edge-tts
audiosegment
scipy
librosa