beta.blog

Spleeter is Deezer source separation library with pretrained models written in Python and uses Tensorflow. It makes it easy to train source separation model (assuming you have a dataset of isolated sources), and provides already trained state of the art model for performing various flavour of separation:

• Vocals (singing voice) / accompaniment separation (2 stems)
• Vocals / drums / bass / other separation (4 stems)
• Vocals / drums / bass / piano / other separation (5 stems)

Since Spleeter is Python based, we may use either pip or pip3 on macOS to install it from the command line:

pip3 install spleeter

Simply run spleeter –help in order to get an overview of the available possibilities and on how to use the tool. The following is an example of how to use the tool to seaparate the voice from the accompaniment sound:

spleeter separate -p spleeter:2stems -o output song.mp3

Depending on your hardware this process will take anywhere between a few seconds to up to a minute or two.

We may either use the 2stems model or the 4stems (or 5stems) one mentioned above. The results are actually quite impressive.