The question of sound servers seems to be a neverending
problem. Today I started that maybe the solution was to give
up or rather to lower the ambitions. I mean we probably will
never find a sound server that satifies both video people,
audio people, the desktop people, game people and that runs
well on all the important flavours of Unix and Linux.
As a person coming from the desktop world my initial
interest was finding a good solution in the desktop context.
Currently we use Esound in GNOME and they use artsd in KDE.
Both sound servers are not exactly something you want to
give a round of applauds, but of the two artsd is
the better one. We have discussed using artds in GNOME, but
the enthusiasm for switching to a sound server that sucks
less, but still sucks has not been overwhelming. So we are
desperatly looking for a solution that we can actually feel
is a good one.
One project I held high hopes for was MAS which
is being developed by Shiman Associates Inc., under the X11
umbrella. Lately I have started being more doubtfull of this
solution as a) they seem intent on bloating it and b) I am
not sure they will finish it unless someone starts sending
them some money. This need not be a problem, the source is
there so anyone could start working on it if it truly is the
solution we have been looking for…..
Another solution that many of the audio developers like
is Jack. There
are two things that makes me a little wary of Jack, one is
that it is closely tied to/dependent on ALSA which makes me
unsure about how well it would work on Solaris or HP-UX.
The design is supposed to be able to use other backends, but
since ALSA is the only backend with the advanced
functionality Jack has been designed for it is a question
how good a solution you get with other low end backends.
Also it has been designed very much with high end audio
applications in mind so there is also a question on how well
it fits the bill for other user and developer groups.
In addition to these there are other projects which I am
not sure is even alive and even if so definetly not ready
for prime time anytime soon, like ASD for instance.
As many of you probably know I am helping out on the GStreamer project, and
your response will probably be that why worry about sound
server when GStreamer based apps can output to all of
Well the reason is simply that when we ship GNOME
2.2 using GStreamer as the backend for our multimedia stuff
it feels deeply disatisfying to tell distros and OS makers
to use whatever sound server/system they can manage to dig
up. Because when we do that we at the same time give up on
making sure that people have a solution that works
consistently well. We risk that on some platforms people
will choose a soundserver which easily creates lagging in
the sound, one that is a resource hog or the most critical
problem atm, the lack of good audio/video sync.
Yet, as I started this longish diary entry with, maybe the
solution is to give up and instead accept that things are
not perfect and there are difference in quality provided
between the platforms, but maybe after all it might be
better than trying to cram a one-size-fits-all solution onto
the world. Especially since the current sizes fit so few :)