8 Feb
2005
8 Feb
'05
3:05 p.m.
Hi all,
I am new to the list and new to Linux audio.
I wrote a simple QT program that uses sound, and it requires the NAS
sound server.
I installed NAS from Mandrake rpm's rebooted the box, and the program
works great.
The problem is that most of my other software, for example Amarok,
stopped producing sound.
Prior to installing NAS, I used KDE and there sound server aRts. When I
start Amarok I get a message from aRts, that it can't open device for
playback.
I thus have two questions:
1. Is there a way to resolve such a situation, in a way that allows two
programs that use different sound servers to play sounds at the same time?
2. A more generic question: Why do we have to use sound servers in
Linux? shouldn't it be the job of the operating system to allow sound
multiplexing through a set of standard API's? The situation is
especially bad, if there is no comfortable solution to the problem in my
first question.
Thanks,
Hillel.
8 Feb
8 Feb
4:29 p.m.
On Tuesday 08 February 2005 20:07, Hillel wrote:
I installed NAS from Mandrake rpm's rebooted the
box, and the program
works great.
The problem is that most of my other software, for example Amarok,
stopped producing sound.
Prior to installing NAS, I used KDE and there sound server aRts. When I
start Amarok I get a message from aRts, that it can't open device for
playback.
I thus have two questions:
1. Is there a way to resolve such a situation, in a way that allows two
programs that use different sound servers to play sounds at the same time?
There are several answers/solutions:
- Buy a soundcard which is allows multiple hardwareaccess such as the sblive
(with emu10k1-chip allowing 32 apps) or something based on the maestro3-chip
(allowing 2 apps) for example...
- Use dmix to do software mixing in alsa.
- Tell aRts to use NAS, this can be done in the control-center.
- Check wether Amarok can use NAS. Or perhaps GStreamer which outputs to NAS.
Arnold
--
There is a theory which states that if ever anyone discovers exactly what the
Universe is for and why it is here, it will instantly disappear and be
replaced by something even more bizarre and inexplicable.
There is another theory which states that this has already happened.
-- Douglas Adams, The Restaurant at the End of the Universe
5:17 p.m.
Arnold Krille wrote:
On Tuesday 08 February 2005 20:07, Hillel wrote:
Thanks for the tips, I tried telling aRts to use NAS. I didn't know what
you mean by control center, so I went through the menu KDE --> Sound -->
Sound System and there I marked Enable the sound system, and Enable
network sound.
Now aRts seems to run fine, and Amarok plays music. The problem is that
if I run my QT program (that uses NAS), while Amarok is playing, then I
hear no sound until I close Amarok. Only when Amarok closes I hear the
sound that was probably queued while Amarok was playing. This is already
fine for me on a practical level, although not a perfect soultion in
general.
Regarding alsa dmix, I don't know the technology, but can it solve the
principle problem of two programs using different sound servers? As I
understand it, the programs are using sound server API's and not ALSA
API's so if two sound server's can't coexist right, no fix in ALSA level
can help.
I installed NAS from Mandrake rpm's rebooted
the box, and the program
works great.
The problem is that most of my other software, for example Amarok,
stopped producing sound.
Prior to installing NAS, I used KDE and there sound server aRts. When I
start Amarok I get a message from aRts, that it can't open device for
playback.
I thus have two questions:
1. Is there a way to resolve such a situation, in a way that allows two
programs that use different sound servers to play sounds at the same time?
There are several answers/solutions:
- Buy a soundcard which is allows multiple hardwareaccess such as the sblive
(with emu10k1-chip allowing 32 apps) or something based on the maestro3-chip
(allowing 2 apps) for example...
- Use dmix to do software mixing in alsa.
- Tell aRts to use NAS, this can be done in the control-center.
5:33 p.m.
Regarding alsa dmix, I don't know the technology,
but can
it solve the principle problem of two programs using
different sound servers? As I understand it, the programs
are using sound server API's and not ALSA API's so if two
sound server's can't coexist right, no fix in ALSA level
can help.
you can stack soundservers, but most often you'll not manage
to run them competively in parallel, except you use DMIX or
you have a special card which supports hardware mixing. Most
consumer cards do not support this.
Best regards
ce
5:35 p.m.
On Tuesday 08 February 2005 22:18, Hillel wrote:
Thanks for the tips, I tried telling aRts to use NAS.
