Computational models for rhythm and applications on human-machine interactions

Abstract

This thesis describes investigations towards the use of computational models for rhythm in the context of human-machine interactions. I propose a model for semiautomatic musical transcription, a model for automatic musical transcription and two interfaces for musical expression, all of them based in concepts related to rhythm. This novel study highlights the importance of using domain-specific knowledge in music information retrieval systems, as it allows not only more accurate systems to be built, but also the development of new ways of musically interacting with computers.