https://github.com/vetrenost/music_separator

Script for music stem separation using Demucs models (htdemucs_ft, mdx_extra_q).
https://github.com/vetrenost/music_separator

audio-separation demucs htdemucs mdx music-processing source-separation

Last synced: about 1 month ago
JSON representation

Script for music stem separation using Demucs models (htdemucs_ft, mdx_extra_q).

• Host: GitHub
• URL: https://github.com/vetrenost/music_separator
• Owner: Vetrenost
• Created: 2025-06-05T13:46:24.000Z (about 1 year ago)
• Default Branch: main
• Last Pushed: 2025-08-23T08:53:56.000Z (10 months ago)
• Last Synced: 2025-08-24T03:14:57.860Z (10 months ago)
• Topics: audio-separation, demucs, htdemucs, mdx, music-processing, source-separation
• Language: Python
• Homepage:
• Size: 84 KB
• Stars: 0
• Watchers: 0
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
# Minusator — Vocal Remover

**Minusator** is a simple Python script that separates vocals from a music track, producing an instrumental version.

---

## Requirements

- Python 3.12+

- [torchaudio](https://pypi.org/project/torchaudio/)

- **demucs.exe** (keep in the same folder as the script)

- **ffmpeg** — download [ffmpeg-release-essentials](https://www.gyan.dev/ffmpeg/builds/) and extract into the script folder

---

## Setup

1. Place the `main.py` script in a convenient folder.

2. Make sure `demucs.exe` and `ffmpeg-7.1.1-essentials_build` are in the same folder.

3. Do **not** change the model in the code.

---

## Usage

1. Open the program and click **Select File**.

2. Choose an audio file (`.mp3`, `.wav`, `.flac`).

3. Wait for processing (may take over a minute).

4. The results will appear in `separated/htdemucs//` next to the script:  

   - `vocals.wav` — vocals  

   - `no_vocals.wav` — instrumental

---

## Notes

- Works best with songs containing clear vocal tracks.

- Processing time depends on the file size and your system performance.