I didn't know what
you mean by control center, so I went through the menu KDE --> Sound -->
Sound System and there I marked Enable the sound system, and Enable
network sound.
Thats the right place but the wrong option...
On the second tab of this dialog you will find a drop-down-box asking you
which Audio Device to use. It should show Autodetect by default, change it to
"Network Audio System".
The option you selected enables aRts' own networking capabilities, which are,
uhm, how to put it... not perfect...
Now aRts seems to run fine, and Amarok plays music.
The problem is that
if I run my QT program (that uses NAS), while Amarok is playing, then I
hear no sound until I close Amarok. Only when Amarok closes I hear the
sound that was probably queued while Amarok was playing. This is already
fine for me on a practical level, although not a perfect soultion in
general.
Thats because aRts autosuspends and releases /dev/dsp if not active.
Regarding alsa dmix, I don't know the technology,
but can it solve the
principle problem of two programs using different sound servers?
for dmix see http://www.alsa-project.org/
It cannot fix the "different soundservers" problem but it fixes the problem of
a sounddevice which can be opened only from one application at a time... So
it fixes more than your problem as it also allows xmms, zinf, realone and
many others to access the soundcard while your favourite soundserver(s) is
running...
Arnold
--
There is a theory which states that if ever anyone discovers exactly what the
Universe is for and why it is here, it will instantly disappear and be
replaced by something even more bizarre and inexplicable.
There is another theory which states that this has already happened.
-- Douglas Adams, The Restaurant at the End of the Universe
6:45 p.m.
Arnold Krille wrote:
On the second tab of this dialog you will find a drop-down-box asking you
which Audio Device to use. It should show Autodetect by default, change it to
"Network Audio System".
Thanks for the help, you were on the spot. When I changed the audio
device to "Network Audio System" everything started working together!,
the caveat is that there is a noticeable audio latency (very noticeable
:-) while playing music with Amarok. I don't know if this is a bug in my
machine or normal latency in this configuration.
Now aRts seems to run fine, and Amarok plays
music. The problem is that
if I run my QT program (that uses NAS), while Amarok is playing, then I
hear no sound until I close Amarok. Only when Amarok closes I hear the
sound that was probably queued while Amarok was playing. This is already
fine for me on a practical level, although not a perfect soultion in
general.
Thats because aRts autosuspends and releases /dev/dsp if not active.
>for dmix see http://www.alsa-project.org/
>It cannot fix the "different soundservers" problem but it fixes the problem
of
>a sounddevice which can be opened only from one application at a time... So
>it fixes more than your problem as it also allows xmms, zinf, realone and
>many others to access the soundcard while your favourite soundserver(s) is
>running...
>
This is what I was looking for. Looks like a much cleaner solution. I
will investigate some more on it :-).
Thanks for the insightful comments.
9 Feb
9 Feb
3:47 a.m.
On Tuesday 08 February 2005 23:46, Hillel wrote:
Have a lot of fun...
Arnold
--
There is a theory which states that if ever anyone discovers exactly what the
Universe is for and why it is here, it will instantly disappear and be
replaced by something even more bizarre and inexplicable.
There is another theory which states that this has already happened.
-- Douglas Adams, The Restaurant at the End of the Universe
Thanks for the help, you were on the spot. When I
changed the audio
device to "Network Audio System" everything started working together!,
the caveat is that there is a noticeable audio latency (very noticeable
:-) while playing music with Amarok. I don't know if this is a bug in my
machine or normal latency in this configuration.
Thats normal in this configuration since now the sound goes through two
soundservers which both have latencies...
>for dmix
see http://www.alsa-project.org/
This is what I was looking for. Looks like a much
cleaner solution. I
will investigate some more on it :-).
8 Feb
8 Feb
5:30 p.m.
Hi,
I am new to the list and new to Linux audio.
welcome :) .
I wrote a simple QT program that uses sound, and it
Mhhh. A simple program to record audio 8-P ?
requires the NAS sound server.
Is it really called NAS, or MAS (media application server)?
I installed NAS from Mandrake rpm's rebooted the
box, and
the program works great.
The problem is that most of my other software, for example
Amarok, stopped producing sound.
Prior to installing NAS, I used KDE and there sound server
aRts. When I start Amarok I get a message from aRts, that
it can't open device for playback.
NAS now blocks the audio device. Try to make artsd play into
NAS instead to the audio device could be a solution.
