microphone and sound access in Ada

microphone and sound access in Ada

[Tony Gair]

I would be very interested in capturing microphone input and also
using sound in an Ada application, (basically to use in a distributed
application for voice conferencing)

The system I am targetting is Linux but I would like to be able to
extend this to windows later.

Re: microphone and sound access in Ada

[Preben Randhol]

Tony Gair wrote:
> I would be very interested in capturing microphone input and also
> using sound in an Ada application, (basically to use in a distributed
> application for voice conferencing)

Check out:

   http://www.speex.org/

For recording I guess one should consider using Alsa
http://www.alsa-project.org/ as OSS is on the way out.

> The system I am targetting is Linux but I would like to be able to
> extend this to windows later.

For sound in general I think a wrapper of the Ogg Vorbis libraries would
be very nice. Ogg files take as little space as MP3, but have better
quality and being free of any patent restrictions. Besides new MP3
players seem also to now add support for ogg. http://www.vorbis.com/
http://www.vorbis.com/faq.psp Besides Ogg works on most platforms.

Do anybody know of software/documentation/theory for matching two sound
files like Language training programs do. I mean that you try to
pronouce the word and the program tells you how well you did compared to
the pronounciations of somebody who speaks the language. You have to
account for different pitches etc... so how do they do it?

-- 
This is Ada95 land. On quiet nights you can hear C programmers debug.

Re: microphone and sound access in Ada

[Adrian Knoth]

Preben Randhol <randhol+news@pvv.org> wrote:

> For recording I guess one should consider using Alsa
> http://www.alsa-project.org/ as OSS is on the way out.

ACK.
 
>> The system I am targetting is Linux but I would like to be able to
>> extend this to windows later.
> For sound in general I think a wrapper of the Ogg Vorbis libraries would
> be very nice. 

Yes. I proposed that some weeks ago, perhaps we should start. Any further
interests?


-- 
mail: adi@thur.de  	http://adi.thur.de	PGP: v2-key via keyserver

Denkste, ich fahr da mit irgendwelchen Sippschaften hin? Damit f�llt
man ja noch auf.	   (Florian Henschel �ber "Urlaub ohne Eltern")

Re: microphone and sound access in Ada

[Preben Randhol]

Adrian Knoth wrote:
> 
> Yes. I proposed that some weeks ago, perhaps we should start. Any further
> interests?

I'm in. Although, I have to read up on how to do C bindings first as I
haven't done this before. I guess GtkAda is a good source for examples
:-)

There are three libraries:

libao: 

   Libao is a cross-platform library that allows programs to output PCM
   audio data to the native audio devices on a wide variety of platforms.
   It currently supports:

       * OSS (Open Sound System)
       * ESD (ESounD)
       * ALSA (Advanced Linux Sound Architecture)
       * Sun audio system (used in Solaris, OpenBSD, and NetBSD)
       * aRts (Analog Realtime Synthesizer) 

(hope this can do input as well)

libogg:

   Libogg contains necessary functionality to create, decode, and work with
   Ogg bitstreams.

libvorbis:

 * libvorbisenc

    Libvorbisenc is a convenient API for setting up an encoding
    environment using libvorbis. Libvorbisenc encapsulates the actions
    needed to set up the encoder properly.

 * Vorbisfile

    The Vorbisfile library provides a convenient high-level API for
    decoding and basic manipulation of all Vorbis I audio streams.
    Libvorbisfile is implemented as a layer on top of Xiph.org's
    reference libogg and libvorbis libraries.

    Vorbisfile can be used along with any ANSI compliant stdio
    implementation for file/stream access, or use custom stream i/o
    routines provided by the embedded environment. Both uses are
    described in detail in this documentation.


-- 
This is Ada95 land. On quiet nights you can hear C programmers debug.

Re: microphone and sound access in Ada

[Adrian Knoth]

Preben Randhol <randhol+news@pvv.org> wrote:

>> Yes. I proposed that some weeks ago, perhaps we should start. Any further
>> interests?
> I'm in. Although, I have to read up on how to do C bindings first as I
> haven't done this before. I guess GtkAda is a good source for examples
> :-)

I think so, too. Most is pragma Import :)
 
> libao: 
>    Libao is a cross-platform library that allows programs to output PCM
>    audio data to the native audio devices on a wide variety of platforms.
>    It currently supports:
> 
>        * OSS (Open Sound System)
>        * ESD (ESounD)
>        * ALSA (Advanced Linux Sound Architecture)
>        * Sun audio system (used in Solaris, OpenBSD, and NetBSD)
>        * aRts (Analog Realtime Synthesizer) 
> 
> (hope this can do input as well)

Well, libao is a generic audio-output-library. We might want to use it
or simply cat to /dev/audio :) 
 
> libogg:
>    Libogg contains necessary functionality to create, decode, and work with
>    Ogg bitstreams.

