BLUES from Music: BLind Underdetermined Extraction of Sources from Music

Michael Syskind Pedersen; Tue Lehn-Schiøler; Jan Larsen

BLUES from Music: BLind Underdetermined Extraction of Sources from Music

Michael Syskind Pedersen, Tue Lehn-Schiøler, Jan Larsen

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

205 Downloads (Pure)

Abstract

In this paper we propose to use an instantaneous ICA method (BLUES) to separate the instruments in a real music stereo recording. We combine two strong separation techniques to segregate instruments from a mixture: ICA and binary time-frequency masking. By combining the methods, we are able to make use of the fact that the sources are differently distributed in both space, time and frequency. Our method is able to segregate an arbitrary number of instruments and the segregated sources are maintained as stereo signals. We have evaluated our method on real stereo recordings, and we can segregate those instruments which are spatially different from other instruments.

Original language	English
Title of host publication	ICA2006
Volume	3889
Publisher	Springer Berlin / Heidelberg
Publication date	2006
Pages	392-399
Publication status	Published - 2006
Event	6th International Conference on Independent Component Analysis and Blind Signal Separation - Charleston, United States Duration: 5 Mar 2006 → 8 Mar 2006

Conference

Conference	6th International Conference on Independent Component Analysis and Blind Signal Separation
Country/Territory	United States
City	Charleston
Period	05/03/2006 → 08/03/2006

Series	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
ISSN	0302-9743

Access to Document

Imm4060

http://www.springerlink.com/(m33ypy45gn3y4p45voyskg55)/app/home/contribution.asp?referrer=parent&backto=issue,49,120;journal,10,3313;linkingpublicationresults,1:105633,1

OpenUrl availability

Full text

Cite this

Pedersen, M. S., Lehn-Schiøler, T., & Larsen, J. (2006). BLUES from Music: BLind Underdetermined Extraction of Sources from Music. In ICA2006 (Vol. 3889, pp. 392-399). Springer Berlin / Heidelberg. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) http://www.springerlink.com/(m33ypy45gn3y4p45voyskg55)/app/home/contribution.asp?referrer=parent&backto=issue,49,120;journal,10,3313;linkingpublicationresults,1:105633,1

@inproceedings{303e7632874d4b08b13f2f48bc79f739,

title = "BLUES from Music: BLind Underdetermined Extraction of Sources from Music",

abstract = "In this paper we propose to use an instantaneous ICA method (BLUES) to separate the instruments in a real music stereo recording. We combine two strong separation techniques to segregate instruments from a mixture: ICA and binary time-frequency masking. By combining the methods, we are able to make use of the fact that the sources are differently distributed in both space, time and frequency. Our method is able to segregate an arbitrary number of instruments and the segregated sources are maintained as stereo signals. We have evaluated our method on real stereo recordings, and we can segregate those instruments which are spatially different from other instruments.",

keywords = "music, ICA, time frequency masking",

author = "Pedersen, {Michael Syskind} and Tue Lehn-Schi{\o}ler and Jan Larsen",

year = "2006",

language = "English",

volume = "3889",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Berlin / Heidelberg",

pages = "392--399",

booktitle = "ICA2006",

note = "6th International Conference on Independent Component Analysis and Blind Signal Separation, ICA 2006 ; Conference date: 05-03-2006 Through 08-03-2006",

}

Pedersen, MS, Lehn-Schiøler, T & Larsen, J 2006, BLUES from Music: BLind Underdetermined Extraction of Sources from Music. in ICA2006. vol. 3889, Springer Berlin / Heidelberg, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 392-399, 6th International Conference on Independent Component Analysis and Blind Signal Separation, Charleston, United States, 05/03/2006. <http://www.springerlink.com/(m33ypy45gn3y4p45voyskg55)/app/home/contribution.asp?referrer=parent&backto=issue,49,120;journal,10,3313;linkingpublicationresults,1:105633,1>

BLUES from Music: BLind Underdetermined Extraction of Sources from Music. / Pedersen, Michael Syskind; Lehn-Schiøler, Tue; Larsen, Jan.
ICA2006. Vol. 3889 Springer Berlin / Heidelberg, 2006. p. 392-399 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

TY - GEN

T1 - BLUES from Music: BLind Underdetermined Extraction of Sources from Music

AU - Pedersen, Michael Syskind

AU - Lehn-Schiøler, Tue

AU - Larsen, Jan

PY - 2006

Y1 - 2006

N2 - In this paper we propose to use an instantaneous ICA method (BLUES) to separate the instruments in a real music stereo recording. We combine two strong separation techniques to segregate instruments from a mixture: ICA and binary time-frequency masking. By combining the methods, we are able to make use of the fact that the sources are differently distributed in both space, time and frequency. Our method is able to segregate an arbitrary number of instruments and the segregated sources are maintained as stereo signals. We have evaluated our method on real stereo recordings, and we can segregate those instruments which are spatially different from other instruments.

AB - In this paper we propose to use an instantaneous ICA method (BLUES) to separate the instruments in a real music stereo recording. We combine two strong separation techniques to segregate instruments from a mixture: ICA and binary time-frequency masking. By combining the methods, we are able to make use of the fact that the sources are differently distributed in both space, time and frequency. Our method is able to segregate an arbitrary number of instruments and the segregated sources are maintained as stereo signals. We have evaluated our method on real stereo recordings, and we can segregate those instruments which are spatially different from other instruments.

KW - music

KW - ICA

KW - time frequency masking

M3 - Article in proceedings

VL - 3889

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 392

EP - 399

BT - ICA2006

PB - Springer Berlin / Heidelberg

T2 - 6th International Conference on Independent Component Analysis and Blind Signal Separation

Y2 - 5 March 2006 through 8 March 2006

ER -