Identifying the classical music composition of an unknown performance with wavelet dispersion vector and neural nets

Stephan Rein, Martin Reisslein

Research output: Contribution to journalArticle

13 Citations (Scopus)

Abstract

As the internet search evolves toward multimedia content based search and information retrieval, audio content identification and retrieval will likely become one of the key components of next generation internet search machines. In this paper we consider the specific problem of identifying the classical music composition of an unknown performance of the composition. We develop and evaluate a wavelet based methodology for this problem. Our methodology combines a novel music information (audio content) descriptor, the wavelet dispersion vector, with neural net assessment of the similarity between unknown query vectors and known (example set) vectors. We define the wavelet dispersion vector as the histogram of the rank orders obtained by the wavelet coefficients of a given wavelet scale among all the coefficients (of all scales at a given time instant). We demonstrate that the wavelet dispersion vector precisely characterizes the audio content of a performance of a classical music composition while achieving good generalization across different performances of the composition. We examine the identification performance of a combination of 39 different wavelets and three different types of neural nets. We find that our wavelet dispersion vector calculated with a biorthogonal wavelet in conjunction with a probabilistic radial basis neural net trained by only three independent example performances correctly identifies approximately 78% of the unknown performances.

Original languageEnglish (US)
Pages (from-to)1629-1655
Number of pages27
JournalInformation Sciences
Volume176
Issue number12
DOIs
StatePublished - Jun 22 2006

Fingerprint

Neural Nets
Music
Wavelets
Neural networks
Unknown
Chemical analysis
Internet
Biorthogonal Wavelets
Rank order
Information retrieval
Methodology
Wavelet Coefficients
Neural nets
Instant
Information Retrieval
Histogram
Descriptors
Multimedia
Retrieval
Likely

Keywords

  • Audio content description
  • Classical music
  • Generalization
  • Identification
  • Internet search machine
  • Music information retrieval
  • Neural net
  • Wavelet

ASJC Scopus subject areas

  • Statistics and Probability
  • Electrical and Electronic Engineering
  • Statistics, Probability and Uncertainty
  • Information Systems and Management
  • Information Systems
  • Computer Science Applications
  • Artificial Intelligence

Cite this

Identifying the classical music composition of an unknown performance with wavelet dispersion vector and neural nets. / Rein, Stephan; Reisslein, Martin.

In: Information Sciences, Vol. 176, No. 12, 22.06.2006, p. 1629-1655.

Research output: Contribution to journalArticle

@article{f4a55dfc55544089a37aec1722f37953,
title = "Identifying the classical music composition of an unknown performance with wavelet dispersion vector and neural nets",
abstract = "As the internet search evolves toward multimedia content based search and information retrieval, audio content identification and retrieval will likely become one of the key components of next generation internet search machines. In this paper we consider the specific problem of identifying the classical music composition of an unknown performance of the composition. We develop and evaluate a wavelet based methodology for this problem. Our methodology combines a novel music information (audio content) descriptor, the wavelet dispersion vector, with neural net assessment of the similarity between unknown query vectors and known (example set) vectors. We define the wavelet dispersion vector as the histogram of the rank orders obtained by the wavelet coefficients of a given wavelet scale among all the coefficients (of all scales at a given time instant). We demonstrate that the wavelet dispersion vector precisely characterizes the audio content of a performance of a classical music composition while achieving good generalization across different performances of the composition. We examine the identification performance of a combination of 39 different wavelets and three different types of neural nets. We find that our wavelet dispersion vector calculated with a biorthogonal wavelet in conjunction with a probabilistic radial basis neural net trained by only three independent example performances correctly identifies approximately 78{\%} of the unknown performances.",
keywords = "Audio content description, Classical music, Generalization, Identification, Internet search machine, Music information retrieval, Neural net, Wavelet",
author = "Stephan Rein and Martin Reisslein",
year = "2006",
month = "6",
day = "22",
doi = "10.1016/j.ins.2005.06.002",
language = "English (US)",
volume = "176",
pages = "1629--1655",
journal = "Information Sciences",
issn = "0020-0255",
publisher = "Elsevier Inc.",
number = "12",

}

TY - JOUR

T1 - Identifying the classical music composition of an unknown performance with wavelet dispersion vector and neural nets

AU - Rein, Stephan

AU - Reisslein, Martin

PY - 2006/6/22

Y1 - 2006/6/22

N2 - As the internet search evolves toward multimedia content based search and information retrieval, audio content identification and retrieval will likely become one of the key components of next generation internet search machines. In this paper we consider the specific problem of identifying the classical music composition of an unknown performance of the composition. We develop and evaluate a wavelet based methodology for this problem. Our methodology combines a novel music information (audio content) descriptor, the wavelet dispersion vector, with neural net assessment of the similarity between unknown query vectors and known (example set) vectors. We define the wavelet dispersion vector as the histogram of the rank orders obtained by the wavelet coefficients of a given wavelet scale among all the coefficients (of all scales at a given time instant). We demonstrate that the wavelet dispersion vector precisely characterizes the audio content of a performance of a classical music composition while achieving good generalization across different performances of the composition. We examine the identification performance of a combination of 39 different wavelets and three different types of neural nets. We find that our wavelet dispersion vector calculated with a biorthogonal wavelet in conjunction with a probabilistic radial basis neural net trained by only three independent example performances correctly identifies approximately 78% of the unknown performances.

AB - As the internet search evolves toward multimedia content based search and information retrieval, audio content identification and retrieval will likely become one of the key components of next generation internet search machines. In this paper we consider the specific problem of identifying the classical music composition of an unknown performance of the composition. We develop and evaluate a wavelet based methodology for this problem. Our methodology combines a novel music information (audio content) descriptor, the wavelet dispersion vector, with neural net assessment of the similarity between unknown query vectors and known (example set) vectors. We define the wavelet dispersion vector as the histogram of the rank orders obtained by the wavelet coefficients of a given wavelet scale among all the coefficients (of all scales at a given time instant). We demonstrate that the wavelet dispersion vector precisely characterizes the audio content of a performance of a classical music composition while achieving good generalization across different performances of the composition. We examine the identification performance of a combination of 39 different wavelets and three different types of neural nets. We find that our wavelet dispersion vector calculated with a biorthogonal wavelet in conjunction with a probabilistic radial basis neural net trained by only three independent example performances correctly identifies approximately 78% of the unknown performances.

KW - Audio content description

KW - Classical music

KW - Generalization

KW - Identification

KW - Internet search machine

KW - Music information retrieval

KW - Neural net

KW - Wavelet

UR - http://www.scopus.com/inward/record.url?scp=33645880131&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33645880131&partnerID=8YFLogxK

U2 - 10.1016/j.ins.2005.06.002

DO - 10.1016/j.ins.2005.06.002

M3 - Article

AN - SCOPUS:33645880131

VL - 176

SP - 1629

EP - 1655

JO - Information Sciences

JF - Information Sciences

SN - 0020-0255

IS - 12

ER -