual_crc; - - struct spng_alloc alloc; - - enum spng_ctx_flags flags; - enum spng_format fmt; - - enum spng_state state; - - unsigned streaming: 1; - unsigned internal_buffer: 1; /* encoding to internal buffer */ - - unsigned inflate: 1; - unsigned deflate: 1; - unsigned encode_only: 1; - unsigned strict: 1; - unsigned discard: 1; - unsigned skip_crc: 1; - unsigned keep_unknown: 1; - unsigned prev_was_idat: 1; - - struct spng__zlib_options image_options; - struct spng__zlib_options text_options; - - spng__undo *undo; - - /* input file contains this chunk */ - struct spng_chunk_bitfield file; - - /* chunk was stored with spng_set_*() */ - struct spng_chunk_bitfield user; - - /* chunk was stored by reading or with spng_set_*() */ - struct spng_chunk_bitfield stored; - - /* used to reset the above in case of an error */ - struct spng_chunk_bitfield prev_stored; - - struct spng_chunk first_idat, last_idat; - - uint32_t max_width, max_height; - - size_t max_chunk_size; - size_t chunk_cache_limit; - size_t chunk_cache_usage; - uint32_t chunk_count_limit; - uint32_t chunk_count_total; - - int crc_action_critical; - int crc_action_ancillary; - - uint32_t optimize_option; - - struct spng_ihdr ihdr; - - struct spng_plte plte; - - struct spng_chrm_int chrm_int; - struct spng_iccp iccp; - - uint32_t gama; - - struct spng_sbit sbit; - - uint8_t srgb_rendering_intent; - - uint32_t n_text; - struct spng_text2 *text_list; - - struct spng_bkgd bkgd; - struct spng_hist hist; - struct spng_trns trns; - struct spng_phys phys; - - uint32_t n_splt; - struct spng_splt *splt_list; - - struct spng_time time; - struct spng_offs offs; - struct spng_exif exif; - - uint32_t n_chunks; - struct spng_unknown_chunk *chunk_list; - - struct spng_subimage subimage[7]; - - z_stream zstream; - unsigned char *scanline_buf, *prev_scanline_buf, *row_buf, *filtered_scanline_buf; - unsigned char *scanline, *prev_scanline, *row, *filtered_scanline; - - /* based on fmt */ - size_t image_size; /* may be zero */ - size_t image_width; - - unsigned bytes_per_pixel; /* derived from ihdr */ - unsigned pixel_size; /* derived from spng_format+ihdr */ - int widest_pass; - int last_pass; /* last non-empty pass */ - - uint16_t *gamma_lut; /* points to either _lut8 or _lut16 */ - uint16_t *gamma_lut16; - uint16_t gamma_lut8[256]; - unsigned char trns_px[8]; - union spng__decode_plte decode_plte; - struct spng_sbit decode_sb; - struct decode_flags decode_flags; - struct spng_row_info row_info; - - struct encode_flags encode_flags; -}; - -static const uint32_t spng_u32max = INT32_MAX; - -static const uint32_t adam7_x_start[7] = { 0, 4, 0, 2, 0, 1, 0 }; -static const uint32_t adam7_y_start[7] = { 0, 0, 4, 0, 2, 0, 1 }; -static const uint32_t adam7_x_delta[7] = { 8, 8, 4, 4, 2, 2, 1 }; -static const uint32_t adam7_y_delta[7] = { 8, 8, 8, 4, 4, 2, 2 }; - -static const uint8_t spng_signature[8] = { 137, 80, 78, 71, 13, 10, 26, 10 }; - -static const uint8_t type_ihdr[4] = { 73, 72, 68, 82 }; -static const uint8_t type_plte[4] = { 80, 76, 84, 69 }; -static const uint8_t type_idat[4] = { 73, 68, 65, 84 }; -static const uint8_t type_iend[4] = { 73, 69, 78, 68 }; - -static const uint8_t type_trns[4] = { 116, 82, 78, 83 }; -static const uint8_t type_chrm[4] = { 99, 72, 82, 77 }; -static const uint8_t type_gama[4] = { 103, 65, 77, 65 }; -static const uint8_t type_iccp[4] = { 105, 67, 67, 80 }; -static const uint8_t type_sbit[4] = { 115, 66, 73, 84 }; -static const uint8_t type_srgb[4] = { 115, 82, 71, 66 }; -static const uint8_t type_text[4] = { 116, 69, 88, 116 }; -static const uint8_t type_ztxt[4] = { 122, 84, 88, 116 }; -static const uint8_t type_itxt[4] = { 105, 84, 88, 116 }; -static const uint8_t type_bkgd[4] = { 98, 75, 71, 68 }; -static const uint8_t type_hist[4] = { 104, 73, 83, 84 }; -static const uint8_t type_phys[4] = { 112, 72, 89, 115 }; -static const uint8_t type_splt[4] = { 115, 80, 76, 84 }; -static const uint8_t type_time[4] = { 116, 73, 77, 69 }; - -static const uint8_t type_offs[4] = { 111, 70, 70, 115 }; -static const uint8_t type_exif[4] = { 101, 88, 73, 102 }; - -static inline void *spng__malloc(spng_ctx *ctx, size_t size) -{ - return ctx->alloc.malloc_fn(size); -} - -static inline void *spng__calloc(spng_ctx *ctx, size_t nmemb, size_t size) -{ - return ctx->alloc.calloc_fn(nmemb, size); -} - -static inline void *spng__realloc(spng_ctx *ctx, void *ptr, size_t size) -{ - return ctx->alloc.realloc_fn(ptr, size); -} - -static inline void spng__free(spng_ctx *ctx, void *ptr) -{ - ctx->alloc.free_fn(ptr); -} - -#if defined(SPNG_USE_MINIZ) -static void *spng__zalloc(void *opaque, size_t items, size_t size) -#else -static void *spng__zalloc(void *opaque, uInt items, uInt size) -#endif -{ - spng_ctx *ctx = opaque; - - if(size > SIZE_MAX / items) return NULL; - - size_t len = (size_t)items * size; - - return spng__malloc(ctx, len); -} - -static void spng__zfree(void *opqaue, void *ptr) -{ - spng_ctx *ctx = opqaue; - spng__free(ctx, ptr); -} - -static inline uint16_t read_u16(const void *src) -{ - const unsigned char *data = src; - - return (data[0] & 0xFFU) << 8 | (data[1] & 0xFFU); -} - -static inline uint32_t read_u32(const void *src) -{ - const unsigned char *data = src; - - return (data[0] & 0xFFUL) << 24 | (data[1] & 0xFFUL) << 16 | - (data[2] & 0xFFUL) << 8 | (data[3] & 0xFFUL); -} - -static inline int32_t read_s32(const void *src) -{ - int32_t ret = (int32_t)read_u32(src); - - return ret; -} - -static inline void write_u16(void *dest, uint16_t x) -{ - unsigned char *data = dest; - - data[0] = x >> 8; - data[1] = x & 0xFF; -} - -static inline void write_u32(void *dest, uint32_t x) -{ - unsigned char *data = dest; - - data[0] = (x >> 24); - data[1] = (x >> 16) & 0xFF; - data[2] = (x >> 8) & 0xFF; - data[3] = x & 0xFF; -} - -static inline void write_s32(void *dest, int32_t x) -{ - uint32_t n = x; - write_u32(dest, n); -} - -/* Returns an iterator for 1,2,4,8-bit samples */ -static struct spng__iter spng__iter_init(unsigned bit_depth, const unsigned char *samples) -{ - struct spng__iter iter = - { - .mask = (uint32_t)(1 << bit_depth) - 1, - .shift_amount = 8 - bit_depth, - .initial_shift = 8 - bit_depth, - .bit_depth = bit_depth, - .samples = samples - }; - - return iter; -} - -/* Returns the current sample unpacked, iterates to the next one */ -static inline uint8_t get_sample(struct spng__iter *iter) -{ - uint8_t x = (iter->samples[0] >> iter->shift_amount) & iter->mask; - - iter->shift_amount -= iter->bit_depth; - - if(iter->shift_amount > 7) - { - iter->shift_amount = iter->initial_shift; - iter->samples++; - } - - return x; -} - -static void u16_row_to_host(void *row, size_t size) -{ - uint16_t *px = row; - size_t i, n = size / 2; - - for(i=0; i < n; i++) - { - px[i] = read_u16(&px[i]); - } -} - -static void u16_row_to_bigendian(void *row, size_t size) -{ - uint16_t *px = (uint16_t*)row; - size_t i, n = size / 2; - - for(i=0; i < n; i++) - { - write_u16(&px[i], px[i]); - } -} - -static void rgb8_row_to_rgba8(const unsigned char *row, unsigned char *out, uint32_t n) -{ - uint32_t i; - for(i=0; i < n; i++) - { - memcpy(out + i * 4, row + i * 3, 3); - out[i*4+3] = 255; - } -} - -static unsigned num_channels(const struct spng_ihdr *ihdr) -{ - switch(ihdr->color_type) - { - case SPNG_COLOR_TYPE_TRUECOLOR: return 3; - case SPNG_COLOR_TYPE_GRAYSCALE_ALPHA: return 2; - case SPNG_COLOR_TYPE_TRUECOLOR_ALPHA: return 4; - case SPNG_COLOR_TYPE_GRAYSCALE: - case SPNG_COLOR_TYPE_INDEXED: - return 1; - default: return 0; - } -} - -/* Calculate scanline width in bits, round up to the nearest byte */ -static int calculate_scanline_width(const struct spng_ihdr *ihdr, uint32_t width, size_t *scanline_width) -{ - if(ihdr == NULL || !width) return SPNG_EINTERNAL; - - size_t res = num_channels(ihdr) * ihdr->bit_depth; - - if(res > SIZE_MAX / width) return SPNG_EOVERFLOW; - res = res * width; - - res += 15; /* Filter byte + 7 for rounding */ - - if(res < 15) return SPNG_EOVERFLOW; - - res /= 8; - - if(res > UINT32_MAX) return SPNG_EOVERFLOW; - - *scanline_width = res; - - return 0; -} - -static int calculate_subimages(struct spng_ctx *ctx) -{ - if(ctx == NULL) return SPNG_EINTERNAL; - - struct spng_ihdr *ihdr = &ctx->ihdr; - struct spng_subimage *sub = ctx->subimage; - - if(ihdr->interlace_method == 1) - { - sub[0].width = (ihdr->width + 7) >> 3; - sub[0].height = (ihdr->height + 7) >> 3; - sub[1].width = (ihdr->width + 3) >> 3; - sub[1].height = (ihdr->height + 7) >> 3; - sub[2].width = (ihdr->width + 3) >> 2; - sub[2].height = (ihdr->height + 3) >> 3; - sub[3].width = (ihdr->width + 1) >> 2; - sub[3].height = (ihdr->height + 3) >> 2; - sub[4].width = (ihdr->width + 1) >> 1; - sub[4].height = (ihdr->height + 1) >> 2; - sub[5].width = ihdr->width >> 1; - sub[5].height = (ihdr->height + 1) >> 1; - sub[6].width = ihdr->width; - sub[6].height = ihdr->height >> 1; - } - else - { - sub[0].width = ihdr->width; - sub[0].height = ihdr->height; - } - - int i; - for(i=0; i < 7; i++) - { - if(sub[i].width == 0 || sub[i].height == 0) continue; - - int ret = calculate_scanline_width(ihdr, sub[i].width, &sub[i].scanline_width); - if(ret) return ret; - - if(sub[ctx->widest_pass].scanline_width < sub[i].scanline_width) ctx->widest_pass = i; - - ctx->last_pass = i; - } - - return 0; -} - -static int check_decode_fmt(const struct spng_ihdr *ihdr, const int fmt) -{ - switch(fmt) - { - case SPNG_FMT_RGBA8: - case SPNG_FMT_RGBA16: - case SPNG_FMT_RGB8: - case SPNG_FMT_PNG: - case SPNG_FMT_RAW: - return 0; - case SPNG_FMT_G8: - case SPNG_FMT_GA8: - if(ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE && ihdr->bit_depth <= 8) return 0; - else return SPNG_EFMT; - case SPNG_FMT_GA16: - if(ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE && ihdr->bit_depth == 16) return 0; - else return SPNG_EFMT; - default: return SPNG_EFMT; - } -} - -static int calculate_image_width(const struct spng_ihdr *ihdr, int fmt, size_t *len) -{ - if(ihdr == NULL || len == NULL) return SPNG_EINTERNAL; - - size_t res = ihdr->width; - unsigned bytes_per_pixel; - - switch(fmt) - { - case SPNG_FMT_RGBA8: - case SPNG_FMT_GA16: - bytes_per_pixel = 4; - break; - case SPNG_FMT_RGBA16: - bytes_per_pixel = 8; - break; - case SPNG_FMT_RGB8: - bytes_per_pixel = 3; - break; - case SPNG_FMT_PNG: - case SPNG_FMT_RAW: - { - int ret = calculate_scanline_width(ihdr, ihdr->width, &res); - if(ret) return ret; - - res -= 1; /* exclude filter byte */ - bytes_per_pixel = 1; - break; - } - case SPNG_FMT_G8: - bytes_per_pixel = 1; - break; - case SPNG_FMT_GA8: - bytes_per_pixel = 2; - break; - default: return SPNG_EINTERNAL; - } - - if(res > SIZE_MAX / bytes_per_pixel) return SPNG_EOVERFLOW; - res = res * bytes_per_pixel; - - *len = res; - - return 0; -} - -static int calculate_image_size(const struct spng_ihdr *ihdr, int fmt, size_t *len) -{ - if(ihdr == NULL || len == NULL) return SPNG_EINTERNAL; - - size_t res = 0; - - int ret = calculate_image_width(ihdr, fmt, &res); - if(ret) return ret; - - if(res > SIZE_MAX / ihdr->height) return SPNG_EOVERFLOW; - res = res * ihdr->height; - - *len = res; - - return 0; -} - -static int increase_cache_usage(spng_ctx *ctx, size_t bytes, int new_chunk) -{ - if(ctx == NULL || !bytes) return SPNG_EINTERNAL; - - if(new_chunk) - { - ctx->chunk_count_total++; - if(ctx->chunk_count_total < 1) return SPNG_EOVERFLOW; - - if(ctx->chunk_count_total > ctx->chunk_count_limit) return SPNG_ECHUNK_LIMITS; - } - - size_t new_usage = ctx->chunk_cache_usage + bytes; - - if(new_usage < ctx->chunk_cache_usage) return SPNG_EOVERFLOW; - - if(new_usage > ctx->chunk_cache_limit) return SPNG_ECHUNK_LIMITS; - - ctx->chunk_cache_usage = new_usage; - - return 0; -} - -static int decrease_cache_usage(spng_ctx *ctx, size_t usage) -{ - if(ctx == NULL || !usage) return SPNG_EINTERNAL; - if(usage > ctx->chunk_cache_usage) return SPNG_EINTERNAL; - - ctx->chunk_cache_usage -= usage; - - return 0; -} - -static int is_critical_chunk(struct spng_chunk *chunk) -{ - if(chunk == NULL) return 0; - if((chunk->type[0] & (1 << 5)) == 0) return 1; - - return 0; -} - -static int decode_err(spng_ctx *ctx, int err) -{ - ctx->state = SPNG_STATE_INVALID; - - return err; -} - -static int encode_err(spng_ctx *ctx, int err) -{ - ctx->state = SPNG_STATE_INVALID; - - return err; -} - -static inline int read_data(spng_ctx *ctx, size_t bytes) -{ - if(ctx == NULL) return SPNG_EINTERNAL; - if(!bytes) return 0; - - if(ctx->streaming && (bytes > SPNG_READ_SIZE)) return SPNG_EINTERNAL; - - int ret = ctx->read_fn(ctx, ctx->stream_user_ptr, ctx->stream_buf, bytes); - - if(ret) - { - if(ret > 0 || ret < SPNG_IO_ERROR) ret = SPNG_IO_ERROR; - - return ret; - } - - ctx->bytes_read += bytes; - if(ctx->bytes_read < bytes) return SPNG_EOVERFLOW; - - return 0; -} - -/* Ensure there is enough space for encoding starting at ctx->write_ptr */ -static int require_bytes(spng_ctx *ctx, size_t bytes) -{ - if(ctx == NULL) return SPNG_EINTERNAL; - - if(ctx->streaming) - { - if(bytes > ctx->stream_buf_size) - { - size_t new_size = ctx->stream_buf_size; - - /* Start at default IDAT size + header + crc */ - if(new_size < (SPNG_WRITE_SIZE + 12)) new_size = SPNG_WRITE_SIZE + 12; - - if(new_size < bytes) new_size = bytes; - - void *temp = spng__realloc(ctx, ctx->stream_buf, new_size); - - if(temp == NULL) return encode_err(ctx, SPNG_EMEM); - - ctx->stream_buf = temp; - ctx->stream_buf_size = bytes; - ctx->write_ptr = ctx->stream_buf; - } - - return 0; - } - - if(!ctx->internal_buffer) return SPNG_ENODST; - - size_t required = ctx->bytes_encoded + bytes; - if(required < bytes) return SPNG_EOVERFLOW; - - if(required > ctx->out_png_size) - { - size_t new_size = ctx->out_png_size; - - /* Start with a size that doesn't require a realloc() 100% of the time */ - if(new_size < (SPNG_WRITE_SIZE * 2)) new_size = SPNG_WRITE_SIZE * 2; - - /* Prefer the next power of two over the requested size */ - while(new_size < required) - { - if(new_size / SIZE_MAX > 2) return encode_err(ctx, SPNG_EOVERFLOW); - - new_size *= 2; - } - - void *temp = spng__realloc(ctx, ctx->out_png, new_size); - - if(temp == NULL) return encode_err(ctx, SPNG_EMEM); - - ctx->out_png = temp; - ctx->out_png_size = new_size; - ctx->write_ptr = ctx->out_png + ctx->bytes_encoded; - } - - return 0; -} - -static int write_data(spng_ctx *ctx, const void *data, size_t bytes) -{ - if(ctx == NULL) return SPNG_EINTERNAL; - if(!bytes) return 0; - - if(ctx->streaming) - { - if(bytes > SPNG_WRITE_SIZE) return SPNG_EINTERNAL; - - int ret = ctx->write_fn(ctx, ctx->stream_user_ptr, (void*)data, bytes); - - if(ret) - { - if(ret > 0 || ret < SPNG_IO_ERROR) ret = SPNG_IO_ERROR; - - return encode_err(ctx, ret); - } - } - else - { - int ret = require_bytes(ctx, bytes); - if(ret) return encode_err(ctx, ret); - - memcpy(ctx->write_ptr, data, bytes); - - ctx->write_ptr += bytes; - } - - ctx->bytes_encoded += bytes; - if(ctx->bytes_encoded < bytes) return SPNG_EOVERFLOW; - - return 0; -} - -static int write_header(spng_ctx *ctx, const uint8_t chunk_type[4], size_t chunk_length, unsigned char **data) -{ - if(ctx == NULL || chunk_type == NULL) return SPNG_EINTERNAL; - if(chunk_length > spng_u32max) return SPNG_EINTERNAL; - - size_t total = chunk_length + 12; - - int ret = require_bytes(ctx, total); - if(ret) return ret; - - uint32_t crc = crc32(0, NULL, 0); - ctx->current_chunk.crc = crc32(crc, chunk_type, 4); - - memcpy(&ctx->current_chunk.type, chunk_type, 4); - ctx->current_chunk.length = (uint32_t)chunk_length; - - if(!data) return SPNG_EINTERNAL; - - if(ctx->streaming) *data = ctx->stream_buf + 8; - else *data = ctx->write_ptr + 8; - - return 0; -} - -static int trim_chunk(spng_ctx *ctx, uint32_t length) -{ - if(length > spng_u32max) return SPNG_EINTERNAL; - if(length > ctx->current_chunk.length) return SPNG_EINTERNAL; - - ctx->current_chunk.length = length; - - return 0; -} - -static int finish_chunk(spng_ctx *ctx) -{ - if(ctx == NULL) return SPNG_EINTERNAL; - - struct spng_chunk *chunk = &ctx->current_chunk; - - unsigned char *header; - unsigned char *chunk_data; - - if(ctx->streaming) - { - chunk_data = ctx->stream_buf + 8; - header = ctx->stream_buf; - } - else - { - chunk_data = ctx->write_ptr + 8; - header = ctx->write_ptr; - } - - write_u32(header, chunk->length); - memcpy(header + 4, chunk->type, 4); - - chunk->crc = crc32(chunk->crc, chunk_data, chunk->length); - - write_u32(chunk_data + chunk->length, chunk->crc); - - if(ctx->streaming) - { - const unsigned char *ptr = ctx->stream_buf; - uint32_t bytes_left = chunk->length + 12; - uint32_t len = 0; - - while(bytes_left) - { - ptr += len; - len = SPNG_WRITE_SIZE; - - if(len > bytes_left) len = bytes_left; - - int ret = write_data(ctx, ptr, len); - if(ret) return ret; - - bytes_left -= len; - } - } - else - { - ctx->bytes_encoded += chunk->length; - if(ctx->bytes_encoded < chunk->length) return SPNG_EOVERFLOW; - - ctx->bytes_encoded += 12; - if(ctx->bytes_encoded < 12) return SPNG_EOVERFLOW; - - ctx->write_ptr += chunk->length + 12; - } - - return 0; -} - -static int write_chunk(spng_ctx *ctx, const uint8_t type[4], const void *data, size_t length) -{ - if(ctx == NULL || type == NULL) return SPNG_EINTERNAL; - if(length && data == NULL) return SPNG_EINTERNAL; - - unsigned char *write_ptr; - - int ret = write_header(ctx, type, length, &write_ptr); - if(ret) return ret; - - if(length) memcpy(write_ptr, data, length); - - return finish_chunk(ctx); -} - -static int write_iend(spng_ctx *ctx) -{ - unsigned char iend_chunk[12] = { 0, 0, 0, 0, 73, 69, 78, 68, 174, 66, 96, 130 }; - return write_data(ctx, iend_chunk, 12); -} - -static int write_unknown_chunks(spng_ctx *ctx, enum spng_location location) -{ - if(!ctx->stored.unknown) return 0; - - const struct spng_unknown_chunk *chunk = ctx->chunk_list; - - uint32_t i; - for(i=0; i < ctx->n_chunks; i++, chunk++) - { - if(chunk->location != location) continue; - - int ret = write_chunk(ctx, chunk->type, chunk->data, chunk->length); - if(ret) return ret; - } - - return 0; -} - -/* Read and check the current chunk's crc, - returns -SPNG_CRC_DISCARD if the chunk should be discarded */ -static inline int read_and_check_crc(spng_ctx *ctx) -{ - if(ctx == NULL) return SPNG_EINTERNAL; - - int ret; - ret = read_data(ctx, 4); - if(ret) return ret; - - ctx->current_chunk.crc = read_u32(ctx->data); - - if(ctx->skip_crc) return 0; - - if(ctx->cur_actual_crc != ctx->current_chunk.crc) - { - if(is_critical_chunk(&ctx->current_chunk)) - { - if(ctx->crc_action_critical == SPNG_CRC_USE) return 0; - } - else - { - if(ctx->crc_action_ancillary == SPNG_CRC_USE) return 0; - if(ctx->crc_action_ancillary == SPNG_CRC_DISCARD) return -SPNG_CRC_DISCARD; - } - - return SPNG_ECHUNK_CRC; - } - - return 0; -} - -/* Read and validate the current chunk's crc and the next chunk header */ -static inline int read_header(spng_ctx *ctx) -{ - if(ctx == NULL) return SPNG_EINTERNAL; - - int ret; - struct spng_chunk chunk = { 0 }; - - ret = read_and_check_crc(ctx); - if(ret) - { - if(ret == -SPNG_CRC_DISCARD) - { - ctx->discard = 1; - } - else return ret; - } - - ret = read_data(ctx, 8); - if(ret) return ret; - - chunk.offset = ctx->bytes_read - 8; - - chunk.length = read_u32(ctx->data); - - memcpy(&chunk.type, ctx->data + 4, 4); - - if(chunk.length > spng_u32max) return SPNG_ECHUNK_STDLEN; - - ctx->cur_chunk_bytes_left = chunk.length; - - if(is_critical_chunk(&chunk) && ctx->crc_action_critical == SPNG_CRC_USE) ctx->skip_crc = 1; - else if(ctx->crc_action_ancillary == SPNG_CRC_USE) ctx->skip_crc = 1; - else ctx->skip_crc = 0; - - if(!ctx->skip_crc) - { - ctx->cur_actual_crc = crc32(0, NULL, 0); - ctx->cur_actual_crc = crc32(ctx->cur_actual_crc, chunk.type, 4); - } - - ctx->current_chunk = chunk; - - return 0; -} - -/* Read chunk bytes and update crc */ -static int read_chunk_bytes(spng_ctx *ctx, uint32_t bytes) -{ - if(ctx == NULL) return SPNG_EINTERNAL; - if(!ctx->cur_chunk_bytes_left || !bytes) return SPNG_EINTERNAL; - if(bytes > ctx->cur_chunk_bytes_left) return SPNG_EINTERNAL; /* XXX: more specific error? */ - - int ret; - - ret = read_data(ctx, bytes); - if(ret) return ret; - - if(!ctx->skip_crc) ctx->cur_actual_crc = crc32(ctx->cur_actual_crc, ctx->data, bytes); - - ctx->cur_chunk_bytes_left -= bytes; - - return ret; -} - -/* read_chunk_bytes() + read_data() with custom output buffer */ -static int read_chunk_bytes2(spng_ctx *ctx, void *out, uint32_t bytes) -{ - if(ctx == NULL) return SPNG_EINTERNAL; - if(!ctx->cur_chunk_bytes_left || !bytes) return SPNG_EINTERNAL; - if(bytes > ctx->cur_chunk_bytes_left) return SPNG_EINTERNAL; /* XXX: more specific error? */ - - int ret; - uint32_t len = bytes; - - if(ctx->streaming && len > SPNG_READ_SIZE) len = SPNG_READ_SIZE; - - while(bytes) - { - if(len > bytes) len = bytes; - - ret = ctx->read_fn(ctx, ctx->stream_user_ptr, out, len); - if(ret) return ret; - - if(!ctx->streaming) memcpy(out, ctx->data, len); - - ctx->bytes_read += len; - if(ctx->bytes_read < len) return SPNG_EOVERFLOW; - - if(!ctx->skip_crc) ctx->cur_actual_crc = crc32(ctx->cur_actual_crc, out, len); - - ctx->cur_chunk_bytes_left -= len; - - out = (char*)out + len; - bytes -= len; - len = SPNG_READ_SIZE; - } - - return 0; -} - -static int discard_chunk_bytes(spng_ctx *ctx, uint32_t bytes) -{ - if(ctx == NULL) return SPNG_EINTERNAL; - if(!bytes) return 0; - - int ret; - - if(ctx->streaming) /* Do small, consecutive reads */ - { - while(bytes) - { - uint32_t len = SPNG_READ_SIZE; - - if(len > bytes) len = bytes; - - ret = read_chunk_bytes(ctx, len); - if(ret) return ret; - - bytes -= len; - } - } - else - { - ret = read_chunk_bytes(ctx, bytes); - if(ret) return ret; - } - - return 0; -} - -static int spng__inflate_init(spng_ctx *ctx, int window_bits) -{ - if(ctx->zstream.state) inflateEnd(&ctx->zstream); - - ctx->inflate = 1; - - ctx->zstream.zalloc = spng__zalloc; - ctx->zstream.zfree = spng__zfree; - ctx->zstream.opaque = ctx; - - if(inflateInit2(&ctx->zstream, window_bits) != Z_OK) return SPNG_EZLIB_INIT; - -#if ZLIB_VERNUM >= 0x1290 && !defined(SPNG_USE_MINIZ) - - int validate = 1; - - if(ctx->flags & SPNG_CTX_IGNORE_ADLER32) validate = 0; - - if(is_critical_chunk(&ctx->current_chunk)) - { - if(ctx->crc_action_critical == SPNG_CRC_USE) validate = 0; - } - else /* ancillary */ - { - if(ctx->crc_action_ancillary == SPNG_CRC_USE) validate = 0; - } - - if(inflateValidate(&ctx->zstream, validate)) return SPNG_EZLIB_INIT; - -#else /* This requires zlib >= 1.2.11 */ - #pragma message ("inflateValidate() not available, SPNG_CTX_IGNORE_ADLER32 will be ignored") -#endif - - return 0; -} - -static int spng__deflate_init(spng_ctx *ctx, struct spng__zlib_options *options) -{ - if(ctx->zstream.state) deflateEnd(&ctx->zstream); - - ctx->deflate = 1; - - z_stream *zstream = &ctx->zstream; - zstream->zalloc = spng__zalloc; - zstream->zfree = spng__zfree; - zstream->opaque = ctx; - zstream->data_type = options->data_type; - - int ret = deflateInit2(zstream, options->compression_level, Z_DEFLATED, options->window_bits, options->mem_level, options->strategy); - - if(ret != Z_OK) return SPNG_EZLIB_INIT; - - return 0; -} - -/* Inflate a zlib stream starting with start_buf if non-NULL, - continuing from the datastream till an end marker, - allocating and writing the inflated stream to *out, - leaving "extra" bytes at the end, final buffer length is *len. - - Takes into account the chunk size and cache limits. -*/ -static int spng__inflate_stream(spng_ctx *ctx, char **out, size_t *len, size_t extra, const void *start_buf, size_t