This would be important for speex.
 
> libvorbis:
>  * libvorbisenc
> 
>     Libvorbisenc is a convenient API for setting up an encoding
>     environment using libvorbis. Libvorbisenc encapsulates the actions
>     needed to set up the encoder properly.

This is a later issue. Sure encoding is important, but playback might be
in first release, because it's easier.
 
>  * Vorbisfile
>     The Vorbisfile library provides a convenient high-level API for
>     decoding and basic manipulation of all Vorbis I audio streams.

Nice. I guess we can start right here. It is linked against libogg, so
we can neglect a binding to it first:

adi@drcomp:~$ ldd /usr/local/lib/libvorbisfile.so.3.0.0 | grep ogg
        libogg.so.0 => /usr/local/lib/libogg.so.0 (0x40033000)

>     Libvorbisfile is implemented as a layer on top of Xiph.org's
>     reference libogg and libvorbis libraries.

Sure.

It might even be fun to write a binding to the ALSA-API, best case would
be a native kernel-binding. Lutz? On the other hand, a wrapper around
the C-library would work as well.

-- 
mail: adi@thur.de  	http://adi.thur.de	PGP: v2-key via keyserver

Liebe deinen n�chsten so wie den ersten!

Re: microphone and sound access in Ada

[Preben Randhol]

Adrian Knoth wrote:
> 
> Well, libao is a generic audio-output-library. We might want to use it
> or simply cat to /dev/audio :) 

I think we want it to use /dev/dsp.

> This is a later issue. Sure encoding is important, but playback might be
> in first release, because it's easier.

Agree.

> 
> Nice. I guess we can start right here. It is linked against libogg, so
> 
> It might even be fun to write a binding to the ALSA-API, best case would
> be a native kernel-binding. Lutz? On the other hand, a wrapper around
> the C-library would work as well.

Yes, however the big problem is the /dev/dsp sharing. I mean two
programs have to fight for /dev/dsp unless one can use esd or similar.
So, I guess the libao might be better if it also works for input. Have
to check it though.

Preben

Re: microphone and sound access in Ada

[Preben Randhol]

Preben Randhol wrote:
> Yes, however the big problem is the /dev/dsp sharing. I mean two
> programs have to fight for /dev/dsp unless one can use esd or similar.
> So, I guess the libao might be better if it also works for input. Have
> to check it though.

On the other hand it is not /dev/dsp we are interested in in recording
*BONK* and libao looks to be for output only. So I think libao for
output and that we need some bindings to alsa as well.

Preben

Re: microphone and sound access in Ada

[Egil Harald H�vik]

"Adrian Knoth" <adi@drcomp.erfurt.thur.de> wrote in message
news:slrnb7e3uf.j6f.adi@drcomp.erfurt.thur.de...
> Preben Randhol <randhol+news@pvv.org> wrote:
>
> > For recording I guess one should consider using Alsa
> > http://www.alsa-project.org/ as OSS is on the way out.
>
> ACK.
>
> >> The system I am targetting is Linux but I would like to be able to
> >> extend this to windows later.
> > For sound in general I think a wrapper of the Ogg Vorbis libraries would
> > be very nice.
>
> Yes. I proposed that some weeks ago, perhaps we should start. Any further
> interests?



I have already done some work on this. I have bindings to libogg, libvorbis
and
libvorbisfile, although I lack some testing. The libvorbisfile binding seems
to
work OK (I saved the output (RAW/PCM) to a file, converted the raw data to
WAV
using Audacity, and played it using WinAmp).

Maybe I should structure and comment the code a bit, before adding license
headers
and releasing them.


~egilhh

Re: microphone and sound access in Ada

[Adrian Knoth]

Egil Harald H�vik <egil.harald.hoevik.nospam@kongsberg.com> wrote:

[Binding to Ogg/Vorbis]
> I have already done some work on this. I have bindings to libogg, libvorbis
> and libvorbisfile, although I lack some testing. 

If you like I can do some of it for you/public.

> The libvorbisfile binding seems to work OK (I saved the output (RAW/PCM) to 
> a file, converted the raw data to WAV using Audacity, and played it using 
> WinAmp).

That's a lot of efforts for simple audio :) You can take the stream and
write it to /dev/audio if the device-settings are appropriate.

So, well, let's concentrate on libao.
 
