Classification in music research

Weihs, Claus; Ligges, Uwe; Mörchen, Fabian; Müllensiefen, Daniel

doi:10.1007/s11634-007-0016-x

Cited by 48 publications

(20 citation statements)

References 97 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In addition to many background chapters of master's theses [39,79,113,132,153,154,188,193,239,361,367,371,418] and doctoral dissertations [9,141,146,280,284,290,320,341,342,381,427,447] at least five reviews are devoted specifically to MGR [23,85,123,241,373], and 19 other reviews discuss related aspects [24,25,51,71,84,100,101,152,181,198,224,233,270,282,315,398,423,441,442]. Many of these reviews compile the variety of feature extraction methods and classification algorithms that have been applied to MGR, and some compare system performance using specific figures of merit (FoM) on particular benchmark datasets.…”

Section: Introductionmentioning

confidence: 99%

A Survey of Evaluation in Music Genre Recognition

Sturm

2014

Adaptive Multimedia Retrieval: Semantics, Context, and Adaptation

View full text Add to dashboard Cite

Abstract. Much work is focused upon music genre recognition (MGR) from audio recordings, symbolic data, and other modalities. While reviews have been written of some of this work before, no survey has been made of the approaches to evaluating approaches to MGR. This paper compiles a bibliography of work in MGR, and analyzes three aspects of evaluation: experimental designs, datasets, and figures of merit.

show abstract

Section: Introductionmentioning

confidence: 99%

A Survey of Evaluation in Music Genre Recognition

Sturm

2014

Adaptive Multimedia Retrieval: Semantics, Context, and Adaptation

View full text Add to dashboard Cite

show abstract

“…Typical applications of classification in music signal analysis can be found in [64]. Instrument classification, relevant for genre classification, is discussed in [30].…”

Section: The Aim Of Feature Processing Is To Reduce the Data And At Tmentioning

confidence: 99%

Huge Music Archives on Mobile Devices

Blume

Bischl

Botteck

et al. 2011

IEEE Signal Process. Mag.

Self Cite

View full text Add to dashboard Cite

T he availability of huge nonvolatile storage capacities such as flash memory allows large music archives to be maintained even in mobile devices. With the increase in size, manual organization of these archives and manual search for specific music becomes very inconvenient. Automated classification makes it possible for the user to organize the available music archives according to different categories, which can be either predefined or user defined, enabling a better overview of these databases.Classification is normally based on a set of features extracted from the music data. Feature extraction and subsequent music classification are very demanding computational tasks, although more and more computational power is available in desktop computers and portable devices such as smartphones and personal digital assistants (PDAs). Today, a variety of music features and classification algorithms are available, which are optimized for various application scenarios, to achieve different classification qualities. This article reviews the many aspects of music classification. Covered in more detail are feature extraction and classification approaches, algorithmic and softwarerelated aspects, and hardware implementation issues for performance and power-sensitive mobile devices.[ Toward an automated dynamic organization ]

show abstract

“…In music classification the raw input time series are seldom the right basis for analysis. Instead, various transformations are in use (see, e.g., Weihs et al, 2007). Since with music frequencies play a dominant role, periodograms are a natural representation for observations.…”

Section: Transformations and Local Modellingmentioning

confidence: 99%

“…E.g. time-series representing music pieces need special distances (Weihs et al 2007). Other important aspects of distance are translation, size, scale and rotation invariance, e.g.…”

Section: Introductionmentioning

confidence: 99%

Distances in Classification

Weihs

Szepannek

2009

Advances in Data Mining. Applications and Theoretical Aspects

Self Cite

View full text Add to dashboard Cite

Abstract. The notion of distance is the most important basis for classification. This is especially true for unsupervised learning, i.e. clustering, since there is no validation mechanism by means of objects with known groups. But also for supervised learning standard distances often do not lead to appropriate results. For every individual problem the adequate distance is to be decided upon. This is demonstrated by means of three practical examples from very different application areas, namely social science, music science, and production economics. In social science, clustering is applied to spatial regions resulting in unconnected clusters. However, connectedness is sometimes important for interpretation, and may have to be taken into account for clustering. In statistical musicology the main problem is often to find an adequate transformation of the input time series as an adequate basis for distance definition. Also, local modelling is proposed in order to account for different subpopulations, e.g. instruments. In production economics often many quality criteria have to be taken into account with very different scaling. In order to find a compromise optimum classification, this leads to a pre-transformation onto the same scale, called desirability.

show abstract

Classification in music research

Cited by 48 publications

References 97 publications

A Survey of Evaluation in Music Genre Recognition

A Survey of Evaluation in Music Genre Recognition

Huge Music Archives on Mobile Devices

Distances in Classification

Contact Info

Product

Resources

About