Linux OSS + pd

[Charles Baker]

    Just to let folks know,
I took a moment or two to look into pd's compaint during
linux_open_audio...it was fairly basic: a bad value for the exponent
where it reconfigures the card's audio buffer. As Larry T. pointed out
to me, the linux full-duplex code doesn't complain during compile or
execution, it just doesn't work as full duplex. Just printed out the OSS
api document, and will try to see what I can do with it...if I am not
duplicating effort....
Also under my NT partition: I *cannot* keep the driver up! It always
starts up correctly (according to the printf's....), but bombs with
"dacs died, switching to alternate timing source". (:-(
This bomb can be easily triggered by toggling audio ON while active
signal is going into card.
My card/driver: Turtle Beach Fiji, driver version 4.04.
I've got the Cygnus gnu stuff for Win32 on my machine...what are you
guys using for NT:
MKS, MSVC, Cygnus, other? I will try to get things up under gdb under
NT, and see what I can learn of this error. PS: I removed driver for my
other sound card from NT, just to simplify. And I have had no problem
with the other audio apps under NT (CoolEdit, Digital Orchetrator Pro,
others).
Well, this is getting to be fun....I'm looking at adding portamento
based on key velocity for  a performance extension to the ringmod
examples: again for performing the "classics"...I know,
c'est tres passe. Tough, I love Mantra...also I performed an old R.
Reynolds piece with the same
performance need a ..uh..'few' years ago. I think it's a cool sound.
(Betcha others do too...)
Oh, and I'll try porting some of my old MAX patches to pd...they were in
the original Essl RTClib,
but since I don't own a power mac, I haven't maintained them...


CharlieB
baker at charlieb.com
baker at ccrma.stanford.edu
cbaker at verity.com