[nas] multiple dsp devices under linux?

Jon Trulson jon at radscan.com
Thu Jan 10 22:05:22 MST 2002

On Thu, 10 Jan 2002, Paul Fox wrote:

> Date: Thu, 10 Jan 2002 08:11:18 -0500
> From: Paul Fox <pgf at foxharp.boston.ma.us>
> To: nas at radscan.com
> Subject: Re: [nas] multiple dsp devices under linux? 
>  > 
>  > >     beep_device: <dsp device>
>  > >         The sound device through which to play the beep sounds.
>  > >         This cannot be the same device which mpg123 uses.  I 
>  > > 	use /dev/dsp1.  Note that most sound drivers provide
>  > > 	multiple dsp devices.
>  > > 
>  > > the last line is news to me.  in particular, i would like
>  > > very much to run the nas server on one audio device, and
>  > > have a second device available for use by non-nas-aware
>  > > applications.  is it really true that most linux sound
>  > > drivers have multiple entry points?  
>  > 
>  > No, when they say device, I'm pretty sure they mean hardware device. 
>  > For example, if you have two physical sound cards, then one will be
>  > /dev/dsp, and the other will be /dev/dsp1. I don't think you are
>  > suggesting getting a second sound card just to support non-NAS program.
> no, i'm not suggesting that.  by the same token i'm kind of
> surprised that _they're_ suggesting a second audio card just
> for beeps.  :-)  (i haven't gotte a response back from those
> developers yet -- maybe the second device is meant to be a
> pc-speaker pseudo sounddriver.)

	Under OSS, nasd (and other apps) seem to use /dev/dsp1 as the
microphone input...  

>  > To suggest a different solution, have you played with audiooss?  Despite
>  > the warning on that config line, you should be able to just specify
>  > /dev/dsp if you are running the program under audiooss.  (Although IMHO,
>  > it's too bad they don't ask for a beep *command*!)
> as it happens, the two programs i'm trying to run are both
> NAS-enabled, and i have to have nas running for one of them,
> but the other (mpg123) is running locally, and it just
> seemed like it would be more efficient (fewer cpu cycles),
> and suffer fewer latency issues (like the too-much-data-
> buffered-when-i-pause problem i mentioned in a different
> email), if it were accessing the device directly rather than
> going through the extra server process.

	I wanted to ask you about that (latency), I'll do that in another

Jon Trulson    mailto:jon at radscan.com
ID: 1A9A2B09, FP: C23F328A721264E7 B6188192EC733962
PGP keys at http://radscan.com/~jon/PGPKeys.txt
#include <std/disclaimer.h>
Bad Color Temperature, Too much Peach.

More information about the Nas mailing list