# facebookresearch/demucs

**Attribution required: if you use, quote, or summarise this content, you must credit and link back to [awesome-repositories.com](https://awesome-repositories.com/repository/facebookresearch-demucs).**

10,236 stars · 1,521 forks · Python · MIT · archived

## Links

- GitHub: https://github.com/facebookresearch/demucs
- awesome-repositories: https://awesome-repositories.com/repository/facebookresearch-demucs.md

## Description

Demucs is a deep learning stem splitter and AI music de-mixing software used to isolate vocals and instruments from a single audio file. It functions as a PyTorch audio source separation tool that splits mixed tracks into individual stems such as drums, bass, and vocals.

The system is a hybrid spectrogram waveform separator that combines spectral and waveform analysis. This approach allows the software to process audio in both frequency and time domains to achieve high-fidelity source separation.

The tool provides capabilities for audio source separation, including acapella track extraction and backing track creation. It supports music production workflows by allowing users to extract single stems and convert audio formats to meet specific playback requirements.

## Tags

### Artificial Intelligence & ML

- [Source Separation Tools](https://awesome-repositories.com/f/artificial-intelligence-ml/audio-source-separation-models/source-separation-tools.md) — Provides software tools for splitting mixed audio recordings into individual instrument and vocal tracks. ([source](https://cdn.jsdelivr.net/gh/facebookresearch/demucs@main/README.md))
- [Audio Source Separation Models](https://awesome-repositories.com/f/artificial-intelligence-ml/audio-source-separation-models.md) — Uses deep learning models to decompose mixed music tracks into individual stems like drums and vocals.
- [Hybrid Spectral-Waveform Separators](https://awesome-repositories.com/f/artificial-intelligence-ml/audio-source-separation-models/hybrid-spectral-waveform-separators.md) — Combines spectral and waveform analysis for high-fidelity source separation.
- [Convolutional Neural Networks](https://awesome-repositories.com/f/artificial-intelligence-ml/convolutional-neural-networks.md) — Implements convolutional neural networks to identify and isolate spatial patterns within audio spectrograms.
- [U-Net Architectures](https://awesome-repositories.com/f/artificial-intelligence-ml/neural-network-architectures/u-net-architectures.md) — Employs a U-Net architecture to restore high-resolution audio details using encoder skip-connections.

### Graphics & Multimedia

- [Audio De-mixing Software](https://awesome-repositories.com/f/graphics-multimedia/audio-de-mixing-software.md) — Automates the creation of acapellas and backing tracks by removing specific instruments from a mix.
- [Hybrid Domain Audio Processing](https://awesome-repositories.com/f/graphics-multimedia/hybrid-domain-audio-processing.md) — Combines spectral and waveform analysis to capture both fine-grained transients and overall spectral shape.
- [Audio Stem Extractors](https://awesome-repositories.com/f/graphics-multimedia/media-processing-analysis/media-manipulation/media-processing-workflows/audio-analysis-synthesis/audio-feature-extraction/audio-track-extraction/audio-stem-extractors.md) — Isolates a single audio source from a mixed recording to create standalone acapellas or backing tracks. ([source](https://cdn.jsdelivr.net/gh/facebookresearch/demucs@main/README.md))
- [Vocal Isolation](https://awesome-repositories.com/f/graphics-multimedia/media-processing-analysis/media-manipulation/media-processing-workflows/audio-analysis-synthesis/audio-feature-extraction/audio-track-extraction/vocal-isolation.md) — Removes all instruments from a song to isolate the vocal performance for remixes or sampling.
- [Vocal-to-Instrumental Converters](https://awesome-repositories.com/f/graphics-multimedia/vocal-artifact-removal/vocal-to-instrumental-converters.md) — Creates instrumental versions of songs by removing the vocal tracks.
- [Time-Frequency Mapping](https://awesome-repositories.com/f/graphics-multimedia/spectral-signal-analysis/time-frequency-mapping.md) — Uses Short-Time Fourier Transforms to transform raw waveforms into time-frequency representations for efficient separation.
- [Time-Domain Signal Refinement](https://awesome-repositories.com/f/graphics-multimedia/time-domain-signal-refinement.md) — Applies final processing directly to the time-domain signal to correct artifacts introduced by spectral processing.