I'd like to recommend the following 15MB:
http://sysexxer.sourceforge.net/files/LinuxAudioBasics.avi
I thus have two questions:
1. Is there a way to resolve such a situation, in a way
that allows two programs that use different sound servers
to play sounds at the same time?
There could only be one layer which can handle all audio stuff
from different applications. In linux, we are on free
software, and so we have choices. There are many audio
layers, but not every audio application can communicate with
each of them.
2. A more generic question: Why do we have to use
sound
servers in Linux? shouldn't it be the job of the operating
system to allow sound multiplexing through a set of
standard API's?
Well, I'm not allowed to comment on this issue, otherwise I
risk to be at least blacklisted by the members of this
list ;-)))) .
The situation is especially bad, if there
is no comfortable solution to the problem in my first
question.
It depends what your application is for.
arts will most probably be dropped by the KDE project, and
maybe Gnome will drop esound. So, your application should not
use one of them.
Instead, MAS and Gstreamer are the two candidates for future
Gnome and KDE versions.
If your application should be independent from Gnome or KDE,
you could make it an direct ALSA application. But this is a
dirty design, because it then would block the audio device
resp. not get access to the device if an other application
already uses the device.
The same if you use the older OSS-device (usually known
as /dev/dsp)
You could use DMIX, an ALSA based sofmixer. But DMIX isn't
installed on much systems. I expect it to be more often found
in the future, but currently it isn't found on common users
machines.
If your application is intended for professional audio work,
you should make it use the JACK soundserver. Most linux
musicians are used to use it for their daily work. There's
already a bunch of applications (Synths and so on) which will
only run on JACK.
Best regards
ce
6:51 p.m.
Christoph Eckert wrote:
I'd like to recommend the following 15MB:
http://sysexxer.sourceforge.net/files/LinuxAudioBasics.avi
Thanks for your help Christoph.
The avi was great, your accent is brilliant and the material was indeed
very informative too :-).
Thanks,
Hillel.
14 Feb
14 Feb
10:44 p.m.
El mar, 08-02-2005 a las 19:53, Hillel escribió:
Christoph Eckert wrote:
Do you perhaps have your speech written? Because I would really like to
translate it to Spanish, if you give permission, of course.
Regards,
Damian-
I'd like to recommend the following 15MB:
http://sysexxer.sourceforge.net/files/LinuxAudioBasics.avi
Thanks for your help Christoph.
The avi was great, your accent is brilliant and the material was indeed
very informative too :-). 
8 Feb
8 Feb
7:13 p.m.
On Tue, 8 Feb 2005, Christoph Eckert wrote:
You could use DMIX, an ALSA based sofmixer. But DMIX
isn't
installed on much systems. I expect it to be more often found
in the future, but currently it isn't found on common users
machines.
Actually this is not quite correct. Dmix is part of standard ALSA and
is thus available on any recent Linux system.
You are of course right that dmix is not commonly configured to be part of
default plugins. But most ALSA apps allow you to specify which soundcard
to use (the usual implicit choice is 'default'). If you specify
"plug:dmix", you'll get dmix output in most cases (unless user has
overridden the asoundrc configuration).
For example:
aplay -Dplug:dmix foo.wav
alsaplayer -o alsa -d plug:dmix
ecasound -i foo.wav -o alsa,plug:dmix
This works on my system with vanilla alsa-lib-1.0.4 (quite old version in
fact) without absolutely no modifications to any ALSA configuration files
(I don't even have a ~/.asoundrc on this machine). So I assume this should
work out-of-the-box on any system that has recent ALSA (SuSE systems, FC2
and newer, Mandrake 10 and so on).
More info at:
http://alsa.opensrc.org/index.php?page=DmixPlugin
PS I think dmix is _the_ solution for "desktop sound mixing".
Together with JACK they form a killer combo. :) I use JACK
when making music (ability to route audio between apps,
transport control), and dmix when listening to mp3s (wide
application support, no need to run nor configure any
servers; still great performance ... also latency-wise!).
--
http://www.eca.cx
Audio software for Linux!
7:49 p.m.
Hi Kai,
Actually this is not quite correct. Dmix is part of
standard ALSA and is thus available on any recent Linux
system.
You are of course right that dmix is not commonly
configured to be part of default plugins.
Yep, you're right. It is installed, but of course I meant that
per default an asoundrc is missing most often.
