Just came across this “black magic” processor by Jan Brähler:
https://www.native-instruments.com/es/reaktor-community/reaktor-user-library/entry/show/10267/
the FFT pitchshift implementations i have seen in SC are either bin-shifting (which sounds more like frequency shifting, and not “musically” useful) or just work with monophonic sources after some process of pitch-tracking.
It seems that Mr. Brähler here has found a nice way to pitch-shift polyphonic sources using FFT sourcery (hear the demo on the link). He mentions:
The core of it is a new fft implementation (still a modded ezfft) for higher overlaps, changable framesizes and low latency. (cost: heavy cpu spikes)
But, trying to look at the ensemble structure, i need to admit i dont understand what is happening at all.
Has anyone in the SC community came up with a comparable solution to pitchshifting with FFT. Or does anyone know how to decrypt mr Brähler’s algorithm and try to implement something similar in SC?