start_len) -{ - int ret = spng__inflate_init(ctx, 15); - if(ret) return ret; - - size_t max = ctx->chunk_cache_limit - ctx->chunk_cache_usage; - - if(ctx->max_chunk_size < max) max = ctx->max_chunk_size; - - if(extra > max) return SPNG_ECHUNK_LIMITS; - max -= extra; - - uint32_t read_size; - size_t size = 8 * 1024; - void *t, *buf = spng__malloc(ctx, size); - - if(buf == NULL) return SPNG_EMEM; - - z_stream *stream = &ctx->zstream; - - if(start_buf != NULL && start_len) - { - stream->avail_in = (uInt)start_len; - stream->next_in = start_buf; - } - else - { - stream->avail_in = 0; - stream->next_in = NULL; - } - - stream->avail_out = (uInt)size; - stream->next_out = buf; - - while(ret != Z_STREAM_END) - { - ret = inflate(stream, Z_NO_FLUSH); - - if(ret == Z_STREAM_END) break; - - if(ret != Z_OK && ret != Z_BUF_ERROR) - { - ret = SPNG_EZLIB; - goto err; - } - - if(!stream->avail_out) /* Resize buffer */ - { - /* overflow or reached chunk/cache limit */ - if( (2 > SIZE_MAX / size) || (size > max / 2) ) - { - ret = SPNG_ECHUNK_LIMITS; - goto err; - } - - size *= 2; - - t = spng__realloc(ctx, buf, size); - if(t == NULL) goto mem; - - buf = t; - - stream->avail_out = (uInt)size / 2; - stream->next_out = (unsigned char*)buf + size / 2; - } - else if(!stream->avail_in) /* Read more chunk bytes */ - { - read_size = ctx->cur_chunk_bytes_left; - if(ctx->streaming && read_size > SPNG_READ_SIZE) read_size = SPNG_READ_SIZE; - - ret = read_chunk_bytes(ctx, read_size); - - if(ret) - { - if(!read_size) ret = SPNG_EZLIB; - - goto err; - } - - stream->avail_in = read_size; - stream->next_in = ctx->data; - } - } - - size = stream->total_out; - - if(!size) - { - ret = SPNG_EZLIB; - goto err; - } - - size += extra; - if(size < extra) goto mem; - - t = spng__realloc(ctx, buf, size); - if(t == NULL) goto mem; - - buf = t; - - (void)increase_cache_usage(ctx, size, 0); - - *out = buf; - *len = size; - - return 0; - -mem: - ret = SPNG_EMEM; -err: - spng__free(ctx, buf); - return ret; -} - -/* Read at least one byte from the IDAT stream */ -static int read_idat_bytes(spng_ctx *ctx, uint32_t *bytes_read) -{ - if(ctx == NULL || bytes_read == NULL) return SPNG_EINTERNAL; - if(memcmp(ctx->current_chunk.type, type_idat, 4)) return SPNG_EIDAT_TOO_SHORT; - - int ret; - uint32_t len; - - while(!ctx->cur_chunk_bytes_left) - { - ret = read_header(ctx); - if(ret) return ret; - - if(memcmp(ctx->current_chunk.type, type_idat, 4)) return SPNG_EIDAT_TOO_SHORT; - } - - if(ctx->streaming) - {/* TODO: estimate bytes to read for progressive reads */ - len = SPNG_READ_SIZE; - if(len > ctx->cur_chunk_bytes_left) len = ctx->cur_chunk_bytes_left; - } - else len = ctx->current_chunk.length; - - ret = read_chunk_bytes(ctx, len); - - *bytes_read = len; - - return ret; -} - -static int read_scanline_bytes(spng_ctx *ctx, unsigned char *dest, size_t len) -{ - if(ctx == NULL || dest == NULL) return SPNG_EINTERNAL; - - int ret = Z_OK; - uint32_t bytes_read; - - z_stream *zstream = &ctx->zstream; - - zstream->avail_out = (uInt)len; - zstream->next_out = dest; - - while(zstream->avail_out != 0) - { - ret = inflate(zstream, Z_NO_FLUSH); - - if(ret == Z_OK) continue; - - if(ret == Z_STREAM_END) /* Reached an end-marker */ - { - if(zstream->avail_out != 0) return SPNG_EIDAT_TOO_SHORT; - } - else if(ret == Z_BUF_ERROR) /* Read more IDAT bytes */ - { - ret = read_idat_bytes(ctx, &bytes_read); - if(ret) return ret; - - zstream->avail_in = bytes_read; - zstream->next_in = ctx->data; - } - else return SPNG_EIDAT_STREAM; - } - - return 0; -} - -static uint8_t paeth(uint8_t a, uint8_t b, uint8_t c) -{ - int16_t p = a + b - c; - int16_t pa = abs(p - a); - int16_t pb = abs(p - b); - int16_t pc = abs(p - c); - - if(pa <= pb && pa <= pc) return a; - else if(pb <= pc) return b; - - return c; -} - -SPNG_TARGET_CLONES("default,avx2") -static void defilter_up(size_t bytes, unsigned char *row, const unsigned char *prev) -{ - size_t i; - for(i=0; i < bytes; i++) - { - row[i] += prev[i]; - } -} - -/* Defilter *scanline in-place. - *prev_scanline and *scanline should point to the first pixel, - scanline_width is the width of the scanline including the filter byte. -*/ -static int defilter_scanline(const unsigned char *prev_scanline, unsigned char *scanline, - size_t scanline_width, unsigned bytes_per_pixel, unsigned filter) -{ - if(prev_scanline == NULL || scanline == NULL || !scanline_width) return SPNG_EINTERNAL; - - size_t i; - scanline_width--; - - if(filter == 0) return 0; - -#ifndef SPNG_DISABLE_OPT - if(filter == SPNG_FILTER_UP) goto no_opt; - - if(bytes_per_pixel == 4) - { - if(filter == SPNG_FILTER_SUB) - defilter_sub4(scanline_width, scanline); - else if(filter == SPNG_FILTER_AVERAGE) - defilter_avg4(scanline_width, scanline, prev_scanline); - else if(filter == SPNG_FILTER_PAETH) - defilter_paeth4(scanline_width, scanline, prev_scanline); - else return SPNG_EFILTER; - - return 0; - } - else if(bytes_per_pixel == 3) - { - if(filter == SPNG_FILTER_SUB) - defilter_sub3(scanline_width, scanline); - else if(filter == SPNG_FILTER_AVERAGE) - defilter_avg3(scanline_width, scanline, prev_scanline); - else if(filter == SPNG_FILTER_PAETH) - defilter_paeth3(scanline_width, scanline, prev_scanline); - else return SPNG_EFILTER; - - return 0; - } -no_opt: -#endif - - if(filter == SPNG_FILTER_UP) - { - defilter_up(scanline_width, scanline, prev_scanline); - return 0; - } - - for(i=0; i < scanline_width; i++) - { - uint8_t x, a, b, c; - - if(i >= bytes_per_pixel) - { - a = scanline[i - bytes_per_pixel]; - b = prev_scanline[i]; - c = prev_scanline[i - bytes_per_pixel]; - } - else /* First pixel in row */ - { - a = 0; - b = prev_scanline[i]; - c = 0; - } - - x = scanline[i]; - - switch(filter) - { - case SPNG_FILTER_SUB: - { - x = x + a; - break; - } - case SPNG_FILTER_AVERAGE: - { - uint16_t avg = (a + b) / 2; - x = x + avg; - break; - } - case SPNG_FILTER_PAETH: - { - x = x + paeth(a,b,c); - break; - } - } - - scanline[i] = x; - } - - return 0; -} - -static int filter_scanline(unsigned char *filtered, const unsigned char *prev_scanline, const unsigned char *scanline, - size_t scanline_width, unsigned bytes_per_pixel, const unsigned filter) -{ - if(prev_scanline == NULL || scanline == NULL || scanline_width <= 1) return SPNG_EINTERNAL; - - if(filter > 4) return SPNG_EFILTER; - if(filter == 0) return 0; - - scanline_width--; - - uint32_t i; - for(i=0; i < scanline_width; i++) - { - uint8_t x, a, b, c; - - if(i >= bytes_per_pixel) - { - a = scanline[i - bytes_per_pixel]; - b = prev_scanline[i]; - c = prev_scanline[i - bytes_per_pixel]; - } - else /* first pixel in row */ - { - a = 0; - b = prev_scanline[i]; - c = 0; - } - - x = scanline[i]; - - switch(filter) - { - case SPNG_FILTER_SUB: - { - x = x - a; - break; - } - case SPNG_FILTER_UP: - { - x = x - b; - break; - } - case SPNG_FILTER_AVERAGE: - { - uint16_t avg = (a + b) / 2; - x = x - avg; - break; - } - case SPNG_FILTER_PAETH: - { - x = x - paeth(a,b,c); - break; - } - } - - filtered[i] = x; - } - - return 0; -} - -static int32_t filter_sum(const unsigned char *prev_scanline, const unsigned char *scanline, - size_t size, unsigned bytes_per_pixel, const unsigned filter) -{ - /* prevent potential over/underflow, bails out at a width of ~8M pixels for RGBA8 */ - if(size > (INT32_MAX / 128)) return INT32_MAX; - - uint32_t i; - int32_t sum = 0; - uint8_t x, a, b, c; - - for(i=0; i < size; i++) - { - if(i >= bytes_per_pixel) - { - a = scanline[i - bytes_per_pixel]; - b = prev_scanline[i]; - c = prev_scanline[i - bytes_per_pixel]; - } - else /* first pixel in row */ - { - a = 0; - b = prev_scanline[i]; - c = 0; - } - - x = scanline[i]; - - switch(filter) - { - case SPNG_FILTER_NONE: - { - break; - } - case SPNG_FILTER_SUB: - { - x = x - a; - break; - } - case SPNG_FILTER_UP: - { - x = x - b; - break; - } - case SPNG_FILTER_AVERAGE: - { - uint16_t avg = (a + b) / 2; - x = x - avg; - break; - } - case SPNG_FILTER_PAETH: - { - x = x - paeth(a,b,c); - break; - } - } - - sum += 128 - abs((int)x - 128); - } - - return sum; -} - -static unsigned get_best_filter(const unsigned char *prev_scanline, const unsigned char *scanline, - size_t scanline_width, unsigned bytes_per_pixel, const int choices) -{ - if(!choices) return SPNG_FILTER_NONE; - - scanline_width--; - - int i; - unsigned int best_filter = 0; - enum spng_filter_choice flag; - int32_t sum, best_score = INT32_MAX; - int32_t filter_scores[5] = { INT32_MAX, INT32_MAX, INT32_MAX, INT32_MAX, INT32_MAX }; - - if( !(choices & (choices - 1)) ) - {/* only one choice/bit is set */ - for(i=0; i < 5; i++) - { - if(choices == 1 << (i + 3)) return i; - } - } - - for(i=0; i < 5; i++) - { - flag = 1 << (i + 3); - - if(choices & flag) sum = filter_sum(prev_scanline, scanline, scanline_width, bytes_per_pixel, i); - else continue; - - filter_scores[i] = abs(sum); - - if(filter_scores[i] < best_score) - { - best_score = filter_scores[i]; - best_filter = i; - } - } - - return best_filter; -} - -/* Scale "sbits" significant bits in "sample" from "bit_depth" to "target" - - "bit_depth" must be a valid PNG depth - "sbits" must be less than or equal to "bit_depth" - "target" must be between 1 and 16 -*/ -static uint16_t sample_to_target(uint16_t sample, unsigned bit_depth, unsigned sbits, unsigned target) -{ - if(bit_depth == sbits) - { - if(target == sbits) return sample; /* No scaling */ - }/* bit_depth > sbits */ - else sample = sample >> (bit_depth - sbits); /* Shift significant bits to bottom */ - - /* Downscale */ - if(target < sbits) return sample >> (sbits - target); - - /* Upscale using left bit replication */ - int8_t shift_amount = target - sbits; - uint16_t sample_bits = sample; - sample = 0; - - while(shift_amount >= 0) - { - sample = sample | (sample_bits << shift_amount); - shift_amount -= sbits; - } - - int8_t partial = shift_amount + (int8_t)sbits; - - if(partial != 0) sample = sample | (sample_bits >> abs(shift_amount)); - - return sample; -} - -static inline void gamma_correct_row(unsigned char *row, uint32_t pixels, int fmt, const uint16_t *gamma_lut) -{ - uint32_t i; - - if(fmt == SPNG_FMT_RGBA8) - { - unsigned char *px; - for(i=0; i < pixels; i++) - { - px = row + i * 4; - - px[0] = gamma_lut[px[0]]; - px[1] = gamma_lut[px[1]]; - px[2] = gamma_lut[px[2]]; - } - } - else if(fmt == SPNG_FMT_RGBA16) - { - for(i=0; i < pixels; i++) - { - uint16_t px[4]; - memcpy(px, row + i * 8, 8); - - px[0] = gamma_lut[px[0]]; - px[1] = gamma_lut[px[1]]; - px[2] = gamma_lut[px[2]]; - - memcpy(row + i * 8, px, 8); - } - } - else if(fmt == SPNG_FMT_RGB8) - { - unsigned char *px; - for(i=0; i < pixels; i++) - { - px = row + i * 3; - - px[0] = gamma_lut[px[0]]; - px[1] = gamma_lut[px[1]]; - px[2] = gamma_lut[px[2]]; - } - } -} - -/* Apply transparency to output row */ -static inline void trns_row(unsigned char *row, - const unsigned char *scanline, - const unsigned char *trns, - unsigned scanline_stride, - struct spng_ihdr *ihdr, - uint32_t pixels, - int fmt) -{ - uint32_t i; - unsigned row_stride; - unsigned depth = ihdr->bit_depth; - - if(fmt == SPNG_FMT_RGBA8) - { - if(ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE) return; /* already applied in the decoding loop */ - - row_stride = 4; - for(i=0; i < pixels; i++, scanline+=scanline_stride, row+=row_stride) - { - if(!memcmp(scanline, trns, scanline_stride)) row[3] = 0; - } - } - else if(fmt == SPNG_FMT_RGBA16) - { - if(ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE) return; /* already applied in the decoding loop */ - - row_stride = 8; - for(i=0; i < pixels; i++, scanline+=scanline_stride, row+=row_stride) - { - if(!memcmp(scanline, trns, scanline_stride)) memset(row + 6, 0, 2); - } - } - else if(fmt == SPNG_FMT_GA8) - { - row_stride = 2; - - if(depth == 16) - { - for(i=0; i < pixels; i++, scanline+=scanline_stride, row+=row_stride) - { - if(!memcmp(scanline, trns, scanline_stride)) memset(row + 1, 0, 1); - } - } - else /* depth <= 8 */ - { - struct spng__iter iter = spng__iter_init(depth, scanline); - - for(i=0; i < pixels; i++, row+=row_stride) - { - if(trns[0] == get_sample(&iter)) row[1] = 0; - } - } - } - else if(fmt == SPNG_FMT_GA16) - { - row_stride = 4; - - if(depth == 16) - { - for(i=0; i< pixels; i++, scanline+=scanline_stride, row+=row_stride) - { - if(!memcmp(scanline, trns, 2)) memset(row + 2, 0, 2); - } - } - else - { - struct spng__iter iter = spng__iter_init(depth, scanline); - - for(i=0; i< pixels; i++, row+=row_stride) - { - if(trns[0] == get_sample(&iter)) memset(row + 2, 0, 2); - } - } - } - else return; -} - -static inline void scale_row(unsigned char *row, uint32_t pixels, int fmt, unsigned depth, const struct spng_sbit *sbit) -{ - uint32_t i; - - if(fmt == SPNG_FMT_RGBA8) - { - unsigned char px[4]; - for(i=0; i < pixels; i++) - { - memcpy(px, row + i * 4, 4); - - px[0] = sample_to_target(px[0], depth, sbit->red_bits, 8); - px[1] = sample_to_target(px[1], depth, sbit->green_bits, 8); - px[2] = sample_to_target(px[2], depth, sbit->blue_bits, 8); - px[3] = sample_to_target(px[3], depth, sbit->alpha_bits, 8); - - memcpy(row + i * 4, px, 4); - } - } - else if(fmt == SPNG_FMT_RGBA16) - { - uint16_t px[4]; - for(i=0; i < pixels; i++) - { - memcpy(px, row + i * 8, 8); - - px[0] = sample_to_target(px[0], depth, sbit->red_bits, 16); - px[1] = sample_to_target(px[1], depth, sbit->green_bits, 16); - px[2] = sample_to_target(px[2], depth, sbit->blue_bits, 16); - px[3] = sample_to_target(px[3], depth, sbit->alpha_bits, 16); - - memcpy(row + i * 8, px, 8); - } - } - else if(fmt == SPNG_FMT_RGB8) - { - unsigned char px[4]; - for(i=0; i < pixels; i++) - { - memcpy(px, row + i * 3, 3); - - px[0] = sample_to_target(px[0], depth, sbit->red_bits, 8); - px[1] = sample_to_target(px[1], depth, sbit->green_bits, 8); - px[2] = sample_to_target(px[2], depth, sbit->blue_bits, 8); - - memcpy(row + i * 3, px, 3); - } - } - else if(fmt == SPNG_FMT_G8) - { - for(i=0; i < pixels; i++) - { - row[i] = sample_to_target(row[i], depth, sbit->grayscale_bits, 8); - } - } - else if(fmt == SPNG_FMT_GA8) - { - for(i=0; i < pixels; i++) - { - row[i*2] = sample_to_target(row[i*2], depth, sbit->grayscale_bits, 8); - } - } -} - -/* Expand to *row using 8-bit palette indices from *scanline */ -static void expand_row(unsigned char *row, - const unsigned char *scanline, - const union spng__decode_plte *decode_plte, - uint32_t width, - int fmt) -{ - uint32_t i = 0; - unsigned char *px; - unsigned char entry; - const struct spng_plte_entry *plte = decode_plte->rgba; - -#if defined(SPNG_ARM) - if(fmt == SPNG_FMT_RGBA8) i = expand_palette_rgba8_neon(row, scanline, decode_plte->raw, width); - else if(fmt == SPNG_FMT_RGB8) - { - i = expand_palette_rgb8_neon(row, scanline, decode_plte->raw, width); - - for(; i < width; i++) - {/* In this case the LUT is 3 bytes packed */ - px = row + i * 3; - entry = scanline[i]; - px[0] = decode_plte->raw[entry * 3 + 0]; - px[1] = decode_plte->raw[entry * 3 + 1]; - px[2] = decode_plte->raw[entry * 3 + 2]; - } - return; - } -#endif - - if(fmt == SPNG_FMT_RGBA8) - { - for(; i < width; i++) - { - px = row + i * 4; - entry = scanline[i]; - px[0] = plte[entry].red; - px[1] = plte[entry].green; - px[2] = plte[entry].blue; - px[3] = plte[entry].alpha; - } - } - else if(fmt == SPNG_FMT_RGB8) - { - for(; i < width; i++) - { - px = row + i * 3; - entry = scanline[i]; - px[0] = plte[entry].red; - px[1] = plte[entry].green; - px[2] = plte[entry].blue; - } - } -} - -/* Unpack 1/2/4/8-bit samples to G8/GA8/GA16 or G16 -> GA16 */ -static void unpack_scanline(unsigned char *out, const unsigned char *scanline, uint32_t width, unsigned bit_depth, int fmt) -{ - struct spng__iter iter = spng__iter_init(bit_depth, scanline); - uint32_t i; - uint16_t sample, alpha = 65535; - - - if(fmt == SPNG_FMT_GA8) goto ga8; - else if(fmt == SPNG_FMT_GA16) goto ga16; - - /* 1/2/4-bit -> 8-bit */ - for(i=0; i < width; i++) out[i] = get_sample(&iter); - - return; - -ga8: - /* 1/2/4/8-bit -> GA8 */ - for(i=0; i < width; i++) - { - out[i*2] = get_sample(&iter); - out[i*2 + 1] = 255; - } - - return; - -ga16: - - /* 16 -> GA16 */ - if(bit_depth == 16) - { - for(i=0; i < width; i++) - { - memcpy(out + i * 4, scanline + i * 2, 2); - memcpy(out + i * 4 + 2, &alpha, 2); - } - return; - } - - /* 1/2/4/8-bit -> GA16 */ - for(i=0; i < width; i++) - { - sample = get_sample(&iter); - memcpy(out + i * 4, &sample, 2); - memcpy(out + i * 4 + 2, &alpha, 2); - } -} - -static int check_ihdr(const struct spng_ihdr *ihdr, uint32_t max_width, uint32_t max_height) -{ - if(ihdr->width > spng_u32max || !ihdr->width) return SPNG_EWIDTH; - if(ihdr->height > spng_u32max || !ihdr->height) return SPNG_EHEIGHT; - - if(ihdr->width > max_width) return SPNG_EUSER_WIDTH; - if(ihdr->height > max_height) return SPNG_EUSER_HEIGHT; - - switch(ihdr->color_type) - { - case SPNG_COLOR_TYPE_GRAYSCALE: - { - if( !(ihdr->bit_depth == 1 || ihdr->bit_depth == 2 || - ihdr->bit_depth == 4 || ihdr->bit_depth == 8 || - ihdr->bit_depth == 16) ) - return SPNG_EBIT_DEPTH; - - break; - } - case SPNG_COLOR_TYPE_TRUECOLOR: - case SPNG_COLOR_TYPE_GRAYSCALE_ALPHA: - case SPNG_COLOR_TYPE_TRUECOLOR_ALPHA: - { - if( !(ihdr->bit_depth == 8 || ihdr->bit_depth == 16) ) - return SPNG_EBIT_DEPTH; - - break; - } - case SPNG_COLOR_TYPE_INDEXED: - { - if( !(ihdr->bit_depth == 1 || ihdr->bit_depth == 2 || - ihdr->bit_depth == 4 || ihdr->bit_depth == 8) ) - return SPNG_EBIT_DEPTH; - - break; - } - default: return SPNG_ECOLOR_TYPE; - } - - if(ihdr->compression_method) return SPNG_ECOMPRESSION_METHOD; - if(ihdr->filter_method) return SPNG_EFILTER_METHOD; - - if(ihdr->interlace_method > 1) return SPNG_EINTERLACE_METHOD; - - return 0; -} - -static int check_plte(const struct spng_plte *plte, const struct spng_ihdr *ihdr) -{ - if(plte == NULL || ihdr == NULL) return 1; - - if(plte->n_entries == 0) return 1; - if(plte->n_entries > 256) return 1; - - if(ihdr->color_type == SPNG_COLOR_TYPE_INDEXED) - { - if(plte->n_entries > (1U << ihdr->bit_depth)) return 1; - } - - return 0; -} - -static int check_sbit(const struct spng_sbit *sbit, const struct spng_ihdr *ihdr) -{ - if(sbit == NULL || ihdr == NULL) return 1; - - if(ihdr->color_type == 0) - { - if(sbit->grayscale_bits == 0) return SPNG_ESBIT; - if(sbit->grayscale_bits > ihdr->bit_depth) return SPNG_ESBIT; - } - else if(ihdr->color_type == 2 || ihdr->color_type == 3) - { - if(sbit->red_bits == 0) return SPNG_ESBIT; - if(sbit->green_bits == 0) return SPNG_ESBIT; - if(sbit->blue_bits == 0) return SPNG_ESBIT; - - uint8_t bit_depth; - if(ihdr->color_type == 3) bit_depth = 8; - else bit_depth = ihdr->bit_depth; - - if(sbit->red_bits > bit_depth) return SPNG_ESBIT; - if(sbit->green_bits > bit_depth) return SPNG_ESBIT; - if(sbit->blue_bits > bit_depth) return SPNG_ESBIT; - } - else if(ihdr->color_type == 4) - { - if(sbit->grayscale_bits == 0) return SPNG_ESBIT; - if(sbit->alpha_bits == 0) return SPNG_ESBIT; - - if(sbit->grayscale_bits > ihdr->bit_depth) return SPNG_ESBIT; - if(sbit->alpha_bits > ihdr->bit_depth) return SPNG_ESBIT; - } - else if(ihdr->color_type == 6) - { - if(sbit->red_bits == 0) return SPNG_ESBIT; - if(sbit->green_bits == 0) return SPNG_ESBIT; - if(sbit->blue_bits == 0) return SPNG_ESBIT; - if(sbit->alpha_bits == 0) return SPNG_ESBIT; - - if(sbit->red_bits > ihdr->bit_depth) return SPNG_ESBIT; - if(sbit->green_bits > ihdr->bit_depth) return SPNG_ESBIT; - if(sbit->blue_bits > ihdr->bit_depth) return SPNG_ESBIT; - if(sbit->alpha_bits > ihdr->bit_depth) return SPNG_ESBIT; - } - - return 0; -} - -static int check_chrm_int(const struct spng_chrm_int *chrm_int) -{ - if(chrm_int == NULL) return 1; - - if(chrm_int->white_point_x > spng_u32max || - chrm_int->white_point_y > spng_u32max || - chrm_int->red_x > spng_u32max || - chrm_int->red_y > spng_u32max || - chrm_int->green_x > spng_u32max || - chrm_int->green_y > spng_u32max || - chrm_int->blue_x > spng_u32max || - chrm_int->blue_y > spng_u32max) return SPNG_ECHRM; - - return 0; -} - -static int check_phys(const struct spng_phys *phys) -{ - if(phys == NULL) return 1; - - if(phys->unit_specifier > 1) return SPNG_EPHYS; - - if(phys->ppu_x > spng_u32max) return SPNG_EPHYS; - if(phys->ppu_y > spng_u32max) return SPNG_EPHYS; - - return 0; -} - -static int check_time(const struct spng_time *time) -{ - if(time == NULL) return 1; - - if(time->month == 0 || time->month > 12) return 1; - if(time->day == 0 || time->day > 31) return 1; - if(time->hour > 23) return 1; - if(time->minute > 59) return 1; - if(time->second > 60) return 1; - - return 0; -} - -static int check_offs(const struct spng_offs *offs) -{ - if(offs == NULL) return 1; - - if(offs->unit_specifier > 1) return 1; - - return 0; -} - -static int check_exif(const struct spng_exif *exif) -{ - if(exif == NULL) return 1; - if(exif->data == NULL) return 1; - - if(exif->length < 4) return SPNG_ECHUNK_SIZE; - if(exif->length > spng_u32max) return SPNG_ECHUNK_STDLEN; - - const uint8_t exif_le[4] = { 73, 73, 42, 0 }; - const uint8_t exif_be[4] = { 77, 77, 0, 42 }; - - if(memcmp(exif->data, exif_le, 4) && memcmp(exif->data, exif_be, 4)) return 1; - - return 0; -} - -/* Validate PNG keyword */ -static int check_png_keyword(const char *str) -{ - if(str == NULL) return 1; - size_t len = strlen(str); - const char *end = str + len; - - if(!len) return 1; - if(len > 79) return 1; - if(str[0] == ' ') return 1; /* Leading space */ - if(end[-1] == ' ') return 1; /* Trailing space */ - if(strstr(str, " ") != NULL) return 1; /* Consecutive spaces */ - - uint8_t c; - while(str != end) - { - memcpy(&c, str, 1); - - if( (c >= 32 && c <= 126) || (c >= 161) ) str++; - else return 1; /* Invalid character */ - } - - return 0; -} - -/* Validate PNG text *str up to 'len' bytes */ -static int check_png_text(const char *str, size_t len) -{/* XXX: are consecutive newlines permitted? */ - if(str == NULL || len == 0) return 1; - - uint8_t c; - size_t i = 0; - while(i < len) - { - memcpy(&c, str + i, 1); - - if( (c >= 32 && c <= 126) || (c >= 161) || c == 10) i++; - else return 1; /* Invalid character */ - } - - return 0; -} - -/* Returns non-zero for standard chunks which are stored without allocating memory */ -static int is_small_chunk(uint8_t type[4]) -{ - if(!memcmp(type, type_plte, 4)) return 1; - else if(!memcmp(type, type_chrm, 4)) return 1; - else if(!memcmp(type, type_gama, 4)) return 1; - else if(!memcmp(type, type_sbit, 4)) return 1; - else if(!memcmp(type, type_srgb, 4)) return 1; - else if(!memcmp(type, type_bkgd, 4)) return 1; - else if(!memcmp(type, type_trns, 4)) return 1; - else if(!memcmp(type, type_hist, 4)) return 1; - else if(!memcmp(type, type_phys, 4)) return 1; - else if(!memcmp(type, type_time, 4)) return 1; - else if(!memcmp(type, type_offs, 4)) return 1; - else return 0; -} - -static int read_ihdr(spng_ctx *ctx) -{ - int ret; - struct spng_chunk *chunk = &ctx->current_chunk; - const unsigned char *data; - - chunk->offset = 8; - chunk->length = 13; - size_t sizeof_sig_ihdr = 29; - - ret = read_data(ctx, sizeof_sig_ihdr); - if(ret) return ret; - - data = ctx->data; - - if(memcmp(data, spng_signature, sizeof(spng_signature))) return SPNG_ESIGNATURE; - - chunk->length = read_u32(data + 8); - memcpy(&chunk->type, data + 12, 4); - - if(chunk->length != 13) return SPNG_EIHDR_SIZE; - if(memcmp(chunk->type, type_ihdr, 4)) return SPNG_ENOIHDR; - - ctx->cur_actual_crc = crc32(0, NULL, 0); - ctx->cur_actual_crc = crc32(ctx->cur_actual_crc, data + 12, 17); - - ctx->ihdr.width = read_u32(data + 16); - ctx->ihdr.height = read_u32(data + 20); - ctx->ihdr.bit_depth = data[24]; - ctx->ihdr.color_type = data[25]; - ctx->ihdr.compression_method = data[26]; - ctx->ihdr.filter_method = data[27]; - ctx->ihdr.interlace_method = data[28]; - - ret = check_ihdr(&ctx->ihdr, ctx->max_width, ctx->max_height); - if(ret) return ret; - - ctx->file.ihdr = 1; - ctx->stored.ihdr = 1; - - if(ctx->ihdr.bit_depth < 8) ctx->bytes_per_pixel = 1; - else ctx->bytes_per_pixel = num_channels(&ctx->ihdr) * (ctx->ihdr.bit_depth / 8); - - ret = calculate_subimages(ctx); - if(ret) return ret; - - return 0; -} - -static void splt_undo(spng_ctx *ctx) -{ - struct spng_splt *splt = &ctx->splt_list[ctx->n_splt - 1]; - - spng__free(ctx, splt->entries); - - decrease_cache_usage(ctx, sizeof(struct spng_splt)); - decrease_cache_usage(ctx, splt->n_entries * sizeof(struct spng_splt_entry)); - - splt->entries = NULL; - - ctx->n_splt--; -} - -static void text_undo(spng_ctx *ctx) -{ - struct spng_text2 *text = &ctx->text_list[ctx->n_text - 1]; - - spng__free(ctx, text->keyword); - if(text->compression_flag) spng__free(ctx, text->text); - - decrease_cache_usage(ctx, text->cache_usage); - decrease_cache_usage(ctx, sizeof(struct spng_text2)); - - text->keyword = NULL; - text->text = NULL; - - ctx->n_text--; -} - -static void chunk_undo(spng_ctx *ctx) -{ - struct spng_unknown_chunk *chunk = &ctx->chunk_list[ctx->n_chunks - 1]; - - spng__free(ctx, chunk->data); - - decrease_cache_usage(ctx, chunk->length); - decrease_cache_usage(ctx, sizeof(struct spng_unknown_chunk)); - - chunk->data = NULL; - - ctx->n_chunks--; -} - -static int read_non_idat_chunks(spng_ctx *ctx) -{ - int ret; - struct spng_chunk chunk; - const unsigned char *data; - - ctx->discard = 0; - ctx->undo = NULL; - ctx->prev_stored = ctx->stored; - - while( !