[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: Crackling sound in games, pulseaudio CPU usage high (F10)



On Tue, 02 Dec 2008 14:20:11 +0000, NM wrote:

> On Mon, 01 Dec 2008 22:31:33 +0000, NMONNET wrote:
> 
>> Thanks a lot, it seems to be working. ☺
> 
> Well actually I now get this filling up my messages:
> 
> Dec  2 15:16:34 ws pulseaudio[30881]: module-alsa-sink.c: ALSA woke us
> up to write new data to the device, but there was actually nothing to
> write! Most likely this is an ALSA driver bug. Please report this issue
> to the PulseAudio developers.


First of all, I had to add myself to the pulse-rt group in order to set 
my Gnome sound to the pulse audio server.

I occasionally get these messages.  Like you, the frequency increased 
dramatically when I added tsched=0 to my /etc/pulse/default.pa file.

Unfortunately for me, it did not reduce the CPU usage significantly when 
running Audacious configured for pulse audio.

Adding tsched=0 did eliminate popping and crackling when running 
Rhythmbox.

Kaffine is rather strange.  When configured to use the GStreamer engine, 
the pulse audio volume monitor shows the client as 'unknown'.  When 
configured to use the xine engine, the pulse audio volume monitor shows 
the client as ALSA (expected).

Basically, any music player that appears to use pulse audio directly 
causes pulse audio to use 13% - 16% CPU on a 2 GHz P6 ancient Dell system
(playing a standard MP3 file).

Attempting to use pulse audio with Xine generates an immediate segfault 
as shown in the /var/log/messages entry below.

xine[25985]: segfault at 0 ip 0072b4dc sp b51bbe30 error 4 in 
libpulse.so.0.7.0[6cf000+6e000]

While I like the concept of pulse audio, there seems to be some 
significant challenges (or at least a lack of understanding on my part).

I'll write up more about my adventures with pulse audio in another 
message.

/mde/


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]