Re: GNOME 2.26 module inclusion discussion heats up


the benefits of using PulseAudio when using OSS don't seem to be
as compelling as when you are using ALSA since OSS has mixing
functionalities lacking in ALSA already built in.

These are misconceptions. Misconceptions about what ALSA does, and
about what PA does. ALSA does have mixing capabilities (it goes by the
name of "dmix"). While dmix might have some drawbacks it is certainly
not nearly as crazy as OSS4-style in-kernel mixing.

It has been discussed gazillions of times before what PA is and why
you'd want to have it or not have it. I am not going to repeat it here
once more. But rest assured: PA is not what you apparently think it

I think the best explanation I have seen you give was on the
libcanberra-discuss mailing list (August 22, 2008):

> It is popular misconception to reduce PA just to something that does
> mixing. If that was the case, then yes, it would be completely
> redundant by OSS4 or dmix.
> One of the primary reasons why I think everyone should adopt PA is the
> "glitch-free" playback model, which you might already have read about
> on Planet GNOME. It's something very technical, and not directly
> visible to the user, but if you ask me, this is an absolute must-have
> for a modern PCM audio system -- and OSS4 certainly doesn't give you
> that. (to make g-f work on OSS will need some substantial hacking from
> your side however, both in PA itself and probably also in OSS).

> Other features PA has, besides nifty effects like positional event
> sounds, and such like is: network transparency, saves/restores
> volumes/devices for streams, allows users to switch streams between
> devices on-the-fly, has an elaborate property system for all
> sterams/devices (i.e. have a nice icon for each stream that shows up
> in the per-application volume control), hotplug, "rescues" streams to
> different devices on hot-unplug; combining multiple audio devices into
> a single one; automatic upmixing/downmixing of surround sound; ...

Here at Sun we've never been really opposed to integrating PulseAudio.
We do, however, tend to avoid integrating new modules unless there is a
very clear need or benefit.  Probably the most exciting new feature,
glitch-free, won't work without significant additional work in OSS
which isn't resourced.  Some features seem nice, like positional event
sounds and better hotplug support, but I am not sure that such features
by themselves is worth the bother of integrating and supporting yet
another audio library.

Whatever this says about the average Solaris user, not a single user has
yet requested we add PulseAudio to Solaris for any reason.  If some user
had ever told us, "please add PulseAudio because it would enable some
important feature for me", then we probably would have already
integrated it.

At any rate, since it seems PulseAudio is becoming a more hard
dependency of GNOME in general, this well might push us to integrate it.


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