(ret = read_header(ctx))) - { - if(ctx->discard) - { - if(ctx->undo) ctx->undo(ctx); - - ctx->stored = ctx->prev_stored; - } - - ctx->discard = 0; - ctx->undo = NULL; - - ctx->prev_stored = ctx->stored; - chunk = ctx->current_chunk; - - if(!memcmp(chunk.type, type_idat, 4)) - { - if(ctx->state < SPNG_STATE_FIRST_IDAT) - { - if(ctx->ihdr.color_type == 3 && !ctx->stored.plte) return SPNG_ENOPLTE; - - ctx->first_idat = chunk; - return 0; - } - - if(ctx->prev_was_idat) - { - /* Ignore extra IDAT's */ - ret = discard_chunk_bytes(ctx, chunk.length); - if(ret) return ret; - - continue; - } - else return SPNG_ECHUNK_POS; /* IDAT chunk not at the end of the IDAT sequence */ - } - - ctx->prev_was_idat = 0; - - if(is_small_chunk(chunk.type)) - { - /* None of the known chunks can be zero length */ - if(!chunk.length) return SPNG_ECHUNK_SIZE; - - /* The largest of these chunks is PLTE with 256 entries */ - ret = read_chunk_bytes(ctx, chunk.length > 768 ? 768 : chunk.length); - if(ret) return ret; - } - - data = ctx->data; - - if(is_critical_chunk(&chunk)) - { - if(!memcmp(chunk.type, type_plte, 4)) - { - if(ctx->file.trns || ctx->file.hist || ctx->file.bkgd) return SPNG_ECHUNK_POS; - if(chunk.length % 3 != 0) return SPNG_ECHUNK_SIZE; - - ctx->plte.n_entries = chunk.length / 3; - - if(check_plte(&ctx->plte, &ctx->ihdr)) return SPNG_ECHUNK_SIZE; /* XXX: EPLTE? */ - - size_t i; - for(i=0; i < ctx->plte.n_entries; i++) - { - ctx->plte.entries[i].red = data[i * 3]; - ctx->plte.entries[i].green = data[i * 3 + 1]; - ctx->plte.entries[i].blue = data[i * 3 + 2]; - } - - ctx->file.plte = 1; - ctx->stored.plte = 1; - } - else if(!memcmp(chunk.type, type_iend, 4)) - { - if(ctx->state == SPNG_STATE_AFTER_IDAT) - { - if(chunk.length) return SPNG_ECHUNK_SIZE; - - ret = read_and_check_crc(ctx); - if(ret == -SPNG_CRC_DISCARD) ret = 0; - - return ret; - } - else return SPNG_ECHUNK_POS; - } - else if(!memcmp(chunk.type, type_ihdr, 4)) return SPNG_ECHUNK_POS; - else return SPNG_ECHUNK_UNKNOWN_CRITICAL; - } - else if(!memcmp(chunk.type, type_chrm, 4)) /* Ancillary chunks */ - { - if(ctx->file.plte) return SPNG_ECHUNK_POS; - if(ctx->state == SPNG_STATE_AFTER_IDAT) return SPNG_ECHUNK_POS; - if(ctx->file.chrm) return SPNG_EDUP_CHRM; - - if(chunk.length != 32) return SPNG_ECHUNK_SIZE; - - ctx->chrm_int.white_point_x = read_u32(data); - ctx->chrm_int.white_point_y = read_u32(data + 4); - ctx->chrm_int.red_x = read_u32(data + 8); - ctx->chrm_int.red_y = read_u32(data + 12); - ctx->chrm_int.green_x = read_u32(data + 16); - ctx->chrm_int.green_y = read_u32(data + 20); - ctx->chrm_int.blue_x = read_u32(data + 24); - ctx->chrm_int.blue_y = read_u32(data + 28); - - if(check_chrm_int(&ctx->chrm_int)) return SPNG_ECHRM; - - ctx->file.chrm = 1; - ctx->stored.chrm = 1; - } - else if(!memcmp(chunk.type, type_gama, 4)) - { - if(ctx->file.plte) return SPNG_ECHUNK_POS; - if(ctx->state == SPNG_STATE_AFTER_IDAT) return SPNG_ECHUNK_POS; - if(ctx->file.gama) return SPNG_EDUP_GAMA; - - if(chunk.length != 4) return SPNG_ECHUNK_SIZE; - - ctx->gama = read_u32(data); - - if(!ctx->gama) return SPNG_EGAMA; - if(ctx->gama > spng_u32max) return SPNG_EGAMA; - - ctx->file.gama = 1; - ctx->stored.gama = 1; - } - else if(!memcmp(chunk.type, type_sbit, 4)) - { - if(ctx->file.plte) return SPNG_ECHUNK_POS; - if(ctx->state == SPNG_STATE_AFTER_IDAT) return SPNG_ECHUNK_POS; - if(ctx->file.sbit) return SPNG_EDUP_SBIT; - - if(ctx->ihdr.color_type == 0) - { - if(chunk.length != 1) return SPNG_ECHUNK_SIZE; - - ctx->sbit.grayscale_bits = data[0]; - } - else if(ctx->ihdr.color_type == 2 || ctx->ihdr.color_type == 3) - { - if(chunk.length != 3) return SPNG_ECHUNK_SIZE; - - ctx->sbit.red_bits = data[0]; - ctx->sbit.green_bits = data[1]; - ctx->sbit.blue_bits = data[2]; - } - else if(ctx->ihdr.color_type == 4) - { - if(chunk.length != 2) return SPNG_ECHUNK_SIZE; - - ctx->sbit.grayscale_bits = data[0]; - ctx->sbit.alpha_bits = data[1]; - } - else if(ctx->ihdr.color_type == 6) - { - if(chunk.length != 4) return SPNG_ECHUNK_SIZE; - - ctx->sbit.red_bits = data[0]; - ctx->sbit.green_bits = data[1]; - ctx->sbit.blue_bits = data[2]; - ctx->sbit.alpha_bits = data[3]; - } - - if(check_sbit(&ctx->sbit, &ctx->ihdr)) return SPNG_ESBIT; - - ctx->file.sbit = 1; - ctx->stored.sbit = 1; - } - else if(!memcmp(chunk.type, type_srgb, 4)) - { - if(ctx->file.plte) return SPNG_ECHUNK_POS; - if(ctx->state == SPNG_STATE_AFTER_IDAT) return SPNG_ECHUNK_POS; - if(ctx->file.srgb) return SPNG_EDUP_SRGB; - - if(chunk.length != 1) return SPNG_ECHUNK_SIZE; - - ctx->srgb_rendering_intent = data[0]; - - if(ctx->srgb_rendering_intent > 3) return SPNG_ESRGB; - - ctx->file.srgb = 1; - ctx->stored.srgb = 1; - } - else if(!memcmp(chunk.type, type_bkgd, 4)) - { - if(ctx->state == SPNG_STATE_AFTER_IDAT) return SPNG_ECHUNK_POS; - if(ctx->file.bkgd) return SPNG_EDUP_BKGD; - - if(ctx->ihdr.color_type == 0 || ctx->ihdr.color_type == 4) - { - if(chunk.length != 2) return SPNG_ECHUNK_SIZE; - - ctx->bkgd.gray = read_u16(data); - } - else if(ctx->ihdr.color_type == 2 || ctx->ihdr.color_type == 6) - { - if(chunk.length != 6) return SPNG_ECHUNK_SIZE; - - ctx->bkgd.red = read_u16(data); - ctx->bkgd.green = read_u16(data + 2); - ctx->bkgd.blue = read_u16(data + 4); - } - else if(ctx->ihdr.color_type == 3) - { - if(chunk.length != 1) return SPNG_ECHUNK_SIZE; - if(!ctx->file.plte) return SPNG_EBKGD_NO_PLTE; - - ctx->bkgd.plte_index = data[0]; - if(ctx->bkgd.plte_index >= ctx->plte.n_entries) return SPNG_EBKGD_PLTE_IDX; - } - - ctx->file.bkgd = 1; - ctx->stored.bkgd = 1; - } - else if(!memcmp(chunk.type, type_trns, 4)) - { - if(ctx->state == SPNG_STATE_AFTER_IDAT) return SPNG_ECHUNK_POS; - if(ctx->file.trns) return SPNG_EDUP_TRNS; - if(!chunk.length) return SPNG_ECHUNK_SIZE; - - if(ctx->ihdr.color_type == 0) - { - if(chunk.length != 2) return SPNG_ECHUNK_SIZE; - - ctx->trns.gray = read_u16(data); - } - else if(ctx->ihdr.color_type == 2) - { - if(chunk.length != 6) return SPNG_ECHUNK_SIZE; - - ctx->trns.red = read_u16(data); - ctx->trns.green = read_u16(data + 2); - ctx->trns.blue = read_u16(data + 4); - } - else if(ctx->ihdr.color_type == 3) - { - if(chunk.length > ctx->plte.n_entries) return SPNG_ECHUNK_SIZE; - if(!ctx->file.plte) return SPNG_ETRNS_NO_PLTE; - - memcpy(ctx->trns.type3_alpha, data, chunk.length); - ctx->trns.n_type3_entries = chunk.length; - } - - if(ctx->ihdr.color_type == 4 || ctx->ihdr.color_type == 6) return SPNG_ETRNS_COLOR_TYPE; - - ctx->file.trns = 1; - ctx->stored.trns = 1; - } - else if(!memcmp(chunk.type, type_hist, 4)) - { - if(!ctx->file.plte) return SPNG_EHIST_NO_PLTE; - if(ctx->state == SPNG_STATE_AFTER_IDAT) return SPNG_ECHUNK_POS; - if(ctx->file.hist) return SPNG_EDUP_HIST; - - if( (chunk.length / 2) != (ctx->plte.n_entries) ) return SPNG_ECHUNK_SIZE; - - size_t k; - for(k=0; k < (chunk.length / 2); k++) - { - ctx->hist.frequency[k] = read_u16(data + k*2); - } - - ctx->file.hist = 1; - ctx->stored.hist = 1; - } - else if(!memcmp(chunk.type, type_phys, 4)) - { - if(ctx->state == SPNG_STATE_AFTER_IDAT) return SPNG_ECHUNK_POS; - if(ctx->file.phys) return SPNG_EDUP_PHYS; - - if(chunk.length != 9) return SPNG_ECHUNK_SIZE; - - ctx->phys.ppu_x = read_u32(data); - ctx->phys.ppu_y = read_u32(data + 4); - ctx->phys.unit_specifier = data[8]; - - if(check_phys(&ctx->phys)) return SPNG_EPHYS; - - ctx->file.phys = 1; - ctx->stored.phys = 1; - } - else if(!memcmp(chunk.type, type_time, 4)) - { - if(ctx->file.time) return SPNG_EDUP_TIME; - - if(chunk.length != 7) return SPNG_ECHUNK_SIZE; - - struct spng_time time; - - time.year = read_u16(data); - time.month = data[2]; - time.day = data[3]; - time.hour = data[4]; - time.minute = data[5]; - time.second = data[6]; - - if(check_time(&time)) return SPNG_ETIME; - - ctx->file.time = 1; - - if(!ctx->user.time) ctx->time = time; - - ctx->stored.time = 1; - } - else if(!memcmp(chunk.type, type_offs, 4)) - { - if(ctx->state == SPNG_STATE_AFTER_IDAT) return SPNG_ECHUNK_POS; - if(ctx->file.offs) return SPNG_EDUP_OFFS; - - if(chunk.length != 9) return SPNG_ECHUNK_SIZE; - - ctx->offs.x = read_s32(data); - ctx->offs.y = read_s32(data + 4); - ctx->offs.unit_specifier = data[8]; - - if(check_offs(&ctx->offs)) return SPNG_EOFFS; - - ctx->file.offs = 1; - ctx->stored.offs = 1; - } - else /* Arbitrary-length chunk */ - { - - if(!memcmp(chunk.type, type_exif, 4)) - { - if(ctx->file.exif) return SPNG_EDUP_EXIF; - if(!chunk.length) return SPNG_EEXIF; - - ctx->file.exif = 1; - - if(ctx->user.exif) goto discard; - - if(increase_cache_usage(ctx, chunk.length, 1)) return SPNG_ECHUNK_LIMITS; - - struct spng_exif exif; - - exif.length = chunk.length; - - exif.data = spng__malloc(ctx, chunk.length); - if(exif.data == NULL) return SPNG_EMEM; - - ret = read_chunk_bytes2(ctx, exif.data, chunk.length); - if(ret) - { - spng__free(ctx, exif.data); - return ret; - } - - if(check_exif(&exif)) - { - spng__free(ctx, exif.data); - return SPNG_EEXIF; - } - - ctx->exif = exif; - - ctx->stored.exif = 1; - } - else if(!memcmp(chunk.type, type_iccp, 4)) - {/* TODO: add test file with color profile */ - if(ctx->file.plte) return SPNG_ECHUNK_POS; - if(ctx->state == SPNG_STATE_AFTER_IDAT) return SPNG_ECHUNK_POS; - if(ctx->file.iccp) return SPNG_EDUP_ICCP; - if(!chunk.length) return SPNG_ECHUNK_SIZE; - - ctx->file.iccp = 1; - - uint32_t peek_bytes = 81 > chunk.length ? chunk.length : 81; - - ret = read_chunk_bytes(ctx, peek_bytes); - if(ret) return ret; - - unsigned char *keyword_nul = memchr(ctx->data, '\0', peek_bytes); - if(keyword_nul == NULL) return SPNG_EICCP_NAME; - - uint32_t keyword_len = keyword_nul - ctx->data; - - if(keyword_len > 79) return SPNG_EICCP_NAME; - - memcpy(ctx->iccp.profile_name, ctx->data, keyword_len); - - if(check_png_keyword(ctx->iccp.profile_name)) return SPNG_EICCP_NAME; - - if(chunk.length < (keyword_len + 2)) return SPNG_ECHUNK_SIZE; - - if(ctx->data[keyword_len + 1] != 0) return SPNG_EICCP_COMPRESSION_METHOD; - - ret = spng__inflate_stream(ctx, &ctx->iccp.profile, &ctx->iccp.profile_len, 0, ctx->data + keyword_len + 2, peek_bytes - (keyword_len + 2)); - - if(ret) return ret; - - ctx->stored.iccp = 1; - } - else if(!memcmp(chunk.type, type_text, 4) || - !memcmp(chunk.type, type_ztxt, 4) || - !memcmp(chunk.type, type_itxt, 4)) - { - if(!chunk.length) return SPNG_ECHUNK_SIZE; - - ctx->file.text = 1; - - if(ctx->user.text) goto discard; - - if(increase_cache_usage(ctx, sizeof(struct spng_text2), 1)) return SPNG_ECHUNK_LIMITS; - - ctx->n_text++; - if(ctx->n_text < 1) return SPNG_EOVERFLOW; - if(sizeof(struct spng_text2) > SIZE_MAX / ctx->n_text) return SPNG_EOVERFLOW; - - void *buf = spng__realloc(ctx, ctx->text_list, ctx->n_text * sizeof(struct spng_text2)); - if(buf == NULL) return SPNG_EMEM; - ctx->text_list = buf; - - struct spng_text2 *text = &ctx->text_list[ctx->n_text - 1]; - memset(text, 0, sizeof(struct spng_text2)); - - ctx->undo = text_undo; - - uint32_t text_offset = 0, language_tag_offset = 0, translated_keyword_offset = 0; - uint32_t peek_bytes = 256; /* enough for 3 80-byte keywords and some text bytes */ - uint32_t keyword_len; - - if(peek_bytes > chunk.length) peek_bytes = chunk.length; - - ret = read_chunk_bytes(ctx, peek_bytes); - if(ret) return ret; - - data = ctx->data; - - const unsigned char *zlib_stream = NULL; - const unsigned char *peek_end = data + peek_bytes; - const unsigned char *keyword_nul = memchr(data, 0, chunk.length > 80 ? 80 : chunk.length); - - if(keyword_nul == NULL) return SPNG_ETEXT_KEYWORD; - - keyword_len = keyword_nul - data; - - if(!memcmp(chunk.type, type_text, 4)) - { - text->type = SPNG_TEXT; - - text->text_length = chunk.length - keyword_len - 1; - - text_offset = keyword_len; - - /* increment past nul if there is a text field */ - if(text->text_length) text_offset++; - } - else if(!memcmp(chunk.type, type_ztxt, 4)) - { - text->type = SPNG_ZTXT; - - if((peek_bytes - keyword_len) <= 2) return SPNG_EZTXT; - - if(keyword_nul[1]) return SPNG_EZTXT_COMPRESSION_METHOD; - - text->compression_flag = 1; - - text_offset = keyword_len + 2; - } - else if(!memcmp(chunk.type, type_itxt, 4)) - { - text->type = SPNG_ITXT; - - /* at least two 1-byte fields, two >=0 length strings, and one byte of (compressed) text */ - if((peek_bytes - keyword_len) < 5) return SPNG_EITXT; - - text->compression_flag = keyword_nul[1]; - - if(text->compression_flag > 1) return SPNG_EITXT_COMPRESSION_FLAG; - - if(keyword_nul[2]) return SPNG_EITXT_COMPRESSION_METHOD; - - language_tag_offset = keyword_len + 3; - - const unsigned char *term; - term = memchr(data + language_tag_offset, 0, peek_bytes - language_tag_offset); - if(term == NULL) return SPNG_EITXT_LANG_TAG; - - if((peek_end - term) < 2) return SPNG_EITXT; - - translated_keyword_offset = term - data + 1; - - zlib_stream = memchr(data + translated_keyword_offset, 0, peek_bytes - translated_keyword_offset); - if(zlib_stream == NULL) return SPNG_EITXT; - if(zlib_stream == peek_end) return SPNG_EITXT; - - text_offset = zlib_stream - data + 1; - text->text_length = chunk.length - text_offset; - } - else return SPNG_EINTERNAL; - - - if(text->compression_flag) - { - /* cache usage = peek_bytes + decompressed text size + nul */ - if(increase_cache_usage(ctx, peek_bytes, 0)) return SPNG_ECHUNK_LIMITS; - - text->keyword = spng__calloc(ctx, 1, peek_bytes); - if(text->keyword == NULL) return SPNG_EMEM; - - memcpy(text->keyword, data, peek_bytes); - - zlib_stream = ctx->data + text_offset; - - ret = spng__inflate_stream(ctx, &text->text, &text->text_length, 1, zlib_stream, peek_bytes - text_offset); - - if(ret) return ret; - - text->text[text->text_length - 1] = '\0'; - text->cache_usage = text->text_length + peek_bytes; - } - else - { - if(increase_cache_usage(ctx, chunk.length + 1, 0)) return SPNG_ECHUNK_LIMITS; - - text->keyword = spng__malloc(ctx, chunk.length + 1); - if(text->keyword == NULL) return SPNG_EMEM; - - memcpy(text->keyword, data, peek_bytes); - - if(chunk.length > peek_bytes) - { - ret = read_chunk_bytes2(ctx, text->keyword + peek_bytes, chunk.length - peek_bytes); - if(ret) return ret; - } - - text->text = text->keyword + text_offset; - - text->text_length = chunk.length - text_offset; - - text->text[text->text_length] = '\0'; - text->cache_usage = chunk.length + 1; - } - - if(check_png_keyword(text->keyword)) return SPNG_ETEXT_KEYWORD; - - text->text_length = strlen(text->text); - - if(text->type != SPNG_ITXT) - { - language_tag_offset = keyword_len; - translated_keyword_offset = keyword_len; - - if(ctx->strict && check_png_text(text->text, text->text_length)) - { - if(text->type == SPNG_ZTXT) return SPNG_EZTXT; - else return SPNG_ETEXT; - } - } - - text->language_tag = text->keyword + language_tag_offset; - text->translated_keyword = text->keyword + translated_keyword_offset; - - ctx->stored.text = 1; - } - else if(!memcmp(chunk.type, type_splt, 4)) - { - if(ctx->state == SPNG_STATE_AFTER_IDAT) return SPNG_ECHUNK_POS; - if(ctx->user.splt) goto discard; /* XXX: could check profile names for uniqueness */ - if(!chunk.length) return SPNG_ECHUNK_SIZE; - - ctx->file.splt = 1; - - /* chunk.length + sizeof(struct spng_splt) + splt->n_entries * sizeof(struct spng_splt_entry) */ - if(increase_cache_usage(ctx, chunk.length + sizeof(struct spng_splt), 1)) return SPNG_ECHUNK_LIMITS; - - ctx->n_splt++; - if(ctx->n_splt < 1) return SPNG_EOVERFLOW; - if(sizeof(struct spng_splt) > SIZE_MAX / ctx->n_splt) return SPNG_EOVERFLOW; - - void *buf = spng__realloc(ctx, ctx->splt_list, ctx->n_splt * sizeof(struct spng_splt)); - if(buf == NULL) return SPNG_EMEM; - ctx->splt_list = buf; - - struct spng_splt *splt = &ctx->splt_list[ctx->n_splt - 1]; - - memset(splt, 0, sizeof(struct spng_splt)); - - ctx->undo = splt_undo; - - void *t = spng__malloc(ctx, chunk.length); - if(t == NULL) return SPNG_EMEM; - - splt->entries = t; /* simplifies error handling */ - data = t; - - ret = read_chunk_bytes2(ctx, t, chunk.length); - if(ret) return ret; - - uint32_t keyword_len = chunk.length < 80 ? chunk.length : 80; - - const unsigned char *keyword_nul = memchr(data, 0, keyword_len); - if(keyword_nul == NULL) return SPNG_ESPLT_NAME; - - keyword_len = keyword_nul - data; - - memcpy(splt->name, data, keyword_len); - - if(check_png_keyword(splt->name)) return SPNG_ESPLT_NAME; - - uint32_t j; - for(j=0; j < (ctx->n_splt - 1); j++) - { - if(!strcmp(ctx->splt_list[j].name, splt->name)) return SPNG_ESPLT_DUP_NAME; - } - - if( (chunk.length - keyword_len) <= 2) return SPNG_ECHUNK_SIZE; - - splt->sample_depth = data[keyword_len + 1]; - - uint32_t entries_len = chunk.length - keyword_len - 2; - if(!entries_len) return SPNG_ECHUNK_SIZE; - - if(splt->sample_depth == 16) - { - if(entries_len % 10 != 0) return SPNG_ECHUNK_SIZE; - splt->n_entries = entries_len / 10; - } - else if(splt->sample_depth == 8) - { - if(entries_len % 6 != 0) return SPNG_ECHUNK_SIZE; - splt->n_entries = entries_len / 6; - } - else return SPNG_ESPLT_DEPTH; - - if(!splt->n_entries) return SPNG_ECHUNK_SIZE; - - size_t list_size = splt->n_entries; - - if(list_size > SIZE_MAX / sizeof(struct spng_splt_entry)) return SPNG_EOVERFLOW; - - list_size *= sizeof(struct spng_splt_entry); - - if(increase_cache_usage(ctx, list_size, 0)) return SPNG_ECHUNK_LIMITS; - - splt->entries = spng__malloc(ctx, list_size); - if(splt->entries == NULL) - { - spng__free(ctx, t); - return SPNG_EMEM; - } - - data = (unsigned char*)t + keyword_len + 2; - - uint32_t k; - if(splt->sample_depth == 16) - { - for(k=0; k < splt->n_entries; k++) - { - splt->entries[k].red = read_u16(data + k * 10); - splt->entries[k].green = read_u16(data + k * 10 + 2); - splt->entries[k].blue = read_u16(data + k * 10 + 4); - splt->entries[k].alpha = read_u16(data + k * 10 + 6); - splt->entries[k].frequency = read_u16(data + k * 10 + 8); - } - } - else if(splt->sample_depth == 8) - { - for(k=0; k < splt->n_entries; k++) - { - splt->entries[k].red = data[k * 6]; - splt->entries[k].green = data[k * 6 + 1]; - splt->entries[k].blue = data[k * 6 + 2]; - splt->entries[k].alpha = data[k * 6 + 3]; - splt->entries[k].frequency = read_u16(data + k * 6 + 4); - } - } - - spng__free(ctx, t); - decrease_cache_usage(ctx, chunk.length); - - ctx->stored.splt = 1; - } - else /* Unknown chunk */ - { - ctx->file.unknown = 1; - - if(!ctx->keep_unknown) goto discard; - if(ctx->user.unknown) goto discard; - - if(increase_cache_usage(ctx, chunk.length + sizeof(struct spng_unknown_chunk), 1)) return SPNG_ECHUNK_LIMITS; - - ctx->n_chunks++; - if(ctx->n_chunks < 1) return SPNG_EOVERFLOW; - if(sizeof(struct spng_unknown_chunk) > SIZE_MAX / ctx->n_chunks) return SPNG_EOVERFLOW; - - void *buf = spng__realloc(ctx, ctx->chunk_list, ctx->n_chunks * sizeof(struct spng_unknown_chunk)); - if(buf == NULL) return SPNG_EMEM; - ctx->chunk_list = buf; - - struct spng_unknown_chunk *chunkp = &ctx->chunk_list[ctx->n_chunks - 1]; - - memset(chunkp, 0, sizeof(struct spng_unknown_chunk)); - - ctx->undo = chunk_undo; - - memcpy(chunkp->type, chunk.type, 4); - - if(ctx->state < SPNG_STATE_FIRST_IDAT) - { - if(ctx->file.plte) chunkp->location = SPNG_AFTER_PLTE; - else chunkp->location = SPNG_AFTER_IHDR; - } - else if(ctx->state >= SPNG_STATE_AFTER_IDAT) chunkp->location = SPNG_AFTER_IDAT; - - if(chunk.length > 0) - { - void *t = spng__malloc(ctx, chunk.length); - if(t == NULL) return SPNG_EMEM; - - ret = read_chunk_bytes2(ctx, t, chunk.length); - if(ret) - { - spng__free(ctx, t); - return ret; - } - - chunkp->length = chunk.length; - chunkp->data = t; - } - - ctx->stored.unknown = 1; - } - -discard: - ret = discard_chunk_bytes(ctx, ctx->cur_chunk_bytes_left); - if(ret) return ret; - } - - } - - return ret; -} - -/* Read chunks before or after the IDAT chunks depending on state */ -static int read_chunks(spng_ctx *ctx, int only_ihdr) -{ - if(ctx == NULL) return SPNG_EINTERNAL; - if(!ctx->state) return SPNG_EBADSTATE; - if(ctx->data == NULL) - { - if(ctx->encode_only) return 0; - else return SPNG_EINTERNAL; - } - - int ret = 0; - - if(ctx->state == SPNG_STATE_INPUT) - { - ret = read_ihdr(ctx); - - if(ret) return decode_err(ctx, ret); - - ctx->state = SPNG_STATE_IHDR; - } - - if(only_ihdr) return 0; - - if(ctx->state == SPNG_STATE_EOI) - { - ctx->state = SPNG_STATE_AFTER_IDAT; - ctx->prev_was_idat = 1; - } - - while(ctx->state < SPNG_STATE_FIRST_IDAT || ctx->state == SPNG_STATE_AFTER_IDAT) - { - ret = read_non_idat_chunks(ctx); - - if(!ret) - { - if(ctx->state < SPNG_STATE_FIRST_IDAT) ctx->state = SPNG_STATE_FIRST_IDAT; - else if(ctx->state == SPNG_STATE_AFTER_IDAT) ctx->state = SPNG_STATE_IEND; - } - else - { - switch(ret) - { - case SPNG_ECHUNK_POS: - case SPNG_ECHUNK_SIZE: /* size != expected size, SPNG_ECHUNK_STDLEN = invalid size */ - case SPNG_EDUP_PLTE: - case SPNG_EDUP_CHRM: - case SPNG_EDUP_GAMA: - case SPNG_EDUP_ICCP: - case SPNG_EDUP_SBIT: - case SPNG_EDUP_SRGB: - case SPNG_EDUP_BKGD: - case SPNG_EDUP_HIST: - case SPNG_EDUP_TRNS: - case SPNG_EDUP_PHYS: - case SPNG_EDUP_TIME: - case SPNG_EDUP_OFFS: - case SPNG_EDUP_EXIF: - case SPNG_ECHRM: - case SPNG_ETRNS_COLOR_TYPE: - case SPNG_ETRNS_NO_PLTE: - case SPNG_EGAMA: - case SPNG_EICCP_NAME: - case SPNG_EICCP_COMPRESSION_METHOD: - case SPNG_ESBIT: - case SPNG_ESRGB: - case SPNG_ETEXT: - case SPNG_ETEXT_KEYWORD: - case SPNG_EZTXT: - case SPNG_EZTXT_COMPRESSION_METHOD: - case SPNG_EITXT: - case SPNG_EITXT_COMPRESSION_FLAG: - case SPNG_EITXT_COMPRESSION_METHOD: - case SPNG_EITXT_LANG_TAG: - case SPNG_EITXT_TRANSLATED_KEY: - case SPNG_EBKGD_NO_PLTE: - case SPNG_EBKGD_PLTE_IDX: - case SPNG_EHIST_NO_PLTE: - case SPNG_EPHYS: - case SPNG_ESPLT_NAME: - case SPNG_ESPLT_DUP_NAME: - case SPNG_ESPLT_DEPTH: - case SPNG_ETIME: - case SPNG_EOFFS: - case SPNG_EEXIF: - case SPNG_EZLIB: - { - if(!ctx->strict && !is_critical_chunk(&ctx->current_chunk)) - { - ret = discard_chunk_bytes(ctx, ctx->cur_chunk_bytes_left); - if(ret) return decode_err(ctx, ret); - - if(ctx->undo) ctx->undo(ctx); - - ctx->stored = ctx->prev_stored; - - ctx->discard = 0; - ctx->undo = NULL; - - continue; - } - else return decode_err(ctx, ret); - - break; - } - default: return decode_err(ctx, ret); - } - } - } - - return ret; -} - -static int read_scanline(spng_ctx *ctx) -{ - int ret, pass = ctx->row_info.pass; - struct spng_row_info *ri = &ctx->row_info; - const struct spng_subimage *sub = ctx->subimage; - size_t scanline_width = sub[pass].scanline_width; - uint32_t scanline_idx = ri->scanline_idx; - - uint8_t next_filter = 0; - - if(scanline_idx == (sub[pass].height - 1) && ri->pass == ctx->last_pass) - { - ret = read_scanline_bytes(ctx, ctx->scanline, scanline_width - 1); - } - else - { - ret = read_scanline_bytes(ctx, ctx->scanline, scanline_width); - if(ret) return ret; - - next_filter = ctx->scanline[scanline_width - 1]; - if(next_filter > 4) ret = SPNG_EFILTER; - } - - if(ret) return ret; - - if(!scanline_idx && ri->filter > 1) - { - /* prev_scanline is all zeros for the first scanline */ - memset(ctx->prev_scanline, 0, scanline_width); - } - - if(ctx->ihdr.bit_depth == 16 && ctx->fmt != SPNG_FMT_RAW) u16_row_to_host(ctx->scanline, scanline_width - 1); - - ret = defilter_scanline(ctx->prev_scanline, ctx->scanline, scanline_width, ctx->bytes_per_pixel, ri->filter); - if(ret) return ret; - - ri->filter = next_filter; - - return 0; -} - -static int update_row_info(spng_ctx *ctx) -{ - int interlacing = ctx->ihdr.interlace_method; - struct spng_row_info *ri = &ctx->row_info; - const struct spng_subimage *sub = ctx->subimage; - - if(ri->scanline_idx == (sub[ri->pass].height - 1)) /* Last scanline */ - { - if(ri->pass == ctx->last_pass) - { - ctx->state = SPNG_STATE_EOI; - - return SPNG_EOI; - } - - ri->scanline_idx = 0; - ri->pass++; - - /* Skip empty passes */ - while( (!sub[ri->pass].width || !sub[ri->pass].height) && (ri->pass < ctx->last_pass)) ri->pass++; - } - else - { - ri->row_num++; - ri->scanline_idx++; - } - - if(interlacing) ri->row_num = adam7_y_start[ri->pass] + ri->scanline_idx * adam7_y_delta[ri->pass]; - - return 0; -} - -int spng_decode_scanline(spng_ctx *ctx, void *out, size_t len) -{ - if(ctx == NULL || out == NULL) return 1; - - if(ctx->state >= SPNG_STATE_EOI) return SPNG_EOI; - - struct decode_flags f = ctx->decode_flags; - - struct spng_row_info *ri = &ctx->row_info; - const struct spng_subimage *sub = ctx->subimage; - - const struct spng_ihdr *ihdr = &ctx->ihdr; - const uint16_t *gamma_lut = ctx->gamma_lut; - unsigned char *trns_px = ctx->trns_px; - const struct spng_sbit *sb = &ctx->decode_sb; - const struct spng_plte_entry *plte = ctx->decode_plte.rgba; - struct spng__iter iter = spng__iter_init(ihdr->bit_depth, ctx->scanline); - - const unsigned char *scanline; - - const int pass = ri->pass; - const int fmt = ctx->fmt; - const size_t scanline_width = sub[pass].scanline_width; - const uint32_t width = sub[pass].width; - uint32_t k; - uint8_t r_8, g_8, b_8, a_8, gray_8; - uint16_t r_16, g_16, b_16, a_16, gray_16; - r_8=0; g_8=0; b_8=0; a_8=0; gray_8=0; - r_16=0; g_16=0; b_16=0; a_16=0; gray_16=0; - size_t pixel_size = 4; /* SPNG_FMT_RGBA8 */ - size_t pixel_offset = 0; - unsigned char *pixel; - unsigned processing_depth = ihdr->bit_depth; - - if(f.indexed) processing_depth = 8; - - if(fmt == SPNG_FMT_RGBA16) pixel_size = 8; - else if(fmt == SPNG_FMT_RGB8) pixel_size = 3; - - if(len < sub[pass].out_width) return SPNG_EBUFSIZ; - - int ret = read_scanline(ctx); - - if(ret) return decode_err(ctx, ret); - - scanline = ctx->scanline; - - for(k=0; k < width; k++) - { - pixel = (unsigned char*)out + pixel_offset; - pixel_offset += pixel_size; - - if(f.same_layout) - { - if(f.zerocopy) break; - - memcpy(out, scanline, scanline_width - 1); - break; - } - - if(f.unpack) - { - unpack_scanline(out, scanline, width, ihdr->bit_depth, fmt); - break; - } - - if(ihdr->color_type == SPNG_COLOR_TYPE_TRUECOLOR) - { - if(ihdr->bit_depth == 16) - { - memcpy(&r_16, scanline + (k * 6), 2); - memcpy(&g_16, scanline + (k * 6) + 2, 2); - memcpy(&b_16, scanline + (k * 6) + 4, 2); - - a_16 = 65535; - } - else /* == 8 */ - { - if(fmt == SPNG_FMT_RGBA8) - { - rgb8_row_to_rgba8(scanline, out, width); - break; - } - - r_8 = scanline[k * 3]; - g_8 = scanline[k * 3 + 1]; - b_8 = scanline[k * 3 + 2]; - - a_8 = 255; - } - } - else if(ihdr->color_type == SPNG_COLOR_TYPE_INDEXED) - { - uint8_t entry = 0; - - if(ihdr->bit_depth == 8) - { - if(fmt & (SPNG_FMT_RGBA8 | SPNG_FMT_RGB8)) - { - expand_row(out, scanline, &ctx->decode_plte, width, fmt); - break; - } - - entry = scanline[k]; - } - else /* < 8 */ - { - entry = get_sample(&iter); - } - - if(fmt & (SPNG_FMT_RGBA8 | SPNG_FMT_RGB8)) - { - pixel[0] = plte[entry].red; - pixel[1] = plte[entry].green; - pixel[2] = plte[entry].blue; - if(fmt == SPNG_FMT_RGBA8) pixel[3] = plte[entry].alpha; - - continue; - } - else /* RGBA16 */ - { - r_16 = plte[entry].red; - g_16 = plte[entry].green; - b_16 = plte[entry].blue; - a_16 = plte[entry].alpha; - - r_16 = (r_16 << 8) | r_16; - g_16 = (g_16 << 8) | g_16; - b_16 = (b_16 << 8) | b_16; - a_16 = (a_16 << 8) | a_16; - - memcpy(pixel, &r_16, 2); - memcpy(pixel + 2, &g_16, 2); - memcpy(pixel + 4, &b_16, 2); - memcpy(pixel + 6, &a_16, 2); - - continue; - } - } - else if(ihdr->color_type == SPNG_COLOR_TYPE_TRUECOLOR_ALPHA) - { - if(ihdr->bit_depth == 16) - { - memcpy(&r_16, scanline + (k * 8), 2); - memcpy(&g_16, scanline + (k * 8) + 2, 2); - memcpy(&b_16, scanline + (k * 8) + 4, 2); - memcpy(&a_16, scanline + (k * 8) + 6, 2); - } - else /* == 8 */ - { - r_8 = scanline[k * 4]; - g_8 = scanline[k * 4 + 1]; - b_8 = scanline[k * 4 + 2]; - a_8 = scanline[k * 4 + 3]; - } - } - else if(ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE) - { - if(ihdr->bit_depth == 16) - { - memcpy(&gray_16, scanline + k * 2, 2); - - if(f.apply_trns && ctx->trns.gray == gray_16) a_16 = 0; - else a_16 = 65535; - - r_16 = gray_16; - g_16 = gray_16; - b_16 = gray_16; - } - else /* <= 8 */ - { - gray_8 = get_sample(&iter); - - if(f.apply_trns && ctx->trns.gray == gray_8) a_8 = 0; - else a_8 = 255; - - r_8 = gray_8; g_8 = gray_8; b_8 = gray_8; - } - } - else if(ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE_ALPHA) - { - if(ihdr->bit_depth == 16) - { - memcpy(&gray_16, scanline + (k * 4), 2); - memcpy(&a_16, scanline + (k * 4) + 2, 2); - - r_16 = gray_16; - g_16 = gray_16; - b_16 = gray_16; - } - else /* == 8 */ - { - gray_8 = scanline[k * 2]; - a_8 = scanline[k * 2 + 1]; - - r_8 = gray_8; - g_8 = gray_8; - b_8 = gray_8; - } - } - - - if(fmt & (SPNG_FMT_RGBA8 | SPNG_FMT_RGB8)) - { - if(ihdr->bit_depth == 16) - { - r_8 = r_16 >> 8; - g_8 = g_16 >> 8; - b_8 = b_16 >> 8; - a_8 = a_16 >> 8; - } - - pixel[0] = r_8; - pixel[1] = g_8; - pixel[2] = b_8; - - if(fmt == SPNG_FMT_RGBA8) pixel[3] = a_8; - } - else if(fmt == SPNG_FMT_RGBA16) - { - if(ihdr->bit_depth != 16) - { - r_16 = r_8; - g_16 = g_8; - b_16 = b_8; - a_16 = a_8; - } - - memcpy(pixel, &r_16, 2); - memcpy(pixel + 2, &g_16, 2); - memcpy(pixel + 4, &b_16, 2); - memcpy(pixel + 6, &a_16, 2); - } - }/* for(k=0; k < width; k++) */ - - if(f.apply_trns) trns_row(out, scanline, trns_px, ctx->bytes_per_pixel, &ctx->ihdr, width, fmt); - - if(f.do_scaling) scale_row(out, width, fmt, processing_depth, sb); - - if(f.apply_gamma) gamma_correct_row(out, width, fmt, gamma_lut); - - /* The previous scanline is always defiltered */ - void *t = ctx->prev_scanline; - ctx->prev_scanline = ctx->scanline; - ctx->scanline = t; - - ret = update_row_info(ctx); - - if(ret == SPNG_EOI) - { - if(ctx->cur_chunk_bytes_left) /* zlib stream ended before an IDAT chunk boundary */ - {/* Discard the rest of the chunk */ - int error = discard_chunk_bytes(ctx, ctx->cur_chunk_bytes_left); - if(error) return decode_err(ctx, error); - } - - ctx->last_idat = ctx->current_chunk; - } - - return ret; -} - -int spng_decode_row(spng_ctx *ctx, void *out, size_t len) -{ - if(ctx == NULL || out == NULL) return 1; - if(ctx->state >= SPNG_STATE_EOI) return SPNG_EOI; - if(len < ctx->image_width) return SPNG_EBUFSIZ; - - const struct spng_ihdr *ihdr = &ctx->ihdr; - int ret, pass = ctx->row_info.pass; - unsigned char *outptr = out; - - if(!ihdr->interlace_method || pass == 6) return spng_decode_scanline(ctx, out, len); - - ret = spng_decode_scanline(ctx, ctx->row, ctx->image_width); - if(ret && ret != SPNG_EOI) return ret; - - uint32_t k; - unsigned pixel_size = 4; /* RGBA8 */ - if(ctx->fmt == SPNG_FMT_RGBA16) pixel_size = 8; - else if(ctx->fmt == SPNG_FMT_RGB8) pixel_size = 3; - else if(ctx->fmt == SPNG_FMT_G8) pixel_size = 1; - else if(ctx->fmt == SPNG_FMT_GA8) pixel_size = 2; - else if(ctx->fmt & (SPNG_FMT_PNG | SPNG_FMT_RAW)) - { - if(ihdr->bit_depth < 8) - { - struct spng__iter iter = spng__iter_init(ihdr->bit_depth, ctx->row); - const uint8_t samples_per_byte = 8 / ihdr->bit_depth; - uint8_t sample; - - for(k=0; k < ctx->subimage[pass].width; k++) - { - sample = get_sample(&iter); - - size_t ioffset = adam7_x_start[pass] + k * adam7_x_delta[pass]; - - sample = sample << (iter.initial_shift - ioffset * ihdr->bit_depth % 8); - - ioffset /= samples_per_byte; - - outptr[ioffset] |= sample; - } - - return 0; - } - else pixel_size = ctx->bytes_per_pixel; - } - - for(k=0; k < ctx->subimage[pass].width; k++) - { - size_t ioffset = (adam7_x_start[pass] + (size_t) k * adam7_x_delta[pass]) * pixel_size; - - memcpy(outptr + ioffset, ctx->row + k * pixel_size, pixel_size); - } - - return 0; -} - -int spng_decode_chunks(spng_ctx *ctx) -{ - if(ctx == NULL) return 1; - if(ctx->encode_only) return SPNG_ECTXTYPE; - if(ctx->state < SPNG_STATE_INPUT) return SPNG_ENOSRC; - if(ctx->state == SPNG_STATE_IEND) return 0; - - return read_chunks(ctx, 0); -} - -int spng_decode_image(spng_ctx *ctx, void *out, size_t len, int fmt, int flags) -{ - if(ctx == NULL) return 1; - if(ctx->encode_only) return SPNG_ECTXTYPE; - if(ctx->state >= SPNG_STATE_EOI) return SPNG_EOI; - - const struct spng_ihdr *ihdr = &ctx->ihdr; - - int ret = read_chunks(ctx, 0); - if(ret) return decode_err(ctx, ret); - - ret = check_decode_fmt(ihdr, fmt); - if(ret) return ret; - - ret = calculate_image_width(ihdr, fmt, &ctx->image_width); - if(ret) return decode_err(ctx, ret); - - if(ctx->image_width > SIZE_MAX / ihdr->height) ctx->image_size = 0; /* overflow */ - else ctx->image_size = ctx->image_width * ihdr->height; - - if( !(flags & SPNG_DECODE_PROGRESSIVE) ) - { - if(out == NULL) return 1; - if(!ctx->image_size) return SPNG_EOVERFLOW; - if(len < ctx->image_size) return SPNG_EBUFSIZ; - } - - uint32_t bytes_read = 0; - - ret = read_idat_bytes(ctx, &bytes_read); - if(ret) return decode_err(ctx, ret); - - if(bytes_read > 1) - { - int valid = read_u16(ctx->data) % 31 ? 0 : 1; - - unsigned flg = ctx->data[1]; - unsigned flevel = flg >> 6; - int compression_level = Z_DEFAULT_COMPRESSION; - - if(flevel == 0) compression_level = 0; /* fastest */ - else if(flevel == 1) compression_level = 1; /* fast */ - else if(flevel == 2) compression_level = 6; /* default */ - else if(flevel == 3) compression_level = 9; /* slowest, max compression */ - - if(valid) ctx->image_options.compression_level = compression_level; - } - - ret = spng__inflate_init(ctx, ctx->image_options.window_bits); - if(ret) return decode_err(ctx, ret); - - ctx->zstream.avail_in = bytes_read; - ctx->zstream.next_in = ctx->data; - - size_t scanline_buf_size = ctx->subimage[ctx->widest_pass].scanline_width; - - scanline_buf_size += 32; - - if(scanline_buf_size < 32) return SPNG_EOVERFLOW; - - ctx->scanline_buf = spng__malloc(ctx, scanline_buf_size); - ctx->prev_scanline_buf = spng__malloc(ctx, scanline_buf_size); - - ctx->scanline = ctx->scanline_buf; - ctx->prev_scanline = ctx->prev_scanline_buf; - - struct decode_flags f = {0}; - - ctx->fmt = fmt; - - if(ihdr->color_type == SPNG_COLOR_TYPE_INDEXED) f.indexed = 1; - - unsigned processing_depth = ihdr->bit_depth; - - if(f.indexed) processing_depth = 8; - - if(ihdr->interlace_method) - { - f.interlaced = 1; - ctx->row_buf = spng__malloc(ctx, ctx->image_width); - ctx->row = ctx->row_buf; - - if(ctx->row == NULL) return decode_err(ctx, SPNG_EMEM); - } - - if(ctx->scanline == NULL || ctx->prev_scanline == NULL) - { - return decode_err(ctx, SPNG_EMEM); - } - - f.do_scaling = 1; - if(f.indexed) f.do_scaling = 0; - - unsigned depth_target = 8; /* FMT_RGBA8, G8 */ - if(fmt == SPNG_FMT_RGBA16) depth_target = 16; - - if(flags & SPNG_DECODE_TRNS && ctx->stored.trns) f.apply_trns = 1; - else flags &= ~SPNG_DECODE_TRNS; - - if(ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE_ALPHA || - ihdr->color_type == SPNG_COLOR_TYPE_TRUECOLOR_ALPHA) flags &= ~SPNG_DECODE_TRNS; - - if(flags & SPNG_DECODE_GAMMA && ctx->stored.gama) f.apply_gamma = 1; - else flags &= ~SPNG_DECODE_GAMMA; - - if(flags & SPNG_DECODE_USE_SBIT && ctx->stored.sbit) f.use_sbit = 1; - else flags &= ~SPNG_DECODE_USE_SBIT; - - if(fmt & (SPNG_FMT_RGBA8 | SPNG_FMT_RGBA16)) - { - if(ihdr->color_type == SPNG_COLOR_TYPE_TRUECOLOR_ALPHA && - ihdr->bit_depth == depth_target) f.same_layout = 1; - } - else if(fmt == SPNG_FMT_RGB8) - { - if(ihdr->color_type == SPNG_COLOR_TYPE_TRUECOLOR && - ihdr->bit_depth == depth_target) f.same_layout = 1; - - f.apply_trns = 0; /* not applicable */ - } - else if(fmt & (SPNG_FMT_PNG | SPNG_FMT_RAW)) - { - f.same_layout = 1; - f.do_scaling = 0; - f.apply_gamma = 0; /* for now */ - f.apply_trns = 0; - } - else if(fmt == SPNG_FMT_G8 && ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE && ihdr->bit_depth <= 8) - { - if(ihdr->bit_depth == depth_target) f.same_layout = 1; - else if(ihdr->bit_depth < 8) f.unpack = 1; - - f.apply_trns = 0; - } - else if(fmt == SPNG_FMT_GA8 && ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE && ihdr->bit_depth <= 8) - { - if(ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE_ALPHA && - ihdr->bit_depth == depth_target) f.same_layout = 1; - else if(ihdr->bit_depth <= 8) f.unpack = 1; - } - else if(fmt == SPNG_FMT_GA16 && ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE && ihdr->bit_depth == 16) - { - if(ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE_ALPHA && - ihdr->bit_depth == depth_target) f.same_layout = 1; - else if(ihdr->bit_depth == 16) f.unpack = 1; - } - - /*if(f.same_layout && !flags && !f.interlaced) f.zerocopy = 1;*/ - - uint16_t *gamma_lut = NULL; - - if(f.apply_gamma) - { - float file_gamma = (float)ctx->gama / 100000.0f; - float max; - - unsigned lut_entries; - - if(fmt & (SPNG_FMT_RGBA8 | SPNG_FMT_RGB8)) - { - lut_entries = 256; - max = 255.0f; - - gamma_lut = ctx->gamma_lut8; - ctx->gamma_lut = ctx->gamma_lut8; - } - else /* SPNG_FMT_RGBA16 */ - { - lut_entries = 65536; - max = 65535.0f; - - ctx->gamma_lut16 = spng__malloc(ctx, lut_entries * sizeof(uint16_t)); - if(ctx->gamma_lut16 == NULL) return decode_err(ctx, SPNG_EMEM); - - gamma_lut = ctx->gamma_lut16; - ctx->gamma_lut = ctx->gamma_lut16; - } - - float screen_gamma = 2.2f; - float exponent = file_gamma * screen_gamma; - - if(FP_ZERO == fpclassify(exponent)) return decode_err(ctx, SPNG_EGAMA); - - exponent = 1.0f / exponent; - - unsigned i; - for(i=0; i < lut_entries; i++) - { - float c = pow((float)i / max, exponent) * max; - if(c > max) c = max; - - gamma_lut[i] = (uint16_t)c; - } - } - - struct spng_sbit *sb = &ctx->decode_sb; - - sb->red_bits = processing_depth; - sb->green_bits = processing_depth; - sb->blue_bits = processing_depth; - sb->alpha_bits = processing_depth; - sb->grayscale_bits = processing_depth; - - if(f.use_sbit) - { - if(ihdr->color_type == 0) - { - sb->grayscale_bits = ctx->sbit.grayscale_bits; - sb->alpha_bits = ihdr->bit_depth; - } - else if(ihdr->color_type == 2 || ihdr->color_type == 3) - { - sb->red_bits = ctx->sbit.red_bits; - sb->green_bits = ctx->sbit.green_bits; - sb->blue_bits = ctx->sbit.blue_bits; - sb->alpha_bits = ihdr->bit_depth; - } - else if(ihdr->color_type == 4) - { - sb->grayscale_bits = ctx->sbit.grayscale_bits; - sb->alpha_bits = ctx->sbit.alpha_bits; - } - else /* == 6 */ - { - sb->red_bits = ctx->sbit.red_bits; - sb->green_bits = ctx->sbit.green_bits; - sb->blue_bits = ctx->sbit.blue_bits; - sb->alpha_bits = ctx->sbit.alpha_bits; - } - } - - if(ihdr->bit_depth == 16 && fmt & (SPNG_FMT_RGBA8 | SPNG_FMT_RGB8)) - {/* samples are scaled down by 8 bits in the decode loop */ - sb->red_bits -= 8; - sb->green_bits -= 8; - sb->blue_bits -= 8; - sb->alpha_bits -= 8; - sb->grayscale_bits -= 8; - - processing_depth = 8; - } - - /* Prevent infinite loops in sample_to_target() */ - if(!depth_target || depth_target > 16 || - !processing_depth || processing_depth > 16 || - !sb->grayscale_bits || sb->grayscale_bits > processing_depth || - !sb->alpha_bits || sb->alpha_bits > processing_depth || - !sb->red_bits || sb->red_bits > processing_depth || - !sb->green_bits || sb->green_bits > processing_depth || - !sb->blue_bits || sb->blue_bits > processing_depth) - { - return decode_err(ctx, SPNG_ESBIT); - } - - if(sb->red_bits == sb->green_bits && - sb->green_bits == sb->blue_bits && - sb->blue_bits == sb->alpha_bits && - sb->alpha_bits == processing_depth && - processing_depth == depth_target) f.do_scaling = 0; - - struct spng_plte_entry *plte = ctx->decode_plte.rgba; - - /* Pre-process palette entries */ - if(f.indexed) - { - uint8_t red, green, blue, alpha; - - uint32_t i; - for(i=0; i < 256; i++) - { - if(f.apply_trns && i < ctx->trns.n_type3_entries) - ctx->plte.entries[i].alpha = ctx->trns.type3_alpha[i]; - else - ctx->plte.entries[i].alpha = 255; - - red = sample_to_target(ctx->plte.entries[i].red, 8, sb->red_bits, 8); - green = sample_to_target(ctx->plte.entries[i].green, 8, sb->green_bits, 8); - blue = sample_to_target(ctx->plte.entries[i].blue, 8, sb->blue_bits, 8); - alpha = sample_to_target(ctx->plte.entries[i].alpha, 8, sb->alpha_bits, 8); - -#if defined(SPNG_ARM) - if(fmt == SPNG_FMT_RGB8 && ihdr->bit_depth == 8) - {/* Working with 3 bytes at a time is more of an ARM thing */ - ctx->decode_plte.rgb[i * 3 + 0] = red; - ctx->decode_plte.rgb[i * 3 + 1] = green; - ctx->decode_plte.rgb[i * 3 + 2] = blue; - continue; - } -#endif - plte[i].red = red; - plte[i].green = green; - plte[i].blue = blue; - plte[i].alpha = alpha; - } - - f.apply_trns = 0; - } - - unsigned char *trns_px = ctx->trns_px; - - if(f.apply_trns) - { - uint16_t mask = ~0; - if(ctx->ihdr.bit_depth < 16) mask = (1 << ctx->ihdr.bit_depth) - 1; - - if(fmt & (SPNG_FMT_RGBA8 | SPNG_FMT_RGBA16)) - { - if(ihdr->color_type == SPNG_COLOR_TYPE_TRUECOLOR) - { - if(ihdr->bit_depth == 16) - { - memcpy(trns_px, &ctx->trns.red, 2); - memcpy(trns_px + 2, &ctx->trns.green, 2); - memcpy(trns_px + 4, &ctx->trns.blue, 2); - } - else - { - trns_px[0] = ctx->trns.red & mask; - trns_px[1] = ctx->trns.green & mask; - trns_px[2] = ctx->trns.blue & mask; - } - } - } - else if(ihdr->color_type == SPNG_COLOR_TYPE_GRAYSCALE) // fmt == SPNG_FMT_GA8 && - { - if(ihdr->bit_depth == 16) - { - memcpy(trns_px, &ctx->trns.gray, 2); - } - else - { - trns_px[0] = ctx->trns.gray & mask; - } - } - } - - ctx->decode_flags = f; - - ctx->state = SPNG_STATE_DECODE_INIT; - - struct spng_row_info *ri = &ctx->row_info; - struct spng_subimage *sub = ctx->subimage; - - while(!sub[ri->pass].width || !sub[ri->pass].height) ri->pass++; - - if(f.interlaced) ri->row_num = adam7_y_start[ri->pass]; - - unsigned pixel_size = 4; /* SPNG_FMT_RGBA8 */ - - if(fmt == SPNG_FMT_RGBA16) pixel_size = 8; - else if(fmt == SPNG_FMT_RGB8) pixel_size = 3; - else if(fmt == SPNG_FMT_G8) pixel_size = 1; - else if(fmt == SPNG_FMT_GA8) pixel_size = 2; - - int i; - for(i=ri->pass; i <= ctx->last_pass; i++) - { - if(!sub[i].scanline_width) continue; - - if(fmt & (SPNG_FMT_PNG | SPNG_FMT_RAW)) sub[i].out_width = sub[i].scanline_width - 1; - else sub[i].out_width = (size_t)sub[i].width * pixel_size; - - if(sub[i].out_width > UINT32_MAX) return decode_err(ctx, SPNG_EOVERFLOW); - } - - /* Read the first filter byte, offsetting all reads by 1 byte. - The scanlines will be aligned with the start of the array with - the next scanline's filter byte at the end, - the last scanline will end up being 1 byte "shorter". */ - ret = read_scanline_bytes(ctx, &ri->filter, 1); - if(ret) return decode_err(ctx, ret); - - if(ri->filter > 4) return decode_err(ctx, SPNG_EFILTER); - - if(flags & SPNG_DECODE_PROGRESSIVE) - { - return 0; - } - - do - { - size_t ioffset = ri->row_num * ctx->image_width; - - ret = spng_decode_row(ctx, (unsigned char*)out + ioffset, ctx->image_width); - }while(!ret); - - if(ret != SPNG_EOI) return decode_err(ctx, ret); - - return 0; -} - -int spng_get_row_info(spng_ctx *ctx, struct spng_row_info *row_info) -{ - if(ctx == NULL || row_info == NULL || ctx->state < SPNG_STATE_DECODE_INIT) return 1; - - if(ctx->state >= SPNG_STATE_EOI) return SPNG_EOI; - - *row_info = ctx->row_info; - - return 0; -} - -static int write_chunks_before_idat(spng_ctx *ctx) -{ - if(ctx == NULL) return SPNG_EINTERNAL; - if(!ctx->encode_only) return SPNG_EINTERNAL; - if(!ctx->stored.ihdr) return SPNG_EINTERNAL; - - int ret; - uint32_t i; - size_t length; - const struct spng_ihdr *ihdr = &ctx->ihdr; - unsigned char *data = ctx->decode_plte.raw; - - ret = write_data(ctx, spng_signature, 8); - if(ret) return ret; - - write_u32(data, ihdr->width); - write_u32(data + 4, ihdr->height); - data[8] = ihdr->bit_depth; - data[9] = ihdr->color_type; - data[10] = ihdr->compression_method; - data[11] = ihdr->filter_method; - data[12] = ihdr->interlace_method; - - ret = write_chunk(ctx, type_ihdr, data, 13); - if(ret) return ret; - - if(ctx->stored.chrm) - { - write_u32(data, ctx->chrm_int.white_point_x); - write_u32(data + 4, ctx->chrm_int.white_point_y); - write_u32(data + 8, ctx->chrm_int.red_x); - write_u32(data + 12, ctx->chrm_int.red_y); - write_u32(data + 16, ctx->chrm_int.green_x); - write_u32(data + 20, ctx->chrm_int.green_y); - write_u32(data + 24, ctx->chrm_int.blue_x); - write_u32(data + 28, ctx->chrm_int.blue_y); - - ret = write_chunk(ctx, type_chrm, data, 32); - if(ret) return ret; - } - - if(ctx->stored.gama) - { - write_u32(data, ctx->gama); - - ret = write_chunk(ctx, type_gama, data, 4); - if(ret) return ret; - } - - if(ctx->stored.iccp) - { - uLongf dest_len = compressBound((uLong)ctx->iccp.profile_len); - - Bytef *buf = spng__malloc(ctx, dest_len); - if(buf == NULL) return SPNG_EMEM; - - ret = compress2(buf, &dest_len, (void*)ctx->iccp.profile, (uLong)ctx->iccp.profile_len, Z_DEFAULT_COMPRESSION); - - if(ret != Z_OK) - { - spng__free(ctx, buf); - return SPNG_EZLIB; - } - - size_t name_len = strlen(ctx->iccp.profile_name); - - length = name_len + 2; - length += dest_len; - - if(dest_len > length) return SPNG_EOVERFLOW; - - unsigned char *cdata = NULL; - - ret = write_header(ctx, type_iccp, length, &cdata); - - if(ret) - { - spng__free(ctx, buf); - return ret; - } - - memcpy(cdata, ctx->iccp.profile_name, name_len + 1); - cdata[name_len + 1] = 0; /* compression method */ - memcpy(cdata + name_len + 2, buf, dest_len); - - spng__free(ctx, buf); - - ret = finish_chunk(ctx); - if(ret) return ret; - } - - if(ctx->stored.sbit) - { - switch(ctx->ihdr.color_type) - { - case SPNG_COLOR_TYPE_GRAYSCALE: - { - length = 1; - - data[0] = ctx->sbit.grayscale_bits; - - break; - } - case SPNG_COLOR_TYPE_TRUECOLOR: - case SPNG_COLOR_TYPE_INDEXED: - { - length = 3; - - data[0] = ctx->sbit.red_bits; - data[1] = ctx->sbit.green_bits; - data[2] = ctx->sbit.blue_bits; - - break; - } - case SPNG_COLOR_TYPE_GRAYSCALE_ALPHA: - { - length = 