Techniques for manipulating (spoken) voice samples?


I see lots of info and examples on working with samples of a pitched nature, and a little bit on drums. What kinds of interesting things have been done with voice clips? I want to get some voice lines from some shows/movies and work them in to some grooves. I used to have a Morphagene and I sometimes got cool sounding stuff with it, but it was really hard to get all the CV adjusted just right and have the samples come out rhythmically.


You might want to give the Clocked Stretch player a try. You can clock the timing and retain independent control over pitch.

Slicing samples and slowing them down is always fun. Especially putting them into delay. For some reason german accents are really good. I had some fun with some Werner Herzog samples recently. Try using really short delay times and high feedback for weird metallic robot voices.

Ring modulation (signal through a bipolar vca with audio rate oscillator modulating the vca) can get some classic sci-fi sound effects, like daleks or star wars communicators.

I have been wanting to implement AM demodulation in the ER-301 to get those am radio type tones, but haven’t gotten to it yet.


What is AM demodulation?

It’s the method of decoding am radio waves. In our situation we couldn’t use radio frequencies but it still works. Make noise mod demod does it and they have a good video about it.

Basically you AM a signal with a higher frequency one, then use that same modulating wave to demodulate it. I’m not sure exactly how it works, but you can get some cool sounds.


Interesting. Thinking the envelope follower might work here. Wonder if it would track fast enough.

I’m wondering if you could invert the modulation signal and modulate the amplitude. Sort of like phase cancellation but just for the modulator.

Next let’s figure out how to demodulate fm signals. See what a dx7 sounds like when you demodulate the frequency. Sounds difficult.

That does sound tough. I remember studying it back in the day but it would take a serious refresher to remember how that works.

I rediscovered this video recently and thought it would be interesting to implement a system like this on the er301

The vocal part is around 3 mins in*

1 Like

Sadly the 301 doesn’t have big tape reels to spin up and go flying.

1 Like

What I would do is similar to the Count to 5. Basically record a voice into 3 separate loopers to make one layered voice. Each looper buffer attached to a different variable speed player + effects with random modulation for speed / position / pitch / panning / lfo speed. It would come out sounding quite weird and spooky I guess.

on non pitched sources (voice, percussions etc…) i tend to love frequency shifting.
you might want to try this custom unit i made a while ago which is my best attempt at frequency shifter:

1 Like

Awesome ideas, I think that’s the sort of thing I’m after! That with some erratic triggering thereof…

“Ring modulation (signal through a bipolar vca with audio rate oscillator modulating the vca)”

Isn’t that actually amplitude modulation? I understand that the two are closely related…

AM demod sounds interesting; I may have to see what I can figure out

I think the difference between ring mod and AM is that ring mod modulates the signal through positive and negative (inverting the signal) whereas AM just brings it to zero and back positive.