A discriminative approach to polyphonic piano note transcription using supervised non-negative matrix factorization

Felix Weninger; Christian Kirst; Bjorn Schuller; Hans Joachim Bungartz

doi:10.1109/ICASSP.2013.6637598

A discriminative approach to polyphonic piano note transcription using supervised non-negative matrix factorization

Felix Weninger, Christian Kirst, Bjorn Schuller, Hans Joachim Bungartz

Technical University of Munich

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

19 Scopus citations

Abstract

We introduce a novel method for the transcription of polyphonic piano music by discriminative training of support vector machines (SVMs). As features, we use pitch activations computed by supervised non-negative matrix factorization from low-level spectral features. Different approaches to low-level feature extraction, NMF dictionary learning and activation feature extraction are analyzed in a large-scale evaluation on eight hours of piano music including synthesized and real recordings. We conclude that the proposed method delivers state-of-the-art results and clearly outperforms SVMs using simple spectral features.

Original language	English
Title of host publication	2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings
Pages	6-10
Number of pages	5
DOIs	https://doi.org/10.1109/ICASSP.2013.6637598
State	Published - 18 Oct 2013
Event	2013 38th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Vancouver, BC, Canada Duration: 26 May 2013 → 31 May 2013

Publication series

Name	ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
ISSN (Print)	1520-6149

Conference

Conference	2013 38th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013
Country/Territory	Canada
City	Vancouver, BC
Period	26/05/13 → 31/05/13

Keywords

Transcription
music information retrieval
non-negative matrix factorization
sparse coding

Access to Document

10.1109/ICASSP.2013.6637598

Cite this

Weninger, F., Kirst, C., Schuller, B., & Bungartz, H. J. (2013). A discriminative approach to polyphonic piano note transcription using supervised non-negative matrix factorization. In 2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings (pp. 6-10). Article 6637598 (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings). https://doi.org/10.1109/ICASSP.2013.6637598

Weninger, Felix ; Kirst, Christian ; Schuller, Bjorn et al. / A discriminative approach to polyphonic piano note transcription using supervised non-negative matrix factorization. 2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings. 2013. pp. 6-10 (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings).

@inproceedings{a2a8eccaf64f40d29d501d60809c5371,

title = "A discriminative approach to polyphonic piano note transcription using supervised non-negative matrix factorization",

abstract = "We introduce a novel method for the transcription of polyphonic piano music by discriminative training of support vector machines (SVMs). As features, we use pitch activations computed by supervised non-negative matrix factorization from low-level spectral features. Different approaches to low-level feature extraction, NMF dictionary learning and activation feature extraction are analyzed in a large-scale evaluation on eight hours of piano music including synthesized and real recordings. We conclude that the proposed method delivers state-of-the-art results and clearly outperforms SVMs using simple spectral features.",

keywords = "Transcription, music information retrieval, non-negative matrix factorization, sparse coding",

author = "Felix Weninger and Christian Kirst and Bjorn Schuller and Bungartz, {Hans Joachim}",

year = "2013",

month = oct,

day = "18",

doi = "10.1109/ICASSP.2013.6637598",

language = "English",

isbn = "9781479903566",

series = "ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings",

pages = "6--10",

booktitle = "2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings",

note = "2013 38th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 ; Conference date: 26-05-2013 Through 31-05-2013",

}

Weninger, F, Kirst, C, Schuller, B & Bungartz, HJ 2013, A discriminative approach to polyphonic piano note transcription using supervised non-negative matrix factorization. in 2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings., 6637598, ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings, pp. 6-10, 2013 38th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013, Vancouver, BC, Canada, 26/05/13. https://doi.org/10.1109/ICASSP.2013.6637598

A discriminative approach to polyphonic piano note transcription using supervised non-negative matrix factorization. / Weninger, Felix; Kirst, Christian; Schuller, Bjorn et al.
2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings. 2013. p. 6-10 6637598 (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings).

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

TY - GEN

T1 - A discriminative approach to polyphonic piano note transcription using supervised non-negative matrix factorization

AU - Weninger, Felix

AU - Kirst, Christian

AU - Schuller, Bjorn

AU - Bungartz, Hans Joachim

PY - 2013/10/18

Y1 - 2013/10/18

N2 - We introduce a novel method for the transcription of polyphonic piano music by discriminative training of support vector machines (SVMs). As features, we use pitch activations computed by supervised non-negative matrix factorization from low-level spectral features. Different approaches to low-level feature extraction, NMF dictionary learning and activation feature extraction are analyzed in a large-scale evaluation on eight hours of piano music including synthesized and real recordings. We conclude that the proposed method delivers state-of-the-art results and clearly outperforms SVMs using simple spectral features.

AB - We introduce a novel method for the transcription of polyphonic piano music by discriminative training of support vector machines (SVMs). As features, we use pitch activations computed by supervised non-negative matrix factorization from low-level spectral features. Different approaches to low-level feature extraction, NMF dictionary learning and activation feature extraction are analyzed in a large-scale evaluation on eight hours of piano music including synthesized and real recordings. We conclude that the proposed method delivers state-of-the-art results and clearly outperforms SVMs using simple spectral features.

KW - Transcription

KW - music information retrieval

KW - non-negative matrix factorization

KW - sparse coding

UR - http://www.scopus.com/inward/record.url?scp=84890485042&partnerID=8YFLogxK

U2 - 10.1109/ICASSP.2013.6637598

DO - 10.1109/ICASSP.2013.6637598

M3 - Conference contribution

AN - SCOPUS:84890485042

SN - 9781479903566

T3 - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings

SP - 6

EP - 10

BT - 2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings

T2 - 2013 38th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013

Y2 - 26 May 2013 through 31 May 2013

ER -

Weninger F, Kirst C, Schuller B , Bungartz HJ. A discriminative approach to polyphonic piano note transcription using supervised non-negative matrix factorization. In 2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings. 2013. p. 6-10. 6637598. (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings). doi: 10.1109/ICASSP.2013.6637598

A discriminative approach to polyphonic piano note transcription using supervised non-negative matrix factorization

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this