Hi,
I'm still having issues on PdE with this error. It happens a lot and glitches audio output (not good).
It's been happening with all the latest builds for a couple of weeks now.
This is with most recent nightly build (today 31st Jan). Debian Wheezy 64b.
There's also a couple of people on the forum with the same issue who are floundering equally.
http://puredata.hurleur.com/viewtopic.php?pid=35002
Any ideas (a search around hasn't turned up anything useful at all) much appreciated, possible solutions would be marvellous.
Cheers,
Julian
Does it happen with Pd-vanilla 0.43.4 or 0.44? I just take that code directly from Miller since I've never done anything with programming audio interfaces. Pd definitely does not place nice with pulseaudio, so make sure its not running when using Pd. I use this to temporarily suspend pulse:
pasuspender pd-extended
Many people who use Pd a lot uninstall pulseaudio entirely.
.hc
On 01/31/2013 11:08 AM, Julian Brooks wrote:
Hi,
I'm still having issues on PdE with this error. It happens a lot and glitches audio output (not good).
It's been happening with all the latest builds for a couple of weeks now.
This is with most recent nightly build (today 31st Jan). Debian Wheezy 64b.
There's also a couple of people on the forum with the same issue who are floundering equally.
http://puredata.hurleur.com/viewtopic.php?pid=35002
Any ideas (a search around hasn't turned up anything useful at all) much appreciated, possible solutions would be marvellous.
Cheers,
Julian
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
Hi Hans,
My Pd is 0.43.2 from debian repos but it doesn't happen with that. I've been messing with Pd-extened quite a lot recently using latest builds and this has only started in the last week or so? First started with the versions that fixed the other strange [list] issue.
Pulseaudio isn't installed on my lappy, so that's out.
Weird.
Best wishes,
Julian
On 31 January 2013 18:12, Hans-Christoph Steiner hans@at.or.at wrote:
Does it happen with Pd-vanilla 0.43.4 or 0.44? I just take that code directly from Miller since I've never done anything with programming audio interfaces. Pd definitely does not place nice with pulseaudio, so make sure its not running when using Pd. I use this to temporarily suspend pulse:
pasuspender pd-extended
Many people who use Pd a lot uninstall pulseaudio entirely.
.hc
On 01/31/2013 11:08 AM, Julian Brooks wrote:
Hi,
I'm still having issues on PdE with this error. It happens a lot and glitches audio output (not good).
It's been happening with all the latest builds for a couple of weeks now.
This is with most recent nightly build (today 31st Jan). Debian Wheezy
64b.
There's also a couple of people on the forum with the same issue who are floundering equally.
http://puredata.hurleur.com/viewtopic.php?pid=35002
Any ideas (a search around hasn't turned up anything useful at all) much appreciated, possible solutions would be marvellous.
Cheers,
Julian
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management ->
http://lists.puredata.info/listinfo/pd-list
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
On Don, 2013-01-31 at 13:12 -0500, Hans-Christoph Steiner wrote:
Does it happen with Pd-vanilla 0.43.4 or 0.44? I just take that code directly from Miller since I've never done anything with programming audio interfaces. Pd definitely does not place nice with pulseaudio,
Pd 0.44 _does_ play nice with pulseaudio (input and output) when using portaudio as backend and "pulse" as devcice. One can also launch Pd with the option '-pa'. N.B.: '-pa' stands for portaudio which is not to be confused with pulseaudio.
so make sure its not running when using Pd. I use this to temporarily suspend pulse:
pasuspender pd-extended
Many people who use Pd a lot uninstall pulseaudio entirely.
Ok, let's clear some dust here. There seems a strong notion that pulseaudio is interfering a lot with other audio applications and is causing lots of troubles. Let me point out a few facts from a standard Ubuntu 12.04 installation:
and thus non-compliant applications like Pd are _not_ blocked from accessing the device. This is even the case when you pause a Youtube movie in the browser.
pulseaudio client is actually playing, you may want to use pasuspender. However, using pasuspender does not harm, even if it is not necessary
so that options passed to pd-extended are bypassed by pasuspender.
any problem. OTOH, it also doesn't harm. However, when pulseaudio is installed and Firefox is playing some Youtube movie, you still are able to grant Pd access to the sound device by using pasuspender. With no pulseaudio installed, the flash-plugin might use ALSA directly and there is no other way to stop it from blocking the sound device than to stop firefox. If you don't know which application is currently using the device, good luck with hunting. My point is: With pulseaudio it is actually easier to guarantee Pd access to the sound device.
client is playing, you can just start it without any troubles.
