[pulseaudio-discuss] null-sink rewinds causing stutter

[Daniel]

Hello Tanu,

that's a nice catch! Indeed this is exactly what I have, just that volume
changing triggers it way more noticably, because of the high frequency of
triggers. And indeed it's also there when recording from an
alsa-sink.monitor. I just thought this would be fine, because the audio out
of the speakers is fine, but the recorded result is broken just as with a
null-sink.

I could reproduce that on 3 different machines, so it's reproducable
anytime. Record any audio from any sink.monitor and change volume of the
stream input or the sink itself and watch it happen.

I am able to grab source code out of a VCS, I can patch and I can compile,
I can package and I can test and analyze. The only thing I cannot do, is
writing code :(

Anyway: Thank you very much.

Best,
Daniel


2013/11/22 Tanu Kaskinen <tanu.kaskinen at linux.intel.com>

> On Fri, 2013-11-22 at 13:56 +0100, Daniel wrote:
> > Dear list,
> >
> > I ask you for some assistance with null-sink and skipping/stutter.
> >
> > On client connect/disconnect and volume changes and many more things,
> > pulseaudio does a buffer rewind to let the changes take effect
> immediately.
> > So far,s o good, on all my soundcards this works like a charm.
> >
> > Now I have some audio stream setup where I play audio to a null-sink and
> > record it from null.monitor. This works really great, until some action
> > like pause/play, volume change, or connecting another audio client is
> > causing a rewind. Then the recorded audio hangs for approx 200ms or even
> > more. On soft volume changes, there are many many rewinds done, and so I
> > have many of these skips.
> >
> > The skipping audio is actually hanging, it does not drop any data. There
> is
> > just a silent break of about 200ms inserted into the stream.
> >
> > Of course it would be great to make rewinding on null sink just work as
> it
> > does on alsa sinks, but in case this is inevitable, I would be fine with
> > some possibility to disable rewinding. I know that all action then are
> > delayed, but that would be okay for me as the stream is delayed about 8
> > seconds anyway, until it reaches the receiver.
>
> There's no reason why those glitches should be inevitable, so probably
> module-null-sink or the rewind handling of monitor sources in general
> has some bug. Now we just need someone to send a fix for the bug...
>
> Not that this will help you much, but this bug might have the same root
> cause: https://bugs.freedesktop.org/show_bug.cgi?id=68107
>
> --
> Tanu
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/pulseaudio-discuss/attachments/20131122/ef39de0d/attachment.html>