make the documentation? #277
Milestone
Comments
|
Hello, for the "5prealpha" release of a few years ago, the documentation is available here: https://cmusphinx.github.io/doc/pocketsphinx/ It will be updated for the first release candidate. To create it, you need Doxygen (https://doxygen.nl/) installed, and you can simply run "make docs" in your build directory, exactly as you say :) |
|
(of course, this is not documented so I will leave this issue open until it is!) |
|
Hi David,
okay took a look and tried writing some documentation as I'd aspire to
write it (attached). It does not reflect what is in the existing api, but
for the life of me I cannot see how one would connect a microphone or open
a file from that. Please take a look, and if you like it perhaps we could
have a conversation about me doing and maintaining the documentation for
pockesphinx5. Note it is full of inaccuracies as my C is quite old and I
only did minimal cpp before switching to Java, and I certainly have an OO
mindset. The use-case approach is however the key, I think, to producing
something people can use when they don't know how it works.
Best wishes and thanks again for taking an interest in resurrecting this,
P.
…On Wed, 10 Aug 2022 at 19:45, David Huggins-Daines ***@***.***> wrote:
(of course, this is not *documented* so I will leave this issue open
until it is!)
—
Reply to this email directly, view it on GitHub
<#277 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ALY3LWJQ4KRMZL4SUHH3MKTVYP2GBANCNFSM56FQ3HWA>
.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
|
Yes, because opening files and attaching to microphones is outside the scope of what a speech recognition engine should do, if you look at the others out there they generally don't do it either. It is extremely platform dependent, and there are a variety of libraries which work way better than what was in PocketSphinx, which didn't work for most people. I'm quite firm on this point as I don't have time to maintain yet another audio library (likewise, anything Java or really any language except C, Python, and JavaScript). I am working on a post for the CMUSphinx website which explains the rationale and provides alternatives. |
|
Okay, it was just an offer of help. Drawing a line around the speech
recognition engine is exactly what an API should do. The doxygen output
lists the functions in an attempt to say how it works; an API should
specify how to use it.
Providing one and only one "test harness" (in C) is what I'm suggesting -
let others create their own platform/language dependent interfaces. For
that to work however, the API needs to be crystal clear - and fit with the
variety of libraries out there.
The only thing I'm quite firm on is that use-cases are the way to structure
the manual.
let me know if I can help,
P
ps Have you played with raspberry pis? At 35 quid it might not be
unreasonable to suggest people set up a pi as a dedicated speech processor
and communicate with that via serial connection.
|
|
...
Gstreamer perhaps as a framework?
https://gstreamer.freedesktop.org/documentation/application-development/index.html?gi-language=c
…On Thu, 11 Aug 2022, 15:02 Peter Wallis, ***@***.***> wrote:
Okay, it was just an offer of help. Drawing a line around the speech
recognition engine is exactly what an API should do. The doxygen output
lists the functions in an attempt to say how it works; an API should
specify how to use it.
Providing one and only one "test harness" (in C) is what I'm suggesting -
let others create their own platform/language dependent interfaces. For
that to work however, the API needs to be crystal clear - and fit with the
variety of libraries out there.
The only thing I'm quite firm on is that use-cases are the way to
structure the manual.
let me know if I can help,
P
ps Have you played with raspberry pis? At 35 quid it might not be
unreasonable to suggest people set up a pi as a dedicated speech processor
and communicate with that via serial connection.
|
|
Yes, GStreamer is nice... the current pocketsphinx code is missing a plugin for it, but it is simply a matter of updating the build system. If you would like to contribute documentation that would be fantastic - sorry if I suggested otherwise, in fact this is the best kind of contribution anyone could make! The main page for documentation is in the pocketsphinx.h header file, and I think this is a good place for it as it means that people can read it without necessarily having to run Doxygen. |
|
pyaudio also works if you want live mic audio |
|
That's great; let's see how it works out. I am not the easiest person to
get on with so if it doesn't work out, well that is to be expected.
I didn't look too closely at streamer but i got the impression that it
would be the other way round ps being (set of) plugins for the gstreamer
framework. See
https://gstreamer.freedesktop.org/documentation/application-development/introduction/gstreamer.html?gi-language=c
Does this make sense?
…On Fri, 12 Aug 2022, 13:42 David Huggins-Daines, ***@***.***> wrote:
Yes, GStreamer is nice... the current pocketsphinx code is missing a
plugin for it, but it is simply a matter of updating the build system.
If you would like to contribute documentation that would be fantastic -
sorry if I suggested otherwise, in fact this is the best kind of
contribution anyone could make! The main page for documentation is in the
pocketsphinx.h header file, and I think this is a good place for it as it
means that people can read it without necessarily having to run Doxygen.
—
Reply to this email directly, view it on GitHub
<#277 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ALY3LWL2SHWGWHBO7VCPIU3VYZBCZANCNFSM56FQ3HWA>
.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
|
Here is a post explaining the removal of audio, with the simplest possible example code: https://cmusphinx.github.io/2022/08/pocketsphinx-continuous/ I'm writing a second one to explain doing VAD once I complete the integration of the VAD code. |
|
Looking good! I installed sox easy enough on this raspberry pi and tomorrow
(between other things) I will try and get "simple.c" running. As a typical
"user" (as opposed to a ASR researcher) one thing I can offer is to
document my experience as a first time user. I will do that and get back
to you.
p
…On Tue, 16 Aug 2022 at 15:06, David Huggins-Daines ***@***.***> wrote:
Here is a post explaining the removal of audio, with the simplest possible
example code: https://cmusphinx.github.io/2022/08/pocketsphinx-continuous/
I'm writing a second one to explain doing VAD once I complete the
integration of the VAD code.