2; - - data[0] = ctx->sbit.grayscale_bits; - data[1] = ctx->sbit.alpha_bits; - - break; - } - case SPNG_COLOR_TYPE_TRUECOLOR_ALPHA: - { - length = 4; - - data[0] = ctx->sbit.red_bits; - data[1] = ctx->sbit.green_bits; - data[2] = ctx->sbit.blue_bits; - data[3] = ctx->sbit.alpha_bits; - - break; - } - default: return SPNG_EINTERNAL; - } - - ret = write_chunk(ctx, type_sbit, data, length); - if(ret) return ret; - } - - if(ctx->stored.srgb) - { - ret = write_chunk(ctx, type_srgb, &ctx->srgb_rendering_intent, 1); - if(ret) return ret; - } - - ret = write_unknown_chunks(ctx, SPNG_AFTER_IHDR); - if(ret) return ret; - - if(ctx->stored.plte) - { - for(i=0; i < ctx->plte.n_entries; i++) - { - data[i * 3 + 0] = ctx->plte.entries[i].red; - data[i * 3 + 1] = ctx->plte.entries[i].green; - data[i * 3 + 2] = ctx->plte.entries[i].blue; - } - - ret = write_chunk(ctx, type_plte, data, ctx->plte.n_entries * 3); - if(ret) return ret; - } - - if(ctx->stored.bkgd) - { - switch(ctx->ihdr.color_type) - { - case SPNG_COLOR_TYPE_GRAYSCALE: - case SPNG_COLOR_TYPE_GRAYSCALE_ALPHA: - { - length = 2; - - write_u16(data, ctx->bkgd.gray); - - break; - } - case SPNG_COLOR_TYPE_TRUECOLOR: - case SPNG_COLOR_TYPE_TRUECOLOR_ALPHA: - { - length = 6; - - write_u16(data, ctx->bkgd.red); - write_u16(data + 2, ctx->bkgd.green); - write_u16(data + 4, ctx->bkgd.blue); - - break; - } - case SPNG_COLOR_TYPE_INDEXED: - { - length = 1; - - data[0] = ctx->bkgd.plte_index; - - break; - } - default: return SPNG_EINTERNAL; - } - - ret = write_chunk(ctx, type_bkgd, data, length); - if(ret) return ret; - } - - if(ctx->stored.hist) - { - length = ctx->plte.n_entries * 2; - - for(i=0; i < ctx->plte.n_entries; i++) - { - write_u16(data + i * 2, ctx->hist.frequency[i]); - } - - ret = write_chunk(ctx, type_hist, data, length); - if(ret) return ret; - } - - if(ctx->stored.trns) - { - if(ctx->ihdr.color_type == SPNG_COLOR_TYPE_GRAYSCALE) - { - write_u16(data, ctx->trns.gray); - - ret = write_chunk(ctx, type_trns, data, 2); - } - else if(ctx->ihdr.color_type == SPNG_COLOR_TYPE_TRUECOLOR) - { - write_u16(data, ctx->trns.red); - write_u16(data + 2, ctx->trns.green); - write_u16(data + 4, ctx->trns.blue); - - ret = write_chunk(ctx, type_trns, data, 6); - } - else if(ctx->ihdr.color_type == SPNG_COLOR_TYPE_INDEXED) - { - ret = write_chunk(ctx, type_trns, ctx->trns.type3_alpha, ctx->trns.n_type3_entries); - } - - if(ret) return ret; - } - - if(ctx->stored.phys) - { - write_u32(data, ctx->phys.ppu_x); - write_u32(data + 4, ctx->phys.ppu_y); - data[8] = ctx->phys.unit_specifier; - - ret = write_chunk(ctx, type_phys, data, 9); - if(ret) return ret; - } - - if(ctx->stored.splt) - { - const struct spng_splt *splt; - unsigned char *cdata = NULL; - - uint32_t k; - for(i=0; i < ctx->n_splt; i++) - { - splt = &ctx->splt_list[i]; - - size_t name_len = strlen(splt->name); - length = name_len + 1; - - if(splt->sample_depth == 8) length += splt->n_entries * 6 + 1; - else if(splt->sample_depth == 16) length += splt->n_entries * 10 + 1; - - ret = write_header(ctx, type_splt, length, &cdata); - if(ret) return ret; - - memcpy(cdata, splt->name, name_len + 1); - cdata += name_len + 2; - cdata[-1] = splt->sample_depth; - - if(splt->sample_depth == 8) - { - for(k=0; k < splt->n_entries; k++) - { - cdata[k * 6 + 0] = splt->entries[k].red; - cdata[k * 6 + 1] = splt->entries[k].green; - cdata[k * 6 + 2] = splt->entries[k].blue; - cdata[k * 6 + 3] = splt->entries[k].alpha; - write_u16(cdata + k * 6 + 4, splt->entries[k].frequency); - } - } - else if(splt->sample_depth == 16) - { - for(k=0; k < splt->n_entries; k++) - { - write_u16(cdata + k * 10 + 0, splt->entries[k].red); - write_u16(cdata + k * 10 + 2, splt->entries[k].green); - write_u16(cdata + k * 10 + 4, splt->entries[k].blue); - write_u16(cdata + k * 10 + 6, splt->entries[k].alpha); - write_u16(cdata + k * 10 + 8, splt->entries[k].frequency); - } - } - - ret = finish_chunk(ctx); - if(ret) return ret; - } - } - - if(ctx->stored.time) - { - write_u16(data, ctx->time.year); - data[2] = ctx->time.month; - data[3] = ctx->time.day; - data[4] = ctx->time.hour; - data[5] = ctx->time.minute; - data[6] = ctx->time.second; - - ret = write_chunk(ctx, type_time, data, 7); - if(ret) return ret; - } - - if(ctx->stored.text) - { - unsigned char *cdata = NULL; - const struct spng_text2 *text; - const uint8_t *text_type_array[4] = { 0, type_text, type_ztxt, type_itxt }; - - for(i=0; i < ctx->n_text; i++) - { - text = &ctx->text_list[i]; - - const uint8_t *text_chunk_type = text_type_array[text->type]; - Bytef *compressed_text = NULL; - size_t keyword_len = 0; - size_t language_tag_len = 0; - size_t translated_keyword_len = 0; - size_t compressed_length = 0; - size_t text_length = 0; - - keyword_len = strlen(text->keyword); - text_length = strlen(text->text); - - length = keyword_len + 1; - - if(text->type == SPNG_ZTXT) - { - length += 1; /* compression method */ - } - else if(text->type == SPNG_ITXT) - { - if(!text->language_tag || !text->translated_keyword) return SPNG_EINTERNAL; - - language_tag_len = strlen(text->language_tag); - translated_keyword_len = strlen(text->translated_keyword); - - length += language_tag_len; - if(length < language_tag_len) return SPNG_EOVERFLOW; - - length += translated_keyword_len; - if(length < translated_keyword_len) return SPNG_EOVERFLOW; - - length += 4; /* compression flag + method + nul for the two strings */ - if(length < 4) return SPNG_EOVERFLOW; - } - - if(text->compression_flag) - { - ret = spng__deflate_init(ctx, &ctx->text_options); - if(ret) return ret; - - z_stream *zstream = &ctx->zstream; - uLongf dest_len = deflateBound(zstream, (uLong)text_length); - - compressed_text = spng__malloc(ctx, dest_len); - - if(compressed_text == NULL) return SPNG_EMEM; - - zstream->next_in = (void*)text->text; - zstream->avail_in = (uInt)text_length; - - zstream->next_out = compressed_text; - zstream->avail_out = dest_len; - - ret = deflate(zstream, Z_FINISH); - - if(ret != Z_STREAM_END) - { - spng__free(ctx, compressed_text); - return SPNG_EZLIB; - } - - compressed_length = zstream->total_out; - - length += compressed_length; - if(length < compressed_length) return SPNG_EOVERFLOW; - } - else - { - text_length = strlen(text->text); - - length += text_length; - if(length < text_length) return SPNG_EOVERFLOW; - } - - ret = write_header(ctx, text_chunk_type, length, &cdata); - if(ret) - { - spng__free(ctx, compressed_text); - return ret; - } - - memcpy(cdata, text->keyword, keyword_len + 1); - cdata += keyword_len + 1; - - if(text->type == SPNG_ITXT) - { - cdata[0] = text->compression_flag; - cdata[1] = 0; /* compression method */ - cdata += 2; - - memcpy(cdata, text->language_tag, language_tag_len + 1); - cdata += language_tag_len + 1; - - memcpy(cdata, text->translated_keyword, translated_keyword_len + 1); - cdata += translated_keyword_len + 1; - } - else if(text->type == SPNG_ZTXT) - { - cdata[0] = 0; /* compression method */ - cdata++; - } - - if(text->compression_flag) memcpy(cdata, compressed_text, compressed_length); - else memcpy(cdata, text->text, text_length); - - spng__free(ctx, compressed_text); - - ret = finish_chunk(ctx); - if(ret) return ret; - } - } - - if(ctx->stored.offs) - { - write_s32(data, ctx->offs.x); - write_s32(data + 4, ctx->offs.y); - data[8] = ctx->offs.unit_specifier; - - ret = write_chunk(ctx, type_offs, data, 9); - if(ret) return ret; - } - - if(ctx->stored.exif) - { - ret = write_chunk(ctx, type_exif, ctx->exif.data, ctx->exif.length); - if(ret) return ret; - } - - ret = write_unknown_chunks(ctx, SPNG_AFTER_PLTE); - if(ret) return ret; - - return 0; -} - -static int write_chunks_after_idat(spng_ctx *ctx) -{ - if(ctx == NULL) return SPNG_EINTERNAL; - - int ret = write_unknown_chunks(ctx, SPNG_AFTER_IDAT); - if(ret) return ret; - - return write_iend(ctx); -} - -/* Compress and write scanline to IDAT stream */ -static int write_idat_bytes(spng_ctx *ctx, const void *scanline, size_t len, int flush) -{ - if(ctx == NULL || scanline == NULL) return SPNG_EINTERNAL; - if(len > UINT_MAX) return SPNG_EINTERNAL; - - int ret = 0; - unsigned char *data = NULL; - z_stream *zstream = &ctx->zstream; - uint32_t idat_length = SPNG_WRITE_SIZE; - - zstream->next_in = scanline; - zstream->avail_in = (uInt)len; - - do - { - ret = deflate(zstream, flush); - - if(zstream->avail_out == 0) - { - ret = finish_chunk(ctx); - if(ret) return encode_err(ctx, ret); - - ret = write_header(ctx, type_idat, idat_length, &data); - if(ret) return encode_err(ctx, ret); - - zstream->next_out = data; - zstream->avail_out = idat_length; - } - - }while(zstream->avail_in); - - if(ret != Z_OK) return SPNG_EZLIB; - - return 0; -} - -static int finish_idat(spng_ctx *ctx) -{ - int ret = 0; - unsigned char *data = NULL; - z_stream *zstream = &ctx->zstream; - uint32_t idat_length = SPNG_WRITE_SIZE; - - while(ret != Z_STREAM_END) - { - ret = deflate(zstream, Z_FINISH); - - if(ret) - { - if(ret == Z_STREAM_END) break; - - if(ret != Z_BUF_ERROR) return SPNG_EZLIB; - } - - if(zstream->avail_out == 0) - { - ret = finish_chunk(ctx); - if(ret) return encode_err(ctx, ret); - - ret = write_header(ctx, type_idat, idat_length, &data); - if(ret) return encode_err(ctx, ret); - - zstream->next_out = data; - zstream->avail_out = idat_length; - } - } - - uint32_t trimmed_length = idat_length - zstream->avail_out; - - ret = trim_chunk(ctx, trimmed_length); - if(ret) return ret; - - return finish_chunk(ctx); -} - -static int encode_scanline(spng_ctx *ctx, const void *scanline, size_t len) -{ - if(ctx == NULL || scanline == NULL) return SPNG_EINTERNAL; - - int ret, pass = ctx->row_info.pass; - uint8_t filter = 0; - struct spng_row_info *ri = &ctx->row_info; - const struct spng_subimage *sub = ctx->subimage; - struct encode_flags f = ctx->encode_flags; - unsigned char *filtered_scanline = ctx->filtered_scanline; - size_t scanline_width = sub[pass].scanline_width; - - if(len < scanline_width - 1) return SPNG_EINTERNAL; - - /* encode_row() interlaces directly to ctx->scanline */ - if(scanline != ctx->scanline) memcpy(ctx->scanline, scanline, scanline_width - 1); - - if(f.to_bigendian) u16_row_to_bigendian(ctx->scanline, scanline_width - 1); - const int requires_previous = f.filter_choice & (SPNG_FILTER_CHOICE_UP | SPNG_FILTER_CHOICE_AVG | SPNG_FILTER_CHOICE_PAETH); - - /* XXX: exclude 'requires_previous' filters by default for first scanline? */ - if(!ri->scanline_idx && requires_previous) - { - /* prev_scanline is all zeros for the first scanline */ - memset(ctx->prev_scanline, 0, scanline_width); - } - - filter = get_best_filter(ctx->prev_scanline, ctx->scanline, scanline_width, ctx->bytes_per_pixel, f.filter_choice); - - if(!filter) filtered_scanline = ctx->scanline; - - filtered_scanline[-1] = filter; - - if(filter) - { - ret = filter_scanline(filtered_scanline, ctx->prev_scanline, ctx->scanline, scanline_width, ctx->bytes_per_pixel, filter); - if(ret) return encode_err(ctx, ret); - } - - ret = write_idat_bytes(ctx, filtered_scanline - 1, scanline_width, Z_NO_FLUSH); - if(ret) return encode_err(ctx, ret); - - /* The previous scanline is always unfiltered */ - void *t = ctx->prev_scanline; - ctx->prev_scanline = ctx->scanline; - ctx->scanline = t; - - ret = update_row_info(ctx); - - if(ret == SPNG_EOI) - { - int error = finish_idat(ctx); - if(error) encode_err(ctx, error); - - if(f.finalize) - { - error = spng_encode_chunks(ctx); - if(error) return encode_err(ctx, error); - } - } - - return ret; -} - -static int encode_row(spng_ctx *ctx, const void *row, size_t len) -{ - if(ctx == NULL || row == NULL) return SPNG_EINTERNAL; - - const int pass = ctx->row_info.pass; - - if(!ctx->ihdr.interlace_method || pass == 6) return encode_scanline(ctx, row, len); - - uint32_t k; - const unsigned pixel_size = ctx->pixel_size; - const unsigned bit_depth = ctx->ihdr.bit_depth; - - if(bit_depth < 8) - { - const unsigned samples_per_byte = 8 / bit_depth; - const uint8_t mask = (1 << bit_depth) - 1; - const unsigned initial_shift = 8 - bit_depth; - unsigned shift_amount = initial_shift; - - unsigned char *scanline = ctx->scanline; - const unsigned char *row_uc = row; - uint8_t sample; - - memset(scanline, 0, ctx->subimage[pass].scanline_width); - - for(k=0; k < ctx->subimage[pass].width; k++) - { - size_t ioffset = adam7_x_start[pass] + k * adam7_x_delta[pass]; - - sample = row_uc[ioffset / samples_per_byte]; - - sample = sample >> (initial_shift - ioffset * bit_depth % 8); - sample = sample & mask; - sample = sample << shift_amount; - - scanline[0] |= sample; - - shift_amount -= bit_depth; - - if(shift_amount > 7) - { - shift_amount = initial_shift; - scanline++; - } - } - - return encode_scanline(ctx, ctx->scanline, len); - } - - for(k=0; k < ctx->subimage[pass].width; k++) - { - size_t ioffset = (adam7_x_start[pass] + (size_t) k * adam7_x_delta[pass]) * pixel_size; - - memcpy(ctx->scanline + k * pixel_size, (unsigned char*)row + ioffset, pixel_size); - } - - return encode_scanline(ctx, ctx->scanline, len); -} - -int spng_encode_scanline(spng_ctx *ctx, const void *scanline, size_t len) -{ - if(ctx == NULL || scanline == NULL) return SPNG_EINVAL; - if(ctx->state >= SPNG_STATE_EOI) return SPNG_EOI; - if(len < (ctx->subimage[ctx->row_info.pass].scanline_width -1) ) return SPNG_EBUFSIZ; - - return encode_scanline(ctx, scanline, len); -} - -int spng_encode_row(spng_ctx *ctx, const void *row, size_t len) -{ - if(ctx == NULL || row == NULL) return SPNG_EINVAL; - if(ctx->state >= SPNG_STATE_EOI) return SPNG_EOI; - if(len < ctx->image_width) return SPNG_EBUFSIZ; - - return encode_row(ctx, row, len); -} - -int spng_encode_chunks(spng_ctx *ctx) -{ - if(ctx == NULL) return 1; - if(!ctx->state) return SPNG_EBADSTATE; - if(ctx->state < SPNG_STATE_OUTPUT) return SPNG_ENODST; - if(!ctx->encode_only) return SPNG_ECTXTYPE; - - int ret = 0; - - if(ctx->state < SPNG_STATE_FIRST_IDAT) - { - if(!ctx->stored.ihdr) return SPNG_ENOIHDR; - - ret = write_chunks_before_idat(ctx); - if(ret) return encode_err(ctx, ret); - - ctx->state = SPNG_STATE_FIRST_IDAT; - } - else if(ctx->state == SPNG_STATE_FIRST_IDAT) - { - return 0; - } - else if(ctx->state == SPNG_STATE_EOI) - { - ret = write_chunks_after_idat(ctx); - if(ret) return encode_err(ctx, ret); - - ctx->state = SPNG_STATE_IEND; - } - else return SPNG_EOPSTATE; - - return 0; -} - -int spng_encode_image(spng_ctx *ctx, const void *img, size_t len, int fmt, int flags) -{ - if(ctx == NULL) return 1; - if(!ctx->state) return SPNG_EBADSTATE; - if(!ctx->encode_only) return SPNG_ECTXTYPE; - if(!ctx->stored.ihdr) return SPNG_ENOIHDR; - if( !(fmt == SPNG_FMT_PNG || fmt == SPNG_FMT_RAW) ) return SPNG_EFMT; - - int ret = 0; - const struct spng_ihdr *ihdr = &ctx->ihdr; - struct encode_flags *encode_flags = &ctx->encode_flags; - - if(ihdr->color_type == SPNG_COLOR_TYPE_INDEXED && !ctx->stored.plte) return SPNG_ENOPLTE; - - ret = calculate_image_width(ihdr, fmt, &ctx->image_width); - if(ret) return encode_err(ctx, ret); - - if(ctx->image_width > SIZE_MAX / ihdr->height) ctx->image_size = 0; /* overflow */ - else ctx->image_size = ctx->image_width * ihdr->height; - - if( !(flags & SPNG_ENCODE_PROGRESSIVE) ) - { - if(img == NULL) return 1; - if(!ctx->image_size) return SPNG_EOVERFLOW; - if(len != ctx->image_size) return SPNG_EBUFSIZ; - } - - ret = spng_encode_chunks(ctx); - if(ret) return encode_err(ctx, ret); - - ret = calculate_subimages(ctx); - if(ret) return encode_err(ctx, ret); - - if(ihdr->bit_depth < 8) ctx->bytes_per_pixel = 1; - else ctx->bytes_per_pixel = num_channels(ihdr) * (ihdr->bit_depth / 8); - - if(spng__optimize(SPNG_FILTER_CHOICE)) - { - /* Filtering would make no difference */ - if(!ctx->image_options.compression_level) - { - encode_flags->filter_choice = SPNG_DISABLE_FILTERING; - } - - /* Palette indices and low bit-depth images do not benefit from filtering */ - if(ihdr->color_type == SPNG_COLOR_TYPE_INDEXED || ihdr->bit_depth < 8) - { - encode_flags->filter_choice = SPNG_DISABLE_FILTERING; - } - } - - /* This is technically the same as disabling filtering */ - if(encode_flags->filter_choice == SPNG_FILTER_CHOICE_NONE) - { - encode_flags->filter_choice = SPNG_DISABLE_FILTERING; - } - - if(!encode_flags->filter_choice && spng__optimize(SPNG_IMG_COMPRESSION_STRATEGY)) - { - ctx->image_options.strategy = Z_DEFAULT_STRATEGY; - } - - ret = spng__deflate_init(ctx, &ctx->image_options); - if(ret) return encode_err(ctx, ret); - - size_t scanline_buf_size = ctx->subimage[ctx->widest_pass].scanline_width; - - scanline_buf_size += 32; - - if(scanline_buf_size < 32) return SPNG_EOVERFLOW; - - ctx->scanline_buf = spng__malloc(ctx, scanline_buf_size); - ctx->prev_scanline_buf = spng__malloc(ctx, scanline_buf_size); - - if(ctx->scanline_buf == NULL || ctx->prev_scanline_buf == NULL) return encode_err(ctx, SPNG_EMEM); - - /* Maintain alignment for pixels, filter at [-1] */ - ctx->scanline = ctx->scanline_buf + 16; - ctx->prev_scanline = ctx->prev_scanline_buf + 16; - - if(encode_flags->filter_choice) - { - ctx->filtered_scanline_buf = spng__malloc(ctx, scanline_buf_size); - if(ctx->filtered_scanline_buf == NULL) return encode_err(ctx, SPNG_EMEM); - - ctx->filtered_scanline = ctx->filtered_scanline_buf + 16; - } - - struct spng_subimage *sub = ctx->subimage; - struct spng_row_info *ri = &ctx->row_info; - - ctx->fmt = fmt; - - z_stream *zstream = &ctx->zstream; - zstream->avail_out = SPNG_WRITE_SIZE; - - ret = write_header(ctx, type_idat, zstream->avail_out, &zstream->next_out); - if(ret) return encode_err(ctx, ret); - - if(ihdr->interlace_method) encode_flags->interlace = 1; - - if(fmt & (SPNG_FMT_PNG | SPNG_FMT_RAW) ) encode_flags->same_layout = 1; - - if(ihdr->bit_depth == 16 && fmt != SPNG_FMT_RAW) encode_flags->to_bigendian = 1; - - if(flags & SPNG_ENCODE_FINALIZE) encode_flags->finalize = 1; - - while(!sub[ri->pass].width || !sub[ri->pass].height) ri->pass++; - - if(encode_flags->interlace) ri->row_num = adam7_y_start[ri->pass]; - - ctx->pixel_size = 4; /* SPNG_FMT_RGBA8 */ - - if(fmt == SPNG_FMT_RGBA16) ctx->pixel_size = 8; - else if(fmt == SPNG_FMT_RGB8) ctx->pixel_size = 3; - else if(fmt == SPNG_FMT_G8) ctx->pixel_size = 1; - else if(fmt == SPNG_FMT_GA8) ctx->pixel_size = 2; - else if(fmt & (SPNG_FMT_PNG | SPNG_FMT_RAW)) ctx->pixel_size = ctx->bytes_per_pixel; - - ctx->state = SPNG_STATE_ENCODE_INIT; - - if(flags & SPNG_ENCODE_PROGRESSIVE) - { - encode_flags->progressive = 1; - - return 0; - } - - do - { - size_t ioffset = ri->row_num * ctx->image_width; - - ret = encode_row(ctx, (unsigned char*)img + ioffset, ctx->image_width); - - }while(!ret); - - if(ret != SPNG_EOI) return encode_err(ctx, ret); - - return 0; -} - -spng_ctx *spng_ctx_new(int flags) -{ - struct spng_alloc alloc = - { - .malloc_fn = malloc, - .realloc_fn = realloc, - .calloc_fn = calloc, - .free_fn = free - }; - - return spng_ctx_new2(&alloc, flags); -} - -spng_ctx *spng_ctx_new2(struct spng_alloc *alloc, int flags) -{ - if(alloc == NULL) return NULL; - if(flags != (flags & SPNG__CTX_FLAGS_ALL)) return NULL; - - if(alloc->malloc_fn == NULL) return NULL; - if(alloc->realloc_fn == NULL) return NULL; - if(alloc->calloc_fn == NULL) return NULL; - if(alloc->free_fn == NULL) return NULL; - - spng_ctx *ctx = alloc->calloc_fn(1, sizeof(spng_ctx)); - if(ctx == NULL) return NULL; - - ctx->alloc = *alloc; - - ctx->max_width = spng_u32max; - ctx->max_height = spng_u32max; - - ctx->max_chunk_size = spng_u32max; - ctx->chunk_cache_limit = SIZE_MAX; - ctx->chunk_count_limit = SPNG_MAX_CHUNK_COUNT; - - ctx->state = SPNG_STATE_INIT; - - ctx->crc_action_critical = SPNG_CRC_ERROR; - ctx->crc_action_ancillary = SPNG_CRC_DISCARD; - - const struct spng__zlib_options image_defaults = - { - .compression_level = Z_DEFAULT_COMPRESSION, - .window_bits = 15, - .mem_level = 8, - .strategy = Z_FILTERED, - .data_type = 0 /* Z_BINARY */ - }; - - const struct spng__zlib_options text_defaults = - { - .compression_level = Z_DEFAULT_COMPRESSION, - .window_bits = 15, - .mem_level = 8, - .strategy = Z_DEFAULT_STRATEGY, - .data_type = 1 /* Z_TEXT */ - }; - - ctx->image_options = image_defaults; - ctx->text_options = text_defaults; - - ctx->optimize_option = ~0; - ctx->encode_flags.filter_choice = SPNG_FILTER_CHOICE_ALL; - - ctx->flags = flags; - - if(flags & SPNG_CTX_ENCODER) ctx->encode_only = 1; - - return ctx; -} - -void spng_ctx_free(spng_ctx *ctx) -{ - if(ctx == NULL) return; - - if(ctx->streaming && ctx->stream_buf != NULL) spng__free(ctx, ctx->stream_buf); - - if(!ctx->user.exif) spng__free(ctx, ctx->exif.data); - - if(!ctx->user.iccp) spng__free(ctx, ctx->iccp.profile); - - uint32_t i; - - if(ctx->splt_list != NULL && !ctx->user.splt) - { - for(i=0; i < ctx->n_splt; i++) - { - spng__free(ctx, ctx->splt_list[i].entries); - } - spng__free(ctx, ctx->splt_list); - } - - if(ctx->text_list != NULL) - { - for(i=0; i< ctx->n_text; i++) - { - if(ctx->user.text) break; - - spng__free(ctx, ctx->text_list[i].keyword); - if(ctx->text_list[i].compression_flag) spng__free(ctx, ctx->text_list[i].text); - } - spng__free(ctx, ctx->text_list); - } - - if(ctx->chunk_list != NULL && !ctx->user.unknown) - { - for(i=0; i< ctx->n_chunks; i++) - { - spng__free(ctx, ctx->chunk_list[i].data); - } - spng__free(ctx, ctx->chunk_list); - } - - if(ctx->deflate) deflateEnd(&ctx->zstream); - else inflateEnd(&ctx->zstream); - - if(!ctx->user_owns_out_png) spng__free(ctx, ctx->out_png); - - spng__free(ctx, ctx->gamma_lut16); - - spng__free(ctx, ctx->row_buf); - spng__free(ctx, ctx->scanline_buf); - spng__free(ctx, ctx->prev_scanline_buf); - spng__free(ctx, ctx->filtered_scanline_buf); - - spng_free_fn *free_fn = ctx->alloc.free_fn; - - memset(ctx, 0, sizeof(spng_ctx)); - - free_fn(ctx); -} - -static int buffer_read_fn(spng_ctx *ctx, void *user, void *data, size_t n) -{ - if(n > ctx->bytes_left) return SPNG_IO_EOF; - - (void)user; - (void)data; - ctx->data = ctx->data + ctx->last_read_size; - - ctx->last_read_size = n; - ctx->bytes_left -= n; - - return 0; -} - -static int file_read_fn(spng_ctx *ctx, void *user, void *data, size_t n) -{ - FILE *file = user; - (void)ctx; - - if(fread(data, n, 1, file) != 1) - { - if(feof(file)) return SPNG_IO_EOF; - else return SPNG_IO_ERROR; - } - - return 0; -} - -static int file_write_fn(spng_ctx *ctx, void *user, void *data, size_t n) -{ - FILE *file = user; - (void)ctx; - - if(fwrite(data, n, 1, file) != 1) return SPNG_IO_ERROR; - - return 0; -} - -int spng_set_png_buffer(spng_ctx *ctx, const void *buf, size_t size) -{ - if(ctx == NULL || buf == NULL) return 1; - if(!ctx->state) return SPNG_EBADSTATE; - if(ctx->encode_only) return SPNG_ECTXTYPE; /* not supported */ - - if(ctx->data != NULL) return SPNG_EBUF_SET; - - ctx->data = buf; - ctx->png_base = buf; - ctx->data_size = size; - ctx->bytes_left = size; - - ctx->read_fn = buffer_read_fn; - - ctx->state = SPNG_STATE_INPUT; - - return 0; -} - -int spng_set_png_stream(spng_ctx *ctx, spng_rw_fn *rw_func, void *user) -{ - if(ctx == NULL || rw_func == NULL) return 1; - if(!ctx->state) return SPNG_EBADSTATE; - - /* SPNG_STATE_OUTPUT shares the same value */ - if(ctx->state >= SPNG_STATE_INPUT) return SPNG_EBUF_SET; - - if(ctx->encode_only) - { - if(ctx->out_png != NULL) return SPNG_EBUF_SET; - - ctx->write_fn = rw_func; - ctx->write_ptr = ctx->stream_buf; - - ctx->state = SPNG_STATE_OUTPUT; - } - else - { - ctx->stream_buf = spng__malloc(ctx, SPNG_READ_SIZE); - if(ctx->stream_buf == NULL) return SPNG_EMEM; - - ctx->read_fn = rw_func; - ctx->data = ctx->stream_buf; - ctx->data_size = SPNG_READ_SIZE; - - ctx->state = SPNG_STATE_INPUT; - } - - ctx->stream_user_ptr = user; - - ctx->streaming = 1; - - return 0; -} - -int spng_set_png_file(spng_ctx *ctx, FILE *file) -{ - if(file == NULL) return 1; - - if(ctx->encode_only) return spng_set_png_stream(ctx, file_write_fn, file); - - return spng_set_png_stream(ctx, file_read_fn, file); -} - -void *spng_get_png_buffer(spng_ctx *ctx, size_t *len, int *error) -{ - int tmp = 0; - error = error ? error : &tmp; - *error = 0; - - if(ctx == NULL || !len) *error = SPNG_EINVAL; - - if(*error) return NULL; - - if(!ctx->encode_only) *error = SPNG_ECTXTYPE; - else if(!ctx->state) *error = SPNG_EBADSTATE; - else if(!ctx->internal_buffer) *error = SPNG_EOPSTATE; - else if(ctx->state < SPNG_STATE_EOI) *error = SPNG_EOPSTATE; - else if(ctx->state != SPNG_STATE_IEND) *error = SPNG_ENOTFINAL; - - if(*error) return NULL; - - ctx->user_owns_out_png = 1; - - *len = ctx->bytes_encoded; - - return ctx->out_png; -} - -int spng_set_image_limits(spng_ctx *ctx, uint32_t width, uint32_t height) -{ - if(ctx == NULL) return 1; - - if(width > spng_u32max || height > spng_u32max) return 1; - - ctx->max_width = width; - ctx->max_height = height; - - return 0; -} - -int spng_get_image_limits(spng_ctx *ctx, uint32_t *width, uint32_t *height) -{ - if(ctx == NULL || width == NULL || height == NULL) return 1; - - *width = ctx->max_width; - *height = ctx->max_height; - - return 0; -} - -int spng_set_chunk_limits(spng_ctx *ctx, size_t chunk_size, size_t cache_limit) -{ - if(ctx == NULL || chunk_size > spng_u32max || chunk_size > cache_limit) return 1; - - ctx->max_chunk_size = chunk_size; - - ctx->chunk_cache_limit = cache_limit; - - return 0; -} - -int spng_get_chunk_limits(spng_ctx *ctx, size_t *chunk_size, size_t *cache_limit) -{ - if(ctx == NULL || chunk_size == NULL || cache_limit == NULL) return 1; - - *chunk_size = ctx->max_chunk_size; - - *cache_limit = ctx->chunk_cache_limit; - - return 0; -} - -int spng_set_crc_action(spng_ctx *ctx, int critical, int ancillary) -{ - if(ctx == NULL) return 1; - if(ctx->encode_only) return SPNG_ECTXTYPE; - - if(critical > 2 || critical < 0) return 1; - if(ancillary > 2 || ancillary < 0) return 1; - - if(critical == SPNG_CRC_DISCARD) return 1; - - ctx->crc_action_critical = critical; - ctx->crc_action_ancillary = ancillary; - - return 0; -} - -int spng_set_option(spng_ctx *ctx, enum spng_option option, int value) -{ - if(ctx == NULL) return 1; - if(!ctx->state) return SPNG_EBADSTATE; - - switch(option) - { - case SPNG_KEEP_UNKNOWN_CHUNKS: - { - ctx->keep_unknown = value ? 