ffmpeg h264_vl42m2m encoded videos do not play back #5150

JooJooBee666 · 2022-08-29T18:12:49Z

Describe the bug

When trying to encode using h264_vl42m2m within ffmpeg, the resulting videos are useless. However, by simply changing to libx264, and nothing else, videos are fine. For example:

ffmpeg -i sample_1920x1080.mp4 -c:v h264_v4l2m2m -b:v 8M -c:a copy sample_1920x1080_test.mp4 - VIDEO NO GOOD
ffmpeg -i sample_1920x1080.mp4 -c:v libx264 -b:v 8M -c:a copy sample_1920x1080_test.mp4 - VIDEO WORKS

Here's the output from the h264_v4l2m2 encoder:

ffmpeg -i sample_1920x1080.mp4 -c:v h264_v4l2m2m -b:v 8M -c:a copy sample_1920x1080_test.mp4
ffmpeg version 4.3.4-0+deb11u1+rpt3 Copyright (c) 2000-2021 the FFmpeg developers
  built with gcc 10 (Debian 10.2.1-6)
  configuration: --prefix=/usr --extra-version=0+deb11u1+rpt3 --toolchain=hardened --incdir=/usr/include/aarch64-linux-gnu --enable-gpl --disable-stripping --enable-avresample --disable-filter=resample --enable-gnutls --enable-ladspa --enable-libaom --enable-libass --enable-libbluray --enable-libbs2b --enable-libcaca --enable-libcdio --enable-libcodec2 --enable-libdav1d --enable-libflite --enable-libfontconfig --enable-libfreetype --enable-libfribidi --enable-libgme --enable-libgsm --enable-libjack --enable-libmp3lame --enable-libmysofa --enable-libopenjpeg --enable-libopenmpt --enable-libopus --enable-libpulse --enable-librabbitmq --enable-librsvg --enable-librubberband --enable-libshine --enable-libsnappy --enable-libsoxr --enable-libspeex --enable-libsrt --enable-libssh --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvorbis --enable-libvpx --enable-libwavpack --enable-libwebp --enable-libx265 --enable-libxml2 --enable-libxvid --enable-libzmq --enable-libzvbi --enable-lv2 --enable-omx --enable-openal --enable-opencl --enable-opengl --enable-sdl2 --disable-mmal --enable-neon --enable-v4l2-request --enable-libudev --enable-epoxy --enable-sand --libdir=/usr/lib/aarch64-linux-gnu --arch=arm64 --enable-pocketsphinx --enable-libdc1394 --enable-libdrm --enable-vout-drm --enable-libiec61883 --enable-chromaprint --enable-frei0r --enable-libx264 --enable-shared
  libavutil      56. 51.100 / 56. 51.100
  libavcodec     58. 91.100 / 58. 91.100
  libavformat    58. 45.100 / 58. 45.100
  libavdevice    58. 10.100 / 58. 10.100
  libavfilter     7. 85.100 /  7. 85.100
  libavresample   4.  0.  0 /  4.  0.  0
  libswscale      5.  7.100 /  5.  7.100
  libswresample   3.  7.100 /  3.  7.100
  libpostproc    55.  7.100 / 55.  7.100
Input #0, mov,mp4,m4a,3gp,3g2,mj2, from 'sample_1920x1080.mp4':
  Metadata:
    major_brand     : isom
    minor_version   : 512
    compatible_brands: isomiso2avc1mp41
    encoder         : Lavf57.83.100
  Duration: 00:00:28.24, start: 0.000000, bitrate: 10836 kb/s
    Stream #0:0(und): Video: h264 (High) (avc1 / 0x31637661), yuv420p, 1920x1080, 10833 kb/s, 23.98 fps, 23.98 tbr, 24k tbn, 47.95 tbc (default)
    Metadata:
      handler_name    : VideoHandler
Stream mapping:
  Stream #0:0 -> #0:0 (h264 (native) -> h264 (h264_v4l2m2m))
Press [q] to stop, [?] for help
[h264_v4l2m2m @ 0x559c591f60]  <<< v4l2_encode_init: fmt=0/-1
[h264_v4l2m2m @ 0x559c591f60] Using device /dev/video11
[h264_v4l2m2m @ 0x559c591f60] driver 'bcm2835-codec' on card 'bcm2835-codec-encode' in mplane mode
[h264_v4l2m2m @ 0x559c591f60] requesting formats: output=YU12 capture=H264
[h264_v4l2m2m @ 0x559c591f60] Failed to set number of B-frames: Invalid argument
[h264_v4l2m2m @ 0x559c591f60] Failed to set gop size: Invalid argument
Output #0, mp4, to 'sample_1920x1080_test.mp4':
  Metadata:
    major_brand     : isom
    minor_version   : 512
    compatible_brands: isomiso2avc1mp41
    encoder         : Lavf58.45.100
    Stream #0:0(und): Video: h264 (h264_v4l2m2m) (avc1 / 0x31637661), yuv420p, 1920x1080, q=-1--1, 8000 kb/s, 23.98 fps, 24k tbn, 23.98 tbc (default)
    Metadata:
      handler_name    : VideoHandler
      encoder         : Lavc58.91.100 h264_v4l2m2m
frame=  677 fps= 21 q=-0.0 Lsize=   22094kB time=00:00:28.19 bitrate=6419.5kbits/s speed=0.871x    
video:22091kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.016396%

Here's the output from the libx264 encoder:

ffmpeg -i sample_1920x1080.mp4 -c:v libx264 -b:v 8M -c:a copy sample_1920x1080_test.mp4 
ffmpeg version 4.3.4-0+deb11u1+rpt3 Copyright (c) 2000-2021 the FFmpeg developers
  built with gcc 10 (Debian 10.2.1-6)
  configuration: --prefix=/usr --extra-version=0+deb11u1+rpt3 --toolchain=hardened --incdir=/usr/include/aarch64-linux-gnu --enable-gpl --disable-stripping --enable-avresample --disable-filter=resample --enable-gnutls --enable-ladspa --enable-libaom --enable-libass --enable-libbluray --enable-libbs2b --enable-libcaca --enable-libcdio --enable-libcodec2 --enable-libdav1d --enable-libflite --enable-libfontconfig --enable-libfreetype --enable-libfribidi --enable-libgme --enable-libgsm --enable-libjack --enable-libmp3lame --enable-libmysofa --enable-libopenjpeg --enable-libopenmpt --enable-libopus --enable-libpulse --enable-librabbitmq --enable-librsvg --enable-librubberband --enable-libshine --enable-libsnappy --enable-libsoxr --enable-libspeex --enable-libsrt --enable-libssh --enable-libtheora --enable-libtwolame --enable-libvidstab --enable-libvorbis --enable-libvpx --enable-libwavpack --enable-libwebp --enable-libx265 --enable-libxml2 --enable-libxvid --enable-libzmq --enable-libzvbi --enable-lv2 --enable-omx --enable-openal --enable-opencl --enable-opengl --enable-sdl2 --disable-mmal --enable-neon --enable-v4l2-request --enable-libudev --enable-epoxy --enable-sand --libdir=/usr/lib/aarch64-linux-gnu --arch=arm64 --enable-pocketsphinx --enable-libdc1394 --enable-libdrm --enable-vout-drm --enable-libiec61883 --enable-chromaprint --enable-frei0r --enable-libx264 --enable-shared
  libavutil      56. 51.100 / 56. 51.100
  libavcodec     58. 91.100 / 58. 91.100
  libavformat    58. 45.100 / 58. 45.100
  libavdevice    58. 10.100 / 58. 10.100
  libavfilter     7. 85.100 /  7. 85.100
  libavresample   4.  0.  0 /  4.  0.  0
  libswscale      5.  7.100 /  5.  7.100
  libswresample   3.  7.100 /  3.  7.100
  libpostproc    55.  7.100 / 55.  7.100
Input #0, mov,mp4,m4a,3gp,3g2,mj2, from 'sample_1920x1080.mp4':
  Metadata:
    major_brand     : isom
    minor_version   : 512
    compatible_brands: isomiso2avc1mp41
    encoder         : Lavf57.83.100
  Duration: 00:00:28.24, start: 0.000000, bitrate: 10836 kb/s
    Stream #0:0(und): Video: h264 (High) (avc1 / 0x31637661), yuv420p, 1920x1080, 10833 kb/s, 23.98 fps, 23.98 tbr, 24k tbn, 47.95 tbc (default)
    Metadata:
      handler_name    : VideoHandler
Stream mapping:
  Stream #0:0 -> #0:0 (h264 (native) -> h264 (libx264))
Press [q] to stop, [?] for help
[libx264 @ 0x55bd668f60] using cpu capabilities: ARMv8 NEON
[libx264 @ 0x55bd668f60] profile High, level 4.0, 4:2:0, 8-bit