—
Reply to this email directly, view it on GitHub
<#277 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ALY3LWOUCAZWQJBHWGM3UIDVZON7VANCNFSM56FQ3HWA>
.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
|
Note that the simple example is ... very simple :) it will seem to do nothing until you hit Control-C, which maybe isn't optimal, so I will fix that in a second. It also requires the -hmm, -lm, and -dict arguments, this too should get fixed in a second. |
|
Sorry I have been distracted the last few days and not done anything other
than look at the code. Hope the VAD development is going smoothly; just
wanted to let you know I am still very interested and will get more active
soon.
p
…On Tue, 16 Aug 2022 at 16:28, David Huggins-Daines ***@***.***> wrote:
Note that the simple example is ... very simple :) it will seem to do
nothing until you hit Control-C, which maybe isn't optimal, so I will fix
that in a second.
—
Reply to this email directly, view it on GitHub
<#277 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ALY3LWLNKACQJBSPQXGRHOLVZOXQ7ANCNFSM56FQ3HWA>
.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
|
No problem at all! Thanks for your offer to help. I am hoping to make a first release candidate today - I will be on vacation for a couple of weeks starting Monday, so that will be a good time to find the problems and write some documentation. |
|
... except I'm on vacation from Monday as well :-) (one week only though)
…On Fri, 19 Aug 2022 at 21:46, David Huggins-Daines ***@***.***> wrote:
No problem at all! Thanks for your offer to help. I am hoping to make a
first release candidate today - I will be on vacation for a couple of weeks
starting Monday, so that will be a good time to find the problems and write
some documentation.
—
Reply to this email directly, view it on GitHub
<#277 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ALY3LWMLPCAXGQFHCA2GIRTVZ7XAXANCNFSM56FQ3HWA>
.
You are receiving this because you authored the thread.Message ID:
***@***.***>
|
|
Hi David,
I see you have got back and got busy. I have had a bit of a fiddle about
with some suggestions on how I'd do documentation and got bogged down with
sox. I need to get back to another project for a bit so thought I'd show
you where I am heading in the hope you approve, and perhaps you can fix my
sox scripting, style, and perhaps make a few corrections with explanations
that I should incorporate. I think my code is self explanatory :-) but the
gist is that in order to draw a line around pocket sphinx vs the
environment, the methods in this would be part of ps, and the main(..)
would be test code to ensure the environment was set up and working. audio
file to speaker; audio file to audio file; mic to audio file... and in the
near future internet streamy thing to file; and microhone to web server.
just one format.
best,
p
…On Sat, 20 Aug 2022 at 10:37, Peter Wallis ***@***.***> wrote:
... except I'm on vacation from Monday as well :-) (one week only though)
On Fri, 19 Aug 2022 at 21:46, David Huggins-Daines <
***@***.***> wrote:
> No problem at all! Thanks for your offer to help. I am hoping to make a
> first release candidate today - I will be on vacation for a couple of weeks
> starting Monday, so that will be a good time to find the problems and write
> some documentation.
>
> —
> Reply to this email directly, view it on GitHub
> <#277 (comment)>,
> or unsubscribe
> <https://github.com/notifications/unsubscribe-auth/ALY3LWMLPCAXGQFHCA2GIRTVZ7XAXANCNFSM56FQ3HWA>
> .
> You are receiving this because you authored the thread.Message ID:
> ***@***.***>
>
|
|
Okay! I am back on it now and have learnt how to use sox. I will continue
to work on it this week (between things) and send you a draft text with
sample code plus some stuff you would need to put in the pocket sphinx
library.
p
…On Sun, 4 Sept 2022 at 08:48, Peter Wallis ***@***.***> wrote:
Hi David,
I see you have got back and got busy. I have had a bit of a fiddle about
with some suggestions on how I'd do documentation and got bogged down with
sox. I need to get back to another project for a bit so thought I'd show
you where I am heading in the hope you approve, and perhaps you can fix my
sox scripting, style, and perhaps make a few corrections with explanations
that I should incorporate. I think my code is self explanatory :-) but the
gist is that in order to draw a line around pocket sphinx vs the
environment, the methods in this would be part of ps, and the main(..)
would be test code to ensure the environment was set up and working. audio
file to speaker; audio file to audio file; mic to audio file... and in the
near future internet streamy thing to file; and microhone to web server.
just one format.
best,
p
On Sat, 20 Aug 2022 at 10:37, Peter Wallis ***@***.***> wrote:
> ... except I'm on vacation from Monday as well :-) (one week only though)
>
> On Fri, 19 Aug 2022 at 21:46, David Huggins-Daines <
> ***@***.***> wrote:
>
>> No problem at all! Thanks for your offer to help. I am hoping to make a
>> first release candidate today - I will be on vacation for a couple of weeks
>> starting Monday, so that will be a good time to find the problems and write
>> some documentation.
>>
>> —
>> Reply to this email directly, view it on GitHub
>> <#277 (comment)>,
>> or unsubscribe
>> <https://github.com/notifications/unsubscribe-auth/ALY3LWMLPCAXGQFHCA2GIRTVZ7XAXANCNFSM56FQ3HWA>
>> .
>> You are receiving this because you authored the thread.Message ID:
>> ***@***.***>
>>
>
|
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
I would like to look at the documentation for the API in C. I am guessing I need to type 'make,' or possibly 'cmake' somewhere so that doxygen produces the html files. I know enough to be dangerous so it has taken close on half a day to not figure this out. Could we have "make docs" a thing in the top level directory possibly? Or possibly
{ > make C-docs }
perhaps ...
The text was updated successfully, but these errors were encountered: