Statistical Music Modeling aimed at Identification and Alignment

This paper describes a methodology for the statistical modeling of music works. Starting from either the representation of the symbolic score or the audio recording of a performance, a HMM is built to represent the corresponding music work. The HMM can be used to identify unknown recordings and to align them with the corresponding score, if available. Experimental evaluation on a collection of classical music showed that the approach allows for a satisfactory level of effectiveness in terms of both identification rate and effectiveness of the alignment. The methodology can be exploited as the core component of a set of tools for accessing and active listening a music collection.