> Maybe I should structure and comment the code a bit, before adding license
> headers and releasing them.

Which license do you prefer? I suggest putting the stuff into CVS and let
the world help you in every possible way.
 

-- 
mail: adi@thur.de  	http://adi.thur.de	PGP: v2-key via keyserver

AIDS: Ab in den Sarg

Re: microphone and sound access in Ada

[Preben Randhol]

Adrian Knoth wrote:
> Which license do you prefer? I suggest putting the stuff into CVS and let
> the world help you in every possible way.

Would be nice if it had a GMGPL (or BSD license) so also commercial
products can use the libraries.

Preben

Re: microphone and sound access in Ada

[Egil Harald H�vik]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1463 bytes --]


"Adrian Knoth" <adi@drcomp.erfurt.thur.de> wrote in message
news:slrnb7e9s9.k3d.adi@drcomp.erfurt.thur.de...
> Egil Harald H�vik <egil.harald.hoevik.nospam@kongsberg.com> wrote:
>
> [Binding to Ogg/Vorbis]
> > I have already done some work on this. I have bindings to libogg,
libvorbis
> > and libvorbisfile, although I lack some testing.
>
> If you like I can do some of it for you/public.
>

I guess that wouldn't be a bad idea  :-)


> > The libvorbisfile binding seems to work OK (I saved the output (RAW/PCM)
to
> > a file, converted the raw data to WAV using Audacity, and played it
using
> > WinAmp).
>
> That's a lot of efforts for simple audio :) You can take the stream and
> write it to /dev/audio if the device-settings are appropriate.
>

Ah, I guess I forgot to mention the little detail of me running Windows...
(With my new hardware installed, I guess my old machine would make a nice
Linux server, though...)


> So, well, let's concentrate on libao.
>
> > Maybe I should structure and comment the code a bit, before adding
license
> > headers and releasing them.
>
> Which license do you prefer? I suggest putting the stuff into CVS and let
> the world help you in every possible way.
>

Yes, I have been thinking about setting up CVS on my old machine (after
installing Linux, that is)
When it comes to licenses, I don't have a preference, but as I understand
it would have to be compatible with the original libraries, right?


~egilhh

Re: microphone and sound access in Ada

[Preben Randhol]

Egil Harald H�vik wrote:

> 
> Yes, I have been thinking about setting up CVS on my old machine (after
> installing Linux, that is)

Or you can use:

   http://savannah.nongnu.org/

   or

   http://sourceforge.net/


> When it comes to licenses, I don't have a preference, but as I understand
> it would have to be compatible with the original libraries, right?


   Xiph.org's Vorbis software libraries (libvorbis and vorbisfile) are
   distributed under a BSD-like license; these libraries may be used by
   any application, Open or proprietary, linked or incorporated in
   whole, so long as acknowldgement is made to Xiph.org Foundation when
   using our codec source in whole or in derived works. See the file
   'COPYING' in the source for all the details.

   All other portions of xiph.org's reference Vorbis software are
   distributed under the (GNU) Library General Public License (LGPL).

   Because the specification is open, third parties are welcome (and
   encouraged) to implement third party Vorbis software under other
   licenses using xiph.org's reference source as an authoritative
   reference.

Preben

Re: microphone and sound access in Ada

[Adrian Knoth]

Egil Harald H�vik <egil.harald.hoevik.nospam@kongsberg.com> wrote:

>> [Binding to Ogg/Vorbis]
>> > I have already done some work on this. I have bindings to libogg,
>> > libvorbis and libvorbisfile, although I lack some testing.
>> If you like I can do some of it for you/public.
> I guess that wouldn't be a bad idea  :-)

If storage is a problem, I can provide the necessary server(s).

[License] 
> When it comes to licenses, I don't have a preference, but as I understand
> it would have to be compatible with the original libraries, right?

I only asked because it sounded a little bit like "no GPL/BSD".

The best thing ever would be to bring the your work into a proper state
(in case it's not already there) and then check-in to xiph.org's CVS
so the Ada-bindings are part of the main distribution, like ncurses
comes along with the Ada-interface. A lot of developers might have a
look at this API.

Concerning the input-stuff I'd suggest not to stick to ALSA directly
but use jackit instead. It's a low-latency-framework on top of ALSA
which is able to share ressources between applications and guarantee
a low overhead in processing which is necessary for realtime-sound-IO.

It's like the ASIO-drivers for Synthesizers/Samplers/Recording-software
under Windows.