[libx264 @ 0x55bd668f60] 264 - core 160 r3011 cde9a93 - H.264/MPEG-4 AVC codec - Copyleft 2003-2020 - http://www.videolan.org/x264.html - options: cabac=1 ref=3 deblock=1:0:0 analyse=0x3:0x113 me=hex subme=7 psy=1 psy_rd=1.00:0.00 mixed_ref=1 me_range=16 chroma_me=1 trellis=1 8x8dct=1 cqm=0 deadzone=21,11 fast_pskip=1 chroma_qp_offset=-2 threads=6 lookahead_threads=1 sliced_threads=0 nr=0 decimate=1 interlaced=0 bluray_compat=0 constrained_intra=0 bframes=3 b_pyramid=2 b_adapt=1 b_bias=0 direct=1 weightb=1 open_gop=0 weightp=2 keyint=250 keyint_min=23 scenecut=40 intra_refresh=0 rc_lookahead=40 rc=abr mbtree=1 bitrate=8000 ratetol=1.0 qcomp=0.60 qpmin=0 qpmax=69 qpstep=4 ip_ratio=1.40 aq=1:1.00
Output #0, mp4, to 'sample_1920x1080_test.mp4':
  Metadata:
    major_brand     : isom
    minor_version   : 512
    compatible_brands: isomiso2avc1mp41
    encoder         : Lavf58.45.100
    Stream #0:0(und): Video: h264 (libx264) (avc1 / 0x31637661), yuv420p, 1920x1080, q=-1--1, 8000 kb/s, 23.98 fps, 24k tbn, 23.98 tbc (default)
    Metadata:
      handler_name    : VideoHandler
      encoder         : Lavc58.91.100 libx264
    Side data:
      cpb: bitrate max/min/avg: 0/0/8000000 buffer size: 0 vbv_delay: N/A
frame=  677 fps=5.2 q=-1.0 Lsize=   27765kB time=00:00:28.11 bitrate=8090.9kbits/s speed=0.216x       
video:27756kB audio:0kB subtitle:0kB other streams:0kB global headers:0kB muxing overhead: 0.032574%
[libx264 @ 0x55bd668f60] frame I:3     Avg QP:20.58  size:457560
[libx264 @ 0x55bd668f60] frame P:177   Avg QP:26.51  size:116193
[libx264 @ 0x55bd668f60] frame B:497   Avg QP:31.01  size: 13043
[libx264 @ 0x55bd668f60] consecutive B-frames:  0.4%  3.0%  6.2% 90.4%
[libx264 @ 0x55bd668f60] mb I  I16..4:  2.9% 60.5% 36.6%
[libx264 @ 0x55bd668f60] mb P  I16..4:  0.7%  4.5%  1.2%  P16..4: 36.3% 26.2% 22.3%  0.0%  0.0%    skip: 8.7%
[libx264 @ 0x55bd668f60] mb B  I16..4:  0.0%  0.0%  0.0%  B16..8: 52.8%  6.5%  2.1%  direct: 1.4%  skip:37.2%  L0:38.7% L1:50.5% BI:10.7%
[libx264 @ 0x55bd668f60] final ratefactor: 25.27
[libx264 @ 0x55bd668f60] 8x8 transform intra:67.9% inter:56.4%
[libx264 @ 0x55bd668f60] coded y,uvDC,uvAC intra: 77.1% 55.9% 6.6% inter: 17.7% 8.9% 0.0%
[libx264 @ 0x55bd668f60] i16 v,h,dc,p: 18% 21% 22% 39%
[libx264 @ 0x55bd668f60] i8 v,h,dc,ddl,ddr,vr,hd,vl,hu: 14% 12% 21%  6% 12%  9% 10%  7%  9%
[libx264 @ 0x55bd668f60] i4 v,h,dc,ddl,ddr,vr,hd,vl,hu: 17% 12% 21%  7% 11%  8% 10%  6%  8%
[libx264 @ 0x55bd668f60] i8c dc,h,v,p: 65% 16% 15%  3%
[libx264 @ 0x55bd668f60] Weighted P-Frames: Y:6.8% UV:0.6%
[libx264 @ 0x55bd668f60] ref P L0: 61.3% 26.3% 10.0%  2.2%  0.2%
[libx264 @ 0x55bd668f60] ref B L0: 95.4%  4.3%  0.4%
[libx264 @ 0x55bd668f60] ref B L1: 98.1%  1.9%
[libx264 @ 0x55bd668f60] kb/s:8052.25

