> Once you have the frequency response, it's just math to convert that to t=
he
> convolution kernel of a finite impulse response filter, which you then
> implement in a dsPIC.  Let's say you sample at 20KHz rate.  A dsPIC can d=
o
> up to 40M multiply-accumulates per second.  That means 2000 MAC per 20KHz
> sample.  That's 100ms of sound, or 33m of propagation.  That's more than
> necessary, so it passes the first level sanity check.  Of course the dsPI=
C
> will have to do a few other things than just running MACs, but there is
> enough headroom for things to work out.

What sort of time delay would this introduce? Without experience of doing t=
his way, it seems to me from your description that it would be 100mS, but a=
m probably way off the mark. My thoughts are that this is going to be a liv=
e situation, and if the participants can get a whisper of a 100mS delay bet=
ween what they say and anything they hear from the PA system, then they are=
 likely to end up stuttering badly.=20
--=20
Scanned by iCritical.

--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.