[MPlayer-dev-eng] [patch] prefer ALSA over OSS

Rich Felker dalias at aerifal.cx
Sun Apr 29 16:40:13 CEST 2007


On Sun, Apr 29, 2007 at 09:38:15AM +0200, Reimar Döffinger wrote:
> Hello,
> On Sat, Apr 28, 2007 at 08:59:20PM -0400, Rich Felker wrote:
> > On Sun, Apr 29, 2007 at 02:10:47AM +0200, Reimar Döffinger wrote:
> > > Well... It wouldn't necessarily be too late for a libalsa-only vs.
> > > clean, portable (as in acceptable to support under e.g. BSDs),
> > > simple and properly documented kernel API for ALSA discussion,
> > > would that one have a chance?
> > 
> > The clean simple properly documented kernel API already exists. It's
> > called OSS. Using the ALSA drivers does not preclude using the proper
> > API that already exists.
> 
> Unfortunately, exist, or more precisely "deteriorate", is the only thing
> it does currently :-(. It doesn't show too much I guess, because audio
> is probably the thing that has made least progress in the PC, probably
> even if you include power supply efficiency, and that is already
> unspeakable...

APIs don't deteriorate, people just add new "requirements" which in
many cases are stupid/useless.

> Still, DTS is at least not even in theory supported by OSS according to
> my linux headers, and I don't see anyone who would do whatever necessary to
> support more things done in hardware, OpenAL or EAX-style, if that ever
> happens to become more popular (maybe not that much more chance with
> that with ALSA either though).

Setting up hardware decoding requires some special considerations
anyway. I have no objection to using a separate interface (although
preferably a good one, not ALSA) for hardware decoding. This has no
bearing on what should be using for normal audio playing.

The whole point of hardware decoding is not to improve quality or
anything beneficial to users. It was a nasty attempt at taking away
users' rights and degree of control by moving the decoding from
software (where they could have complete control over it and access to
the algorithms) to hardware (where it was supposedly locked away...
yeah right...)

Rich



More information about the MPlayer-dev-eng mailing list