← Glossary

Demucs

An open-source music source separation model developed by Meta's research team, widely used to split songs into vocals, drums, bass, and other instruments.

Demucs is an open-source music source separation model developed by Meta's research team. It separates mixed recordings into vocals, drums, bass, and “other” stems. Its code is released under the MIT license, but certain pretrained versions, including HT Demucs, were trained partly on MUSDB18-HQ, a dataset restricted to educational and noncommercial use. The project's author has stated that pretrained weights are provided for scientific purposes only. Code licensing and model-weight usage rights should be evaluated separately.

Links: Demucs on GitHub · Compare with AudioShake · Can Demucs be used commercially?

Where this comes up at AudioShake
No items found.