Short-term Feature Space and Music Genre Classification

Marques, Gonçalo; Langlois, Thibault; Gouyon, Fabien; Lopes, Miguel; Sordo, Mohamed

doi:10.1080/09298215.2011.573563

Cited by 12 publications

(14 citation statements)

References 9 publications

(11 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…By applying the method of irrelevant transformations to an MIR system, we seek to determine whether the system is actually considering the music at all. In this respect, our aims in using the method of irrelevant transformations are more aligned with the work of Porter and Neuringer with pigeons [28], Watanabe and Nemoto with sparrows [29], Chase with koi carp [30], and Marques et al [7], [8] with MDS.…”

Section: Contrasting With Robustness Evaluationmentioning

confidence: 99%

A Simple Method to Determine if a Music Information Retrieval System is a “Horse”

Sturm

2014

IEEE Trans. Multimedia

101

View full text Add to dashboard Cite

We propose and demonstrate a simple method to explain the figure of merit (FoM) of a music information retrieval (MIR) system evaluated in a dataset, specifically, whether the FoM comes from the system using characteristics confounded with the "ground truth" of the dataset. Akin to the controlled experiments designed to test the supposed mathematical ability of the famous horse "Clever Hans," we perform two experiments to show how three state-of-the-art MIR systems produce excellent FoM in spite of not using musical knowledge. This provides avenues for improving MIR systems, as well as their evaluation. We make available a reproducible research package so that others can apply the same method to evaluating other MIR systems.Index Terms-2-WORK system performance, 5-CONT content description and annotation, 5-SEAR multimedia search and retrieval.Bob L. Sturm (S'06-M'09) received the Ph.D. degree in electrical and computer engineering from the

show abstract

Section: Contrasting With Robustness Evaluationmentioning

confidence: 99%

A Simple Method to Determine if a Music Information Retrieval System is a “Horse”

Sturm

2014

IEEE Trans. Multimedia

101

View full text Add to dashboard Cite

show abstract

“…Often identified from extremely brief excerpts (Gjerdingen & Perrott, 2008), genre is equally the subject of automatic recognition systems (e.g. Marques, Langlois, Gouyon, Lopes, & Sordo, 2011). However, the genres themselves-also called musical styles in the analysis and history literature-are not easy to define, since no natural genre taxonomies exist and they are defined as cultural constructs rather than by any intrinsic musical properties (e.g.…”

Section: Audio As a Basis For Modelling Expressed Emotions In Musicmentioning

confidence: 99%

Are the Emotions Expressed in Music Genre-specific? An Audio-based Evaluation of Datasets Spanning Classical, Film, Pop and Mixed Genres

Eerola

2011

Journal of New Music Research

View full text Add to dashboard Cite

Empirical studies of emotions in music have described the role of individual musical features in recognizing particular emotions. However, no attempts have been made as yet to establish if there is a link between particular emotions and a specific genre. Here this is investigated by analysing nine separate datasets that represent categories ranging from classical (three sets), and film music (two), to popular music (two), and mixed genre (two). A total of 39 musical features were extracted from the audio. Models were then constructed from these to explain self-reports of valence and arousal, by using multiple and Random Forest regression. The models were fully validated across the datasets, suggesting low generalizability between the genres for valence (16% variance was accounted for) and moderately good generalizability between the genres for arousal (43%). In contrast, the generalizability within genres was considerably higher (43% and 62% respectively), which suggests that emotions, especially those that express valence, operate differently depending on the musical genre. The most reliable musical features of affects across genres were identified, yielding a ranked set of features most likely to operate across the genre. In conclusion, the implications of the findings, and the genre-specificity of emotions in music are discussed.

show abstract

“…This point of view, however, is not evident in much of the MGR literature, e.g., the three reviews devoted specifically to MGR (Aucouturier and Pachet 2003;Scaringella et al 2006;Fu et al 2011), the work of Tzanetakis and Cook (2002), Barbedo and Lopes (2008), Bergstra et al (2006a), Holzapfel and Stylianou (2008), Marques et al (2011b), Panagakis et al (2010a), Benetos and Kotropoulos (2010), and so on. It is thus not idiosyncratic to claim that one purpose of MGR could be to identify, discriminate between, and learn the criteria of music genres in order to produce genre labels that are indistinguishable from those humans would produce.…”

Section: Argumentsmentioning

confidence: 85%

Classification accuracy is not enough

Sturm

2013

J Intell Inf Syst

View full text Add to dashboard Cite

We argue that an evaluation of system behavior at the level of the music is required to usefully address the fundamental problems of music genre recognition (MGR), and indeed other tasks of music information retrieval, such as autotagging. A recent review of works in MGR since 1995 shows that most (82 %) measure the capacity of a system to recognize genre by its classification accuracy. After reviewing evaluation in MGR, we show that neither classification accuracy, nor recall and precision, nor confusion tables, necessarily reflect the capacity of a system to recognize genre in musical signals. Hence, such figures of merit cannot be used to reliably rank, promote or discount the genre recognition performance of MGR systems if genre recognition (rather than identification by irrelevant confounding factors) is the objective. This motivates the development of a richer experimental toolbox for evaluating any system designed to intelligently extract information from music signals.

show abstract

Short-term Feature Space and Music Genre Classification

Cited by 12 publications

References 9 publications

A Simple Method to Determine if a Music Information Retrieval System is a “Horse”

A Simple Method to Determine if a Music Information Retrieval System is a “Horse”

Are the Emotions Expressed in Music Genre-specific? An Audio-based Evaluation of Datasets Spanning Classical, Film, Pop and Mixed Genres

Classification accuracy is not enough

Contact Info

Product

Resources

About