Skip to main content

Demucs Nvidia

Demucs is an music separation tool that has potential for a karaoke setup. 

https://github.com/facebookresearch/demucs

https://www.youtube.com/watch?v=9QnFMKWEFcI&t=585s

https://docs.google.com/document/d/1XMmLrz-Tct1Hdb_PatcwEeBrV9Wrt15wHB1xhkB2oiY/edit

Installation on PC with Nvidia

  1. Firstly install Anaconda. Download Anaconda for Windows https://www.anaconda.com/products/distribution 
  2. Install PyTorch. https://pytorch.org/get-started/locally/. Select the correct version of pytorch.
  3. Install ffmpeg. https://www.gyan.dev/ffmpeg/builds/ choose Full, and extract the ffmpeg files then add it to path.image.png

Demucs

After installing the prerequesties.

Open "Anaconda terminal" and type 

python.exe -m pip install -U demucs
pip install PySoundFile 

Running Demucs

demucs "C:\path\to\music\file.mp3"

This will run demucs with CUDA GPU acceleration, make sure to put the path in double quote. The extracted file will be found in the directory where you run the command eg. the default Anaconda prompt starts in ~/separated