On Mon, 2009-08-03 at 10:34 +1000, david wrote: > How do you dump PulseAudio?
In Ubuntu 8.10, the main thing to do is run gstreamer-properties, and
select ALSA as the default output plugin instead of PulseAudio.
In Ubuntu 9.04, the situation is stickier, as the default ALSA device
('default') is also rerouted through PulseAudio, which breaks all but
the most basic of all applications (e.g. most media editing apps and
games break as a result).
To fix this, edit /usr/share/alsa/alsa.conf, and comment out the line
that refers to "/usr/share/alsa/pulse.conf".
I presume recent versions of Fedora will have the same behaviour as
Ubuntu 9.04.
> pulseaudio also causes problems with Kino so what does it do that is
> actually useful?
I'm no expert on the subject, but basically any application that tries
to use anything outside the "Safe ALSA subset" [1] (a subset
suspiciously invented by Lennart himself) will break if it is using the
PulseAudio virtual device, rather than real hardware.
That said, Pulse does do some useful things. Networked sound is one of
them -- my Bluetooth headset doesn't work pair with more than one device
at a time for some weird reason. So to avoid re-pairing all the time, I
can share my headset on the network, and have my desktop computer stream
audio to my headset via my Eee PC, which stays paired with it all the
time.
While that's nice, in practice, the Avahi-based sharing of audio devices
is flaky, and will often fail to reappear if the Pulse daemon restarts
for any reason (which it does often, because it segfaults multiple times
during any session).
Per-application volume controls are also provided by Pulse, which are
nice, although I'd prefer it if the functionality was built into dmix
instead. Also, live switching of audio streams between audio devices
should be handy in theory, although in practice it just segfaults the
Pulse daemon nearly every time you switch (and it hasn't changed at all
in that regard since I first started using it in 2007).
I used to love PulseAudio. Nowadays I think the sound system in Haiku
(which can use OSS4 as a backend) is much better -- and it certainly
doesn't crash several times a day (which is incredible, given that Haiku
hasn't even had an alpha release), and supports per-application volume
controls.
[1] http://0pointer.de/blog/projects/guide-to-sound-apis.html
signature.asc
Description: This is a digitally signed message part
-- SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/ Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
