Writing down an idea at 3 AM so that I don't forget it ...

… and it’s kind of a cool idea other people might try:

Get a sample of speech. I like recording shortwave radio stations where I don’t know the language being spoken.

Get a sampler playing it, and drive an envelope follower.

Envelope follower gate clocks your favorite sequencer. Fill it with some sequences & snapshot them with 8Face.

Sequencer drives a synth voice. Use that as the carrier for a vocoder, and use the speech sample as the modulator.


:thinking: That sounds interesting.

It’s kinda’ the opposite of when I made a recording of macOS synthesized speech - in various accents - to use in a sampler and then “distress” it, to sound like one of those mysterious shortwave “numbers stations”.

I’ve done this kind of thing through a really long reverb impulse response. See if I can find a link to the one I used.

edit to add: Found it.

Teufelsberg NSA listening station impulse responses: Link removed as it no longer works.

I’ve had a similar 3am idea. To make upside-down audio. Use any audio sample as the modulator for a vocoder, and white noise as the carrier. But invert all the frequency bands. It did not work . . .

Not exactly the topic, but here is a recent AI and ML code synthesizing and making “music” based on midi files and Audio. Don’t ask me how it works exactly, i am still trying to understand it myself

sort of polyglot Peter Frampton for modular

simpliciter contains an envelope follower CV out