An artificial intelligence approach to musical grouping analysis

Publisher Website
Google Scholar
Abstract
Work on modelling music cognition now seems to be at the stage where theories concerning the ongoing processes of musical understanding are required. This paper describes and critically assesses algorithms for two possible approaches to modelling how listeners process the ‘musical surface’ of discrete pitches to produce high level grouping structures in memory, being (i) syntactic processing using a grammar of chord function, and (ii) knowledgebased recognition of stereotypical phrase-level structures, using a system of frames (Minsky 1977). An integrated and resource limited system is described, which combines a parser of musical chord function with schema-based expectations concerning grouping structures. Major problems include describing how real-world knowledge of musical context is brought to bear on the listening experience, and integrating the processing of grouping structure within a broader theory of musical cognition.