[Audacity-nyquist] Reverse in Nyquist

[Roger B. D. <rb...@cs...>]

I'll try to write and post a reverse function for Nyquist sometime soon.
The SHAPE idea was a good one, but Nyquist limits tables to (I think) 1M
samples to avoid disaster when someone passes a huge sound to OSC or =
SHAPE.
You can get away with large sounds when writing to disk, but tables =
evaluate
sounds in memory and are best suited to small things like wavetables.
Reverse would work well on a sound file, where you have random access, =
so
you can read from the end to the front, incrementally. I think with
Audacity, you'll have to read the entire sound to be reversed into =
memory in
order to get access to the end, so you'll need 4 bytes-of-RAM/sample =
plus a
couple percent overhead.  -Roger

PS (STRETCH x my-sound) has no effect on my-sound if my-sound is of type
SOUND. STRETCH changes the evaluation environment so that *behaviors* =
such
as PWL, LFO, OSC, PLUCK, etc. return longer duration sounds, but sounds
themselves are not changed.

For example:

(setf my-sound (osc c4)) ; computes 1 sec of middle C
  ;; it's a bad practice to assign sounds to globals, but we did it =
anyway
(play (stretch 2 my-sound)) ; plays my-sound without change, 1 sec =
duration

Alternatively:

(play (stretch 2 (osc c4))) ; evaluate OSC in an environment with a=20
  ;; stretch factor of 2, so the result is 2 sec of middle C