Pulseaudio problems

Lennart Poettering mzerqung at 0pointer.de
Tue Nov 20 17:08:53 UTC 2007


On Tue, 20.11.07 08:32, Paulo Cavalcanti (promac at gmail.com) wrote:

> >> Hmm? I don't really understand what you edited in that file and what
> >> you achieved with this?
> 
> This is how I send mplayer output to my TV:
> 
> /usr/bin/xinit /usr/bin/xterm -ut -e $HOME/bin/$PROG \
>              $MOVIE -- /usr/bin/X :1 -layout TV &
> 
> With F8, I get not sound at all, unless I execute the above command as root.
> Then I created an audio group for myself and added to
> /etc/security/console.perms.d/50-default.perms:
> 
> # roma
> <sound>=/dev/dsp* /dev/audio* /dev/midi* \
>         /dev/mixer* /dev/sequencer* \
>         /dev/sound/* /dev/beep \
>         /dev/snd/* /dev/adsp*

In F8 permissions to the audio devices are no longer managed by
pam_console, but by HAL. You apparently start MPlayer on a new
virtual terminal. If you want PA to work on that new terminal you of
course have to start it much the same way as you now start the X
server on it.

> >> Yes, I will have another look on the whole Xine situation.
> 
> xine and gxine do not work with pulse plugin (the connection is lost
> after a few
> seconds), but totem does (it is based on xine, right?). What is the
> difference?

Totem uses GStreamer by default, last time I checked. And GST is
supported very well by PA.

> > arecord -D hw:0.0 -d 0 -f S16_LE -c2 -r48000 | aplay -D pulse &
> >
> > for capturing from line in.
> 
> >> Hmm? what did you notice?
> 
> There is a slight hiccup (missing of sound) from time to time.
> It is very fast,and it does not seem to be related to the amount of
> processing on my computer: Core 2 duo, 2.1GHz, 2GB ram, x86_64, F8

Hmm? so "hw:0.0" seems not to be controlled by PA? What is controlled
by PA then? If you connect two different audio devices like this via a
pipe than you will very likely get hiccups eventually. Different sound
cards have different quartzes with small deviations in the
frequency. This, while "hw:0,0" and the pulse output might be in sync
at the beginning they start to deviate more and more the longer they
play. And eventually you get an "hiccup" when the deviation becomes
too large and either the buffers overran or underran.

Lennart

-- 
Lennart Poettering                        Red Hat, Inc.
lennart [at] poettering [dot] net         ICQ# 11060553
http://0pointer.net/lennart/           GnuPG 0x1A015CC4




More information about the fedora-devel-list mailing list