I had a "dead simple" idea while playing today. Many people have very low quality (or broken) mikrophones and/or at least one noisy fan in their computers'. I though "hey, maybe thats a distinguishable frequency!" and hey: most of the time it is! (check attached image)

What I want to do:
  1. Hook the audio stream of one player (not all of them because of performance)
  2. Transform (Fast-Fourier)
  3. Do some frequency analysis (segmentation + classification)
  4. Apply some filtering
  5. Transform back (Fast-Fourier again)
  6. Listen to clean and crisp audio, that does not suck


I'm pretty confident, that I can make this work in realtime, even it's 44k samples/s. The thing, that bites me most is "can i modify the audio-stream of a single player?" and put it back into TeamSpeak in a clean way by creating a Plug-In? The other option I can think of ain't legal...

Cheers,
Chris