Improving access to digital music through content‐based analysis

The purpose of this paper is to report recent advances on a collaborative project that aims to develop content‐based methods for music information retrieval (MIR) as an alternative to standard text‐based modes of access to digital music libraries.

The paper describes current practices and ongoing research, and it discusses potential applications for future use.

Content‐based MIR approaches can extend and enhance the capabilities of traditional text‐based discovery and delivery systems and thus support the work of expert users such as musicians and musicologists. Examples of technologies developed in the context of the project include novel methods for automatic chord identification, motif finding, the visualization of musical structure, and retrieval of musical variations using harmonic and structural information.

The paper looks at new, non‐verbal modes of interaction with digital music archives based on musically substantive features such as chords, motifs, rhythms, etc. By building more sophisticated dimensions of interactivity into a discovery‐and‐delivery system, these tools could give the end‐user a more meaningful and rewarding experience. The tools potentially would be less costly and more scalable than textual annotation and markup, and their applicability extends beyond digital libraries to other music services.

This article discusses the advantages and challenges posed by audio‐based MIR and shows, via project‐specific examples, its relevance to supporting the needs of digital music library users.