> I remember how happy I once was when I found libao to avoid
> the parallel ALSA/OSS existance problem when coding a very
> small application that basically just wanted to deliver
> output. I wonder how hard it would be to write one library
> that does the multiplexing and autodetection/config parsing
> once and for all, and maybe also auto-wrapped the
> callback/vs/push model situation in both directions, such
> that an application could choose which model it wanted?
> New options on the planet could then in the future just be
> added to this layer. Some global library config would
> allow the user to define default output devices and maybe
> even do more sophisticated things like passing backend
> specific parameters for a specific program...
I also had this idea some time ago. Some disadvantages:
* Someone is needed who does this job
* It added a further layer to the already existing mess
* Many many applications need to be rewritten
* It is a dirty hack regardless how well it will be written
> > BTW: For me ALSA direct access (and therefore blocking
> > the device) seems to be a bit ugly for a multitasking and
> > multi user operating system like linux is. Is ALSA direct
> > access really an option (regardless that DMIX can help
> > with this)?
>
> I use it all the time since I own an SB Live!. On one
> other box I occasionally do work on I have one of those
> evil non-hardware-mixing cards, and it iritates the hell
> out of me that I can no start jack while the stream from my
> local radio station is running.
Yeah, but users want hardware as cheap as possible, even
musicians. AC '97 doesn't offer hardware mixing, my terratec
USB card doesn't and I'm not sure if the AC '97 successor HDA
does.
> > An desktop independent soundserver available on each
> > linux machine could help a lot. JACK could be a possible
> > solution.
>
> Why can't ALSA solve this once and for all somehow? It
> seems a bit of an overkill to have a full-fledged
> soundserver running when something like dmix could also do
> it?
Why not? JACK even enables users to easily record a web radio
stream while listening to it. Even normal users can benefit
from it. CPU usage? Forget about it. Contemporary machines
are bored while the user listens to ogg files while reading a
web page (or even lenghthy discussions on LAD ;-) .
> (BTW, have to try dmix now, really...)
Me too. I'd like to try kasound which has been released on
sourceforge just these days (Thanks dp for the hint).
> What I mean is
> if you have your soundserver, you are still left with
> rewritting every single app you ever want to use to use the
> soundserver, otherwise the problem still remains. There are
> a hell of a lot of tools around that use either plain OSS
> or direct ALSA.
I really dislike the OSS based applications.
Thanks for your thoughts & best regards
Received on Tue Jun 21 04:15:12 2005
This archive was generated by hypermail 2.1.8 : Tue Jun 21 2005 - 04:15:12 EEST