Supervised nonnegative matrix factorization using active-period-aware structured l1-norm for music transcription

Yu Morikawa; Masahiro Yukawa; Hisakazu Kikuchi

doi:10.1109/APSIPA.2015.7415510

Supervised nonnegative matrix factorization using active-period-aware structured l1-norm for music transcription

Yu Morikawa, Masahiro Yukawa, Hisakazu Kikuchi

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

Abstract

An active-period-aware supervised nonnegative matrix factorization (NMF) approach for music transcription is proposed. Supervised NMF relies on a set of known spectrograms associated with all musical instruments that may possibly be involved with given music data; this is supported by the availability of large database of a variety of musical instruments. It is free from the source-number determination problem and this is a significant advantage over the unsupervised NMF approaches. The proposed approach is composed of three steps. Step 1: Apply the existing supervised NMF algorithm. Step 2: Estimate the 'active' periods (during which musical sounds are present) based on the outcomes of Step 1. Step 3: Optimize a refined cost function reflecting the estimate of active periods. The awareness of active periods leads to avoidance of the so-called octave-errors which is a central issue of the existing supervised NMF method. Simulation results show the efficacy of the proposed approach.1

Original language	English
Title of host publication	2015 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2015
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	14-18
Number of pages	5
ISBN (Electronic)	9789881476807
DOIs	https://doi.org/10.1109/APSIPA.2015.7415510
Publication status	Published - 2016 Feb 19
Event	2015 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2015 - Hong Kong, Hong Kong Duration: 2015 Dec 16 → 2015 Dec 19

Other

Other	2015 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2015
Country	Hong Kong
City	Hong Kong
Period	15/12/16 → 15/12/19

ASJC Scopus subject areas

Artificial Intelligence
Modelling and Simulation
Signal Processing

Access to Document

10.1109/APSIPA.2015.7415510

Fingerprint Dive into the research topics of 'Supervised nonnegative matrix factorization using active-period-aware structured l1-norm for music transcription'. Together they form a unique fingerprint.

Cite this

Morikawa, Y., Yukawa, M., & Kikuchi, H. (2016). Supervised nonnegative matrix factorization using active-period-aware structured l1-norm for music transcription. In 2015 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2015 (pp. 14-18). [7415510] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/APSIPA.2015.7415510

Supervised nonnegative matrix factorization using active-period-aware structured l1-norm for music transcription. / Morikawa, Yu; Yukawa, Masahiro; Kikuchi, Hisakazu.

2015 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2015. Institute of Electrical and Electronics Engineers Inc., 2016. p. 14-18 7415510.

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

Morikawa, Y, Yukawa, M & Kikuchi, H 2016, Supervised nonnegative matrix factorization using active-period-aware structured l1-norm for music transcription. in 2015 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2015., 7415510, Institute of Electrical and Electronics Engineers Inc., pp. 14-18, 2015 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2015, Hong Kong, Hong Kong, 15/12/16. https://doi.org/10.1109/APSIPA.2015.7415510

Morikawa Y, Yukawa M, Kikuchi H. Supervised nonnegative matrix factorization using active-period-aware structured l1-norm for music transcription. In 2015 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2015. Institute of Electrical and Electronics Engineers Inc. 2016. p. 14-18. 7415510 https://doi.org/10.1109/APSIPA.2015.7415510

@inproceedings{edd5076e92194ac2abe60c2106ede605,

title = "Supervised nonnegative matrix factorization using active-period-aware structured l1-norm for music transcription",

abstract = "An active-period-aware supervised nonnegative matrix factorization (NMF) approach for music transcription is proposed. Supervised NMF relies on a set of known spectrograms associated with all musical instruments that may possibly be involved with given music data; this is supported by the availability of large database of a variety of musical instruments. It is free from the source-number determination problem and this is a significant advantage over the unsupervised NMF approaches. The proposed approach is composed of three steps. Step 1: Apply the existing supervised NMF algorithm. Step 2: Estimate the 'active' periods (during which musical sounds are present) based on the outcomes of Step 1. Step 3: Optimize a refined cost function reflecting the estimate of active periods. The awareness of active periods leads to avoidance of the so-called octave-errors which is a central issue of the existing supervised NMF method. Simulation results show the efficacy of the proposed approach.1",

author = "Yu Morikawa and Masahiro Yukawa and Hisakazu Kikuchi",

year = "2016",

month = feb,

day = "19",

doi = "10.1109/APSIPA.2015.7415510",

language = "English",

pages = "14--18",

booktitle = "2015 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2015",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

note = "2015 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2015 ; Conference date: 16-12-2015 Through 19-12-2015",

}

TY - GEN

T1 - Supervised nonnegative matrix factorization using active-period-aware structured l1-norm for music transcription

AU - Morikawa, Yu

AU - Yukawa, Masahiro

AU - Kikuchi, Hisakazu

PY - 2016/2/19

Y1 - 2016/2/19

N2 - An active-period-aware supervised nonnegative matrix factorization (NMF) approach for music transcription is proposed. Supervised NMF relies on a set of known spectrograms associated with all musical instruments that may possibly be involved with given music data; this is supported by the availability of large database of a variety of musical instruments. It is free from the source-number determination problem and this is a significant advantage over the unsupervised NMF approaches. The proposed approach is composed of three steps. Step 1: Apply the existing supervised NMF algorithm. Step 2: Estimate the 'active' periods (during which musical sounds are present) based on the outcomes of Step 1. Step 3: Optimize a refined cost function reflecting the estimate of active periods. The awareness of active periods leads to avoidance of the so-called octave-errors which is a central issue of the existing supervised NMF method. Simulation results show the efficacy of the proposed approach.1

AB - An active-period-aware supervised nonnegative matrix factorization (NMF) approach for music transcription is proposed. Supervised NMF relies on a set of known spectrograms associated with all musical instruments that may possibly be involved with given music data; this is supported by the availability of large database of a variety of musical instruments. It is free from the source-number determination problem and this is a significant advantage over the unsupervised NMF approaches. The proposed approach is composed of three steps. Step 1: Apply the existing supervised NMF algorithm. Step 2: Estimate the 'active' periods (during which musical sounds are present) based on the outcomes of Step 1. Step 3: Optimize a refined cost function reflecting the estimate of active periods. The awareness of active periods leads to avoidance of the so-called octave-errors which is a central issue of the existing supervised NMF method. Simulation results show the efficacy of the proposed approach.1

UR - http://www.scopus.com/inward/record.url?scp=84986208159&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84986208159&partnerID=8YFLogxK

U2 - 10.1109/APSIPA.2015.7415510

DO - 10.1109/APSIPA.2015.7415510

M3 - Conference contribution

AN - SCOPUS:84986208159

SP - 14

EP - 18

BT - 2015 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2015

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2015 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2015

Y2 - 16 December 2015 through 19 December 2015

ER -

Supervised nonnegative matrix factorization using active-period-aware structured l1-norm for music transcription

Abstract

Other

ASJC Scopus subject areas

Access to Document

Other files and links

Cite this