[pulseaudio-discuss] problem with webrtc and pulseaudio

Tanu Kaskinen tanuk at iki.fi
Mon Feb 20 14:59:29 UTC 2017


On Sun, 2017-02-19 at 14:02 -0300, Fatima Castiglione Maldonado 发
wrote:
> > Ok, that suggests that the problem is not with chrome specifically. But
> > in your first mail you said that firefox works fine...
> 
> Firefox works ok, have just tested it again to be sure.
> 
> > If you record with parecord without having chrome running at the same
> > time, is the sound still bad? If it is, you can check if the audio is
> > better without pulseaudio in theI middle with this command:
> 
> I have just recorded with parecord without having chrome running at the
> same time, and it sounds ok.
> 
> > pasuspender -- arecord --device=plughw:0 --format=cd test.wav
> > If that produces bad audio too, then it's a kernel driver problem.
> 
> And have just recorded with pasuspender and it also sounds ok.

Ok, so chrome does something that the other programs don't do. Maybe
chrome asks for so low latency that pulseaudio or the clients (chrome
and parecord) can't keep up with the frequent wakeups, or maybe chrome
uses a different sample rate than firefox, and the sound card doesn't
work properly with that rate.

To get some data to support or disprove those hypotheses, could you run
"pactl list" twice: when chrome is using the mic, and when firefox is
using the mic. The output will show what latency and what rate the
microphone is configured to. Since sometimes chrome seems to fail to
create the stream (when you get just silence), make sure you capture a
case when chrome manages to create the stream.

-- 
Tanu

https://www.patreon.com/tanuk


More information about the pulseaudio-discuss mailing list