I'm really not sure what to try from this point. I am all the way up to date packages and firmware. I would expect this encoder to just work like all the others but maybe we're not there yet. Any ideas on how to get this outputting usable hardware encoded videos here?

Thanks!

Steps to reproduce the behaviour

try:
ffmpeg -i sample_1920x1080.mp4 -c:v h264_v4l2m2m -b:v 8M -c:a copy sample_1920x1080_test.mp4
...and watch as it consumes power and outputs garbage.

Device (s)

Raspberry Pi 4 Mod. B

System

Raspberry Pi 4B, 8GB model

Raspberry Pi reference 2022-04-04
Generated using pi-gen, https://github.com/RPi-Distro/pi-gen, 27a8050c3c06e567c794620394a8c2d74262a516, stage4

Aug 23 2022 14:23:34
Copyright (c) 2012 Broadcom
version bf881c824bac214626a9f78094d614f7db462c34 (clean) (release) (start)

Linux pi5 5.15.61-v8+ #1578 SMP PREEMPT Wed Aug 24 11:51:18 BST 2022 aarch64 GNU/Linux

Logs

No response

Additional context

No response

The text was updated successfully, but these errors were encountered:

popcornmix · 2022-09-10T14:51:30Z

@jc-kynesim can you reproduce this?

jc-kynesim · 2022-09-12T16:57:45Z

I now understand what is going on - the mp4 muxer appears to require the SPS/PPS extradata at init time if it is going to take it from the context, but the h/w encoder won't provide it before it has been given its first frame (so after that). However it can also be passed along attached to the encoded packet as sidedata so I'll fix the encoder to do that and fix the mp4 muxer to respect it.

6by9 · 2022-09-13T08:23:57Z

so I'll fix the encoder to do that and fix the mp4 muxer to respect it

Encoder being the FFmpeg encode side, or does it need a kernel change?

jc-kynesim · 2022-09-13T09:54:59Z

Just ffmpeg - it all works fine if you have a raw h264 stream as the output type from ffmpeg - it is just separating off SPS/PPS for the mp4 muxer that has issues and that is all internal to ffmpeg (I do make the nasty assumption that SPS/PPS turns up on its own in the 1st capture frame which isn't a part of the V4L2 spec but I recall you saying was always true for our encoder).

6by9 · 2022-09-13T10:29:17Z

There is a V4L2 control V4L2_CID_MPEG_VIDEO_HEADER_MODE. FFmpeg chooses V4L2_MPEG_VIDEO_HEADER_MODE_SEPARATE to ask for the the header bytes to be separate from the 1st frame. The default is that they are joined with the first frame.
https://www.kernel.org/doc/html/latest/userspace-api/media/v4l/ext-ctrls-codec.html

jc-kynesim · 2022-09-13T10:54:45Z

Ah! Missed that. Excellent.

jc-kynesim · 2022-09-15T10:42:19Z

Fix now pushed to my ffmpeg repo (https://github.com/jc-kynesim/rpi-ffmpeg.git branch test/4.3.4/rpi_main if you want to build it yourself). It'll be in the next ffmpeg update. Also fixes the same issue in mkv generation.

