[sdiy] PIC Quantizer project

[john mahoney]

----- Original Message ----- 
From: "harrybissell" <harrybissell at prodigy.net>
>
> No...what I meant is that you should have equal amounts of CV input
mapped to scale steps.   If you examine the Wiard quantized bank, you
will see that there are some scales that map 4, 5, 6, 8... input steps
per output note (IE 256 input states mapped equally to maybe 61 output
states.
>
> If the steps are not equally weighted... a triangle wave driving
(lets say) would produce an output of notes of unequal length... an
unnatural result.
>
> I tried this with the theremin and it, well.... sucked !    I would
not want John Blacet to duplicate my failures :^P

That triangle wave example is interesting, and I can see where you'd
want that. On the other hand...

Say the target scale was a major scale, where there are half steps and
whole steps between notes. A sweeping triangle will be closer to some
notes for a longer duration than others. How would you reconcile that?
The way I'm picturing it, it's like 1970's school bussing, where some
notes are moved away from a "correct" pitch that they are close to, in
order to "fill the quota" of a target pitch that's somewhere up or
down the scale.

The Theremin example seems to me like the opposite situation from the
triangle. In this case, you are supposedly (hopefully, wishfully!)
going to get close to the target pitch with your hand. Therefore, you
want to quantize to the nearest pitch, regardless of what percentage
of the scale's frequency range that pitch occupies. (This is hard to
describe, huh?) However, you have tried it and I have not, so I bow to
your experience.

I'm sensing a need for 2 different quantizing modes. One is "nearest
pitch," while the other is... I don't know... "equal spectrum" or
something. Oh, there, you called it "equally weighted."

I should play with this on my PSIM. Probably won't, though...
--
john