Upgraded to v1.0!
MegaTTS 3 but with voice cloning!
Text-to-music generation with melody conditioning
Audio production style transfer (EQ, compression)
Separate music into vocal and instrumental stems
LA-2A compressor/limiter emulation with a neural TCN
Synthesize guitar audio from MIDI using DDSP
Synthesize piano audio from MIDI with 10 learned timbres
Generate controlled error responses for testing HARP handling
Meta AudioSeal Watermarking
Extract timestamped chords from music on CPU
Compact English text-to-speech on CPU