JooJooBee666 · 2022-09-15T15:27:18Z

@jc-kynesim What are the build parameters used to build from source? I'm not having any luck.

jc-kynesim · 2022-09-15T15:48:43Z

You should find build/install instructions in pi-util/BUILD.txt - if they don't work or are unclear please tell me

JooJooBee666 · 2022-09-15T15:54:44Z

Ahh, didn't know that was there. I'll try that and let you know.

JooJooBee666 · 2022-09-15T16:11:44Z

No luck. Ran make celan and manually cleaned up the out folder but it's not working. Not sure what to do about this error:

pi@pi5:~/source/rpi-ffmpeg $ pi-util/conf_native.sh --noshared
Configure for native build
M/C aarch64
Static libs
Destination directory: /home/pi/source/rpi-ffmpeg/out/arm64-bullseye-4.3.4-static-rel
Out of tree builds are impossible with config.h in source dir.

jc-kynesim · 2022-09-15T16:13:05Z

Run "git clean -dxf" - that will get rid of the old in-tree objects etc.

JooJooBee666 · 2022-09-15T21:35:31Z

Well, finally got back around to this. It completed compiling but after I ran a make install but it didn't actually install ffmpeg anywhere. Now its just completely broken (i.e. no ffmpeg in path). I tried both shared and static. What did I miss?

Update: NM, I see that this doesn't build something that is installed with make install and it just needs to be manually copied. This version produces playable videos! 🥂

JooJooBee666 · 2022-09-16T00:41:01Z

I just noted that 4.3.4 was up on the repos. Is that this version already or do we need to wait a bit longer?

Also, something I did here today completely broke the camera. libcamera-jpeg says:

0:04:09.668224540] [1723]  INFO Camera camera_manager.cpp:293 libcamera v0.0.0+3866-0c55e522
[0:04:09.688411074] [1727] ERROR V4L2 v4l2_device.cpp:349 'imx477 10-001a': Unable to set controls: Device or resource busy

I ran in to this problem before and reported it but it got closed because I couldn't replicate. Well, here its replicated again. What part of this could have permanently broken (until a new image is loaded) the camera?

jc-kynesim · 2022-09-16T08:17:18Z

a) I'm pretty sure that the BUILD.txt I pointed you at made it clear that you needed to get rid of the previous version and copy the result if you wanted it globally accessible.
b) 4.3.4 is the version used by PiOS and in the branch I pointed you at
c) Dunno about the camera issue - maybe its been grabbed by libcamera in some way? not directly my area.

JooJooBee666 · 2022-09-16T13:01:14Z

Yeah, I didn't finish reading the build.txt but noted that afterward.

Regarding the version number, I was just trying to clarify since that is also the version number on your repo but I confirmed that that version is also broken still.

Regarding the camera, I wasn't expecting you to know, but was hoping maybe something would make sense when you heard it. I'm just reloading.....AGAIN 😠

ViktuK359 · 2022-09-28T07:40:31Z

Fix now pushed to my ffmpeg repo (https://github.com/jc-kynesim/rpi-ffmpeg.git branch test/4.3.4/rpi_main if you want to build it yourself). It'll be in the next ffmpeg update. Also fixes the same issue in mkv generation.

And how update for RPi4 HAOS?

pelwell · 2022-09-28T07:49:33Z

That's a question for the HAOS devs, surely?

roger- · 2022-10-28T14:26:34Z

Does anyone have a binary with this fix incorporated?

hackneyb · 2022-11-19T23:57:16Z

Fix now pushed to my ffmpeg repo (https://github.com/jc-kynesim/rpi-ffmpeg.git branch test/4.3.4/rpi_main if you want to build it yourself). It'll be in the next ffmpeg update.