1 : 0; - break; - } - case SPNG_IMG_COMPRESSION_LEVEL: - { - ctx->image_options.compression_level = value; - break; - } - case SPNG_IMG_WINDOW_BITS: - { - ctx->image_options.window_bits = value; - break; - } - case SPNG_IMG_MEM_LEVEL: - { - ctx->image_options.mem_level = value; - break; - } - case SPNG_IMG_COMPRESSION_STRATEGY: - { - ctx->image_options.strategy = value; - break; - } - case SPNG_TEXT_COMPRESSION_LEVEL: - { - ctx->text_options.compression_level = value; - break; - } - case SPNG_TEXT_WINDOW_BITS: - { - ctx->text_options.window_bits = value; - break; - } - case SPNG_TEXT_MEM_LEVEL: - { - ctx->text_options.mem_level = value; - break; - } - case SPNG_TEXT_COMPRESSION_STRATEGY: - { - ctx->text_options.strategy = value; - break; - } - case SPNG_FILTER_CHOICE: - { - if(value & ~SPNG_FILTER_CHOICE_ALL) return 1; - ctx->encode_flags.filter_choice = value; - break; - } - case SPNG_CHUNK_COUNT_LIMIT: - { - if(value < 0) return 1; - if(value > (int)ctx->chunk_count_total) return 1; - ctx->chunk_count_limit = value; - break; - } - case SPNG_ENCODE_TO_BUFFER: - { - if(value < 0) return 1; - if(!ctx->encode_only) return SPNG_ECTXTYPE; - if(ctx->state >= SPNG_STATE_OUTPUT) return SPNG_EOPSTATE; - - if(!value) break; - - ctx->internal_buffer = 1; - ctx->state = SPNG_STATE_OUTPUT; - - break; - } - default: return 1; - } - - /* Option can no longer be overriden by the library */ - if(option < 32) ctx->optimize_option &= ~(1 << option); - - return 0; -} - -int spng_get_option(spng_ctx *ctx, enum spng_option option, int *value) -{ - if(ctx == NULL || value == NULL) return 1; - if(!ctx->state) return SPNG_EBADSTATE; - - switch(option) - { - case SPNG_KEEP_UNKNOWN_CHUNKS: - { - *value = ctx->keep_unknown; - break; - } - case SPNG_IMG_COMPRESSION_LEVEL: - { - *value = ctx->image_options.compression_level; - break; - } - case SPNG_IMG_WINDOW_BITS: - { - *value = ctx->image_options.window_bits; - break; - } - case SPNG_IMG_MEM_LEVEL: - { - *value = ctx->image_options.mem_level; - break; - } - case SPNG_IMG_COMPRESSION_STRATEGY: - { - *value = ctx->image_options.strategy; - break; - } - case SPNG_TEXT_COMPRESSION_LEVEL: - { - *value = ctx->text_options.compression_level; - break; - } - case SPNG_TEXT_WINDOW_BITS: - { - *value = ctx->text_options.window_bits; - break; - } - case SPNG_TEXT_MEM_LEVEL: - { - *value = ctx->text_options.mem_level; - break; - } - case SPNG_TEXT_COMPRESSION_STRATEGY: - { - *value = ctx->text_options.strategy; - break; - } - case SPNG_FILTER_CHOICE: - { - *value = ctx->encode_flags.filter_choice; - break; - } - case SPNG_CHUNK_COUNT_LIMIT: - { - *value = ctx->chunk_count_limit; - break; - } - case SPNG_ENCODE_TO_BUFFER: - { - if(ctx->internal_buffer) *value = 1; - else *value = 0; - - break; - } - default: return 1; - } - - return 0; -} - -int spng_decoded_image_size(spng_ctx *ctx, int fmt, size_t *len) -{ - if(ctx == NULL || len == NULL) return 1; - - int ret = read_chunks(ctx, 1); - if(ret) return ret; - - ret = check_decode_fmt(&ctx->ihdr, fmt); - if(ret) return ret; - - return calculate_image_size(&ctx->ihdr, fmt, len); -} - -int spng_get_ihdr(spng_ctx *ctx, struct spng_ihdr *ihdr) -{ - if(ctx == NULL) return 1; - int ret = read_chunks(ctx, 1); - if(ret) return ret; - if(ihdr == NULL) return 1; - - *ihdr = ctx->ihdr; - - return 0; -} - -int spng_get_plte(spng_ctx *ctx, struct spng_plte *plte) -{ - SPNG_GET_CHUNK_BOILERPLATE(plte); - - *plte = ctx->plte; - - return 0; -} - -int spng_get_trns(spng_ctx *ctx, struct spng_trns *trns) -{ - SPNG_GET_CHUNK_BOILERPLATE(trns); - - *trns = ctx->trns; - - return 0; -} - -int spng_get_chrm(spng_ctx *ctx, struct spng_chrm *chrm) -{ - SPNG_GET_CHUNK_BOILERPLATE(chrm); - - chrm->white_point_x = (double)ctx->chrm_int.white_point_x / 100000.0; - chrm->white_point_y = (double)ctx->chrm_int.white_point_y / 100000.0; - chrm->red_x = (double)ctx->chrm_int.red_x / 100000.0; - chrm->red_y = (double)ctx->chrm_int.red_y / 100000.0; - chrm->blue_y = (double)ctx->chrm_int.blue_y / 100000.0; - chrm->blue_x = (double)ctx->chrm_int.blue_x / 100000.0; - chrm->green_x = (double)ctx->chrm_int.green_x / 100000.0; - chrm->green_y = (double)ctx->chrm_int.green_y / 100000.0; - - return 0; -} - -int spng_get_chrm_int(spng_ctx *ctx, struct spng_chrm_int *chrm) -{ - SPNG_GET_CHUNK_BOILERPLATE(chrm); - - *chrm = ctx->chrm_int; - - return 0; -} - -int spng_get_gama(spng_ctx *ctx, double *gamma) -{ - double *gama = gamma; - SPNG_GET_CHUNK_BOILERPLATE(gama); - - *gama = (double)ctx->gama / 100000.0; - - return 0; -} - -int spng_get_gama_int(spng_ctx *ctx, uint32_t *gama_int) -{ - uint32_t *gama = gama_int; - SPNG_GET_CHUNK_BOILERPLATE(gama); - - *gama_int = ctx->gama; - - return 0; -} - -int spng_get_iccp(spng_ctx *ctx, struct spng_iccp *iccp) -{ - SPNG_GET_CHUNK_BOILERPLATE(iccp); - - *iccp = ctx->iccp; - - return 0; -} - -int spng_get_sbit(spng_ctx *ctx, struct spng_sbit *sbit) -{ - SPNG_GET_CHUNK_BOILERPLATE(sbit); - - *sbit = ctx->sbit; - - return 0; -} - -int spng_get_srgb(spng_ctx *ctx, uint8_t *rendering_intent) -{ - uint8_t *srgb = rendering_intent; - SPNG_GET_CHUNK_BOILERPLATE(srgb); - - *srgb = ctx->srgb_rendering_intent; - - return 0; -} - -int spng_get_text(spng_ctx *ctx, struct spng_text *text, uint32_t *n_text) -{ - if(ctx == NULL) return 1; - int ret = read_chunks(ctx, 0); - if(ret) return ret; - if(!ctx->stored.text) return SPNG_ECHUNKAVAIL; - if(n_text == NULL) return 1; - - if(text == NULL) - { - *n_text = ctx->n_text; - return 0; - } - - if(*n_text < ctx->n_text) return 1; - - uint32_t i; - for(i=0; i< ctx->n_text; i++) - { - text[i].type = ctx->text_list[i].type; - memcpy(&text[i].keyword, ctx->text_list[i].keyword, strlen(ctx->text_list[i].keyword) + 1); - text[i].compression_method = 0; - text[i].compression_flag = ctx->text_list[i].compression_flag; - text[i].language_tag = ctx->text_list[i].language_tag; - text[i].translated_keyword = ctx->text_list[i].translated_keyword; - text[i].length = ctx->text_list[i].text_length; - text[i].text = ctx->text_list[i].text; - } - - return ret; -} - -int spng_get_bkgd(spng_ctx *ctx, struct spng_bkgd *bkgd) -{ - SPNG_GET_CHUNK_BOILERPLATE(bkgd); - - *bkgd = ctx->bkgd; - - return 0; -} - -int spng_get_hist(spng_ctx *ctx, struct spng_hist *hist) -{ - SPNG_GET_CHUNK_BOILERPLATE(hist); - - *hist = ctx->hist; - - return 0; -} - -int spng_get_phys(spng_ctx *ctx, struct spng_phys *phys) -{ - SPNG_GET_CHUNK_BOILERPLATE(phys); - - *phys = ctx->phys; - - return 0; -} - -int spng_get_splt(spng_ctx *ctx, struct spng_splt *splt, uint32_t *n_splt) -{ - if(ctx == NULL) return 1; - int ret = read_chunks(ctx, 0); - if(ret) return ret; - if(!ctx->stored.splt) return SPNG_ECHUNKAVAIL; - if(n_splt == NULL) return 1; - - if(splt == NULL) - { - *n_splt = ctx->n_splt; - return 0; - } - - if(*n_splt < ctx->n_splt) return 1; - - memcpy(splt, ctx->splt_list, ctx->n_splt * sizeof(struct spng_splt)); - - return 0; -} - -int spng_get_time(spng_ctx *ctx, struct spng_time *time) -{ - SPNG_GET_CHUNK_BOILERPLATE(time); - - *time = ctx->time; - - return 0; -} - -int spng_get_unknown_chunks(spng_ctx *ctx, struct spng_unknown_chunk *chunks, uint32_t *n_chunks) -{ - if(ctx == NULL) return 1; - int ret = read_chunks(ctx, 0); - if(ret) return ret; - if(!ctx->stored.unknown) return SPNG_ECHUNKAVAIL; - if(n_chunks == NULL) return 1; - - if(chunks == NULL) - { - *n_chunks = ctx->n_chunks; - return 0; - } - - if(*n_chunks < ctx->n_chunks) return 1; - - memcpy(chunks, ctx->chunk_list, sizeof(struct spng_unknown_chunk)); - - return 0; -} - -int spng_get_offs(spng_ctx *ctx, struct spng_offs *offs) -{ - SPNG_GET_CHUNK_BOILERPLATE(offs); - - *offs = ctx->offs; - - return 0; -} - -int spng_get_exif(spng_ctx *ctx, struct spng_exif *exif) -{ - SPNG_GET_CHUNK_BOILERPLATE(exif); - - *exif = ctx->exif; - - return 0; -} - -int spng_set_ihdr(spng_ctx *ctx, struct spng_ihdr *ihdr) -{ - SPNG_SET_CHUNK_BOILERPLATE(ihdr); - - if(ctx->stored.ihdr) return 1; - - ret = check_ihdr(ihdr, ctx->max_width, ctx->max_height); - if(ret) return ret; - - ctx->ihdr = *ihdr; - - ctx->stored.ihdr = 1; - ctx->user.ihdr = 1; - - return 0; -} - -int spng_set_plte(spng_ctx *ctx, struct spng_plte *plte) -{ - SPNG_SET_CHUNK_BOILERPLATE(plte); - - if(!ctx->stored.ihdr) return 1; - - if(check_plte(plte, &ctx->ihdr)) return 1; - - ctx->plte.n_entries = plte->n_entries; - - memcpy(ctx->plte.entries, plte->entries, plte->n_entries * sizeof(struct spng_plte_entry)); - - ctx->stored.plte = 1; - ctx->user.plte = 1; - - return 0; -} - -int spng_set_trns(spng_ctx *ctx, struct spng_trns *trns) -{ - SPNG_SET_CHUNK_BOILERPLATE(trns); - - if(!ctx->stored.ihdr) return SPNG_ENOIHDR; - - if(ctx->ihdr.color_type == SPNG_COLOR_TYPE_GRAYSCALE) - { - ctx->trns.gray = trns->gray; - } - else if(ctx->ihdr.color_type == SPNG_COLOR_TYPE_TRUECOLOR) - { - ctx->trns.red = trns->red; - ctx->trns.green = trns->green; - ctx->trns.blue = trns->blue; - } - else if(ctx->ihdr.color_type == SPNG_COLOR_TYPE_INDEXED) - { - if(!ctx->stored.plte) return SPNG_ETRNS_NO_PLTE; - if(trns->n_type3_entries > ctx->plte.n_entries) return 1; - - ctx->trns.n_type3_entries = trns->n_type3_entries; - memcpy(ctx->trns.type3_alpha, trns->type3_alpha, trns->n_type3_entries); - } - else return SPNG_ETRNS_COLOR_TYPE; - - ctx->stored.trns = 1; - ctx->user.trns = 1; - - return 0; -} - -int spng_set_chrm(spng_ctx *ctx, struct spng_chrm *chrm) -{ - SPNG_SET_CHUNK_BOILERPLATE(chrm); - - struct spng_chrm_int chrm_int; - - chrm_int.white_point_x = (uint32_t)(chrm->white_point_x * 100000.0); - chrm_int.white_point_y = (uint32_t)(chrm->white_point_y * 100000.0); - chrm_int.red_x = (uint32_t)(chrm->red_x * 100000.0); - chrm_int.red_y = (uint32_t)(chrm->red_y * 100000.0); - chrm_int.green_x = (uint32_t)(chrm->green_x * 100000.0); - chrm_int.green_y = (uint32_t)(chrm->green_y * 100000.0); - chrm_int.blue_x = (uint32_t)(chrm->blue_x * 100000.0); - chrm_int.blue_y = (uint32_t)(chrm->blue_y * 100000.0); - - if(check_chrm_int(&chrm_int)) return SPNG_ECHRM; - - ctx->chrm_int = chrm_int; - - ctx->stored.chrm = 1; - ctx->user.chrm = 1; - - return 0; -} - -int spng_set_chrm_int(spng_ctx *ctx, struct spng_chrm_int *chrm_int) -{ - SPNG_SET_CHUNK_BOILERPLATE(chrm_int); - - if(check_chrm_int(chrm_int)) return SPNG_ECHRM; - - ctx->chrm_int = *chrm_int; - - ctx->stored.chrm = 1; - ctx->user.chrm = 1; - - return 0; -} - -int spng_set_gama(spng_ctx *ctx, double gamma) -{ - SPNG_SET_CHUNK_BOILERPLATE(ctx); - - uint32_t gama = gamma * 100000.0; - - if(!gama) return 1; - if(gama > spng_u32max) return 1; - - ctx->gama = gama; - - ctx->stored.gama = 1; - ctx->user.gama = 1; - - return 0; -} - -int spng_set_gama_int(spng_ctx *ctx, uint32_t gamma) -{ - SPNG_SET_CHUNK_BOILERPLATE(ctx); - - if(!gamma) return 1; - if(gamma > spng_u32max) return 1; - - ctx->gama = gamma; - - ctx->stored.gama = 1; - ctx->user.gama = 1; - - return 0; -} - -int spng_set_iccp(spng_ctx *ctx, struct spng_iccp *iccp) -{ - SPNG_SET_CHUNK_BOILERPLATE(iccp); - - if(check_png_keyword(iccp->profile_name)) return SPNG_EICCP_NAME; - if(!iccp->profile_len) return SPNG_ECHUNK_SIZE; - if(iccp->profile_len > spng_u32max) return SPNG_ECHUNK_STDLEN; - - if(ctx->iccp.profile && !ctx->user.iccp) spng__free(ctx, ctx->iccp.profile); - - ctx->iccp = *iccp; - - ctx->stored.iccp = 1; - ctx->user.iccp = 1; - - return 0; -} - -int spng_set_sbit(spng_ctx *ctx, struct spng_sbit *sbit) -{ - SPNG_SET_CHUNK_BOILERPLATE(sbit); - - if(check_sbit(sbit, &ctx->ihdr)) return 1; - - if(!ctx->stored.ihdr) return 1; - - ctx->sbit = *sbit; - - ctx->stored.sbit = 1; - ctx->user.sbit = 1; - - return 0; -} - -int spng_set_srgb(spng_ctx *ctx, uint8_t rendering_intent) -{ - SPNG_SET_CHUNK_BOILERPLATE(ctx); - - if(rendering_intent > 3) return 1; - - ctx->srgb_rendering_intent = rendering_intent; - - ctx->stored.srgb = 1; - ctx->user.srgb = 1; - - return 0; -} - -int spng_set_text(spng_ctx *ctx, struct spng_text *text, uint32_t n_text) -{ - if(!n_text) return 1; - SPNG_SET_CHUNK_BOILERPLATE(text); - - uint32_t i; - for(i=0; i < n_text; i++) - { - if(check_png_keyword(text[i].keyword)) return SPNG_ETEXT_KEYWORD; - if(!text[i].length) return 1; - if(text[i].length > UINT_MAX) return 1; - if(text[i].text == NULL) return 1; - - if(text[i].type == SPNG_TEXT) - { - if(ctx->strict && check_png_text(text[i].text, text[i].length)) return 1; - } - else if(text[i].type == SPNG_ZTXT) - { - if(ctx->strict && check_png_text(text[i].text, text[i].length)) return 1; - - if(text[i].compression_method != 0) return SPNG_EZTXT_COMPRESSION_METHOD; - } - else if(text[i].type == SPNG_ITXT) - { - if(text[i].compression_flag > 1) return SPNG_EITXT_COMPRESSION_FLAG; - if(text[i].compression_method != 0) return SPNG_EITXT_COMPRESSION_METHOD; - if(text[i].language_tag == NULL) return SPNG_EITXT_LANG_TAG; - if(text[i].translated_keyword == NULL) return SPNG_EITXT_TRANSLATED_KEY; - } - else return 1; - - } - - struct spng_text2 *text_list = spng__calloc(ctx, sizeof(struct spng_text2), n_text); - - if(!text_list) return SPNG_EMEM; - - if(ctx->text_list != NULL) - { - for(i=0; i < ctx->n_text; i++) - { - if(ctx->user.text) break; - - spng__free(ctx, ctx->text_list[i].keyword); - if(ctx->text_list[i].compression_flag) spng__free(ctx, ctx->text_list[i].text); - } - spng__free(ctx, ctx->text_list); - } - - for(i=0; i < n_text; i++) - { - text_list[i].type = text[i].type; - /* Prevent issues with spng_text.keyword[80] going out of scope */ - text_list[i].keyword = text_list[i].user_keyword_storage; - memcpy(text_list[i].user_keyword_storage, text[i].keyword, strlen(text[i].keyword)); - text_list[i].text = text[i].text; - text_list[i].text_length = text[i].length; - - if(text[i].type == SPNG_ZTXT) - { - text_list[i].compression_flag = 1; - } - else if(text[i].type == SPNG_ITXT) - { - text_list[i].compression_flag = text[i].compression_flag; - text_list[i].language_tag = text[i].language_tag; - text_list[i].translated_keyword = text[i].translated_keyword; - } - } - - ctx->text_list = text_list; - ctx->n_text = n_text; - - ctx->stored.text = 1; - ctx->user.text = 1; - - return 0; -} - -int spng_set_bkgd(spng_ctx *ctx, struct spng_bkgd *bkgd) -{ - SPNG_SET_CHUNK_BOILERPLATE(bkgd); - - if(!ctx->stored.ihdr) return 1; - - if(ctx->ihdr.color_type == 0 || ctx->ihdr.color_type == 4) - { - ctx->bkgd.gray = bkgd->gray; - } - else if(ctx->ihdr.color_type == 2 || ctx->ihdr.color_type == 6) - { - ctx->bkgd.red = bkgd->red; - ctx->bkgd.green = bkgd->green; - ctx->bkgd.blue = bkgd->blue; - } - else if(ctx->ihdr.color_type == 3) - { - if(!ctx->stored.plte) return SPNG_EBKGD_NO_PLTE; - if(bkgd->plte_index >= ctx->plte.n_entries) return SPNG_EBKGD_PLTE_IDX; - - ctx->bkgd.plte_index = bkgd->plte_index; - } - - ctx->stored.bkgd = 1; - ctx->user.bkgd = 1; - - return 0; -} - -int spng_set_hist(spng_ctx *ctx, struct spng_hist *hist) -{ - SPNG_SET_CHUNK_BOILERPLATE(hist); - - if(!ctx->stored.plte) return SPNG_EHIST_NO_PLTE; - - ctx->hist = *hist; - - ctx->stored.hist = 1; - ctx->user.hist = 1; - - return 0; -} - -int spng_set_phys(spng_ctx *ctx, struct spng_phys *phys) -{ - SPNG_SET_CHUNK_BOILERPLATE(phys); - - if(check_phys(phys)) return SPNG_EPHYS; - - ctx->phys = *phys; - - ctx->stored.phys = 1; - ctx->user.phys = 1; - - return 0; -} - -int spng_set_splt(spng_ctx *ctx, struct spng_splt *splt, uint32_t n_splt) -{ - if(!n_splt) return 1; - SPNG_SET_CHUNK_BOILERPLATE(splt); - - uint32_t i; - for(i=0; i < n_splt; i++) - { - if(check_png_keyword(splt[i].name)) return SPNG_ESPLT_NAME; - if( !(splt[i].sample_depth == 8 || splt[i].sample_depth == 16) ) return SPNG_ESPLT_DEPTH; - } - - if(ctx->stored.splt && !ctx->user.splt) - { - for(i=0; i < ctx->n_splt; i++) - { - if(ctx->splt_list[i].entries != NULL) spng__free(ctx, ctx->splt_list[i].entries); - } - spng__free(ctx, ctx->splt_list); - } - - ctx->splt_list = splt; - ctx->n_splt = n_splt; - - ctx->stored.splt = 1; - ctx->user.splt = 1; - - return 0; -} - -int spng_set_time(spng_ctx *ctx, struct spng_time *time) -{ - SPNG_SET_CHUNK_BOILERPLATE(time); - - if(check_time(time)) return SPNG_ETIME; - - ctx->time = *time; - - ctx->stored.time = 1; - ctx->user.time = 1; - - return 0; -} - -int spng_set_unknown_chunks(spng_ctx *ctx, struct spng_unknown_chunk *chunks, uint32_t n_chunks) -{ - if(!n_chunks) return 1; - SPNG_SET_CHUNK_BOILERPLATE(chunks); - - uint32_t i; - for(i=0; i < n_chunks; i++) - { - if(chunks[i].length > spng_u32max) return SPNG_ECHUNK_STDLEN; - if(chunks[i].length && chunks[i].data == NULL) return 1; - - switch(chunks[i].location) - { - case SPNG_AFTER_IHDR: - case SPNG_AFTER_PLTE: - case SPNG_AFTER_IDAT: - break; - default: return SPNG_ECHUNK_POS; - } - } - - if(ctx->stored.unknown && !ctx->user.unknown) - { - for(i=0; i < ctx->n_chunks; i++) - { - spng__free(ctx, ctx->chunk_list[i].data); - } - spng__free(ctx, ctx->chunk_list); - } - - ctx->chunk_list = chunks; - ctx->n_chunks = n_chunks; - - ctx->stored.unknown = 1; - ctx->user.unknown = 1; - - return 0; -} - -int spng_set_offs(spng_ctx *ctx, struct spng_offs *offs) -{ - SPNG_SET_CHUNK_BOILERPLATE(offs); - - if(check_offs(offs)) return SPNG_EOFFS; - - ctx->offs = *offs; - - ctx->stored.offs = 1; - ctx->user.offs = 1; - - return 0; -} - -int spng_set_exif(spng_ctx *ctx, struct spng_exif *exif) -{ - SPNG_SET_CHUNK_BOILERPLATE(exif); - - if(check_exif(exif)) return SPNG_EEXIF; - - if(ctx->exif.data != NULL && !ctx->user.exif) spng__free(ctx, ctx->exif.data); - - ctx->exif = *exif; - - ctx->stored.exif = 1; - ctx->user.exif = 1; - - return 0; -} - -const char *spng_strerror(int err) -{ - switch(err) - { - case SPNG_IO_EOF: return "end of stream"; - case SPNG_IO_ERROR: return "stream error"; - case SPNG_OK: return "success"; - case SPNG_EINVAL: return "invalid argument"; - case SPNG_EMEM: return "out of memory"; - case SPNG_EOVERFLOW: return "arithmetic overflow"; - case SPNG_ESIGNATURE: return "invalid signature"; - case SPNG_EWIDTH: return "invalid image width"; - case SPNG_EHEIGHT: return "invalid image height"; - case SPNG_EUSER_WIDTH: return "image width exceeds user limit"; - case SPNG_EUSER_HEIGHT: return "image height exceeds user limit"; - case SPNG_EBIT_DEPTH: return "invalid bit depth"; - case SPNG_ECOLOR_TYPE: return "invalid color type"; - case SPNG_ECOMPRESSION_METHOD: return "invalid compression method"; - case SPNG_EFILTER_METHOD: return "invalid filter method"; - case SPNG_EINTERLACE_METHOD: return "invalid interlace method"; - case SPNG_EIHDR_SIZE: return "invalid IHDR chunk size"; - case SPNG_ENOIHDR: return "missing IHDR chunk"; - case SPNG_ECHUNK_POS: return "invalid chunk position"; - case SPNG_ECHUNK_SIZE: return "invalid chunk length"; - case SPNG_ECHUNK_CRC: return "invalid chunk checksum"; - case SPNG_ECHUNK_TYPE: return "invalid chunk type"; - case SPNG_ECHUNK_UNKNOWN_CRITICAL: return "unknown critical chunk"; - case SPNG_EDUP_PLTE: return "duplicate PLTE chunk"; - case SPNG_EDUP_CHRM: return "duplicate cHRM chunk"; - case SPNG_EDUP_GAMA: return "duplicate gAMA chunk"; - case SPNG_EDUP_ICCP: return "duplicate iCCP chunk"; - case SPNG_EDUP_SBIT: return "duplicate sBIT chunk"; - case SPNG_EDUP_SRGB: return "duplicate sRGB chunk"; - case SPNG_EDUP_BKGD: return "duplicate bKGD chunk"; - case SPNG_EDUP_HIST: return "duplicate hIST chunk"; - case SPNG_EDUP_TRNS: return "duplicate tRNS chunk"; - case SPNG_EDUP_PHYS: return "duplicate pHYs chunk"; - case SPNG_EDUP_TIME: return "duplicate tIME chunk"; - case SPNG_EDUP_OFFS: return "duplicate oFFs chunk"; - case SPNG_EDUP_EXIF: return "duplicate eXIf chunk"; - case SPNG_ECHRM: return "invalid cHRM chunk"; - case SPNG_EPLTE_IDX: return "invalid palette (PLTE) index"; - case SPNG_ETRNS_COLOR_TYPE: return "tRNS chunk with incompatible color type"; - case SPNG_ETRNS_NO_PLTE: return "missing palette (PLTE) for tRNS chunk"; - case SPNG_EGAMA: return "invalid gAMA chunk"; - case SPNG_EICCP_NAME: return "invalid iCCP profile name"; - case SPNG_EICCP_COMPRESSION_METHOD: return "invalid iCCP compression method"; - case SPNG_ESBIT: return "invalid sBIT chunk"; - case SPNG_ESRGB: return "invalid sRGB chunk"; - case SPNG_ETEXT: return "invalid tEXt chunk"; - case SPNG_ETEXT_KEYWORD: return "invalid tEXt keyword"; - case SPNG_EZTXT: return "invalid zTXt chunk"; - case SPNG_EZTXT_COMPRESSION_METHOD: return "invalid zTXt compression method"; - case SPNG_EITXT: return "invalid iTXt chunk"; - case SPNG_EITXT_COMPRESSION_FLAG: return "invalid iTXt compression flag"; - case SPNG_EITXT_COMPRESSION_METHOD: return "invalid iTXt compression method"; - case SPNG_EITXT_LANG_TAG: return "invalid iTXt language tag"; - case SPNG_EITXT_TRANSLATED_KEY: return "invalid iTXt translated key"; - case SPNG_EBKGD_NO_PLTE: return "missing palette for bKGD chunk"; - case SPNG_EBKGD_PLTE_IDX: return "invalid palette index for bKGD chunk"; - case SPNG_EHIST_NO_PLTE: return "missing palette for hIST chunk"; - case SPNG_EPHYS: return "invalid pHYs chunk"; - case SPNG_ESPLT_NAME: return "invalid suggested