"Wait your turn" polyphony

[earthtonesaudio]

The recent "frequency shifter" thread with it's mentions of other analog pitch modulation techniques (PLL, ring mod) got me thinking there might be a way to do quasi-polyphonic pitch effects a/la Whammy or POG using relatively simple* analog techniques.
*Depending on how you define "simple" of course.   


Start with a "core" modulation device of your choosing, be it a simple diode rectifier, or a flip-flop divider, or a PLL, to give you whatever pitch interval you're after.  The main limitation with these things is that more than one note (or too much harmonic content) causes intermodulation distortion or other noise/junk that one way or another causes them to be single-note devices.

What if... instead of inputting a full audio signal, you chopped up the signal into different frequency bands and processed each band separately?  (Yes, I know it's been said... you just split the signal into a hundred narrow frequency bands, feed those into a hundred separate full-wave rectifiers or what have you, then sum the hundred sources with a mixer... not very practical; difficult to fit in a 1590A.) 

But rather than process each band with a separate circuit, how about use a single circuit to process each band separately in time?  Imagine your signal comes in, is buffered, then split into separate frequency bands.  Then, a multiplexer chooses one frequency band at a time, and sends it on to the rest of the circuit.  Then it de-selects that frequency, moves on to the next one, and repeats the process.  All of this happens very quickly so it sounds continuous.  You'd get middle C one octave up, then C#, then D, and so on across the range, and back to the start.  It wouldn't be perfect but it might be quirky in a good way.

[aziltz]

like a fast arpeggiator.   i doubt it would be continuous, but could be cool.

its kinda like playing monophonic into a frequency shifter where the shift follows some pattern. 

only now we can play chords in and theoretically get a single shifted note added that bounces around in pitch.

[earthtonesaudio]

You certainly could go that route, but I was envisioning a different approach.

Perhaps a little more detail on the signal conditioning:
Each "narrow band" would be a separate fundamental extractor, "tuned" to the range covered by its pass-band. 
You would need at least 3 to 5 bands per octave, maybe 4 octaves for guitar, for a total of say 20 bands.

-I'm suggesting you would sweep through these bands very fast, a full cycle of all bands within the space of few microseconds or so.  At this speed, you wouldn't hear any pattern or rhythm.

There is the issue of lower notes taking longer to process... I have to think about this some more.

[nelson]

You certainly could go that route, but I was envisioning a different approach.

Perhaps a little more detail on the signal conditioning:
Each "narrow band" would be a separate fundamental extractor, "tuned" to the range covered by its pass-band. 
You would need at least 3 to 5 bands per octave, maybe 4 octaves for guitar, for a total of say 20 bands.

-I'm suggesting you would sweep through these bands very fast, a full cycle of all bands within the space of few microseconds or so.  At this speed, you wouldn't hear any pattern or rhythm.

There is the issue of lower notes taking longer to process... I have to think about this some more.

You're basically talking about sampling different fundamental extractors one at a time and feeding them to a single pitch shifting circuit block.

I don't think this would work. You'd still need to somehow present a simple fundamental to the pitch shifter circuit. With your idea you'd have a pretty strange waveform made up of parts of the other fundamentals stitched together. That might be a unique effect in and of itself though.

At least, that's what I think would happen, I could be wrong.

[earthtonesaudio]

Nelson,

You're right.  Dang!