Group Delay Function from All-Pole Models for Musical Instrument Recognition

Alexander Diment; Rajan Padmanabhan; Toni Heittola; Tuomas Virtanen

doi:10.1007/978-3-319-12976-1_37

Group Delay Function from All-Pole Models for Musical Instrument Recognition

Alexander Diment, Rajan Padmanabhan, Toni Heittola, Tuomas Virtanen

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

2 Citations (Scopus)

Abstract

In this work, the feature based on the group delay function from all-pole models (APGD) is proposed for pitched musical instrument recognition. Conventionally, the spectrum-related features take into account merely the magnitude information, whereas the phase is often overlooked due to the complications related to its interpretation. However, there is often additional information concealed in the phase, which could be beneficial for recognition. The APGD is an elegant approach to inferring phase information, which lacks of the issues related to interpreting the phase and does not require extensive parameter adjustment. Having shown applicability for speech-related problems, it is now explored in terms of instrument recognition. The evaluation is performed with various instrument sets and shows noteworthy absolute accuracy gains of up to 7% compared to the baseline mel-frequency cepstral coefficients (MFCCs) case. Combined with the MFCCs and with feature selection, APGD demonstrates superiority over the baseline with all the evaluated sets.

Translated title of the contribution	Group Delay Function from All-Pole Models for Musical Instrument Recognition
Original language	English
Title of host publication	Sound, Music, and Motion, 10th International Symposium, CMMR 2013, Marseille, France, October 15-18, 2013. Revised Selected Papers
Editors	Mitsuko Aramaki, Olivier Derrien, Richard Kronland-Martinet, Solvi Ystad
Publisher	Springer International Publishing
Pages	606-618
Number of pages	18
ISBN (Electronic)	978-3-319-12976-1
ISBN (Print)	978-3-319-12975-4
DOIs	https://doi.org/10.1007/978-3-319-12976-1_37
Publication status	Published - 2014
Publication type	A4 Article in a conference publication
Event	International Symposium on Computer Music Multidisciplinary Research - Duration: 1 Jan 1900 → …

Conference

Conference	International Symposium on Computer Music Multidisciplinary Research
Period	1/01/00 → …

Keywords

Musical instrument recognition
Phase spectrum
Music information retrieval
All-pole group delay feature

Publication forum classification

Publication forum level 1

Access to Document

10.1007/978-3-319-12976-1_37

Cite this

Diment, A., Padmanabhan, R., Heittola, T., & Virtanen, T. (2014). Group Delay Function from All-Pole Models for Musical Instrument Recognition. In M. Aramaki, O. Derrien, R. Kronland-Martinet, & S. Ystad (Eds.), Sound, Music, and Motion, 10th International Symposium, CMMR 2013, Marseille, France, October 15-18, 2013. Revised Selected Papers (pp. 606-618). Springer International Publishing. https://doi.org/10.1007/978-3-319-12976-1_37

Diment, Alexander ; Padmanabhan, Rajan ; Heittola, Toni ; Virtanen, Tuomas. / Group Delay Function from All-Pole Models for Musical Instrument Recognition. Sound, Music, and Motion, 10th International Symposium, CMMR 2013, Marseille, France, October 15-18, 2013. Revised Selected Papers. editor / Mitsuko Aramaki ; Olivier Derrien ; Richard Kronland-Martinet ; Solvi Ystad. Springer International Publishing, 2014. pp. 606-618

@inproceedings{9e2d8466a02b46eab33cdb0fed8c2dcb,

title = "Group Delay Function from All-Pole Models for Musical Instrument Recognition",

abstract = "In this work, the feature based on the group delay function from all-pole models (APGD) is proposed for pitched musical instrument recognition. Conventionally, the spectrum-related features take into account merely the magnitude information, whereas the phase is often overlooked due to the complications related to its interpretation. However, there is often additional information concealed in the phase, which could be beneficial for recognition. The APGD is an elegant approach to inferring phase information, which lacks of the issues related to interpreting the phase and does not require extensive parameter adjustment. Having shown applicability for speech-related problems, it is now explored in terms of instrument recognition. The evaluation is performed with various instrument sets and shows noteworthy absolute accuracy gains of up to 7% compared to the baseline mel-frequency cepstral coefficients (MFCCs) case. Combined with the MFCCs and with feature selection, APGD demonstrates superiority over the baseline with all the evaluated sets.",

keywords = "Musical instrument recognition, Phase spectrum, Music information retrieval, All-pole group delay feature",

author = "Alexander Diment and Rajan Padmanabhan and Toni Heittola and Tuomas Virtanen",

note = "Contribution: organisation=sgn,FACT1=1<br/>Portfolio EDEND: 2015-02-10; International Symposium on Computer Music Multidisciplinary Research ; Conference date: 01-01-1900",