Roman
On 02/01/2013 06:29 AM, Roman Haefeli wrote:
On Don, 2013-01-31 at 13:12 -0500, Hans-Christoph Steiner wrote:
Does it happen with Pd-vanilla 0.43.4 or 0.44? I just take that code directly from Miller since I've never done anything with programming audio interfaces. Pd definitely does not place nice with pulseaudio,
Pd 0.44 _does_ play nice with pulseaudio (input and output) when using portaudio as backend and "pulse" as devcice. One can also launch Pd with the option '-pa'. N.B.: '-pa' stands for portaudio which is not to be confused with pulseaudio.
The situation is improved, but not quite usable on my machine. I'm on Linux Mint Maya amd64 (Ubuntu/precise). Many other apps are happy to play at the same time, like Youtube, browser+flash, Twinkle voip phone, skype, mumble, etc. If I have rhythmbox playing music, and start 'pd -pa', Pd outputs chunky audio and I get tons of these in the console:
ALSA lib pcm.c:7339:(snd_pcm_recover) underrun occurred ALSA lib pcm.c:7339:(snd_pcm_recover) underrun occurred ALSA lib pcm.c:7339:(snd_pcm_recover) underrun occurred
After 10ish seconds, it settles in and works decently. Until I close the Test Audio patch, then I get tons of these again until I turn off DSP:
ALSA lib pcm.c:7339:(snd_pcm_recover) underrun occurred
so make sure its not running when using Pd. I use this to temporarily suspend pulse:
pasuspender pd-extended
Many people who use Pd a lot uninstall pulseaudio entirely.
Ok, let's clear some dust here. There seems a strong notion that pulseaudio is interfering a lot with other audio applications and is causing lots of troubles. Let me point out a few facts from a standard Ubuntu 12.04 installation:
- When no pulseaudio client is playing, the sound device is not occupied
and thus non-compliant applications like Pd are _not_ blocked from accessing the device. This is even the case when you pause a Youtube movie in the browser.
- pasuspender is actually not necessary in most cases. Only when some
pulseaudio client is actually playing, you may want to use pasuspender. However, using pasuspender does not harm, even if it is not necessary
- Use "pasuspender -- pd-extended" instead of "pasuspender pd-extended",
so that options passed to pd-extended are bypassed by pasuspender.
- Uninstalling pulseaudio is completely unnecessary, as it doesn't solve
any problem. OTOH, it also doesn't harm. However, when pulseaudio is installed and Firefox is playing some Youtube movie, you still are able to grant Pd access to the sound device by using pasuspender. With no pulseaudio installed, the flash-plugin might use ALSA directly and there is no other way to stop it from blocking the sound device than to stop firefox. If you don't know which application is currently using the device, good luck with hunting. My point is: With pulseaudio it is actually easier to guarantee Pd access to the sound device.
- Jackd (qjackctl, respectively) behaves similar to Pd: When no pulse
client is playing, you can just start it without any troubles.
Yeah, that sounds like my experience as well. One thing I've heard is that pulseaudio can use jack as the backend instead of ALSA. That sounds like an ideal situation. Anyone messed with that?
.hc
On Fre, 2013-02-01 at 10:37 -0500, Hans-Christoph Steiner wrote:
[...]
Yeah, that sounds like my experience as well. One thing I've heard is that pulseaudio can use jack as the backend instead of ALSA. That sounds like an ideal situation. Anyone messed with that?
Yeah, there are some tutorials around about how to install the jack-sink plugin for pulseaudio. It worked well for me. But it had the side-effect that the volume control keys didn't work anymore.