palette name"; - case SPNG_ESPLT_DUP_NAME: return "duplicate suggested palette (sPLT) name"; - case SPNG_ESPLT_DEPTH: return "invalid suggested palette (sPLT) sample depth"; - case SPNG_ETIME: return "invalid tIME chunk"; - case SPNG_EOFFS: return "invalid oFFs chunk"; - case SPNG_EEXIF: return "invalid eXIf chunk"; - case SPNG_EIDAT_TOO_SHORT: return "IDAT stream too short"; - case SPNG_EIDAT_STREAM: return "IDAT stream error"; - case SPNG_EZLIB: return "zlib error"; - case SPNG_EFILTER: return "invalid scanline filter"; - case SPNG_EBUFSIZ: return "invalid buffer size"; - case SPNG_EIO: return "i/o error"; - case SPNG_EOF: return "end of file"; - case SPNG_EBUF_SET: return "buffer already set"; - case SPNG_EBADSTATE: return "non-recoverable state"; - case SPNG_EFMT: return "invalid format"; - case SPNG_EFLAGS: return "invalid flags"; - case SPNG_ECHUNKAVAIL: return "chunk not available"; - case SPNG_ENCODE_ONLY: return "encode only context"; - case SPNG_EOI: return "reached end-of-image state"; - case SPNG_ENOPLTE: return "missing PLTE for indexed image"; - case SPNG_ECHUNK_LIMITS: return "reached chunk/cache limits"; - case SPNG_EZLIB_INIT: return "zlib init error"; - case SPNG_ECHUNK_STDLEN: return "chunk exceeds maximum standard length"; - case SPNG_EINTERNAL: return "internal error"; - case SPNG_ECTXTYPE: return "invalid operation for context type"; - case SPNG_ENOSRC: return "source PNG not set"; - case SPNG_ENODST: return "PNG output not set"; - case SPNG_EOPSTATE: return "invalid operation for state"; - case SPNG_ENOTFINAL: return "PNG not finalized"; - default: return "unknown error"; - } -} - -const char *spng_version_string(void) -{ - return SPNG_VERSION_STRING; -} - -#if defined(_MSC_VER) - #pragma warning(pop) -#endif - -/* The following SIMD optimizations are derived from libpng source code. */ - -/* -* PNG Reference Library License version 2 -* -* Copyright (c) 1995-2019 The PNG Reference Library Authors. -* Copyright (c) 2018-2019 Cosmin Truta. -* Copyright (c) 2000-2002, 2004, 2006-2018 Glenn Randers-Pehrson. -* Copyright (c) 1996-1997 Andreas Dilger. -* Copyright (c) 1995-1996 Guy Eric Schalnat, Group 42, Inc. -* -* The software is supplied "as is", without warranty of any kind, -* express or implied, including, without limitation, the warranties -* of merchantability, fitness for a particular purpose, title, and -* non-infringement. In no event shall the Copyright owners, or -* anyone distributing the software, be liable for any damages or -* other liability, whether in contract, tort or otherwise, arising -* from, out of, or in connection with the software, or the use or -* other dealings in the software, even if advised of the possibility -* of such damage. -* -* Permission is hereby granted to use, copy, modify, and distribute -* this software, or portions hereof, for any purpose, without fee, -* subject to the following restrictions: -* -* 1. The origin of this software must not be misrepresented; you -* must not claim that you wrote the original software. If you -* use this software in a product, an acknowledgment in the product -* documentation would be appreciated, but is not required. -* -* 2. Altered source versions must be plainly marked as such, and must -* not be misrepresented as being the original software. -* -* 3. This Copyright notice may not be removed or altered from any -* source or altered source distribution. -*/ - -#if defined(SPNG_X86) - -#ifndef SPNG_SSE - #define SPNG_SSE 1 -#endif - -#if defined(__GNUC__) && !defined(__clang__) - #if SPNG_SSE == 3 - #pragma GCC target("ssse3") - #elif SPNG_SSE == 4 - #pragma GCC target("sse4.1") - #else - #pragma GCC target("sse2") - #endif -#endif - -/* SSE2 optimised filter functions - * Derived from filter_neon_intrinsics.c - * - * Copyright (c) 2018 Cosmin Truta - * Copyright (c) 2016-2017 Glenn Randers-Pehrson - * Written by Mike Klein and Matt Sarett - * Derived from arm/filter_neon_intrinsics.c - * - * This code is derived from libpng source code. - * For conditions of distribution and use, see the disclaimer - * and license above. - */ - -#include -#include -#include - -/* Functions in this file look at most 3 pixels (a,b,c) to predict the 4th (d). - * They're positioned like this: - * prev: c b - * row: a d - * The Sub filter predicts d=a, Avg d=(a+b)/2, and Paeth predicts d to be - * whichever of a, b, or c is closest to p=a+b-c. - */ - -static __m128i load4(const void* p) -{ - int tmp; - memcpy(&tmp, p, sizeof(tmp)); - return _mm_cvtsi32_si128(tmp); -} - -static void store4(void* p, __m128i v) -{ - int tmp = _mm_cvtsi128_si32(v); - memcpy(p, &tmp, sizeof(int)); -} - -static __m128i load3(const void* p) -{ - uint32_t tmp = 0; - memcpy(&tmp, p, 3); - return _mm_cvtsi32_si128(tmp); -} - -static void store3(void* p, __m128i v) -{ - int tmp = _mm_cvtsi128_si32(v); - memcpy(p, &tmp, 3); -} - -static void defilter_sub3(size_t rowbytes, unsigned char *row) -{ - /* The Sub filter predicts each pixel as the previous pixel, a. - * There is no pixel to the left of the first pixel. It's encoded directly. - * That works with our main loop if we just say that left pixel was zero. - */ - size_t rb = rowbytes; - - __m128i a, d = _mm_setzero_si128(); - - while(rb >= 4) - { - a = d; d = load4(row); - d = _mm_add_epi8(d, a); - store3(row, d); - - row += 3; - rb -= 3; - } - - if(rb > 0) - { - a = d; d = load3(row); - d = _mm_add_epi8(d, a); - store3(row, d); - } -} - -static void defilter_sub4(size_t rowbytes, unsigned char *row) -{ - /* The Sub filter predicts each pixel as the previous pixel, a. - * There is no pixel to the left of the first pixel. It's encoded directly. - * That works with our main loop if we just say that left pixel was zero. - */ - size_t rb = rowbytes+4; - - __m128i a, d = _mm_setzero_si128(); - - while(rb > 4) - { - a = d; d = load4(row); - d = _mm_add_epi8(d, a); - store4(row, d); - - row += 4; - rb -= 4; - } -} - -static void defilter_avg3(size_t rowbytes, unsigned char *row, const unsigned char *prev) -{ - /* The Avg filter predicts each pixel as the (truncated) average of a and b. - * There's no pixel to the left of the first pixel. Luckily, it's - * predicted to be half of the pixel above it. So again, this works - * perfectly with our loop if we make sure a starts at zero. - */ - - size_t rb = rowbytes; - - const __m128i zero = _mm_setzero_si128(); - - __m128i b; - __m128i a, d = zero; - - while(rb >= 4) - { - __m128i avg; - b = load4(prev); - a = d; d = load4(row ); - - /* PNG requires a truncating average, so we can't just use _mm_avg_epu8 */ - avg = _mm_avg_epu8(a,b); - /* ...but we can fix it up by subtracting off 1 if it rounded up. */ - avg = _mm_sub_epi8(avg, _mm_and_si128(_mm_xor_si128(a, b), - _mm_set1_epi8(1))); - d = _mm_add_epi8(d, avg); - store3(row, d); - - prev += 3; - row += 3; - rb -= 3; - } - - if(rb > 0) - { - __m128i avg; - b = load3(prev); - a = d; d = load3(row ); - - /* PNG requires a truncating average, so we can't just use _mm_avg_epu8 */ - avg = _mm_avg_epu8(a, b); - /* ...but we can fix it up by subtracting off 1 if it rounded up. */ - avg = _mm_sub_epi8(avg, _mm_and_si128(_mm_xor_si128(a, b), - _mm_set1_epi8(1))); - - d = _mm_add_epi8(d, avg); - store3(row, d); - } -} - -static void defilter_avg4(size_t rowbytes, unsigned char *row, const unsigned char *prev) -{ - /* The Avg filter predicts each pixel as the (truncated) average of a and b. - * There's no pixel to the left of the first pixel. Luckily, it's - * predicted to be half of the pixel above it. So again, this works - * perfectly with our loop if we make sure a starts at zero. - */ - size_t rb = rowbytes+4; - - const __m128i zero = _mm_setzero_si128(); - __m128i b; - __m128i a, d = zero; - - while(rb > 4) - { - __m128i avg; - b = load4(prev); - a = d; d = load4(row ); - - /* PNG requires a truncating average, so we can't just use _mm_avg_epu8 */ - avg = _mm_avg_epu8(a,b); - /* ...but we can fix it up by subtracting off 1 if it rounded up. */ - avg = _mm_sub_epi8(avg, _mm_and_si128(_mm_xor_si128(a, b), - _mm_set1_epi8(1))); - - d = _mm_add_epi8(d, avg); - store4(row, d); - - prev += 4; - row += 4; - rb -= 4; - } -} - -/* Returns |x| for 16-bit lanes. */ -#if (SPNG_SSE >= 3) && !defined(_MSC_VER) -__attribute__((target("ssse3"))) -#endif -static __m128i abs_i16(__m128i x) -{ -#if SPNG_SSE >= 3 - return _mm_abs_epi16(x); -#else - /* Read this all as, return x<0 ? -x : x. - * To negate two's complement, you flip all the bits then add 1. - */ - __m128i is_negative = _mm_cmplt_epi16(x, _mm_setzero_si128()); - - /* Flip negative lanes. */ - x = _mm_xor_si128(x, is_negative); - - /* +1 to negative lanes, else +0. */ - x = _mm_sub_epi16(x, is_negative); - return x; -#endif -} - -/* Bytewise c ? t : e. */ -static __m128i if_then_else(__m128i c, __m128i t, __m128i e) -{ -#if SPNG_SSE >= 4 - return _mm_blendv_epi8(e, t, c); -#else - return _mm_or_si128(_mm_and_si128(c, t), _mm_andnot_si128(c, e)); -#endif -} - -static void defilter_paeth3(size_t rowbytes, unsigned char *row, const unsigned char *prev) -{ - /* Paeth tries to predict pixel d using the pixel to the left of it, a, - * and two pixels from the previous row, b and c: - * prev: c b - * row: a d - * The Paeth function predicts d to be whichever of a, b, or c is nearest to - * p=a+b-c. - * - * The first pixel has no left context, and so uses an Up filter, p = b. - * This works naturally with our main loop's p = a+b-c if we force a and c - * to zero. - * Here we zero b and d, which become c and a respectively at the start of - * the loop. - */ - size_t rb = rowbytes; - const __m128i zero = _mm_setzero_si128(); - __m128i c, b = zero, - a, d = zero; - - while(rb >= 4) - { - /* It's easiest to do this math (particularly, deal with pc) with 16-bit - * intermediates. - */ - __m128i pa,pb,pc,smallest,nearest; - c = b; b = _mm_unpacklo_epi8(load4(prev), zero); - a = d; d = _mm_unpacklo_epi8(load4(row ), zero); - - /* (p-a) == (a+b-c - a) == (b-c) */ - - pa = _mm_sub_epi16(b, c); - - /* (p-b) == (a+b-c - b) == (a-c) */ - pb = _mm_sub_epi16(a, c); - - /* (p-c) == (a+b-c - c) == (a+b-c-c) == (b-c)+(a-c) */ - pc = _mm_add_epi16(pa, pb); - - pa = abs_i16(pa); /* |p-a| */ - pb = abs_i16(pb); /* |p-b| */ - pc = abs_i16(pc); /* |p-c| */ - - smallest = _mm_min_epi16(pc, _mm_min_epi16(pa, pb)); - - /* Paeth breaks ties favoring a over b over c. */ - nearest = if_then_else(_mm_cmpeq_epi16(smallest, pa), a, - if_then_else(_mm_cmpeq_epi16(smallest, pb), b, c)); - - /* Note `_epi8`: we need addition to wrap modulo 255. */ - d = _mm_add_epi8(d, nearest); - store3(row, _mm_packus_epi16(d, d)); - - prev += 3; - row += 3; - rb -= 3; - } - - if(rb > 0) - { - /* It's easiest to do this math (particularly, deal with pc) with 16-bit - * intermediates. - */ - __m128i pa, pb, pc, smallest, nearest; - c = b; b = _mm_unpacklo_epi8(load3(prev), zero); - a = d; d = _mm_unpacklo_epi8(load3(row ), zero); - - /* (p-a) == (a+b-c - a) == (b-c) */ - pa = _mm_sub_epi16(b, c); - - /* (p-b) == (a+b-c - b) == (a-c) */ - pb = _mm_sub_epi16(a, c); - - /* (p-c) == (a+b-c - c) == (a+b-c-c) == (b-c)+(a-c) */ - pc = _mm_add_epi16(pa, pb); - - pa = abs_i16(pa); /* |p-a| */ - pb = abs_i16(pb); /* |p-b| */ - pc = abs_i16(pc); /* |p-c| */ - - smallest = _mm_min_epi16(pc, _mm_min_epi16(pa, pb)); - - /* Paeth breaks ties favoring a over b over c. */ - nearest = if_then_else(_mm_cmpeq_epi16(smallest, pa), a, - if_then_else(_mm_cmpeq_epi16(smallest, pb), b, c)); - - /* Note `_epi8`: we need addition to wrap modulo 255. */ - d = _mm_add_epi8(d, nearest); - store3(row, _mm_packus_epi16(d, d)); - } -} - -static void defilter_paeth4(size_t rowbytes, unsigned char *row, const unsigned char *prev) -{ - /* Paeth tries to predict pixel d using the pixel to the left of it, a, - * and two pixels from the previous row, b and c: - * prev: c b - * row: a d - * The Paeth function predicts d to be whichever of a, b, or c is nearest to - * p=a+b-c. - * - * The first pixel has no left context, and so uses an Up filter, p = b. - * This works naturally with our main loop's p = a+b-c if we force a and c - * to zero. - * Here we zero b and d, which become c and a respectively at the start of - * the loop. - */ - size_t rb = rowbytes+4; - - const __m128i zero = _mm_setzero_si128(); - __m128i pa, pb, pc, smallest, nearest; - __m128i c, b = zero, - a, d = zero; - - while(rb > 4) - { - /* It's easiest to do this math (particularly, deal with pc) with 16-bit - * intermediates. - */ - c = b; b = _mm_unpacklo_epi8(load4(prev), zero); - a = d; d = _mm_unpacklo_epi8(load4(row ), zero); - - /* (p-a) == (a+b-c - a) == (b-c) */ - pa = _mm_sub_epi16(b, c); - - /* (p-b) == (a+b-c - b) == (a-c) */ - pb = _mm_sub_epi16(a, c); - - /* (p-c) == (a+b-c - c) == (a+b-c-c) == (b-c)+(a-c) */ - pc = _mm_add_epi16(pa, pb); - - pa = abs_i16(pa); /* |p-a| */ - pb = abs_i16(pb); /* |p-b| */ - pc = abs_i16(pc); /* |p-c| */ - - smallest = _mm_min_epi16(pc, _mm_min_epi16(pa, pb)); - - /* Paeth breaks ties favoring a over b over c. */ - nearest = if_then_else(_mm_cmpeq_epi16(smallest, pa), a, - if_then_else(_mm_cmpeq_epi16(smallest, pb), b, c)); - - /* Note `_epi8`: we need addition to wrap modulo 255. */ - d = _mm_add_epi8(d, nearest); - store4(row, _mm_packus_epi16(d, d)); - - prev += 4; - row += 4; - rb -= 4; - } -} - -#endif /* SPNG_X86 */ - - -#if defined(SPNG_ARM) - -/* NEON optimised filter functions - * Derived from filter_neon_intrinsics.c - * - * Copyright (c) 2018 Cosmin Truta - * Copyright (c) 2014,2016 Glenn Randers-Pehrson - * Written by James Yu , October 2013. - * Based on filter_neon.S, written by Mans Rullgard, 2011. - * - * This code is derived from libpng source code. - * For conditions of distribution and use, see the disclaimer - * and license in this file. - */ - -#define png_aligncast(type, value) ((void*)(value)) -#define png_aligncastconst(type, value) ((const void*)(value)) - -/* libpng row pointers are not necessarily aligned to any particular boundary, - * however this code will only work with appropriate alignment. mips/mips_init.c - * checks for this (and will not compile unless it is done). This code uses - * variants of png_aligncast to avoid compiler warnings. - */ -#define png_ptr(type,pointer) png_aligncast(type *,pointer) -#define png_ptrc(type,pointer) png_aligncastconst(const type *,pointer) - -/* The following relies on a variable 'temp_pointer' being declared with type - * 'type'. This is written this way just to hide the GCC strict aliasing - * warning; note that the code is safe because there never is an alias between - * the input and output pointers. - */ -#define png_ldr(type,pointer)\ - (temp_pointer = png_ptr(type,pointer), *temp_pointer) - - -#if defined(_MSC_VER) && !defined(__clang__) && defined(_M_ARM64) - #include -#else - #include -#endif - -static void defilter_sub3(size_t rowbytes, unsigned char *row) -{ - unsigned char *rp = row; - unsigned char *rp_stop = row + rowbytes; - - uint8x16_t vtmp = vld1q_u8(rp); - uint8x8x2_t *vrpt = png_ptr(uint8x8x2_t, &vtmp); - uint8x8x2_t vrp = *vrpt; - - uint8x8x4_t vdest; - vdest.val[3] = vdup_n_u8(0); - - for (; rp < rp_stop;) - { - uint8x8_t vtmp1, vtmp2; - uint32x2_t *temp_pointer; - - vtmp1 = vext_u8(vrp.val[0], vrp.val[1], 3); - vdest.val[0] = vadd_u8(vdest.val[3], vrp.val[0]); - vtmp2 = vext_u8(vrp.val[0], vrp.val[1], 6); - vdest.val[1] = vadd_u8(vdest.val[0], vtmp1); - - vtmp1 = vext_u8(vrp.val[1], vrp.val[1], 1); - vdest.val[2] = vadd_u8(vdest.val[1], vtmp2); - vdest.val[3] = vadd_u8(vdest.val[2], vtmp1); - - vtmp = vld1q_u8(rp + 12); - vrpt = png_ptr(uint8x8x2_t, &vtmp); - vrp = *vrpt; - - vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[0]), 0); - rp += 3; - vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[1]), 0); - rp += 3; - vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[2]), 0); - rp += 3; - vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[3]), 0); - rp += 3; - } -} - -static void defilter_sub4(size_t rowbytes, unsigned char *row) -{ - unsigned char *rp = row; - unsigned char *rp_stop = row + rowbytes; - - uint8x8x4_t vdest; - vdest.val[3] = vdup_n_u8(0); - - for (; rp < rp_stop; rp += 16) - { - uint32x2x4_t vtmp = vld4_u32(png_ptr(uint32_t,rp)); - uint8x8x4_t *vrpt = png_ptr(uint8x8x4_t,&vtmp); - uint8x8x4_t vrp = *vrpt; - uint32x2x4_t *temp_pointer; - uint32x2x4_t vdest_val; - - vdest.val[0] = vadd_u8(vdest.val[3], vrp.val[0]); - vdest.val[1] = vadd_u8(vdest.val[0], vrp.val[1]); - vdest.val[2] = vadd_u8(vdest.val[1], vrp.val[2]); - vdest.val[3] = vadd_u8(vdest.val[2], vrp.val[3]); - - vdest_val = png_ldr(uint32x2x4_t, &vdest); - vst4_lane_u32(png_ptr(uint32_t,rp), vdest_val, 0); - } -} - -static void defilter_avg3(size_t rowbytes, unsigned char *row, const unsigned char *prev_row) -{ - unsigned char *rp = row; - const unsigned char *pp = prev_row; - unsigned char *rp_stop = row + rowbytes; - - uint8x16_t vtmp; - uint8x8x2_t *vrpt; - uint8x8x2_t vrp; - uint8x8x4_t vdest; - vdest.val[3] = vdup_n_u8(0); - - vtmp = vld1q_u8(rp); - vrpt = png_ptr(uint8x8x2_t,&vtmp); - vrp = *vrpt; - - for (; rp < rp_stop; pp += 12) - { - uint8x8_t vtmp1, vtmp2, vtmp3; - - uint8x8x2_t *vppt; - uint8x8x2_t vpp; - - uint32x2_t *temp_pointer; - - vtmp = vld1q_u8(pp); - vppt = png_ptr(uint8x8x2_t,&vtmp); - vpp = *vppt; - - vtmp1 = vext_u8(vrp.val[0], vrp.val[1], 3); - vdest.val[0] = vhadd_u8(vdest.val[3], vpp.val[0]); - vdest.val[0] = vadd_u8(vdest.val[0], vrp.val[0]); - - vtmp2 = vext_u8(vpp.val[0], vpp.val[1], 3); - vtmp3 = vext_u8(vrp.val[0], vrp.val[1], 6); - vdest.val[1] = vhadd_u8(vdest.val[0], vtmp2); - vdest.val[1] = vadd_u8(vdest.val[1], vtmp1); - - vtmp2 = vext_u8(vpp.val[0], vpp.val[1], 6); - vtmp1 = vext_u8(vrp.val[1], vrp.val[1], 1); - - vtmp = vld1q_u8(rp + 12); - vrpt = png_ptr(uint8x8x2_t,&vtmp); - vrp = *vrpt; - - vdest.val[2] = vhadd_u8(vdest.val[1], vtmp2); - vdest.val[2] = vadd_u8(vdest.val[2], vtmp3); - - vtmp2 = vext_u8(vpp.val[1], vpp.val[1], 1); - - vdest.val[3] = vhadd_u8(vdest.val[2], vtmp2); - vdest.val[3] = vadd_u8(vdest.val[3], vtmp1); - - vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[0]), 0); - rp += 3; - vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[1]), 0); - rp += 3; - vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[2]), 0); - rp += 3; - vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[3]), 0); - rp += 3; - } -} - -static void defilter_avg4(size_t rowbytes, unsigned char *row, const unsigned char *prev_row) -{ - unsigned char *rp = row; - unsigned char *rp_stop = row + rowbytes; - const unsigned char *pp = prev_row; - - uint8x8x4_t vdest; - vdest.val[3] = vdup_n_u8(0); - - for (; rp < rp_stop; rp += 16, pp += 16) - { - uint32x2x4_t vtmp; - uint8x8x4_t *vrpt, *vppt; - uint8x8x4_t vrp, vpp; - uint32x2x4_t *temp_pointer; - uint32x2x4_t vdest_val; - - vtmp = vld4_u32(png_ptr(uint32_t,rp)); - vrpt = png_ptr(uint8x8x4_t,&vtmp); - vrp = *vrpt; - vtmp = vld4_u32(png_ptrc(uint32_t,pp)); - vppt = png_ptr(uint8x8x4_t,&vtmp); - vpp = *vppt; - - vdest.val[0] = vhadd_u8(vdest.val[3], vpp.val[0]); - vdest.val[0] = vadd_u8(vdest.val[0], vrp.val[0]); - vdest.val[1] = vhadd_u8(vdest.val[0], vpp.val[1]); - vdest.val[1] = vadd_u8(vdest.val[1], vrp.val[1]); - vdest.val[2] = vhadd_u8(vdest.val[1], vpp.val[2]); - vdest.val[2] = vadd_u8(vdest.val[2], vrp.val[2]); - vdest.val[3] = vhadd_u8(vdest.val[2], vpp.val[3]); - vdest.val[3] = vadd_u8(vdest.val[3], vrp.val[3]); - - vdest_val = png_ldr(uint32x2x4_t, &vdest); - vst4_lane_u32(png_ptr(uint32_t,rp), vdest_val, 0); - } -} - -static uint8x8_t paeth_arm(uint8x8_t a, uint8x8_t b, uint8x8_t c) -{ - uint8x8_t d, e; - uint16x8_t p1, pa, pb, pc; - - p1 = vaddl_u8(a, b); /* a + b */ - pc = vaddl_u8(c, c); /* c * 2 */ - pa = vabdl_u8(b, c); /* pa */ - pb = vabdl_u8(a, c); /* pb */ - pc = vabdq_u16(p1, pc); /* pc */ - - p1 = vcleq_u16(pa, pb); /* pa <= pb */ - pa = vcleq_u16(pa, pc); /* pa <= pc */ - pb = vcleq_u16(pb, pc); /* pb <= pc */ - - p1 = vandq_u16(p1, pa); /* pa <= pb && pa <= pc */ - - d = vmovn_u16(pb); - e = vmovn_u16(p1); - - d = vbsl_u8(d, b, c); - e = vbsl_u8(e, a, d); - - return e; -} - -static void defilter_paeth3(size_t rowbytes, unsigned char *row, const unsigned char *prev_row) -{ - unsigned char *rp = row; - const unsigned char *pp = prev_row; - unsigned char *rp_stop = row + rowbytes; - - uint8x16_t vtmp; - uint8x8x2_t *vrpt; - uint8x8x2_t vrp; - uint8x8_t vlast = vdup_n_u8(0); - uint8x8x4_t vdest; - vdest.val[3] = vdup_n_u8(0); - - vtmp = vld1q_u8(rp); - vrpt = png_ptr(uint8x8x2_t,&vtmp); - vrp = *vrpt; - - for (; rp < rp_stop; pp += 12) - { - uint8x8x2_t *vppt; - uint8x8x2_t vpp; - uint8x8_t vtmp1, vtmp2, vtmp3; - uint32x2_t *temp_pointer; - - vtmp = vld1q_u8(pp); - vppt = png_ptr(uint8x8x2_t,&vtmp); - vpp = *vppt; - - vdest.val[0] = paeth_arm(vdest.val[3], vpp.val[0], vlast); - vdest.val[0] = vadd_u8(vdest.val[0], vrp.val[0]); - - vtmp1 = vext_u8(vrp.val[0], vrp.val[1], 3); - vtmp2 = vext_u8(vpp.val[0], vpp.val[1], 3); - vdest.val[1] = paeth_arm(vdest.val[0], vtmp2, vpp.val[0]); - vdest.val[1] = vadd_u8(vdest.val[1], vtmp1); - - vtmp1 = vext_u8(vrp.val[0], vrp.val[1], 6); - vtmp3 = vext_u8(vpp.val[0], vpp.val[1], 6); - vdest.val[2] = paeth_arm(vdest.val[1], vtmp3, vtmp2); - vdest.val[2] = vadd_u8(vdest.val[2], vtmp1); - - vtmp1 = vext_u8(vrp.val[1], vrp.val[1], 1); - vtmp2 = vext_u8(vpp.val[1], vpp.val[1], 1); - - vtmp = vld1q_u8(rp + 12); - vrpt = png_ptr(uint8x8x2_t,&vtmp); - vrp = *vrpt; - - vdest.val[3] = paeth_arm(vdest.val[2], vtmp2, vtmp3); - vdest.val[3] = vadd_u8(vdest.val[3], vtmp1); - - vlast = vtmp2; - - vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[0]), 0); - rp += 3; - vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[1]), 0); - rp += 3; - vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[2]), 0); - rp += 3; - vst1_lane_u32(png_ptr(uint32_t,rp), png_ldr(uint32x2_t,&vdest.val[3]), 0); - rp += 3; - } -} - -static void defilter_paeth4(size_t rowbytes, unsigned char *row, const unsigned char *prev_row) -{ - unsigned char *rp = row; - unsigned char *rp_stop = row + rowbytes; - const unsigned char *pp = prev_row; - - uint8x8_t vlast = vdup_n_u8(0); - uint8x8x4_t vdest; - vdest.val[3] = vdup_n_u8(0); - - for (; rp < rp_stop; rp += 16, pp += 16) - { - uint32x2x4_t vtmp; - uint8x8x4_t *vrpt, *vppt; - uint8x8x4_t vrp, vpp; - uint32x2x4_t *temp_pointer; - uint32x2x4_t vdest_val; - - vtmp = vld4_u32(png_ptr(uint32_t,rp)); - vrpt = png_ptr(uint8x8x4_t,&vtmp); - vrp = *vrpt; - vtmp = vld4_u32(png_ptrc(uint32_t,pp)); - vppt = png_ptr(uint8x8x4_t,&vtmp); - vpp = *vppt; - - vdest.val[0] = paeth_arm(vdest.val[3], vpp.val[0], vlast); - vdest.val[0] = vadd_u8(vdest.val[0], vrp.val[0]); - vdest.val[1] = paeth_arm(vdest.val[0], vpp.val[1], vpp.val[0]); - vdest.val[1] = vadd_u8(vdest.val[1], vrp.val[1]); - vdest.val[2] = paeth_arm(vdest.val[1], vpp.val[2], vpp.val[1]); - vdest.val[2] = vadd_u8(vdest.val[2], vrp.val[2]); - vdest.val[3] = paeth_arm(vdest.val[2], vpp.val[3], vpp.val[2]); - vdest.val[3] = vadd_u8(vdest.val[3], vrp.val[3]); - - vlast = vpp.val[3]; - - vdest_val = png_ldr(uint32x2x4_t, &vdest); - vst4_lane_u32(png_ptr(uint32_t,rp), vdest_val, 0); - } -} - -/* NEON optimised palette expansion functions - * Derived from palette_neon_intrinsics.c - * - * Copyright (c) 2018-2019 Cosmin Truta - * Copyright (c) 2017-2018 Arm Holdings. All rights reserved. - * Written by Richard Townsend , February 2017. - * - * This code is derived from libpng source code. - * For conditions of