Melody Transformation with Semiotic Patterns

Abstract

This paper presents a music generation method based on the extraction of a semiotic structure from a template piece followed by generation into this semiotic structure using a statistical model of a corpus. To describe the semiotic structure of a template piece, a pattern discovery method is applied, covering the template piece with significant patterns using melodic viewpoints at varying levels of abstraction. Melodies are generated into this structure using a stochastic optimization method. A selection of melodies was performed in a public concert, and audience evaluation results show that the method generates good coherent melodies.