But most ALSA
apps allow you to specify which soundcard to use (the usual
implicit choice is 'default'). If you specify "plug:dmix",
you'll get dmix output in most cases (unless user has
overridden the asoundrc configuration).
For example:
aplay -Dplug:dmix foo.wav
alsaplayer -o alsa -d plug:dmix
ecasound -i foo.wav -o alsa,plug:dmix
Thanks for pointing this out. I'm still not experienced enough
concerning ALSA and DMIX. I tried to write an asoundrc (which
should not only enable DMIX but also order my cards in a
certain order), but at some point I gave up ;-) .
This works on my system with vanilla alsa-lib-1.0.4
(quite
old version in fact) without absolutely no modifications to
any ALSA configuration files (I don't even have a
~/.asoundrc on this machine). So I assume this should work
out-of-the-box on any system that has recent ALSA (SuSE
systems, FC2 and newer, Mandrake 10 and so on).
More info at:
http://alsa.opensrc.org/index.php?page=DmixPlugin
Yeah, I already read it. But still it requires manual work.
So, I thouht about a small script which could create an
asoundrc automatically, then I tried to tweak the alsaconf
script, the I saw it needs some rewriting and refactoring,
and then I gave up ;-) .
PS I think dmix is _the_ solution for "desktop
sound
mixing". Together with JACK they form a killer combo. :)
Do you use both at the same time? Jack is running all day and
using DMIX, while maybe artsd or esound is running in
parallel for desktop audio?
I
use JACK when making music (ability to route audio between
apps, transport control), and dmix when listening to mp3s
(wide application support, no need to run nor configure any
servers; still great performance ... also latency-wise!).
Is DMIX transparent to any ALSA aware application, or does an
application need some coding to use DMIX, or is it needed to
invoke the application always with the plug:DMIX parameter?
It would be cool if it were transparent to the user.
Thanks & best regards
ce
8:16 p.m.
On Wed, 9 Feb 2005, Christoph Eckert wrote:
Well actually the above three examples work without any manual settings!
Only if you encounter unexpected problems (I didn't on my machines), you
may need to consult the Wiki and other dmix documentation.
No, I've configured jackd to access soundcard without any plugins in
between (the 'hw:0' ALSA sound device). Switching is still quite easy as I
just have to stop any mp3 players ('stop' should be enough if the player
is well written), launch JACK, and start making music. :)
For example:
aplay -Dplug:dmix foo.wav
alsaplayer -o alsa -d plug:dmix
ecasound -i foo.wav -o alsa,plug:dmix
http://alsa.opensrc.org/index.php?page=DmixPlugin
Yeah, I already read it. But
still it requires manual work. PS I think
dmix is _the_ solution for "desktop sound
mixing". Together with JACK they form a killer combo. :)
Do you use both at
the same time? Jack is running all day and
using DMIX, while maybe artsd or esound is running in
parallel for desktop audio? Is DMIX transparent to any ALSA aware application, or
does an
application need some coding to use DMIX, or is it needed to
invoke the application always with the plug:DMIX parameter?
It's completely transparent, dmix is just another sound device
from application's point of view. Without a custom asoundrc, you
have to instruct the app to use 'plug:dmix' as the sound device,
but this is really quite normal.
For example I have two soundcards on my desktop Linux machine.
Even without dmix I anyway have to instruct apps to select which of
the cards to use. Of course, not all apps provide an easy way to select
the soundcard, but this problem is not specific to dmix alone, but
is a more general usability problem.
--
http://www.eca.cx
Audio software for Linux!
9 Feb
9 Feb
3:37 a.m.
Kai Vehmanen wrote:
No, I've configured jackd to access soundcard without any plugins in
between (the 'hw:0' ALSA sound device). Switching is still quite easy as I
just have to stop any mp3 players ('stop' should be enough if the player
is well written), launch JACK, and start making music. :)
xmms has a plugin that will play into jack. I use it to import mp3 into
ardour sometimes when I want to do something with someone else's backing
track.
--
Our enemies are innovative and resourceful - and so are we
They never stop thinking about new ways to harm our country and our people - and neither
do we.
-- George W. Bush
7317
days inactive
7323
days old
linux-audio-user@lists.linuxaudio.org
14 comments
6 participants
participants (6)
-
Arnold Krille -
Hillel -
Kai Vehmanen -
Lisandro Damián Nicanor Pérez Meyer -
mchristoph.eckert＠t-online.de -
Noah Roberts