Guzdial seems to hint at applications to audio compression, but my mind turned to speech synthesis and recognition. I'd like to maximize a handful of audio clip

Guzdial seems to hint at applications to audio compression, but my mind turned to speech synthesis and recognition. I'd like to maximize a handful of audio clips of human voice and look for patterns.