[alsa-devel] 4.0-rc3: sound broken in chromium but not in mpg123

[Takashi Iwai]

At Mon, 16 Mar 2015 13:37:19 +0100,
Pavel Machek wrote:
> 
> On Sun 2015-03-15 09:23:15, Takashi Iwai wrote:
> > At Sun, 15 Mar 2015 09:15:44 +0100,
> > Pavel Machek wrote:
> > > 
> > > On Sun 2015-03-15 09:12:47, Takashi Iwai wrote:
> > > > At Sat, 14 Mar 2015 21:11:09 +0100,
> > > > Pavel Machek wrote:
> > > > > 
> > > > > Hi!
> > > > > 
> > > > > Intel desktop machine:
> > > > > 
> > > > > 00:1b.0 Audio device: Intel Corporation NM10/ICH7 Family High
> > > > > Definition Audio Controller (rev 01)
> > > > > 
> > > > > Sound worked ok in both in previous kernel versions.
> > > > > 
> > > > > Any ideas?
> > > > 
> > > > How broken?  Did you get any error messages?
> > > 
> > > I get some funny, faint, rhytmic noises where audio should
> > > be... Actually, I somehow suspect something is wrong with time during
> > > video playback.
> > 
> > Did you try audio playback without video on chromium?
> 
> Yes. Does not work, either; I tried
> http://www.thenakedscientists.com/HTML/podcasts/ .
> 
> > > > If the behavior depends on the application (e.g. ALSA vs OSS), I'd
> > > > suspect rather a device permission problem at first.  There are no
> > > > significant changes between 3.19 and 4.0-rc3 regarding HD-audio.
> > > 
> > > My system contains pulseaudio, so it is more complex than ALSA vs OSS
> > > :-(.
> > 
> > OK, that was missing information :)
> > Then there should be any difference in the kernel driver side -- it's
> > only a single PA process who accesses the sound device in the very
> > same manner no matter what clients are used (after all it's the
> > purpose of PA).
> 
> > > Audio seems to work ok on thinkpad x60 with very very similar config.
> > 
> > OK, then another thing I would check is to try different sample rates
> > (e.g. via playing WAV files over PA).
> 
> Like this?
> 
> pavel at amd:/usr/share$ pacat ./sounds/alsa/Front_Right.wav
> pavel at amd:/usr/share$ pacat --rate=48000 ./sounds/alsa/Front_Right.wav
> pavel at amd:/usr/share$ pacat --rate=44100 ./sounds/alsa/Front_Right.wav
> pavel at amd:/usr/share$ pacat --rate=32000 ./sounds/alsa/Front_Right.wav
> pavel at amd:/usr/share$ pacat --rate=16000 ./sounds/alsa/Front_Right.wav
> pavel at amd:/usr/share$ pacat --rate=8000 ./sounds/alsa/Front_Right.wav
> 
> This seems to have behaved as expected.

Hmm.  You might see a difference when comparing
/proc/asound/card*/pcm0p/sub0/* files between working and broken
cases (while playing something).

Other than that, I have no idea.  A bisection would be needed, I'm
afraid.


Takashi