Re: Polypaudio for Gnome 2.10, the next steps

[Lennart Poettering <mztabzr 0pointer de>]

On Tue, 23.11.04 09:19, Damien Sandras (dsandras seconix com) wrote:

Hi!

> I don't know if Polypaudio is good in low-latency cases, but the
> reason

Polypaudio is good in low latency cases. That was one of the key
design prinicples. The client has full access to the playback buffer
in polypaudio, it may set the size, seek around, overwrite, pause, at
its free will. A client may define the buffering metrics very
flexibly. The user may configure the soundcard's playback buffers when
loading the driver modules into polypaudio freely. The polypaudio
daemon makes use of SCHED_FIFO if this is available to make sure that
even very short playback buffers don't hurt. There's no artificial
buffering in Polypaudio, memory is usually not copied around (i.e. the
read() call on the sound card uses the same buffer as the write() call
to the socket. In case you use mmap()ed sound IO the DMA buffer is
used with with the write() directly.) Short: polypaudio does
everything i could think of to make sure that no extra latency is
introduced. And in case latency is still too bad for you, you might
even load your own module into the polypaudio process, to let it do
VoIP for you. This would eliminate the latency cost of IPC.  (However,
I don't think that this is required, latency should be good enough.)

A Polypaudio driver for GnomeMeeting would be a nice asset.

> So *any* replacement able to cope with both input and output with
> reasonable delay is good for me.

Lennart

-- 
name { Lennart Poettering } loc { Hamburg - Germany }
mail { mzft (at) 0pointer (dot) de } gpg { 1A015CC4 }  
www { http://0pointer.de/lennart/ } icq# { 11060553 }