Re: Proposed Modules, My Take

[Mike Hearn <mike navi cx>]

On Sun, 16 Jan 2005 14:07:25 +0100, gabor wrote:
> please understand that dmix does not work. period.

It mostly works here. I have an AC97/i810 audio chipset. Each
alsalib upgrade has resolved more bugs for me.

> maybe it's bad design, or simply buggy code, but it does not work.
> and it's like that for a looong time..

Yes. It's been buggy for a long time.
 
> maybe it's good enough to play 2 mp3 files at the same time, but the
> multimedia needs does not end there.

You're right, there is a much larger world of audio out there. However
people who have cards that don't do hardware mixing probably don't care.

> so currently the ONLY way to have software mixing is to use a sound
> server.

dmix *is* a sound server, essentially. So the real question is, should the
sound server have public API that programs are supposed to use, or should
it be an implementation detail of the operating systems low level sound
API (ALSA on linux, DirectSound on win32, CoreAudio on MacOS, OSS on
FreeBSD etc).

> i'm just tired of this 'software mixing is not a problem. dmix works'
> argument. which is not true.

I'd suggest upgrading or doing some debugging ... ? I agree it's buggy, it
is here too, but it's not unusable. And I do believe that it will become
stable.