@jc-kynesim Would you have an estimate on how long these releases take and which version will contain this fix? My path to have this working is wait for the ffmpeg update and then for HAOS to uptake that ... I'd appreciate you shedding some light on this for me. Thanks

popcornmix · 2022-11-20T10:57:16Z

There was a recent ffmpeg update in RPiOS (timestamp shows it was built on Nov 7).

LaurenceGough · 2022-11-21T00:23:40Z

Is there any way to get that ffmpeg update installed using apt on a Pi? I tried installing the deb files manually but it's one dependency after another missing and it was getting a bit crazy, obviously this was what apt was made for.

apt sources is the standard: deb https://archive.raspberrypi.org/debian/ bullseye main

Many thanks

Edit:

I managed it somehow (7:4.3.5-0+deb11u1+rpt2)
but the same issue remains.

jherby2k · 2022-11-21T18:50:39Z

Fix now pushed to my ffmpeg repo (https://github.com/jc-kynesim/rpi-ffmpeg.git branch test/4.3.4/rpi_main if you want to build it yourself). It'll be in the next ffmpeg update. Also fixes the same issue in mkv generation.

@jc-kynesim This does not appear to have trickled up, as nightly builds still don't include these patches. They also target a very old version of ffmpeg. Any update for us?

mantaalex · 2023-01-03T20:34:40Z

still doesn't work for some people @jc-kynesim do you have some news? thanks on advance

jc-kynesim · 2023-01-04T10:51:48Z

When you say 'still does not work for some people' - does that imply that it does work for others?
Assuming so, what are the differences between those it does work for and those it doesn't? I'll happily debug but I need an actual failing example on PiOS.
The initial example given in this thread works for me with the current PiOS distribution ffmpeg.

hackneyb · 2023-01-06T13:39:41Z

When you say 'still does not work for some people' - does that imply that it does work for others? Assuming so, what are the differences between those it does work for and those it doesn't? I'll happily debug but I need an actual failing example on PiOS. The initial example given in this thread works for me with the current PiOS distribution ffmpeg.

Not a developer, but there are extensive threads detailing what doesn't work (here: blakeblackshear/frigate#3780) and an official bug fix request at the ffmpeg page here: https://trac.ffmpeg.org/ticket/10060. The specific case that locks "some people" out of using acceleration is an RPi4 64 bit scenario with ffmpeg packaged from your repository as part of Frigate. Frigate runs as an add-on (in a container) in Home Assistant OS with the affected ffmpeg version (4.3.3?) and acceleration used by cameras does not work. Add-on container images can be manually patched when they launch to fix but this is temporary until the next restart and beyond the hobbyist / retail user (me) capabilities. I'm sure I could pull it off if I dedicated an evening to it but it's a band-aid. Since Home Assistant can be ran in multiple configurations, only the typical HAOS + add-on install mode is not permanently fixable, as far as I know. If you run HA on non-restricted linux distribution then ffmpeg that Frigate uses can be patched easier, if you know what you're doing.

More info here, with many messy examples of what works and what doesn't: blakeblackshear/frigate#3780

@jherby2k has a good handle on where it is reproducible and where it isn't, and took the initiative to open the ffmpeg request which in theory will trickle down to your project which will fix the ffmpeg version the Frigate project ultimately uses downstream. I think the last idea here was if you can help influence the ffmpeg project bug fix request, or debug and fix it in parallel in your project. The definition of the issue is obtainable, and I can help get you the info on what I'm running to help, but the Frigate project link above is teeming with examples.

Thanks @jc-kynesim

6by9 · 2023-01-06T14:21:10Z

Is that FFmpeg bug report valid for this issue? It's referencing DECODE, whereas this is ENCODE.

Accurate bug reports are critical if you want developers to investigate them.

hackneyb · 2023-01-07T00:56:02Z

Is that FFmpeg bug report valid for this issue? It's referencing DECODE, whereas this is ENCODE.

Accurate bug reports are critical if you want developers to investigate them.