distribution and use, see the disclaimer - * and license in this file. - * - * Related: https://developer.arm.com/documentation/101964/latest/Color-palette-expansion - * - * The functions were refactored to iterate forward. - * - */ - -/* Expands a palettized row into RGBA8. */ -static uint32_t expand_palette_rgba8_neon(unsigned char *row, const unsigned char *scanline, const unsigned char *plte, uint32_t width) -{ - const uint32_t scanline_stride = 4; - const uint32_t row_stride = scanline_stride * 4; - const uint32_t count = width / scanline_stride; - const uint32_t *palette = (const uint32_t*)plte; - - if(!count) return 0; - - uint32_t i; - uint32x4_t cur; - for(i=0; i < count; i++, scanline += scanline_stride) - { - cur = vld1q_dup_u32 (palette + scanline[0]); - cur = vld1q_lane_u32(palette + scanline[1], cur, 1); - cur = vld1q_lane_u32(palette + scanline[2], cur, 2); - cur = vld1q_lane_u32(palette + scanline[3], cur, 3); - vst1q_u32((uint32_t*)(row + i * row_stride), cur); - } - - return count * scanline_stride; -} - -/* Expands a palettized row into RGB8. */ -static uint32_t expand_palette_rgb8_neon(unsigned char *row, const unsigned char *scanline, const unsigned char *plte, uint32_t width) -{ - const uint32_t scanline_stride = 8; - const uint32_t row_stride = scanline_stride * 3; - const uint32_t count = width / scanline_stride; - - if(!count) return 0; - - uint32_t i; - uint8x8x3_t cur; - for(i=0; i < count; i++, scanline += scanline_stride) - { - cur = vld3_dup_u8 (plte + 3 * scanline[0]); - cur = vld3_lane_u8(plte + 3 * scanline[1], cur, 1); - cur = vld3_lane_u8(plte + 3 * scanline[2], cur, 2); - cur = vld3_lane_u8(plte + 3 * scanline[3], cur, 3); - cur = vld3_lane_u8(plte + 3 * scanline[4], cur, 4); - cur = vld3_lane_u8(plte + 3 * scanline[5], cur, 5); - cur = vld3_lane_u8(plte + 3 * scanline[6], cur, 6); - cur = vld3_lane_u8(plte + 3 * scanline[7], cur, 7); - vst3_u8(row + i * row_stride, cur); - } - - return count * scanline_stride; -} - -#endif /* SPNG_ARM */ diff --git a/cmake/externals/libspng/spng.h b/cmake/externals/libspng/spng.h deleted file mode 100644 index 8f946337b..000000000 --- a/cmake/externals/libspng/spng.h +++ /dev/null @@ -1,537 +0,0 @@ -/* SPDX-License-Identifier: BSD-2-Clause */ -#ifndef SPNG_H -#define SPNG_H - -#ifdef __cplusplus -extern "C" { -#endif - -#if (defined(_WIN32) || defined(__CYGWIN__)) && !defined(SPNG_STATIC) - #if defined(SPNG__BUILD) - #define SPNG_API __declspec(dllexport) - #else - #define SPNG_API __declspec(dllimport) - #endif -#else - #define SPNG_API -#endif - -#if defined(_MSC_VER) - #define SPNG_CDECL __cdecl -#else - #define SPNG_CDECL -#endif - -#include -#include -#include - -#define SPNG_VERSION_MAJOR 0 -#define SPNG_VERSION_MINOR 7 -#define SPNG_VERSION_PATCH 4 - -enum spng_errno -{ - SPNG_IO_ERROR = -2, - SPNG_IO_EOF = -1, - SPNG_OK = 0, - SPNG_EINVAL, - SPNG_EMEM, - SPNG_EOVERFLOW, - SPNG_ESIGNATURE, - SPNG_EWIDTH, - SPNG_EHEIGHT, - SPNG_EUSER_WIDTH, - SPNG_EUSER_HEIGHT, - SPNG_EBIT_DEPTH, - SPNG_ECOLOR_TYPE, - SPNG_ECOMPRESSION_METHOD, - SPNG_EFILTER_METHOD, - SPNG_EINTERLACE_METHOD, - SPNG_EIHDR_SIZE, - SPNG_ENOIHDR, - SPNG_ECHUNK_POS, - SPNG_ECHUNK_SIZE, - SPNG_ECHUNK_CRC, - SPNG_ECHUNK_TYPE, - SPNG_ECHUNK_UNKNOWN_CRITICAL, - SPNG_EDUP_PLTE, - SPNG_EDUP_CHRM, - SPNG_EDUP_GAMA, - SPNG_EDUP_ICCP, - SPNG_EDUP_SBIT, - SPNG_EDUP_SRGB, - SPNG_EDUP_BKGD, - SPNG_EDUP_HIST, - SPNG_EDUP_TRNS, - SPNG_EDUP_PHYS, - SPNG_EDUP_TIME, - SPNG_EDUP_OFFS, - SPNG_EDUP_EXIF, - SPNG_ECHRM, - SPNG_EPLTE_IDX, - SPNG_ETRNS_COLOR_TYPE, - SPNG_ETRNS_NO_PLTE, - SPNG_EGAMA, - SPNG_EICCP_NAME, - SPNG_EICCP_COMPRESSION_METHOD, - SPNG_ESBIT, - SPNG_ESRGB, - SPNG_ETEXT, - SPNG_ETEXT_KEYWORD, - SPNG_EZTXT, - SPNG_EZTXT_COMPRESSION_METHOD, - SPNG_EITXT, - SPNG_EITXT_COMPRESSION_FLAG, - SPNG_EITXT_COMPRESSION_METHOD, - SPNG_EITXT_LANG_TAG, - SPNG_EITXT_TRANSLATED_KEY, - SPNG_EBKGD_NO_PLTE, - SPNG_EBKGD_PLTE_IDX, - SPNG_EHIST_NO_PLTE, - SPNG_EPHYS, - SPNG_ESPLT_NAME, - SPNG_ESPLT_DUP_NAME, - SPNG_ESPLT_DEPTH, - SPNG_ETIME, - SPNG_EOFFS, - SPNG_EEXIF, - SPNG_EIDAT_TOO_SHORT, - SPNG_EIDAT_STREAM, - SPNG_EZLIB, - SPNG_EFILTER, - SPNG_EBUFSIZ, - SPNG_EIO, - SPNG_EOF, - SPNG_EBUF_SET, - SPNG_EBADSTATE, - SPNG_EFMT, - SPNG_EFLAGS, - SPNG_ECHUNKAVAIL, - SPNG_ENCODE_ONLY, - SPNG_EOI, - SPNG_ENOPLTE, - SPNG_ECHUNK_LIMITS, - SPNG_EZLIB_INIT, - SPNG_ECHUNK_STDLEN, - SPNG_EINTERNAL, - SPNG_ECTXTYPE, - SPNG_ENOSRC, - SPNG_ENODST, - SPNG_EOPSTATE, - SPNG_ENOTFINAL, -}; - -enum spng_text_type -{ - SPNG_TEXT = 1, - SPNG_ZTXT = 2, - SPNG_ITXT = 3 -}; - -enum spng_color_type -{ - SPNG_COLOR_TYPE_GRAYSCALE = 0, - SPNG_COLOR_TYPE_TRUECOLOR = 2, - SPNG_COLOR_TYPE_INDEXED = 3, - SPNG_COLOR_TYPE_GRAYSCALE_ALPHA = 4, - SPNG_COLOR_TYPE_TRUECOLOR_ALPHA = 6 -}; - -enum spng_filter -{ - SPNG_FILTER_NONE = 0, - SPNG_FILTER_SUB = 1, - SPNG_FILTER_UP = 2, - SPNG_FILTER_AVERAGE = 3, - SPNG_FILTER_PAETH = 4 -}; - -enum spng_filter_choice -{ - SPNG_DISABLE_FILTERING = 0, - SPNG_FILTER_CHOICE_NONE = 8, - SPNG_FILTER_CHOICE_SUB = 16, - SPNG_FILTER_CHOICE_UP = 32, - SPNG_FILTER_CHOICE_AVG = 64, - SPNG_FILTER_CHOICE_PAETH = 128, - SPNG_FILTER_CHOICE_ALL = (8|16|32|64|128) -}; - -enum spng_interlace_method -{ - SPNG_INTERLACE_NONE = 0, - SPNG_INTERLACE_ADAM7 = 1 -}; - -/* Channels are always in byte-order */ -enum spng_format -{ - SPNG_FMT_RGBA8 = 1, - SPNG_FMT_RGBA16 = 2, - SPNG_FMT_RGB8 = 4, - - /* Partially implemented, see documentation */ - SPNG_FMT_GA8 = 16, - SPNG_FMT_GA16 = 32, - SPNG_FMT_G8 = 64, - - /* No conversion or scaling */ - SPNG_FMT_PNG = 256, - SPNG_FMT_RAW = 512 /* big-endian (everything else is host-endian) */ -}; - -enum spng_ctx_flags -{ - SPNG_CTX_IGNORE_ADLER32 = 1, /* Ignore checksum in DEFLATE streams */ - SPNG_CTX_ENCODER = 2 /* Create an encoder context */ -}; - -enum spng_decode_flags -{ - SPNG_DECODE_USE_TRNS = 1, /* Deprecated */ - SPNG_DECODE_USE_GAMA = 2, /* Deprecated */ - SPNG_DECODE_USE_SBIT = 8, /* Undocumented */ - - SPNG_DECODE_TRNS = 1, /* Apply transparency */ - SPNG_DECODE_GAMMA = 2, /* Apply gamma correction */ - SPNG_DECODE_PROGRESSIVE = 256 /* Initialize for progressive reads */ -}; - -enum spng_crc_action -{ - /* Default for critical chunks */ - SPNG_CRC_ERROR = 0, - - /* Discard chunk, invalid for critical chunks. - Since v0.6.2: default for ancillary chunks */ - SPNG_CRC_DISCARD = 1, - - /* Ignore and don't calculate checksum. - Since v0.6.2: also ignores checksums in DEFLATE streams */ - SPNG_CRC_USE = 2 -}; - -enum spng_encode_flags -{ - SPNG_ENCODE_PROGRESSIVE = 1, /* Initialize for progressive writes */ - SPNG_ENCODE_FINALIZE = 2, /* Finalize PNG after encoding image */ -}; - -struct spng_ihdr -{ - uint32_t width; - uint32_t height; - uint8_t bit_depth; - uint8_t color_type; - uint8_t compression_method; - uint8_t filter_method; - uint8_t interlace_method; -}; - -struct spng_plte_entry -{ - uint8_t red; - uint8_t green; - uint8_t blue; - - uint8_t alpha; /* Reserved for internal use */ -}; - -struct spng_plte -{ - uint32_t n_entries; - struct spng_plte_entry entries[256]; -}; - -struct spng_trns -{ - uint16_t gray; - - uint16_t red; - uint16_t green; - uint16_t blue; - - uint32_t n_type3_entries; - uint8_t type3_alpha[256]; -}; - -struct spng_chrm_int -{ - uint32_t white_point_x; - uint32_t white_point_y; - uint32_t red_x; - uint32_t red_y; - uint32_t green_x; - uint32_t green_y; - uint32_t blue_x; - uint32_t blue_y; -}; - -struct spng_chrm -{ - double white_point_x; - double white_point_y; - double red_x; - double red_y; - double green_x; - double green_y; - double blue_x; - double blue_y; -}; - -struct spng_iccp -{ - char profile_name[80]; - size_t profile_len; - char *profile; -}; - -struct spng_sbit -{ - uint8_t grayscale_bits; - uint8_t red_bits; - uint8_t green_bits; - uint8_t blue_bits; - uint8_t alpha_bits; -}; - -struct spng_text -{ - char keyword[80]; - int type; - - size_t length; - char *text; - - uint8_t compression_flag; /* iTXt only */ - uint8_t compression_method; /* iTXt, ztXt only */ - char *language_tag; /* iTXt only */ - char *translated_keyword; /* iTXt only */ -}; - -struct spng_bkgd -{ - uint16_t gray; /* Only for gray/gray alpha */ - uint16_t red; - uint16_t green; - uint16_t blue; - uint16_t plte_index; /* Only for indexed color */ -}; - -struct spng_hist -{ - uint16_t frequency[256]; -}; - -struct spng_phys -{ - uint32_t ppu_x, ppu_y; - uint8_t unit_specifier; -}; - -struct spng_splt_entry -{ - uint16_t red; - uint16_t green; - uint16_t blue; - uint16_t alpha; - uint16_t frequency; -}; - -struct spng_splt -{ - char name[80]; - uint8_t sample_depth; - uint32_t n_entries; - struct spng_splt_entry *entries; -}; - -struct spng_time -{ - uint16_t year; - uint8_t month; - uint8_t day; - uint8_t hour; - uint8_t minute; - uint8_t second; -}; - -struct spng_offs -{ - int32_t x, y; - uint8_t unit_specifier; -}; - -struct spng_exif -{ - size_t length; - char *data; -}; - -struct spng_chunk -{ - size_t offset; - uint32_t length; - uint8_t type[4]; - uint32_t crc; -}; - -enum spng_location -{ - SPNG_AFTER_IHDR = 1, - SPNG_AFTER_PLTE = 2, - SPNG_AFTER_IDAT = 8, -}; - -struct spng_unknown_chunk -{ - uint8_t type[4]; - size_t length; - void *data; - enum spng_location location; -}; - -enum spng_option -{ - SPNG_KEEP_UNKNOWN_CHUNKS = 1, - - SPNG_IMG_COMPRESSION_LEVEL, - SPNG_IMG_WINDOW_BITS, - SPNG_IMG_MEM_LEVEL, - SPNG_IMG_COMPRESSION_STRATEGY, - - SPNG_TEXT_COMPRESSION_LEVEL, - SPNG_TEXT_WINDOW_BITS, - SPNG_TEXT_MEM_LEVEL, - SPNG_TEXT_COMPRESSION_STRATEGY, - - SPNG_FILTER_CHOICE, - SPNG_CHUNK_COUNT_LIMIT, - SPNG_ENCODE_TO_BUFFER, -}; - -typedef void* SPNG_CDECL spng_malloc_fn(size_t size); -typedef void* SPNG_CDECL spng_realloc_fn(void* ptr, size_t size); -typedef void* SPNG_CDECL spng_calloc_fn(size_t count, size_t size); -typedef void SPNG_CDECL spng_free_fn(void* ptr); - -struct spng_alloc -{ - spng_malloc_fn *malloc_fn; - spng_realloc_fn *realloc_fn; - spng_calloc_fn *calloc_fn; - spng_free_fn *free_fn; -}; - -struct spng_row_info -{ - uint32_t scanline_idx; - uint32_t row_num; /* deinterlaced row index */ - int pass; - uint8_t filter; -}; - -typedef struct spng_ctx spng_ctx; - -typedef int spng_read_fn(spng_ctx *ctx, void *user, void *dest, size_t length); -typedef int spng_write_fn(spng_ctx *ctx, void *user, void *src, size_t length); - -typedef int spng_rw_fn(spng_ctx *ctx, void *user, void *dst_src, size_t length); - -SPNG_API spng_ctx *spng_ctx_new(int flags); -SPNG_API spng_ctx *spng_ctx_new2(struct spng_alloc *alloc, int flags); -SPNG_API void spng_ctx_free(spng_ctx *ctx); - -SPNG_API int spng_set_png_buffer(spng_ctx *ctx, const void *buf, size_t size); -SPNG_API int spng_set_png_stream(spng_ctx *ctx, spng_rw_fn *rw_func, void *user); -SPNG_API int spng_set_png_file(spng_ctx *ctx, FILE *file); - -SPNG_API void *spng_get_png_buffer(spng_ctx *ctx, size_t *len, int *error); - -SPNG_API int spng_set_image_limits(spng_ctx *ctx, uint32_t width, uint32_t height); -SPNG_API int spng_get_image_limits(spng_ctx *ctx, uint32_t *width, uint32_t *height); - -SPNG_API int spng_set_chunk_limits(spng_ctx *ctx, size_t chunk_size, size_t cache_size); -SPNG_API int spng_get_chunk_limits(spng_ctx *ctx, size_t *chunk_size, size_t *cache_size); - -SPNG_API int spng_set_crc_action(spng_ctx *ctx, int critical, int ancillary); - -SPNG_API int spng_set_option(spng_ctx *ctx, enum spng_option option, int value); -SPNG_API int spng_get_option(spng_ctx *ctx, enum spng_option option, int *value); - -SPNG_API int spng_decoded_image_size(spng_ctx *ctx, int fmt, size_t *len); - -/* Decode */ -SPNG_API int spng_decode_image(spng_ctx *ctx, void *out, size_t len, int fmt, int flags); - -/* Progressive decode */ -SPNG_API int spng_decode_scanline(spng_ctx *ctx, void *out, size_t len); -SPNG_API int spng_decode_row(spng_ctx *ctx, void *out, size_t len); -SPNG_API int spng_decode_chunks(spng_ctx *ctx); - -/* Encode/decode */ -SPNG_API int spng_get_row_info(spng_ctx *ctx, struct spng_row_info *row_info); - -/* Encode */ -SPNG_API int spng_encode_image(spng_ctx *ctx, const void *img, size_t len, int fmt, int flags); - -/* Progressive encode */ -SPNG_API int spng_encode_scanline(spng_ctx *ctx, const void *scanline, size_t len); -SPNG_API int spng_encode_row(spng_ctx *ctx, const void *row, size_t len); -SPNG_API int spng_encode_chunks(spng_ctx *ctx); - -SPNG_API int spng_get_ihdr(spng_ctx *ctx, struct spng_ihdr *ihdr); -SPNG_API int spng_get_plte(spng_ctx *ctx, struct spng_plte *plte); -SPNG_API int spng_get_trns(spng_ctx *ctx, struct spng_trns *trns); -SPNG_API int spng_get_chrm(spng_ctx *ctx, struct spng_chrm *chrm); -SPNG_API int spng_get_chrm_int(spng_ctx *ctx, struct spng_chrm_int *chrm_int); -SPNG_API int spng_get_gama(spng_ctx *ctx, double *gamma); -SPNG_API int spng_get_gama_int(spng_ctx *ctx, uint32_t *gama_int); -SPNG_API int spng_get_iccp(spng_ctx *ctx, struct spng_iccp *iccp); -SPNG_API int spng_get_sbit(spng_ctx *ctx, struct spng_sbit *sbit); -SPNG_API int spng_get_srgb(spng_ctx *ctx, uint8_t *rendering_intent); -SPNG_API int spng_get_text(spng_ctx *ctx, struct spng_text *text, uint32_t *n_text); -SPNG_API int spng_get_bkgd(spng_ctx *ctx, struct spng_bkgd *bkgd); -SPNG_API int spng_get_hist(spng_ctx *ctx, struct spng_hist *hist); -SPNG_API int spng_get_phys(spng_ctx *ctx, struct spng_phys *phys); -SPNG_API int spng_get_splt(spng_ctx *ctx, struct spng_splt *splt, uint32_t *n_splt); -SPNG_API int spng_get_time(spng_ctx *ctx, struct spng_time *time); -SPNG_API int spng_get_unknown_chunks(spng_ctx *ctx, struct spng_unknown_chunk *chunks, uint32_t *n_chunks); - -/* Official extensions */ -SPNG_API int spng_get_offs(spng_ctx *ctx, struct spng_offs *offs); -SPNG_API int spng_get_exif(spng_ctx *ctx, struct spng_exif *exif); - - -SPNG_API int spng_set_ihdr(spng_ctx *ctx, struct spng_ihdr *ihdr); -SPNG_API int spng_set_plte(spng_ctx *ctx, struct spng_plte *plte); -SPNG_API int spng_set_trns(spng_ctx *ctx, struct spng_trns *trns); -SPNG_API int spng_set_chrm(spng_ctx *ctx, struct spng_chrm *chrm); -SPNG_API int spng_set_chrm_int(spng_ctx *ctx, struct spng_chrm_int *chrm_int); -SPNG_API int spng_set_gama(spng_ctx *ctx, double gamma); -SPNG_API int spng_set_gama_int(spng_ctx *ctx, uint32_t gamma); -SPNG_API int spng_set_iccp(spng_ctx *ctx, struct spng_iccp *iccp); -SPNG_API int spng_set_sbit(spng_ctx *ctx, struct spng_sbit *sbit); -SPNG_API int spng_set_srgb(spng_ctx *ctx, uint8_t rendering_intent); -SPNG_API int spng_set_text(spng_ctx *ctx, struct spng_text *text, uint32_t n_text); -SPNG_API int spng_set_bkgd(spng_ctx *ctx, struct spng_bkgd *bkgd); -SPNG_API int spng_set_hist(spng_ctx *ctx, struct spng_hist *hist); -SPNG_API int spng_set_phys(spng_ctx *ctx, struct spng_phys *phys); -SPNG_API int spng_set_splt(spng_ctx *ctx, struct spng_splt *splt, uint32_t n_splt); -SPNG_API int spng_set_time(spng_ctx *ctx, struct spng_time *time); -SPNG_API int spng_set_unknown_chunks(spng_ctx *ctx, struct spng_unknown_chunk *chunks, uint32_t n_chunks); - -/* Official extensions */ -SPNG_API int spng_set_offs(spng_ctx *ctx, struct spng_offs *offs); -SPNG_API int spng_set_exif(spng_ctx *ctx, struct spng_exif *exif); - - -SPNG_API const char *spng_strerror(int err); -SPNG_API const char *spng_version_string(void); - -#ifdef __cplusplus -} -#endif - -#endif /* SPNG_H */ diff --git a/cmake/externals/opencv.cmake b/cmake/externals/opencv.cmake index e3fa7e904..7dc49e6fb 100644 --- a/cmake/externals/opencv.cmake +++ b/cmake/externals/opencv.cmake @@ -1,16 +1,171 @@ +set(BUILD_ANDROID_EXAMPLES OFF CACHE INTERNAL "") +set(BUILD_ANDROID_PROJECTS OFF CACHE INTERNAL "") +set(BUILD_ANDROID_SERVICE OFF CACHE INTERNAL "") +set(BUILD_DOCS OFF CACHE INTERNAL "") +set(BUILD_FAT_JAVA_LIB OFF CACHE INTERNAL "") +set(BUILD_IPP_IW OFF CACHE INTERNAL "") +set(BUILD_ITT OFF CACHE INTERNAL "") +set(BUILD_JASPER OFF CACHE INTERNAL "") +set(BUILD_JAVA OFF CACHE INTERNAL "") +set(BUILD_JPEG OFF CACHE INTERNAL "") +set(BUILD_OBJC OFF CACHE INTERNAL "") +set(BUILD_OPENJPEG OFF CACHE INTERNAL "") +set(BUILD_PNG OFF CACHE INTERNAL "") +set(BUILD_opencv_apps OFF CACHE INTERNAL "") +set(BUILD_opencv_calib3d OFF CACHE INTERNAL "") +set(BUILD_opencv_dnn OFF CACHE INTERNAL "") +set(BUILD_opencv_features2d OFF CACHE INTERNAL "") +set(BUILD_opencv_flann OFF CACHE INTERNAL "") +set(BUILD_opencv_gapi OFF CACHE INTERNAL "") +set(BUILD_opencv_highgui OFF CACHE INTERNAL "") +set(BUILD_opencv_imgcodecs OFF CACHE INTERNAL "") +set(BUILD_opencv_java OFF CACHE INTERNAL "") +set(BUILD_opencv_js OFF CACHE INTERNAL "") +set(BUILD_opencv_ml OFF CACHE INTERNAL "") +set(BUILD_opencv_objc OFF CACHE INTERNAL "") +set(BUILD_opencv_objdetect OFF CACHE INTERNAL "") +set(BUILD_opencv_photo OFF CACHE INTERNAL "") +set(BUILD_opencv_python2 OFF CACHE INTERNAL "") +set(BUILD_opencv_python3 OFF CACHE INTERNAL "") +set(BUILD_opencv_stitching OFF CACHE INTERNAL "") +set(BUILD_opencv_ts OFF CACHE INTERNAL "") +set(BUILD_opencv_video OFF CACHE INTERNAL "") +set(BUILD_opencv_videoio OFF CACHE INTERNAL "") +set(BUILD_opencv_world OFF CACHE INTERNAL "") +set(BUILD_OPENEXR OFF CACHE INTERNAL "") +set(BUILD_TBB OFF CACHE INTERNAL "") +set(BUILD_TIFF OFF CACHE INTERNAL "") +set(BUILD_WEBP OFF CACHE INTERNAL "") +set(BUILD_WITH_STATIC_CRT OFF CACHE INTERNAL "") +if(OCOS_BUILD_APPLE_FRAMEWORK) + # tell OpenCV to build zlib so we can link to the static library + set(BUILD_ZLIB ON CACHE INTERNAL "") +else() + set(BUILD_ZLIB OFF CACHE INTERNAL "") +endif() +set(ENABLE_FAST_MATH OFF CACHE INTERNAL "") +set(ENABLE_PRECOMPILED_HEADERS OFF CACHE INTERNAL "") +set(WITH_ANDROID_MEDIANDK OFF CACHE INTERNAL "") +set(WITH_AVFOUNDATION OFF CACHE INTERNAL "") +set(WITH_CAP_IOS OFF CACHE INTERNAL "") +set(WITH_CAROTENE OFF CACHE INTERNAL "") +set(WITH_CLP OFF CACHE INTERNAL "") +set(WITH_CPUFEATURES OFF CACHE INTERNAL "") +set(WITH_DIRECTX OFF CACHE INTERNAL "") +set(WITH_DSHOW OFF CACHE INTERNAL "") +set(WITH_EIGEN OFF CACHE INTERNAL "") +set(WITH_FFMPEG OFF CACHE INTERNAL "") +set(WITH_GDCM OFF CACHE INTERNAL "") +set(WITH_GSTREAMER OFF CACHE INTERNAL "") +set(WITH_GTK OFF CACHE INTERNAL "") +set(WITH_HALIDE OFF CACHE INTERNAL "") +set(WITH_HPX OFF CACHE INTERNAL "") +set(WITH_IMGCODEC_HDR OFF CACHE INTERNAL "") +set(WITH_IMGCODEC_PFM OFF CACHE INTERNAL "") +set(WITH_IMGCODEC_PXM OFF CACHE INTERNAL "") +set(WITH_IMGCODEC_SUNRASTER OFF CACHE INTERNAL "") +set(WITH_INF_ENGINE OFF CACHE INTERNAL "") +set(WITH_IPP OFF CACHE INTERNAL "") +set(WITH_ITT OFF CACHE INTERNAL "") +set(WITH_JASPER OFF CACHE INTERNAL "") +set(WITH_JPEG OFF CACHE INTERNAL "") +set(WITH_MSMF OFF CACHE INTERNAL "") +set(WITH_NGRAPH OFF CACHE INTERNAL "") +set(WITH_ONNX OFF CACHE INTERNAL "") +set(WITH_OPENCL OFF CACHE INTERNAL "") +set(WITH_OPENCL_SVM OFF CACHE INTERNAL "") +set(WITH_OPENEXR OFF CACHE INTERNAL "") +set(WITH_OPENJPEG OFF CACHE INTERNAL "") +set(WITH_OPENMP OFF CACHE INTERNAL "") +set(WITH_OPENVX OFF CACHE INTERNAL "") +set(WITH_PNG OFF CACHE INTERNAL "") +set(WITH_PROTOBUF OFF CACHE INTERNAL "") +set(WITH_PTHREADS_PF OFF CACHE INTERNAL "") +set(WITH_QUIRC OFF CACHE INTERNAL "") +set(WITH_TBB OFF CACHE INTERNAL "") +set(WITH_TENGINE OFF CACHE INTERNAL "") +set(WITH_TIFF OFF CACHE INTERNAL "") +set(WITH_V4L OFF CACHE INTERNAL "") +set(WITH_VULKAN OFF CACHE INTERNAL "") +set(WITH_WEBP OFF CACHE INTERNAL "") +set(WITH_WIN32UI OFF CACHE INTERNAL "") + if (OCOS_ENABLE_OPENCV_CODECS) - # Include the subdirectories for libspng and libjpeg - add_subdirectory(${PROJECT_SOURCE_DIR}/cmake/externals/libspng) - add_subdirectory(${PROJECT_SOURCE_DIR}/cmake/externals/libjpeg) + set(BUILD_opencv_imgcodecs ON CACHE INTERNAL "") + + set(BUILD_JPEG ON CACHE INTERNAL "") + set(BUILD_PNG ON CACHE INTERNAL "") + + set(WITH_JPEG ON CACHE INTERNAL "") + set(WITH_PNG ON CACHE INTERNAL "") +endif() + +set(BUILD_SHARED_LIBS OFF CACHE INTERNAL "") +set(BUILD_DOCS OFF CACHE INTERNAL "") +set(BUILD_EXAMPLES OFF CACHE INTERNAL "") +set(BUILD_TESTS OFF CACHE INTERNAL "") + +if(IOS) + # copy what OpenCV's platforms/ios/build_framework.py does and set CPU_BASELINE=DETECT + # https://github.com/opencv/opencv/blob/4223495e6cd67011f86b8ecd9be1fa105018f3b1/platforms/ios/build_framework.py#L253 + set(CPU_BASELINE DETECT) +endif() + +if (CMAKE_SYSTEM_NAME MATCHES "Darwin") + # error _png_do_read_transformations in liblibpng.a(pngrtran.c.o) not found for architecture arm64 + # workaround to disable NEON optimizations + add_definitions(-DPNG_ARM_NEON_IMPLEMENTATION=0) + add_definitions(-DPNG_ARM_NEON_OPT=0) +endif() - # Specify where the static libraries are built - set(LIBRARY_OUTPUT_PATH ${CMAKE_CURRENT_BINARY_DIR}/cvout) +if (MSVC AND CMAKE_GENERATOR_PLATFORM) + string(TOLOWER ${CMAKE_GENERATOR_PLATFORM} _GEN_PLATFORM) + if (${_GEN_PLATFORM} MATCHES "arm|arm64") + set(OPENCV_SKIP_SYSTEM_PROCESSOR_DETECTION ON) + endif() +endif() - # Add the libraries - add_library(libspng STATIC IMPORTED) - add_library(libjpeg STATIC IMPORTED) +FetchContent_Declare( + opencv + GIT_REPOSITORY https://github.com/opencv/opencv.git + SOURCE_DIR opencv_source + GIT_TAG 4.5.4 + GIT_SHALLOW TRUE + -DBUILD_DOCS:BOOL=FALSE + -DBUILD_EXAMPLES:BOOL=FALSE + -DBUILD_TESTS:BOOL=FALSE + -DBUILD_SHARED_LIBS:BOOL=FALSE + -DCMAKE_INSTALL_PREFIX:PATH=${CMAKE_CURRENT_BINARY_DIR}/opencv + -DCV_TRACE:BOOL=FALSE + PATCH_COMMAND git checkout . && git apply --whitespace=fix --ignore-space-change --ignore-whitespace ${CMAKE_CURRENT_SOURCE_DIR}/cmake/externals/opencv-no-rtti.patch +) + +string(TOLOWER "opencv" lcName) +if(NOT ${lcName}_POPULATED) + FetchContent_Populate(opencv) + + # Set variablles to paths of libspng and libjpeg (we only these 2 folders for our vision operators in Extensions) + set(libspng ${opencv_source}/3rdparty/libspng/) + set(libjpeg ${opencv_source}/3rdparty/libjpeg/) + + # Get a list of all files in the folders + file(GLOB_RECURSE LIBSPNG_FILES "${libspng}/*") + file(GLOB_RECURSE LIBJPEG_FILES "${libjpeg}/*") +endif() + +set(opencv_INCLUDE_DIRS "") +list(APPEND opencv_INCLUDE_DIRS ${OPENCV_CONFIG_FILE_INCLUDE_DIR}) +list(APPEND opencv_INCLUDE_DIRS + ${OPENCV_MODULE_opencv_core_LOCATION}/include + ${OPENCV_MODULE_opencv_imgproc_LOCATION}/include) +set(opencv_LIBS "") +list(APPEND opencv_LIBS opencv_core opencv_imgproc) + +if (OCOS_ENABLE_OPENCV_CODECS) + list(APPEND opencv_INCLUDE_DIRS ${OPENCV_MODULE_opencv_imgcodecs_LOCATION}/include) + list(APPEND opencv_LIBS opencv_imgcodecs) - # Set the properties for the libraries - set_property(TARGET libspng PROPERTY IMPORTED_LOCATION ${CMAKE_CURRENT_BINARY_DIR}/cmake/externals/cvout/libspng.a) - set_property(TARGET libjpeg PROPERTY IMPORTED_LOCATION ${CMAKE_CURRENT_BINARY_DIR}/cmake/externals/cvout/libjpeg.a) -endif() \ No newline at end of file + # Add libspng and libjpeg files to our build target + list(APPEND opencv_LIBS LIBSPNG_FILES) + list(APPEND opencv_LIBS LIBJPEG_FILES) +endif() From 065e329a4a5fffb6e958453bb06f5ca8aacb6097 Mon Sep 17 00:00:00 2001 From: Sayan Shaw Date: Wed, 15 May 2024 16:27:40 -0700 Subject: [PATCH 3/3] readd unset --- cmake/externals/opencv.cmake | 3 +++ 1 file changed, 3 insertions(+) diff --git a/cmake/externals/opencv.cmake b/cmake/externals/opencv.cmake index ae1a9883c..470694b2c 100644 --- a/cmake/externals/opencv.cmake +++ b/cmake/externals/opencv.cmake @@ -171,3 +171,6 @@ if (OCOS_ENABLE_OPENCV_CODECS) list(APPEND opencv_LIBS LIBSPNG_FILES) list(APPEND opencv_LIBS LIBJPEG_FILES) endif() + +# unset it to avoid affecting other projects. +unset(EXECUTABLE_OUTPUT_PATH CACHE) \ No newline at end of file