🚀 go-pugleaf

In article <1aek39INN17s@agate.berkeley.edu> izumi@pinoko.berkeley.edu  
(Izumi Ohzawa) writes:
> In article <1aehh5INNa0r@stephen.cs.ubc.ca> sritchie@cs.ubc.ca (Stuart  
Ritchie) writes:
> >In article <1ae4klINNpph@agate.berkeley.edu> izumi@pinoko.berkeley.edu  
writes:
> >>My DSP gets interrupted 150,000 times per second with each sample
> >>being 12 or 16 bits, and 68040 doesn't even notice this unless
> >>something interesting happens in the input samples.
> >
> >Is the DSP fielding interrupts at this rate, or is this while polling?
> >(polling is cheating)
> 
> Interrupts, not polling.  Some CD-quality stereo recording apps run
> with the DSP interrupted 88,200 times per sec to get 44.1k x 2  
samples/sec.
> 
> You get get a lot more than that, although doing anything with the
> obtained samples gets tougher as the data rate increases.
> 
The DSP is the easiest way to get serious quantities of data into the  
machine.  I currently have an A/D running that simultaneously samples up  
to 16 channels at selectable rates.  The current configuration is 16  
channels with each channel at 16KHz which produces 256Ksamples per second.   
This is approximately equivalent to 6 channels of 44.1 data.  Building  
devices that accepted data at these rates would be much more difficult  
without the DSP.

Bob Brennan
Carnegie Mellon University &
Applied Speech Technologies
rab@cs.cmu.edu