I get what you're saying, it would be hilarious if it was the reverse. Though I believe this issue (5150) describes the fix enough for people to take the solution, apply it to their custom compiled versions of ffmpeg and restore the functionality in Frigate. Frigate uses ffmpeg to decode https://docs.frigate.video/configuration/hardware_acceleration but maybe its not the act of decoding that's broken, but some other common requirement for encoding/decoding to take place. I could be totally wrong (disclaimer).

Does this provide more of a clue why people think this issue for encoding is linked with accelerated decoding issue in Frigate? jc-kynesim/rpi-ffmpeg#53 (comment)

6by9 · 2023-01-07T20:59:38Z

#5150 (comment)

I now understand what is going on - the mp4 muxer appears to require the SPS/PPS extradata at init time if it is going to take it from the context, but the h/w encoder won't provide it before it has been given its first frame (so after that). However it can also be passed along attached to the encoded packet as sidedata so I'll fix the encoder to do that and fix the mp4 muxer to respect it.

#5150 (comment)

Just ffmpeg - it all works fine if you have a raw h264 stream as the output type from ffmpeg - it is just separating off SPS/PPS for the mp4 muxer that has issues and that is all internal to ffmpeg (I do make the nasty assumption that SPS/PPS turns up on its own in the 1st capture frame which isn't a part of the V4L2 spec but I recall you saying was always true for our encoder).

#5150 (comment)

Fix now pushed to my ffmpeg repo (https://github.com/jc-kynesim/rpi-ffmpeg.git branch test/4.3.4/rpi_main if you want to build it yourself). It'll be in the next ffmpeg update. Also fixes the same issue in mkv generation.

That's the sum total of the fix for this issue related to encoding videos.

If using https://github.com/jc-kynesim/rpi-ffmpeg.git also fixes something else, then great, but it is not related to this issue regardless of what others think.
Telling the FFmpeg devs to look at this issue will not help them solve any decode related issue.

blakeblackshear · 2023-01-07T21:10:57Z

As the frigate maintainer, I can confirm that this issue is not relevant.

hackneyb · 2023-01-07T23:26:16Z

If using https://github.com/jc-kynesim/rpi-ffmpeg.git also fixes something else, then great, but it is not related to this issue regardless of what others think. Telling the FFmpeg devs to look at this issue will not help them solve any decode related issue.

Thanks for taking the time to do this, then. Clarity is good.

hackneyb · 2023-01-07T23:27:02Z

As the frigate maintainer, I can confirm that this issue is not relevant.

thank you

kylehendricks mentioned this issue Sep 7, 2022

[Support]: hwaccel stopped working on raspberry pi after 5.15.61 kernel update blakeblackshear/frigate#3780

Closed

JooJooBee666 closed this as completed Sep 17, 2022

idontcare99999 mentioned this issue Sep 22, 2022

[HW Accel Support]: Bullseye RASPI 64bit v0.11-rc(x) blakeblackshear/frigate#3868

Closed

hackneyb mentioned this issue Nov 20, 2022

h264 hardware acceleration in ffmpeg broken after 5.15.61 kernel update home-assistant/operating-system#2238

