Restyle, inpaint, or continue music from a prompt
Magenta DDSP timbre transfer and synthesis
Generates MIDI from a text description
ITU-R BS.1770-4 loudness and range metering
Spectral-gating noise reduction
Emulates an LA-2A compressor with a small TCN
OpenAI Whisper speech-to-text transcription
Restores degraded speech of any kind
Edit English speech, keeping the surrounding take
Zero-shot speech and singing conversion and TTS
Unified audio generation, editing, and voice cloning
Exercises every control, track, and label type
Removes room reverb from mono speech
Transcribes solo piano performance to MIDI
SpeechBrain Tacotron2 TTS on LJSpeech
Turns rhythm and timbre prompts into drum parts
Predicts and applies a per-stem effects chain
Edit music with source and target descriptions
Zero-shot singing voice synthesis from a reference
Text-controlled music restoration and mastering