LinuxMusicians

I experimented a bit more.

The creators supposed, that someone who sings couldn't be as tonedeaf to sing one octave off.

In an actual speech the spectrum (or spectra in english?) can be much wider than this, that's make the glithces in the output.

IMHO

Thanks for all your help. I managed to make it work. As I see, this zita really would do what I seek for. But not very well at the moment. Maybe becouse of this: "AT1 does not include formant correction, so it should be used to correct small errors only and not to really transpose a song. "...

Hello, thanks for the reply. It needs libzita resampler... Could you put me in direction? I've found the video that you linked, but it is the method I tried, and don't liked. The input voice and carrier is mixed togather somehow. Makes it too robotic. My goal is something like the Gregory's work: ht...

Hello everyone! Autotalent is able to snap the input sound to the closest note, or to a scale, if it is adjusted. VocProc can do something similar. But as I noticed, it somehow modulates togather the two signals. The output tone depends on the carrier's tone as well. I'd like to make sing peoples wh...