A multimodal approach to song-level style identification in pop/rock using similarity metrics

Research output: Contribution to conferencePaper

Abstract

This paper presents a multimodal approach to style identification in pop/rock music. Considering the intuitive feelings of similarity from the listener's perspective, this study focuses on features that are computed using similarity metrics for melodies, harmonies, and audio signals for style identification. Support vector machine is used as a binary classifier to determine if two songs are created by the same artist given their similarity distances in the three aspects. Experiments are conducted using songs of four well-known pop/rock bands from 6 albums. The preliminary result shows that the approach achieves the best result in correct rate of 85% using only seven similarity metrics.

Original languageEnglish (US)
Pages321-324
Number of pages4
DOIs
StatePublished - Jan 1 2013
Externally publishedYes
Event2013 12th International Conference on Machine Learning and Applications, ICMLA 2013 - Miami, FL, United States
Duration: Dec 4 2013Dec 7 2013

Other

Other2013 12th International Conference on Machine Learning and Applications, ICMLA 2013
CountryUnited States
CityMiami, FL
Period12/4/1312/7/13

Fingerprint

Rocks
Support vector machines
Classifiers
Experiments

Keywords

  • Gaussian mixture models
  • melodic contour
  • music similarity
  • n-grams
  • Style

ASJC Scopus subject areas

  • Computer Science Applications
  • Human-Computer Interaction

Cite this

Chuan, C-H. (2013). A multimodal approach to song-level style identification in pop/rock using similarity metrics. 321-324. Paper presented at 2013 12th International Conference on Machine Learning and Applications, ICMLA 2013, Miami, FL, United States. https://doi.org/10.1109/ICMLA.2013.143

A multimodal approach to song-level style identification in pop/rock using similarity metrics. / Chuan, Ching-Hua.

2013. 321-324 Paper presented at 2013 12th International Conference on Machine Learning and Applications, ICMLA 2013, Miami, FL, United States.

Research output: Contribution to conferencePaper

Chuan, C-H 2013, 'A multimodal approach to song-level style identification in pop/rock using similarity metrics' Paper presented at 2013 12th International Conference on Machine Learning and Applications, ICMLA 2013, Miami, FL, United States, 12/4/13 - 12/7/13, pp. 321-324. https://doi.org/10.1109/ICMLA.2013.143
Chuan C-H. A multimodal approach to song-level style identification in pop/rock using similarity metrics. 2013. Paper presented at 2013 12th International Conference on Machine Learning and Applications, ICMLA 2013, Miami, FL, United States. https://doi.org/10.1109/ICMLA.2013.143
Chuan, Ching-Hua. / A multimodal approach to song-level style identification in pop/rock using similarity metrics. Paper presented at 2013 12th International Conference on Machine Learning and Applications, ICMLA 2013, Miami, FL, United States.4 p.
@conference{654ceadcfac649c2bdb598620315c4cf,
title = "A multimodal approach to song-level style identification in pop/rock using similarity metrics",
abstract = "This paper presents a multimodal approach to style identification in pop/rock music. Considering the intuitive feelings of similarity from the listener's perspective, this study focuses on features that are computed using similarity metrics for melodies, harmonies, and audio signals for style identification. Support vector machine is used as a binary classifier to determine if two songs are created by the same artist given their similarity distances in the three aspects. Experiments are conducted using songs of four well-known pop/rock bands from 6 albums. The preliminary result shows that the approach achieves the best result in correct rate of 85{\%} using only seven similarity metrics.",
keywords = "Gaussian mixture models, melodic contour, music similarity, n-grams, Style",
author = "Ching-Hua Chuan",
year = "2013",
month = "1",
day = "1",
doi = "10.1109/ICMLA.2013.143",
language = "English (US)",
pages = "321--324",
note = "2013 12th International Conference on Machine Learning and Applications, ICMLA 2013 ; Conference date: 04-12-2013 Through 07-12-2013",

}

TY - CONF

T1 - A multimodal approach to song-level style identification in pop/rock using similarity metrics

AU - Chuan, Ching-Hua

PY - 2013/1/1

Y1 - 2013/1/1

N2 - This paper presents a multimodal approach to style identification in pop/rock music. Considering the intuitive feelings of similarity from the listener's perspective, this study focuses on features that are computed using similarity metrics for melodies, harmonies, and audio signals for style identification. Support vector machine is used as a binary classifier to determine if two songs are created by the same artist given their similarity distances in the three aspects. Experiments are conducted using songs of four well-known pop/rock bands from 6 albums. The preliminary result shows that the approach achieves the best result in correct rate of 85% using only seven similarity metrics.

AB - This paper presents a multimodal approach to style identification in pop/rock music. Considering the intuitive feelings of similarity from the listener's perspective, this study focuses on features that are computed using similarity metrics for melodies, harmonies, and audio signals for style identification. Support vector machine is used as a binary classifier to determine if two songs are created by the same artist given their similarity distances in the three aspects. Experiments are conducted using songs of four well-known pop/rock bands from 6 albums. The preliminary result shows that the approach achieves the best result in correct rate of 85% using only seven similarity metrics.

KW - Gaussian mixture models

KW - melodic contour

KW - music similarity

KW - n-grams

KW - Style

UR - http://www.scopus.com/inward/record.url?scp=84899416541&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84899416541&partnerID=8YFLogxK

U2 - 10.1109/ICMLA.2013.143

DO - 10.1109/ICMLA.2013.143

M3 - Paper

SP - 321

EP - 324

ER -