Resynthesis using wavelet transform - why is this almost impossible to find?

2022-10-30 13:29:48

Now THIS is something. If such ideas were floating around as early as in the late 80's, why is it that we don't see an option of wavelet resynthesis in current sound editors? It's not just things like noise reduction. It's also things like time scaling, pitch shifting, vocal formant preservation when altering a singer's melody, spectral morphing, analyzing sound timbres, converting sound to graphics and back again, and many more possibilities.
Why is it that one can't easily find sound examples of such things on the web?
Why is it that everything seems to be realized with Fourier transforms and that I'm out of luck if I want to hear the difference between Fourier resynthesis and wavelet resynthesis or Fourier time scaling and wavelet time scaling?
What am I supposed to do if I'm not a programmer and I can't just take a PDF and turn the presented ideas into audible sound?
https://www.youtube.com/watch?v=Aqpf5prJjEo

Re: Resynthesis using wavelet transform - why is this almost impossible to find?

Reply #1 – 2022-10-30 13:49:50

I have wondered about this myself.  I assume it's because the existing codec (etc) methods are so well established.

I have postulated that a wavelet transformation codec would make it possible to tempo- or pitch-shift by a simple manipulation of the wavelet components in the compressed data stream, instead of fancy subsequent processing of the decoded audio data.