Roman
On Fri, 01 Feb 2013 19:45:12 +0100, Roman Haefeli reduzent@gmail.com
wrote:
On Fre, 2013-02-01 at 10:37 -0500, Hans-Christoph Steiner wrote:
[...]
Yeah, that sounds like my experience as well. One thing I've heard is
that pulseaudio can use jack as the backend instead of ALSA. That sounds
like an ideal situation. Anyone messed with that?Yeah, there are some tutorials around about how to install the jack-sink plugin for pulseaudio. It worked well for me. But it had the side-effect that the volume control keys didn't work anymore.
Roman
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management ->
http://lists.puredata.info/listinfo/pd-list
Just install pulseaudio-module-jack
Use jackdbus instead of jackd (jackd2 in Debian based) to have the module
activate jack sink and source automatically when starting jackdbus.
Some jack control apps will want to start jackdbus. jack_control is a
command line tool for starting jackdbus. Just do:
jack_control start|stop
If using qjackctl, make sure dbus is enabled in Setup -> Misc
Restart PA after installing the module. Written by David Henningson as a
starting point for integrating jack with PA.
jack is able to grab the audio device from PA, but this feature is a
little buggy, so it won't work every time. It happens through dbus, and
it's some code that the author Poettering implemented in both PA and jack.
jack is set to have higher priority, and PA is supposed to let go of the
card, and let jack grab it.
Just to comment on why I ended up un-installing portaudio: I booted my Pi with a clean new Raspbian distro, compiled Pd (after installing some packages like git and alsa libs), ran it and found that pulseaudio was running. So I exit Pd and likk pulseaudio via "pulseaudio -kill". Then start Pd and find it _still_ can't open audio. Why? ell, looking back I found that pulseaudio was running again! So I re-killed it and tried a third time. Same result: for some reason, the mere act of starting Pd was causing pulseaudio to start itself up and grab the audio device so that Pd couldn't get it. There are similar plot sequences in old Three Stooges and Marx Brothers movies, but I wasn't exactly laughing. What if I were about to have to walk on stage with this thing? So I googled around and read through the many-page man page for pulseaudio and found no explanation of why this was happening, nor any suggestions for how to disable pulseaudio other than the "-kill" command I had already tried. After a while I lost patience and uninstalled pulse and I suggest that everyone else do so as well.
cheers Miller
Many people who use Pd a lot uninstall pulseaudio entirely.
Ok, let's clear some dust here. There seems a strong notion that pulseaudio is interfering a lot with other audio applications and is causing lots of troubles. Let me point out a few facts from a standard Ubuntu 12.04 installation:
- When no pulseaudio client is playing, the sound device is not occupied
and thus non-compliant applications like Pd are _not_ blocked from accessing the device. This is even the case when you pause a Youtube movie in the browser.
- pasuspender is actually not necessary in most cases. Only when some
pulseaudio client is actually playing, you may want to use pasuspender. However, using pasuspender does not harm, even if it is not necessary
- Use "pasuspender -- pd-extended" instead of "pasuspender pd-extended",
so that options passed to pd-extended are bypassed by pasuspender.
- Uninstalling pulseaudio is completely unnecessary, as it doesn't solve
any problem. OTOH, it also doesn't harm. However, when pulseaudio is installed and Firefox is playing some Youtube movie, you still are able to grant Pd access to the sound device by using pasuspender. With no pulseaudio installed, the flash-plugin might use ALSA directly and there is no other way to stop it from blocking the sound device than to stop firefox. If you don't know which application is currently using the device, good luck with hunting. My point is: With pulseaudio it is actually easier to guarantee Pd access to the sound device.
- Jackd (qjackctl, respectively) behaves similar to Pd: When no pulse
client is playing, you can just start it without any troubles.
