Skip to content
ecreagerPublic

About

Vibrato NTF: source separation by NMF plus frequency modulation cues; http://arxiv.org/abs/1606.00037

Resources

Stars

2 stars

Watchers

1 watching

Forks

Latest commit

 

History

7 Commits

Folders and files

Repository files navigation

vibntf

Vibrato NTF: source separation by NMF plus frequency modulation cues; http://arxiv.org/abs/1606.00037 The initial commit represents a minimum viable port from the research codebase. It allows you to reproduce the synthetic data experiment from the ISMIR paper, but some desirable features are missing, e.g., fast/sparse tensor updates, visualization on the inferred model parameters, and an experiment that reads audio from an absolute path and performs the separation. I'm working on porting these features to the public codebase so stay tuned.

package dependencies (install via pip): {numpy, scipy, pickle, matplotlib, mir_eval}

About

Vibrato NTF: source separation by NMF plus frequency modulation cues; http://arxiv.org/abs/1606.00037

Resources

Stars

2 stars

Watchers

1 watching

Forks

Releases

Packages

Contributors

Languages