I didn't do deep investigations on jackit, I only felt attracted by a
brief overview located on their website. If it provides a mixer-API
for changing volumes than it would suffice. I won't suggest using jackit
if we'll also need to interface to ALSA directly only for volume-change :)

On the other hand: if there is a libao(ut), perhaps there is also a
libai(n)?

-- 
mail: adi@thur.de  	http://adi.thur.de	PGP: v2-key via keyserver

Gib Gates keine Chance!

Re: microphone and sound access in Ada

[Preben Randhol]

Adrian Knoth wrote:
> 
> Concerning the input-stuff I'd suggest not to stick to ALSA directly
> but use jackit instead. It's a low-latency-framework on top of ALSA
> which is able to share ressources between applications and guarantee
> a low overhead in processing which is necessary for realtime-sound-IO.
>
> It's like the ASIO-drivers for Synthesizers/Samplers/Recording-software
> under Windows.
> 
> I didn't do deep investigations on jackit, I only felt attracted by a
> brief overview located on their website. If it provides a mixer-API
> for changing volumes than it would suffice. I won't suggest using jackit
> if we'll also need to interface to ALSA directly only for volume-change :)

The only problem is that you also have arts and esd servers. One cannot
know which the user prefers (depending on other programs she uses).

> On the other hand: if there is a libao(ut), perhaps there is also a
> libai(n)?

Haven't found one, but I'll ask around.

-- 
Preben Randhol ------------------------ http://www.pvv.org/~randhol/ --
�There are three things you can do to a woman. You can love her, suffer
 for her, or turn her into literature.�  - Justine, by Lawrence Durrell

Re: microphone and sound access in Ada

[Adrian Knoth]

Preben Randhol <randhol+news@pvv.org> wrote:

[ALSA, jackit and so on]
> The only problem is that you also have arts and esd servers. One cannot
> know which the user prefers (depending on other programs she uses).

Same problem under Windows where the ASIO-drivers cannot be used in
conjunction with the Windows-MultimediaExtensions (MME).

I feel not well about arts and esd. They are only necessary for
non-ALSA-platforms where the underlaying sound-architecture, i.e. OSS,
is blocking, so they cannot play their error-wavs while xmms is
running :)

Have a look at:
http://www.op.net/~pbd/alsa-audio.html  (ALSA Devel-FAQ)

and especially at the last paragraph.

-- 
mail: adi@thur.de  	http://adi.thur.de	PGP: v2-key via keyserver

Lieber 'ne Latte in der Hose als ein Brett vorm Kopf.

Re: microphone and sound access in Ada

[Preben Randhol]

Adrian Knoth wrote:
> 
> Same problem under Windows where the ASIO-drivers cannot be used in
> conjunction with the Windows-MultimediaExtensions (MME).
> 
> I feel not well about arts and esd. They are only necessary for
> non-ALSA-platforms where the underlaying sound-architecture, i.e. OSS,
> is blocking, so they cannot play their error-wavs while xmms is
> running :)

Don't like esd or arts either.

> Have a look at:
> http://www.op.net/~pbd/alsa-audio.html  (ALSA Devel-FAQ)
> 
> and especially at the last paragraph.

I see. I also see that Debian has it in unstable. I think jackit is
probably a better way to go. have you checked if it does input as well
as output?

Preben

Re: microphone and sound access in Ada

[Adrian Knoth]

Preben Randhol <randhol+news@pvv.org> wrote:

>> http://www.op.net/~pbd/alsa-audio.html  (ALSA Devel-FAQ)
>> and especially at the last paragraph.
> I see. I also see that Debian has it in unstable. I think jackit is
> probably a better way to go. 

I checked the jackit-API and wondered: "That's all?" It seems to be
straight-forward to implement.

> have you checked if it does input as well as output?

It does. The API is well documented. The main goal of jackit is
professionell interconnection of soundhardware and multiple apps.

It does not support volume-changes, but for that question it is possible
to use a native mixer-application from the OS. Drawback is that the
application won't have a volume-slider, but professionell Audio-solutions
like Cubase don't have it as well. They change the volume by lowering the
peak of their data without interfearing the mixer of the soundcard.


-- 
mail: adi@thur.de  	http://adi.thur.de	PGP: v2-key via keyserver

Mit Apachen ist gut quatschen

Re: microphone and sound access in Ada

[Preben Randhol]

Adrian Knoth wrote:
> Preben Randhol <randhol+news@pvv.org> wrote:
> 
>>> http://www.op.net/~pbd/alsa-audio.html  (ALSA Devel-FAQ)
>>> and especially at the last paragraph.
>> I see. I also see that Debian has it in unstable. I think jackit is
>> probably a better way to go. 
> 
> I checked the jackit-API and wondered: "That's all?" It seems to be
> straight-forward to implement.

