 |  |
Independent Component Analysis and Signal Separation, International Conference on
Abstract: We present a novel method for blind separation of instruments in polyphonic music based on a non-negative matrix factor 2-D deconvolution algorithm. Using a model which is convolutive in both time and frequency we factorize a spectrogram representation of music into components corresponding to individual instruments. Based on this factorization we separate the instruments using spectrogram masking. The proposed algorithm has applications in computational auditory scene analysis, music information retrieval, and automatic music transcription.
Demonstration:
| Mixed Signal | Separated Instruments | |
|---|---|---|
| Guitar + Flute | Guitar | Flute |
|
|
|
Talk: The slides included below were used for my talk at the ICA conference, and includes an audio demonstration.
- See also:
Sparse Non-negative Matrix Factor 2-D Deconvolution- Sparse Non-negative Matrix Factor 2-D Deconvolution for Automatic Transcription of Polyphonic Music
- Nonnegative Matrix Factor 2-D Deconvolution for Blind Single Channel Source Separation
- Files:
imm4061.pdf
nmf2d_ica.pps
- Cite:
- Mikkel N. Schmidt and Morten Mørup, Nonnegative Matrix Factor 2-D Deconvolution for Blind Single Channel Source Separation, Independent Component Analysis and Signal Separation, International Conference on, 2006
- BibTeX:
- @inproceedings{schmidt06nmf2d,
title = "Nonnegative Matrix Factor {2-D} Deconvolution for Blind Single Channel Source Separation",
author = "Mikkel N. Schmidt and Morten Mørup",
booktitle = "Independent Component Analysis and Signal Separation, International Conference on",
month = "apr",
pages = "700--707",
publisher = "Springer",
series = "Lecture Notes in Computer Science (LNCS)",
volume = "3889",
year = "2006"
}
Mikkel N. Schmidt | Technical University of Denmark | Email: mns(a)imm.dtu.dk