COVERS_SONGS

Set of dissimilarity matrices of covers

    D = COVERS_SONGS(N)

   INPUT
     N - Integer scalar or vector 1:36 determining desired dataset

   OUTPUT
     D - Dataset with dissimilarity matrix of desired dataset. If N is a
          vector D is a cell array of dissimilarity matrices.
          They are 205 x 205 and are built from the same 21-class problem.

Description

This is one of the datasets used by Liem and Hanjalic [1]. It is build  from a set of dissimilarity matrices between musical fragments in a study  on the detection of covers.

Reference(s)

[1] Cynthia C. S. Liem, Alan Hanjalic: Cover Song Retrieval: A Comparative Study of System Component Choices. ISMIR 2009: 573-578.

See also

datasets, distools, prdisdata, covers_beatles, covers_beethoven, covers80,