As I've mentioned, I working on a sort of "harmonizer" patch that takes an incoming signal and pitch-shifts it by appropriate amounts based on what keys are pressed on a 25-key MIDI keyboard. My first attempt uses 25 separate pitch-shifters, derived from the patch in the docs area, with the most recent velocity sent by each key determining the volume of the signal fed through that object.
(I don't know if this makes a difference, but I've replaced the half-steps-based calculations of the original with the simple frequency ratios of just intonation.)
Unfortunately, this appears to be beyond the capabilities of my somewhat lame laptop (266 MHz Pentium under Windows 98 (I'd rather use Linux but so far have had no success with the USB-MIDI driver)). When audio is turned on, the system slows down so radically that, from watching the numbers attached to the outlets of the notein object, a key press is processed some 2-3 seconds later, and what audio gets out is brief instants of distortion.
A couple of possibilities strike me:
- Is it possible to, in a sense, power down the objects when they are
not being used (that is, when it has most recently received a non-zero control input)?
- Do less processing-intensive pitch shifters exist that the one I've
modified (which uses the tape-head emulation to do the shifts)?
- Am I just more ambitious than my laptop can stand?
Thanks for any enlightenment.