Hello everyone,

I'm actually programming a vocal sampler that can record vocals, slice them by pitch and volume and afterwards trigger and re-pitch the slices using MIDI note events.
To re-pitch them according to MIDI note numbers in the right way, I firstly need to make the sung vocals extremely even in tonal pitch.
That means, I need to analyze the pitch over time and afterwards de-pitch it to note C5.
I know how to analyze it, but I don't find a way to de-pitch it millisecond by millisecond using a kind of FFT-Pitchshifter or something.

Do you guys have any advices?

Hi,

If I understood you correctly, I would try using [fiddle~] or [sigmund~] to analyse get pitch and attack information and [pvoc~] to shift it according to some math expression I'd write.

Take care,
Gilberto

Ok, thanks! What about artifacts? Are there any?

Hi Elden, it's hard to tell, I have used [pvoc~] very seldom and only within a very small range, but the results were pretty much okay. The best thing would be to give it a try and see if it suits you or not.

Take care,
Gilberto

you could also try [helmhotz~] to get pitch and [soundtouch~] to shift the pitch, both externals by Katja Vetter. I think you can find binaries of these on Pd's website.

You might use a vocoder, and feed it monotonic pitch data.