Roman
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
On Fri, 01 Feb 2013 18:29:15 +0100, Miller Puckette msp@ucsd.edu wrote:
Just to comment on why I ended up un-installing portaudio: I booted my Pi with a clean new Raspbian distro, compiled Pd (after installing some packages like git and alsa libs), ran it and found that pulseaudio was running. So I exit Pd and likk pulseaudio via "pulseaudio -kill". Then
start Pd and find it _still_ can't open audio. Why? ell, looking back I found
that pulseaudio was running again! So I re-killed it and tried a third time. Same result: for some reason, the mere act of starting Pd was causing pulseaudio to start itself up and grab the audio device so that Pd
couldn't get it. There are similar plot sequences in old Three Stooges and Marx Brothers movies, but I wasn't exactly laughing. What if I were about to have to walk on stage with this thing? So I googled around and read
through the many-page man page for pulseaudio and found no explanation of why
this was happening, nor any suggestions for how to disable pulseaudio other
than the "-kill" command I had already tried. After a while I lost patience
and uninstalled pulse and I suggest that everyone else do so as well.cheers Miller
Pulseaudio respawns itself, unless you've configured it not to. You can
disable it temporarily with the tool pasuspender.
The syntax would be:
pasuspender -- pd
On Fri, 01 Feb 2013 18:34:49 +0100, Kaj Ailomaa zequence@mousike.me
wrote:
On Fri, 01 Feb 2013 18:29:15 +0100, Miller Puckette msp@ucsd.edu wrote:
Just to comment on why I ended up un-installing portaudio: I booted my Pi with a clean new Raspbian distro, compiled Pd (after installing some packages like git and alsa libs), ran it and found that pulseaudio was running. So I exit Pd and likk pulseaudio via "pulseaudio -kill".
Then start Pd and find it _still_ can't open audio. Why? ell, looking back I
found that pulseaudio was running again! So I re-killed it and tried a third time. Same result: for some reason, the mere act of starting Pd was causing pulseaudio to start itself up and grab the audio device so that Pd
couldn't get it. There are similar plot sequences in old Three Stooges and Marx Brothers movies, but I wasn't exactly laughing. What if I were about to have to walk on stage with this thing? So I googled around and read
through the many-page man page for pulseaudio and found no explanation of why
this was happening, nor any suggestions for how to disable pulseaudio other
than the "-kill" command I had already tried. After a while I lost patience
and uninstalled pulse and I suggest that everyone else do so as well.cheers Miller
Pulseaudio respawns itself, unless you've configured it not to. You can
disable it temporarily with the tool pasuspender. The syntax would be: pasuspender -- pd
..ah, sorry, Didn't read through the previous posts very carefully.
Anyway, you can turn off PA autospawning in a configuration file.
Pulseaudio respawns itself, unless you've configured it not to. You can disable it temporarily with the tool pasuspender. The syntax would be: pasuspender -- pd
Good to know - but even though I now theoretically know how to deal with this, the whole thing gives me the creeps - what if some update changes the rules and pulse starts re-spawning itself again? I'll feel safer with it uninstalled - then it will have a much harder time restarting itself on my machine.
cheers Miller
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
I think I read somewhere that Romero intends to make a film about this ;)
Cheers, Pierre.
2013/2/1 Miller Puckette msp@ucsd.edu
Pulseaudio respawns itself, unless you've configured it not to. You can disable it temporarily with the tool pasuspender. The syntax would be: pasuspender -- pd
Good to know - but even though I now theoretically know how to deal with this, the whole thing gives me the creeps - what if some update changes the rules and pulse starts re-spawning itself again? I'll feel safer with it uninstalled - then it will have a much harder time restarting itself on my machine.
cheers Miller
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management ->
http://lists.puredata.info/listinfo/pd-list
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list
On 02/01/2013 06:29 PM, Miller Puckette wrote:
Just to comment on why I ended up un-installing portaudio: I booted my Pi with a clean new Raspbian distro, compiled Pd (after installing some packages like git and alsa libs), ran it and found that pulseaudio was running. So I exit Pd and likk pulseaudio via "pulseaudio -kill". Then start Pd and find it _still_ can't open audio. Why? ell, looking back I found that pulseaudio was running again! So I re-killed it and tried a third time. Same result: for some reason, the mere act of starting Pd was causing pulseaudio to start itself up and grab the audio device so that Pd couldn't get it. There are similar plot sequences in old Three Stooges and Marx Brothers movies, but I wasn't exactly laughing. What if I were about to have to walk on stage with this thing? So I googled around and read through the many-page man page for pulseaudio and found no explanation of why this was happening, nor any suggestions for how to disable pulseaudio other than the "-kill" command I had already tried. After a while I lost patience and uninstalled pulse and I suggest that everyone else do so as well.
