[pulseaudio-discuss] PulseSink Properties

pl bossart bossart.nospam at gmail.com
Thu Jul 28 11:34:08 PDT 2011

On Wed, Jul 27, 2011 at 12:55 AM, Vallabha Hampiholi
<vallabha.pa at googlemail.com> wrote:
> We are now running PA with high resolution timers enabled in the kernel.
> Yet I still can hear the audio glitches.
> Attached are new traces.

I: alsa-sink.c: Using 16.0 fragments of size 8192 bytes (46.44ms),
buffer size is 131072 bytes (743.04ms)
I: alsa-sink.c: Time scheduling watermark is 20.00ms

This seems really wrong. If you are using timer-based scheduling,
looks like you are setting a very small ALSA period size for no good
reason. Try using 2 fragments in daemon.conf to reduce the number of
ALSA interrupts.
Try also using frace and pytimechart (available on gitorious) to
figure out what is going on in your system, there's nothing like a
visual representation of thread activity to understand possible
overhead or the reason for wakes/underflows.

More information about the pulseaudio-discuss mailing list