Blind demixing using Non-negative Matrix Factorization
This is a commandline program that lets you decompose a sound into additive components in the spectral domain.
what it does
- generate the sound’s spectrogram
- decompose the spectrogram into N layers
- resynthesize each layer and the residual to a separate soundfile
usage:
nmfdemix [options] filename #components
-o , --overlap _overlap factor (default 4)_
-c , --criteria _stopping criteria (default 0.001)_
-i , --maxiterations _max number of iterations (default 1000)_
-f , --fftsize _fftsize (default 4096)_
-m , --matrixout _output W and H matrices_
-s , --silent _don't be verbose_
-h , print help