year = "2014",

doi = "10.1007/978-3-319-12976-1_37",

language = "English",

isbn = "978-3-319-12975-4",

pages = "606--618",

editor = "Mitsuko Aramaki and Olivier Derrien and Richard Kronland-Martinet and Solvi Ystad",

booktitle = "Sound, Music, and Motion, 10th International Symposium, CMMR 2013, Marseille, France, October 15-18, 2013. Revised Selected Papers",

publisher = "Springer International Publishing",

}

Diment, A, Padmanabhan, R, Heittola, T & Virtanen, T 2014, Group Delay Function from All-Pole Models for Musical Instrument Recognition. in M Aramaki, O Derrien, R Kronland-Martinet & S Ystad (eds), Sound, Music, and Motion, 10th International Symposium, CMMR 2013, Marseille, France, October 15-18, 2013. Revised Selected Papers. Springer International Publishing, pp. 606-618, International Symposium on Computer Music Multidisciplinary Research, 1/01/00. https://doi.org/10.1007/978-3-319-12976-1_37

Group Delay Function from All-Pole Models for Musical Instrument Recognition. / Diment, Alexander; Padmanabhan, Rajan; Heittola, Toni ; Virtanen, Tuomas.

Sound, Music, and Motion, 10th International Symposium, CMMR 2013, Marseille, France, October 15-18, 2013. Revised Selected Papers. ed. / Mitsuko Aramaki; Olivier Derrien; Richard Kronland-Martinet; Solvi Ystad. Springer International Publishing, 2014. p. 606-618.

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

TY - GEN

T1 - Group Delay Function from All-Pole Models for Musical Instrument Recognition

AU - Diment, Alexander

AU - Padmanabhan, Rajan

AU - Heittola, Toni

AU - Virtanen, Tuomas

N1 - Contribution: organisation=sgn,FACT1=1<br/>Portfolio EDEND: 2015-02-10

PY - 2014

Y1 - 2014

N2 - In this work, the feature based on the group delay function from all-pole models (APGD) is proposed for pitched musical instrument recognition. Conventionally, the spectrum-related features take into account merely the magnitude information, whereas the phase is often overlooked due to the complications related to its interpretation. However, there is often additional information concealed in the phase, which could be beneficial for recognition. The APGD is an elegant approach to inferring phase information, which lacks of the issues related to interpreting the phase and does not require extensive parameter adjustment. Having shown applicability for speech-related problems, it is now explored in terms of instrument recognition. The evaluation is performed with various instrument sets and shows noteworthy absolute accuracy gains of up to 7% compared to the baseline mel-frequency cepstral coefficients (MFCCs) case. Combined with the MFCCs and with feature selection, APGD demonstrates superiority over the baseline with all the evaluated sets.

AB - In this work, the feature based on the group delay function from all-pole models (APGD) is proposed for pitched musical instrument recognition. Conventionally, the spectrum-related features take into account merely the magnitude information, whereas the phase is often overlooked due to the complications related to its interpretation. However, there is often additional information concealed in the phase, which could be beneficial for recognition. The APGD is an elegant approach to inferring phase information, which lacks of the issues related to interpreting the phase and does not require extensive parameter adjustment. Having shown applicability for speech-related problems, it is now explored in terms of instrument recognition. The evaluation is performed with various instrument sets and shows noteworthy absolute accuracy gains of up to 7% compared to the baseline mel-frequency cepstral coefficients (MFCCs) case. Combined with the MFCCs and with feature selection, APGD demonstrates superiority over the baseline with all the evaluated sets.

KW - Musical instrument recognition

KW - Phase spectrum

KW - Music information retrieval

KW - All-pole group delay feature

U2 - 10.1007/978-3-319-12976-1_37

DO - 10.1007/978-3-319-12976-1_37

M3 - Conference contribution

SN - 978-3-319-12975-4

SP - 606

EP - 618

BT - Sound, Music, and Motion, 10th International Symposium, CMMR 2013, Marseille, France, October 15-18, 2013. Revised Selected Papers

A2 - Aramaki, Mitsuko

A2 - Derrien, Olivier

A2 - Kronland-Martinet, Richard

A2 - Ystad, Solvi

PB - Springer International Publishing

T2 - International Symposium on Computer Music Multidisciplinary Research

Y2 - 1 January 1900

ER -

Diment A, Padmanabhan R, Heittola T , Virtanen T. Group Delay Function from All-Pole Models for Musical Instrument Recognition. In Aramaki M, Derrien O, Kronland-Martinet R, Ystad S, editors, Sound, Music, and Motion, 10th International Symposium, CMMR 2013, Marseille, France, October 15-18, 2013. Revised Selected Papers. Springer International Publishing. 2014. p. 606-618 https://doi.org/10.1007/978-3-319-12976-1_37

Group Delay Function from All-Pole Models for Musical Instrument Recognition

Abstract

Conference

Keywords

Publication forum classification

Access to Document

Fingerprint

Cite this