On Thu, September 13, 2012 5:00 pm, Paul Davis wrote:
On Thu, Sep 13, 2012 at 6:51 PM, Len Ovens
<len(a)ovenwerks.net> wrote:
It depends how pulse is set up. If pulse is also connected to the
internal
card (that is the internal card is not turned off in pulse) Then the
pulse
jacksink/source modules will hold jack to whatever that internal card is
doing.
there is some confusion here. the pulse jacksink/source modules connect
pulse to JACK, not vice-versa. JACK will be using whatever hardware it was
told to use. the only situation in which JACK routes audio via Pulse is if
the user tells JACK to use an ALSA PCM device that is actually a pulse
pseudo-device (eg. "default" on a system set up with Pulse as the
default).
If pulse has the module-jack-source loaded, then jack feeds audio to
pulse. and if module-jack-sink is loaded then pulse feeds audio to jack.
Normally they are both loaded at the same time by module-jackdbus-detect.
It is possible to route audio in through one device to jack through pulse
and out to another device. Pulse has routing too... much harder to use
though.
I have a netbook with intel HDA audio. When I use Jack to run it the
lowest latency I can get jack to start at is -p128. At that latency the
wireless driver causes xruns on a regular basis. I also have an ART USB
audio interface. If I have manually unloaded the module-jackdbus-detect
module from pulse, The USB audio interface will start with jack with the
latency at -p 64 and no xruns. If I reload the same module I can now only
start the USB sound at -p 128 and I get the same xruns...on the USB card.
now if I configure pulse to ignore the internal sound interface...even
with the module-jackdbus-detect module loaded and streaming sound through
pulse and then jack... I can start jack at -p 64 with no problem. Pulse
does in some cases affect jack. I have repeated this more than once it is
always the same. Low latency kernel, jack is starting in RT.
In the same way that I do not know the internals of jack, I don't know the
internals of pulse. I am sure you know more than I do about both. But it
is hard to argue with experimental data.
--
Len Ovens
www.OvenWerks.net