Generative grammars and the computer-aided composition of music

The application of computers in music has focused almost exclusively on problems of sound synthesis. The application of computers in the process of music composition, ie. the generation of sound structures, remains largely unexplored. This thesis describes a computer software system designed to usefully aid composers in the process of music composition by automating part of the composition process. The composition system described uses generative grammars to automate the generation of music structures. The core of the system described is two facilities. These are 1) a facility for formally and explicitly defining the grammars of music languages, ie. the GGDL programming language, and 2) a facility for using GGDL language definitions to automatically generate utterances in the specified languages, ie. the GGDL—Generator. An implementation of these facilities has been integrated with programs to enable sound synthesis and the graphic editing of music structures. The system, implemented on a network of computers at the Department of Computer Science, Unversity of Edinburgh, is described. The thesis presents and evaluates some of the practical results obtained using the GGDL computer aided composition system. It is shown how the system may be used to compose macroand micro—sound structures. An automated digital sound synthesis instrument developed using generative grammars is also described.