online audio sound music track separation separator karaoke split vocals bass guitar tracks machine learning free research

Music Track Separator

musicsep before musicsep after

Split a song into each of its vocal and musical instrument tracks using AI!

Rebalance without the stems! Create tracks for karaoke! Hear what you couldn't hear before!
Method partially based on the excellent work described in:

[1] Rafii, Zafar, et al. "An overview of lead and accompaniment separation in music." IEEE/ACM Transactions on Audio, Speech, and Language Processing 26.8 (2018): 1307-1335.
[2] Sigsep, Open-Unmix - Music Source Separation for PyTorch. (2021). GitHub Repository. https://github.com/sigsep/open-unmix-pytorch
[3] Meseguer-Brocal, Gabriel, and Geoffroy Peeters. "Conditioned-U-Net: Introducing a control mechanism in the U-Net for multiple source separations." arXiv preprint arXiv:1907.01277 (2019).
[4] Yuan, Weitao, et al. "A skip attention mechanism for monaural singing voice separation." IEEE Signal Processing Letters 26.10 (2019): 1481-1485.
Terms of Service
Privacy Policy