since i was just playing around with pulseaudio recently, here some hints:
pulseaudio from starting again and again.
applications that want to use pulseaudio can use it, but others are not effected. (i recently decided to make this my default configuration on desktop machines: it allows me to watch vimeo and at the same time fuddle with Pd :-))
fgamsdr IOhannes
On Fre, 2013-02-01 at 18:43 +0100, IOhannes zmölnig wrote:
On 02/01/2013 06:29 PM, Miller Puckette wrote:
Just to comment on why I ended up un-installing portaudio: I booted my Pi with a clean new Raspbian distro, compiled Pd (after installing some packages like git and alsa libs), ran it and found that pulseaudio was running. So I exit Pd and likk pulseaudio via "pulseaudio -kill". Then start Pd and find it _still_ can't open audio. Why? ell, looking back I found that pulseaudio was running again! So I re-killed it and tried a third time. Same result: for some reason, the mere act of starting Pd was causing pulseaudio to start itself up and grab the audio device so that Pd couldn't get it. There are similar plot sequences in old Three Stooges and Marx Brothers movies, but I wasn't exactly laughing. What if I were about to have to walk on stage with this thing? So I googled around and read through the many-page man page for pulseaudio and found no explanation of why this was happening, nor any suggestions for how to disable pulseaudio other than the "-kill" command I had already tried. After a while I lost patience and uninstalled pulse and I suggest that everyone else do so as well.
since i was just playing around with pulseaudio recently, here some hints:
- in /etc/pulse/client.conf, you can set "autospawn = no" to prevent
pulseaudio from starting again and again.
Alternatively, one can put that into ~/.pulse/client.conf which doesn't require root and does not get overwritten by any future package updates.
Roman
On 02/01/2013 07:25 PM, Roman Haefeli wrote:
since i was just playing around with pulseaudio recently, here some hints:
- in /etc/pulse/client.conf, you can set "autospawn = no" to prevent
pulseaudio from starting again and again.
Alternatively, one can put that into ~/.pulse/client.conf which doesn't require root
true. but otoh, putting it into /etc will make it apply to *all* users (which might be desired or not)
and does not get overwritten by any future package updates.
if a package update overwrites configuration in /etc/ then it is broken. at least it should ask.
gfmadr IOhannes
On Don, 2013-01-31 at 13:12 -0500, Hans-Christoph Steiner wrote:
Does it happen with Pd-vanilla 0.43.4 or 0.44? I just take that code directly from Miller since I've never done anything with programming audio interfaces. Pd definitely does not place nice with pulseaudio, so make sure its not running when using Pd.
Let me also add, that when Pd is already running with ALSA (or Jack, for that matter), there is no way for pulseaudio to interfere, as the sound device is already occupied by Pd and not accessible for pulseaudio. I would say at this point pulseaudio can be ruled out as a cause for Julian Brooks' troubles.
Roman
Hi Roman,
Cheers for clearing this up (and the previous email), info is always good.
Still no idea what this error message is though?
Hmm - oh well.
Best wishes,
Julian
On 1 February 2013 11:36, Roman Haefeli reduzent@gmail.com wrote:
On Don, 2013-01-31 at 13:12 -0500, Hans-Christoph Steiner wrote:
Does it happen with Pd-vanilla 0.43.4 or 0.44? I just take that code
directly
from Miller since I've never done anything with programming audio
interfaces.
Pd definitely does not place nice with pulseaudio, so make sure its not running when using Pd.
Let me also add, that when Pd is already running with ALSA (or Jack, for that matter), there is no way for pulseaudio to interfere, as the sound device is already occupied by Pd and not accessible for pulseaudio. I would say at this point pulseaudio can be ruled out as a cause for Julian Brooks' troubles.
Roman
Pd-list@iem.at mailing list UNSUBSCRIBE and account-management -> http://lists.puredata.info/listinfo/pd-list