Wavetables and Bill's bits

[Roman Sowa]

Hi all,

I've been off the keyboard for a while, so let me reply to some messages

to Don and partly to Harvey:
I know sqaure has rich harmonic content, but if you think of how
lookup table works, it won't be a problem for square wave.
Of course narrow pulses will die of aliasing, but for square it's not so
bad.
Why bother with lookup for simple waves? And why not? I'm free from
temperature issues and tuning 50 oscilators. I can do morphing like you
mentioned in your message. And like I said, it was a silly idea, just an
idea.
If I wanted just sqr, saw, tri I wouldn't use lookup, but other tehniques
(i.e. phase accumulator gives direct sqr and saw)

Harvey:
I didn't mean postfiltering! I ment filtering the sample *before* using it
as a lookup table - thus no aliasing surprises will come up.
BTW I can't see any problems with sweeping

Paul:
The whole thing was supose to be voltage controlled
(ADC at DSP input). ADC data would be - after exp conversion - put
directly to DCO. No sweeping problems at all, or maybe I'm missing
something.
Aliasing? see previous replies...
Sine wave out of DSP? sine is excellent made with lookup. I played once
with DSP56002 - from 1 to over a 100 sweeping sines produced via
256 words lookup table and they sounded better than any analog
wave-shaped-from-triangle sines I've heard.
Another way is to make sine generator out of sinus z-transform.
Tried this too, but wide range sweeping is not too easy.

Roman