[Bug 692953] alsa modules are silent or noisy after several hours of use

[GStreamer (bugzilla.gnome.org)]

https://bugzilla.gnome.org/show_bug.cgi?id=692953
  GStreamer | gst-plugins-base | 1.0.6

--- Comment #21 from Thomas DEBESSE <thomas.debesse at rcf.fr> 2013-08-22 09:50:43 UTC ---
Robert Krakora wrote:
> Check this out:
> http://gstreamer-devel.966125.n4.nabble.com/Timestamp-drift-when-using-dmix-dsnoop-plugins-with-alsasrc-sink-td4658874.html

That's interesting, it's very similar to my problem, but with some differences
:

1. For him these three usecases has different effects :

* alsasrc device="default"
* alsasrc device="dsnoop:hw:0"
* alsasrc device="hw:0"

For me, these three usecases has same effect.

2. For him, the problem occurs according to the usecase and we hear right now
if it does not work well.

For me the problem occurs in any case and can not be heard immediately, it
takes hours to check.

3. He have warnings when the pipeline goes wrong.

Except for the pipeline of yesterday (jackaudiosrc ! … ! alsasink), I have no
warnings. Perhaps with this particular pipeline I also get the same bug that
he, in addition.

But these two bugs are probably related.

4. He can publish small log

He had attached a small "GST_DEBUG=5" log (57K uncompressed) that shows his
bug.

Yesterday I have found a computer that can run my pipeline with "GST_DEBUG=5"
and which is fast enough to output debug information without interfering with
the sound processing. The bug has not yet occurred, and I have already more
than 1.1Gb of compressed log (with xz). If unfortunately I will not be exactly
next to the machine to hear when the bug will happen, to say "I can cut the
log, the problem is just at the end", I'll have to search through billions of
lines.

Robert Krakora wrote:
> Have you tried adding "slave-method=2 provide-clock=true"
> properties to the alsasink plugin in the pipeline?  Try...
> "jackaudiosrc ! something ! alsasink slave-method=2 provide-clock=true"

Sebastian Dröge wrote:
> Also it can help to use provide-clock=false on alsasrc
> and alsasink. In that case it will use the systemclock
> for the pipeline and will try to use the driver
> timestamps of both devices to properly synchronize
> everything.

OK I have two soundcards that are not used yet, I'll try both.
Sebastian, which slave-method I have to use to test the "provide-clock=false" ?

-- 
Configure bugmail: https://bugzilla.gnome.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.
You are the assignee for the bug.