[PlanetCCRMA] Pd + PulseAudio?

[Fernando Lopez-Lezcano]

On Wed, 2010-03-24 at 15:04 -0400, Peter Kirn wrote:
> Okay, using the launch parameters:
> pd -rt -jack -inchannels 2 -outchannels 2
> 
> Instantly get some DIO errors, though only on launch:
> audio I/O error history:
> seconds ago	error type
>      2.84	A/D/A sync
>      2.99	unknown
>      2.99	unknown
> 
> Ah-HAH! Sure enough, Fernando, you're right -- that fixes the problem.
> Pd *instantly* autoconnects (those little glitches in the errorlog
> show that is immediately tries to connect and succeeds), patches
> properly to system out, the lot.
> 
> So what's going on here, exactly? Why does Pd work with the explicit
> inchannels/outchannels flags but not without them? 

I don't know. I guess it is just the way the current i/o code in Pd is
written. The only way to really know is to read the code (and maybe fix
it and send in the changes to Miller, that's how open source works :-)

And I may add that in general I prefer programs that do not try to
second guess what I want to do. Specially in something like Pd. How many
channels are you going to use? Where should they connect when you start
Pd? That is something you can know but the program, in the more general
case, cannot. 

> JACK shows up the
> same way in the Pd GUI.
> 
> And why did the AK1 not have the same problem with its 4 outs, 2 ins?
> (wouldn't it seem that'd be more likely to fail, not less, especially
> since the AK1 lacks any alsamixer-visible parameters?)

That I don't know. 

> Having blamed pulseaudio, the issues here seem to be related to Pd,
> since once JACK is the audio system, PA is indeed out of the picture.

That is correct. The problems you were having had nothing to do with PA.
I use Jack for all my "serious" audio needs so I never see problems with
PA (now, _before_ Jack was able to negotiate card ownership with PA,
that was indeed problematic - but that is in the past now). 

> Also, Fernando, since I just manually started qjackctl and Pd, are
> there any sort of scripts people tend to use to start JACK (maybe
> without the GUI), then Pd configured properly? I'd love to be able to
> recommend a solution for people, as I know a lot of people just give
> up before they even get where I got -- and automating the process
> would be part of the appeal of setting up a dedicated Linux config.

Sorry, I don't have a set of scripts. What you would need depends on
what you want to do. 

-- Fernando


> Cheers, and truly, thanks for your immense patience and help, Fernando
> and Niels and everyone.
> Peter
> 
> On Wed, Mar 24, 2010 at 2:49 PM, Niels Mayer <nielsmayer at gmail.com> wrote:
> > On Tue, Mar 23, 2010 at 8:26 PM, Peter Kirn <peter at createdigitalmedia.net>
> > wrote:
> >>
> >> Looking at AlsaMixer as you suggest, I definitely see the HDA Intel
> >> card. Master and PCM have volume levels; Headphone and Speaker show up
> >> but do not have mixer values.
> >
> > ...
> >>
> >> "control open "hw:intel" (No such device)
> >>
> >> Cannot initialize driver"
> >
> > Hunches:
> > Something is unexpectedly grabbing your hw:intel controls.
> > Consider that X/Gdm launch pulseaudio, so you can have those cute noises
> > when you mistype your password; it will therefore grab the hw:intel
> > controls. By totally removing pulseaudio, this can't happen, ever.
> > Consider also that these intel cards often include a semi-shared HDMI audio
> > portion e.g.
> > ##  1 [SB             ]: HDA-Intel - HDA ATI SB
> > ##                       HDA ATI SB at 0xfb7f4000 irq 16
> > ##  2 [HDMI           ]: HDA-Intel - HDA ATI HDMI
> > ##                       HDA ATI HDMI at 0xfb9fc000 irq 19
> > Perhaps also consider disabling the HDMI portion to prevent interference
> > from that (as well as not loading extra drivers that you're not using).
> > (from /etc/modprobe/*.conf: )
> > alias snd-card-1 snd-hda-intel
> > options snd-hda-intel index=1,2 enable=1,0
> > As a half-step towards preventing problems, you can also even w/ pulse
> > running, disable all system sounds and all error warning sounds. That causes
> > a whole class of sound system lockups in and of itself. However, once you've
> > gone that far, you might as well go all the way and follow my original
> > suggestion: http://old.nabble.com/uninstall-pulseaudio-to-increase-audio-app-stability-across-updates-%28was-Re%3A-yum-update%29-tp27759501p27759501.html
> > -- Niels
> > http://nielsmayer.com