pitch shifting....the analog way???

Magnus Danielson magnus at analogue.org
Mon Sep 14 20:34:03 CEST 1998


>>>>> "MC" == Martin Czech <martin.czech at intermetall.de> writes:

 >> Hi list...
 >> just wondered if there was any way to pitch shift with analog stuff...Even

 MC> It is possible, was done with a tape recorder and rotating heads,
 MC> could time stretch as well as pitch shift, but of course with lots of
 MC> artefacts. The name of the machine was Temphon, I think. In the 50's.
 MC> Maybe one could try BBDs, but this will be noisy.
 MC> WHat's the advantage of analog here?
 MC> Better get a used LXP5, this even has midi control.

The closest to analog that I can imagine is an Eventide H949 which
admitedly contains a lot of digital content but which for most of the
time has a analog feel to it (yes, CV in and frequency tracking!).
Sadly enough it also can carry some of the age diseaces that needs to
be cured.

Another thing that springs to mind is the Hammond Vibrato unit as
found in the A-100s for instance. It basically contains of a multi-tap
delay filter (RLC-based) with a capacitive rotary scanner (attached to
the synchronious motor assembly). This is a quite odd way to create a
vibrato, they will actually wiggle the pitch with the rotating scanner
connections. This should be a quite simple thing to clone for those
twisted that way, and hey, no mechanics is needed! A suitable planning
of the use of CMOS muxes, two VCAs a modulation oscillator (triangle),
a op-amp and a counter for the muxes and you are away.

I think that one might need quite a few delay-stages to make a good
approximation, but I think it basically can be done in mainly analog
by this method. The pitch shift effect of this kind of cursuit can be
done both up and down ("rotation" direction of the scanner controls
this) but the quality, bandwidth and usable depth of modulation is
most probably a direct function of the number of delay stages.

Compressing on input and expanding on output is badly needed I guess.

Lowpass filtering input and output can also prove important in
attaining a good result.

Crude but intreaguing, eh? ;)

Cheers,
Magnus (feelin a bit whacky today)



More information about the Synth-diy mailing list