Singer Identification In Rembetiko Music

In this paper, the problem of the automatic identification of a singer is investigated using methods known from speaker identification. Ways for using world models are presented and the usage of Cepstral Mean Subtraction (CMS) is evaluated. In order to minimize the difference due to musical style we use a novel data set, consisting of samples from greekRembetiko music, being very similar in style. The data set also explores for the first time the influence of the recording quality, by including many historical gramophone recordings. Experimental evaluations show the benefits of world models for frame selection and CMS, resulting in an average classification accuracy of about 81% among 21 different singers.

QC 20161031

Country

Sweden

Related Organizations

University of Crete
Greece
Royal Institute of Technology
Sweden

Keywords

Other Engineering and Technologies, Artist Identification; Gaussian Mixture Model; Music Information Retrieval, Annan teknik

Impact byBIP!

	selected citations These citations are derived from selected sources. This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).	0
	popularity This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.	Average
	influence This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).	Average
	impulse This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.	Average