17 Dec
2009
17 Dec
'09
8:40 a.m.
It's also true to say that elsewhere, on Windows, if you want to have
low-latency, you also loose the software mixer. AFAIK ASIO drivers and
KS implementations are single-client like alsa.
To me, it makes perfect sense to have alsa as a HAL, and get people
using pulseaudio or jack at application level.
Victor
On 17 Dec 2009, at 12:29, Paul Davis wrote:
On Thu, Dec 17, 2009 at 6:47 AM, Kjetil S. Matheussen
<k.s.matheussen(a)notam02.no> wrote:
No. Not providing mixing for all devices is a
design fault
in alsa. I'm going to install OSS as soon as I get time.
Kjetil, although you're absolutely welcome to your own opinions on
this, I would note two things:
1) ALSA has long provided "dmix" user space mixing without a server.
Its a very, very clever
solution that is, unfortunately, so clever that it often doesn't
work for the most demanding
situations. For normal desktop use, it has always worked well.
However, as well as it works
simply goes to emphasize all the desirable details of a desktop
sound system that dmix
by itself does not provide: per-app volume control, easy device
switching, and more. As
a result ...
2) many people would say that its not ALSA's job to provide this
kind of functionality but
instead to act as a kind of HAL for audio. The job of mixing
should (and can) be done
by a user-space implementation (PulseAudio or JACK depending on
your workflow
needs), and ALSA can simply be "the device drivers". This is
certainly how the
audio ecosystem is shaping up on Linux now, particularly with
the increasing
emphasis among those paying for system development because
their work
generally involves low power mobile devices where the desktop
issues are absent.
--p
_______________________________________________
Linux-audio-dev mailing list
Linux-audio-dev(a)lists.linuxaudio.org
http://lists.linuxaudio.org/mailman/listinfo/linux-audio-dev