3 ms·

Given a large enough neural network, I'm quite confident music could be compressed much - MUCH - more than traditional methods. https://arstechnica.com/informa

• fsiefken 3y ago
  1.5 kbps with Vocos and Meta's EnCodec, voice samples here, but I wonder how music would sound. https://gemelo-ai.github.io/vocos/ https://gemelo-ai.github.io/vocos/