Mailinglist Archive: opensuse-packaging (139 mails)
| < Previous | Next > |
Re: [opensuse-packaging] PulseAudio plans for suse 11.0
- From: Wade Berrier <wberrier@xxxxxxxxxx>
- Date: Thu, 13 Mar 2008 14:32:23 -0600
- Message-id: <1205440343.10300.69.camel@xxxxxxxxxxxxx>
On Thu, 2008-03-13 at 09:30 +0100, Adrian Schröter wrote:
On Thursday 13 March 2008 00:43:08 wrote Wade Berrier:
On Wed, 2008-03-12 at 22:59 +0100, Adrian Schröter wrote:
On Wednesday 12 March 2008 22:46:51 wrote Wade Berrier:
Hi,
On Wed, 2008-03-12 at 18:07 +0100, Ludwig Nussel wrote:
Marcus Rueckert wrote:
On 2008-03-12 17:45:05 +0100, Christian Morales Vega wrote:
In any case what openSUSE will do is something must be cleared.
For what I understood Gnome is going to use PulseAudio yes or
yes. Upstream will port their apps to PulseAudio and openSUSE
can't patch all of them, correct?
GStreamer, xine... they can't have a default output for KDE and
another one for Gnome. KDE4 will use Phonon, that can use xine or
gstreamer like backend. Will KDE4 apps use PulseAudio when Phonon
uses gstreamer and ALSA directly when Phonon uses xine? In the
first case apps will have *two* per application volume controls
(Phonon and PulseAudio)?
Without a clear path this could be very confusing.
ideally pulseaudio and phonon would be just wrapper for that
functionally inside alsa.
Quoting the fedora-devel mail:
6. If you have an application that uses ALSA, please make sure that
it doesn't hardcode ALSA driver names (i.e. something like "hw:0"),
it should use "default" instead, which is now being redirected to
"pulse", our plugin for libasound. Hardcoding ALSA device names
(besides "default") is a bug in your application anyway, so here
you have yet another reason to fix that!
So there is an alsa plugin and if an application normally uses alsa
it will transparently use pulseaudio. No application change
required. Sounds sane to me. I wonder whether that also works for
applications that support surround sound such as quake4 though.
I'm not sure either... but, if it fails, this would be an instance
where the real quake binary could be wrapped in a shell wrapper that
also runs 'pasuspender' beforehand so that quake could use the hardware
directly.
I brought this up on opensuse-packaging with the hope of preconfiguring
all apps (patching our packages to: create wrappers, modify default
config files, etc... ) so they would work with pulseaudio.
The goal: to be running any desktop environment, start up any
application that uses sound, have sound actually work (even when using
a cheap sound card that does no hardware mixing), and get all the
features of pulseaudio.
(I'm currently using one of those cheap cards, but am almost frustrated
at the point of buying a sound card that can mix 32 streams in hardware
just so I can hear pidgin beeps and hear flash video at the same time
:)
Was there any problem in the last years ?
Not so much when I had a card with at least 4 ports for hardware mixing,
but when using a 1 port card, definitely yes. There always seem to be
oss apps that hog my sound card.
Plus, as a gnome user, some apps still expect to output sound to esd
(instead of going to gstreamer as they should).
hm, this sounds like esd still used /dev/dsp and caused this ?
I haven't any problems anymore since 2-3 years after backporting the KDEMM
(pre-Phonon) stuff to avoid fixing arts bugs ;)
And surely some headaches come from me trying to configure everything to
work with pulse, but the features of pulse are important to me. I think
it's doable to have both: features of pulse, all apps working.
I can only remeber that some sound cards, which do not support mixing
where not configure with dmix. But beside of that, I think we got rid off
all problems since that we dropped sound server in KDE 3 and used plain
alsa.
From that perspective, I fear more steps backward than forward with
introducing a sound daemon again ...
I agree there may be steps taken backwards in transition, but the
feature set of pulse makes it possible to move ahead of where alsa alone
is capable of.
just out of interest, when do you see these features used ? For an advanced
user, who want to play remotely on his HiFi ?
Or is it something like NMM as universal managing all multimedia devices
around you, switching all video / audio playbacks/recordings from device to
device ?
The features I use/like:
1. Play my local sound across network with a nicer sound system, with
no latency issues, as well as synchronous output to multiple computers
(I have a computer upstairs and downstairs, and I can output the same
music to both)
2. Plug in my usb soundcard with headphones and use the gui app to
route all sound to usb sound card (which gets remembered next time I
plug the usb sound device in)
3. Individual sound settings for each app, so I can adjust volume for
apps that don't have software volume control (in which all settings are
remembered the next time that app plays)
4. ability to use pasuspender for closed source apps/games that want to
hog my sound device. Even if I have music playing, or any other app
that has a hold on the sound device, I can suspend the sound server, and
let the non-compatible app take control of the sound device.
I really do understand that not everyone wants to use pulse :) As stated
in my original mail, I'm not interested in debating over this, but was
wondering if pulse was going to be distribution wide.
(I was hoping it was going to be distribution wide, then there would be
a better chance of 3rd party vendors' packages supporting pulse: ie
packman.)
Anyway, now that the answer to that question is clear, I would still
like to see all apps working in whatever desktop, with and without a
sound server.
Wade
| < Previous | Next > |