Closed

NickM-27 mentioned this issue Nov 21, 2022

v4l2m2m hwaccel not working on rpi kernel 5.15.61+ BtbN/FFmpeg-Builds#217

Closed

mattdsteele mentioned this issue Feb 26, 2023

h264_v4l2m2m as a hardware encoder for streaming owncast/owncast#1379

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

ffmpeg h264_vl42m2m encoded videos do not play back #5150

ffmpeg h264_vl42m2m encoded videos do not play back #5150

JooJooBee666 commented Aug 29, 2022 •

edited

Loading

popcornmix commented Sep 10, 2022

jc-kynesim commented Sep 12, 2022

6by9 commented Sep 13, 2022

jc-kynesim commented Sep 13, 2022

6by9 commented Sep 13, 2022

jc-kynesim commented Sep 13, 2022

jc-kynesim commented Sep 15, 2022

JooJooBee666 commented Sep 15, 2022 •

edited

Loading

jc-kynesim commented Sep 15, 2022

JooJooBee666 commented Sep 15, 2022

JooJooBee666 commented Sep 15, 2022 •

edited

Loading

jc-kynesim commented Sep 15, 2022

JooJooBee666 commented Sep 15, 2022 •

edited

Loading

JooJooBee666 commented Sep 16, 2022 •

edited

Loading

jc-kynesim commented Sep 16, 2022 •

edited

Loading

JooJooBee666 commented Sep 16, 2022 •

edited

Loading

ViktuK359 commented Sep 28, 2022

pelwell commented Sep 28, 2022

roger- commented Oct 28, 2022

hackneyb commented Nov 19, 2022

popcornmix commented Nov 20, 2022 •

edited

Loading

LaurenceGough commented Nov 21, 2022 •

edited

Loading

jherby2k commented Nov 21, 2022

mantaalex commented Jan 3, 2023

jc-kynesim commented Jan 4, 2023

hackneyb commented Jan 6, 2023 •

edited

Loading

6by9 commented Jan 6, 2023

hackneyb commented Jan 7, 2023 •

edited

Loading

6by9 commented Jan 7, 2023

blakeblackshear commented Jan 7, 2023

hackneyb commented Jan 7, 2023

hackneyb commented Jan 7, 2023

ffmpeg h264_vl42m2m encoded videos do not play back #5150

ffmpeg h264_vl42m2m encoded videos do not play back #5150

Comments

JooJooBee666 commented Aug 29, 2022 • edited Loading

Describe the bug

Steps to reproduce the behaviour

Device (s)

System

Logs

Additional context

popcornmix commented Sep 10, 2022

jc-kynesim commented Sep 12, 2022

6by9 commented Sep 13, 2022

jc-kynesim commented Sep 13, 2022

6by9 commented Sep 13, 2022

jc-kynesim commented Sep 13, 2022

jc-kynesim commented Sep 15, 2022

JooJooBee666 commented Sep 15, 2022 • edited Loading

jc-kynesim commented Sep 15, 2022

JooJooBee666 commented Sep 15, 2022

JooJooBee666 commented Sep 15, 2022 • edited Loading

jc-kynesim commented Sep 15, 2022

JooJooBee666 commented Sep 15, 2022 • edited Loading

JooJooBee666 commented Sep 16, 2022 • edited Loading

jc-kynesim commented Sep 16, 2022 • edited Loading

JooJooBee666 commented Sep 16, 2022 • edited Loading

ViktuK359 commented Sep 28, 2022

pelwell commented Sep 28, 2022

roger- commented Oct 28, 2022

hackneyb commented Nov 19, 2022

popcornmix commented Nov 20, 2022 • edited Loading

LaurenceGough commented Nov 21, 2022 • edited Loading

jherby2k commented Nov 21, 2022

mantaalex commented Jan 3, 2023

jc-kynesim commented Jan 4, 2023

hackneyb commented Jan 6, 2023 • edited Loading

6by9 commented Jan 6, 2023

hackneyb commented Jan 7, 2023 • edited Loading

6by9 commented Jan 7, 2023

blakeblackshear commented Jan 7, 2023

hackneyb commented Jan 7, 2023

hackneyb commented Jan 7, 2023

JooJooBee666 commented Aug 29, 2022 •

edited

Loading

JooJooBee666 commented Sep 15, 2022 •

edited

Loading

JooJooBee666 commented Sep 15, 2022 •

edited

Loading

JooJooBee666 commented Sep 15, 2022 •

edited

Loading

JooJooBee666 commented Sep 16, 2022 •

edited

Loading

jc-kynesim commented Sep 16, 2022 •

edited

Loading

JooJooBee666 commented Sep 16, 2022 •

edited

Loading

popcornmix commented Nov 20, 2022 •

edited

Loading

LaurenceGough commented Nov 21, 2022 •

edited

Loading

hackneyb commented Jan 6, 2023 •

edited

Loading

hackneyb commented Jan 7, 2023 •

edited

Loading