Good!
>
>> have you checked if it does input as well as output?
> 
> It does. The API is well documented. The main goal of jackit is
> professionell interconnection of soundhardware and multiple apps.

Ah finally!

> It does not support volume-changes, but for that question it is possible
> to use a native mixer-application from the OS. Drawback is that the
> application won't have a volume-slider, but professionell Audio-solutions
> like Cubase don't have it as well. They change the volume by lowering the
> peak of their data without interfearing the mixer of the soundcard.

Yes, but it still would be nice to be able to change the recording
volume and listing volume from some apps. But I guess we can make
bindings only to this part of ALSA and equivalent later.

Preben

Re: microphone and sound access in Ada

[Egil Harald H�vik]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 1386 bytes --]


"Adrian Knoth" <adi@drcomp.erfurt.thur.de> wrote in message
news:slrnb7ejkf.lp0.adi@drcomp.erfurt.thur.de...
> Egil Harald H�vik <egil.harald.hoevik.nospam@kongsberg.com> wrote:
>
> >> [Binding to Ogg/Vorbis]
> >> > I have already done some work on this. I have bindings to libogg,
> >> > libvorbis and libvorbisfile, although I lack some testing.
> >> If you like I can do some of it for you/public.
> > I guess that wouldn't be a bad idea  :-)
>
> If storage is a problem, I can provide the necessary server(s).
>

Storage is not an issue :-)  but a server with dedicated/static IP is.


> [License]
> > When it comes to licenses, I don't have a preference, but as I
understand
> > it would have to be compatible with the original libraries, right?
>
> I only asked because it sounded a little bit like "no GPL/BSD".
>

Absolutely not. I'm just not very into the whole license business, and
assumed that the licences on the original libraries limited my
choices (but I haven't actually read every word in those licenses).



> The best thing ever would be to bring the your work into a proper state
> (in case it's not already there) and then check-in to xiph.org's CVS
> so the Ada-bindings are part of the main distribution, like ncurses
> comes along with the Ada-interface. A lot of developers might have a
> look at this API.
>

That's an interesting option.



~egilhh

Re: microphone and sound access in Ada

[Adrian Knoth]

Egil Harald H�vik <egil.harald.hoevik.nospam@kongsberg.com> wrote:

>> If storage is a problem, I can provide the necessary server(s).
> Storage is not an issue :-)  but a server with dedicated/static IP is.

I have both, but if you think about dyndns it would be ok.
 

-- 
mail: adi@thur.de  	http://adi.thur.de	PGP: v2-key via keyserver

Vorbeugen ist besser als sich auf die Schuhe zu kotzen.

Re: microphone and sound access in Ada

[Preben Randhol]

Egil Harald H�vik wrote:
> I have already done some work on this. I have bindings to libogg,
> libvorbis and libvorbisfile, although I lack some testing. The
> libvorbisfile binding seems to work OK (I saved the output (RAW/PCM)
> to a file, converted the raw data to WAV using Audacity, and played it
> using WinAmp).

So if libao was bound then you could play directly couldn't you?

> Maybe I should structure and comment the code a bit, before adding
> license headers and releasing them.

Would be very nice!

Preben

Re: microphone and sound access in Ada

[Egil Harald H�vik]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 861 bytes --]


"Preben Randhol" <randhol+news@pvv.org> wrote in message
news:slrnb7eb3s.c9b.randhol+news@kiuk0152.chembio.ntnu.no...
> Egil Harald H�vik wrote:
> > I have already done some work on this. I have bindings to libogg,
> > libvorbis and libvorbisfile, although I lack some testing. The
> > libvorbisfile binding seems to work OK (I saved the output (RAW/PCM)
> > to a file, converted the raw data to WAV using Audacity, and played it
> > using WinAmp).
>
> So if libao was bound then you could play directly couldn't you?
>

I guess so. First, I looked at DirectX/DirectSound (Yes, I'm on a
Windows platform), but with too little documentation (or maybe I
just didn't look the right place), I kind of gave up that idea.
I didn't even know about libao until a week or so ago, when I
looked through the xiph pages, and started reading about ogg123.



~egilhh

Re: microphone and sound access in Ada

[tmoran]

> I would be very interested in capturing microphone input and also
> using sound in an Ada application, (basically to use in a distributed
> application for voice conferencing)
>
> The system I am targetting is Linux but I would like to be able to
> extend this to windows later.
  Claw has both audio and video, but it's currently in alpha.  It also
runs only on Windows.