Section: New Results

Music Content Processing and Music Information Retrieval

Acoustic modeling, non-negative matrix factorisation, music language modeling, music structure

Music structure representation and decomposition

Participants : Frédéric Bimbot, Corentin Guichaoua, Anaik Olivero.

Main collaboration: E. Vincent (EPI PAROLE, Inria Nancy), E. Deruty (external consultant)

Interest has been steadily growing in semantic audio and music information retrieval for the description of music structure, i.e. the global organization of music pieces in terms of large-scale structural units. Our group has defined a detailed methodology for the semiotic description of music structure, based on concepts and criteria which are formulated as generically as possible, i.e. without resorting to intrinsic properties of the musical content, but rather on relationships between musical elements resulting in well-identifiable patterns. The essential principles and practices developed during an annotation effort deployed by our research group on audio data, in the context of the Quaero project, has led to the public release of over 380 annotations of pop songs from three different data sets (http://musicdata.gforge.inria.fr/structureAnnotation.html ) documented by a technical report which includes a few case studies and a concise statistical overview of the annotated data [31] . From the algorithmic point of view, we are currently exploring tree-based representation of music structure where a sequence of musical elements and their relationships are modeled hierarchically as the derivation of a context-free grammar. Parsimony criteria and specific cost functions adapted to music patterns are used to learn the grammar rules and the possibility of distorting the rules is introduced to account for variability accross different repetitions of musical segments.