| This project will investigate the effectiveness of genetic algorithms (GAs) in suggesting musically coherent extensions to incomplete melodic material, and the usefulness of such suggestions to composers working in computer assisted compositional environments. In particular it will explore the effectiveness in teaching melody construction. Compositional environments are traditionally passive mediums within which the composer works. This cross disciplinary project explores the potential of the computer as an active partner in the creative process. Furthermore, it applies genetic algorithm research to a previously unexplored task; that of solving temporal musical problems with a largely ineffable aesthetic goal. The computer will use genetic algorithms to provide melodic extensions of a limited scope, constrained by a set of compositional heuristics derived from the theories of western diatonic music, and subsequently further directed by feedback from the user's acceptance or rejection of the suggestions. The user will interact with the system in a familiar environment displaying music in common music notation and as audio. The success of the process will be based upon the aesthetic appropriateness of the suggested melodic extensions and the value to the composer of the suggestions in assisting the compositional process. | |