vibrato

[Harry Bissell]

Having just done a markey study on available pitch shifters... there is a
correlation between how long the delay is, and the amount of glitch.
Unfortunatly... the short delays that have "fast" tracking of the input
produce major glitches that get worse with amount of shift (octaves are
terrible...). The long delays are smooth, but very slow in response.

And (IMHO) using a BBD for this would be like Fr. Guido Sarducci comment
about the invention of "Peas on a Cob"

"Its like taking something that tastes BAD, and then making it hard to eat
TOO !!!"

H^) harry   (actual mileage may vary....)

"Mike I." wrote:

> It should be possible to build a pitch shifter (harmonizer) using a pair
> of long BBD's. A steady pitch shift will result for the length of time
> that a "sample rate differential" can be maintained between the input
> and output of the delay. eg. for a one octave up transposition, the
> sample rate for signals at the input is R KHz, while the sample rate for
> signals coming out is 2R KHz. A voltage controlled 1/x clock would
> probably be best for this, with the clock being modulated by an upramp
> (upshift) or a downramp (downshift). One way of doing this would be to
> use a variable amplitude ramp generator with a period of about 20 mS,
> another way would be with a fixed amplitude ramp generator with a
> variable period (faster LFO gives larger shift). A problem is the
> "retrace glitch" which occurs when the ramp resets. A possible solution
> for this is to use 2 parallel BBD's, modulated by ramps that are 90
> degrees out of phase. A "crossfade" circuit alternates between the 2
> outputs, so the BBD output glitches are alternately blanked. Doing it
> with DSP is the most sensible way to go, since there are better methods
> of avoiding glitches. I think there were some attempts to make BBD
> pitch shifters in the 1970's. Mike