Music composition system based on subjective evaluation

This paper aims at the construction of the interactive music composition system which generates 16-bars music reflecting users feeling and subjectivity for music. The interactive Genetic Algorithm is introduced to this system. This paper tries to reduce users load in the evaluation of musical works. To this end, this paper uses fitness evaluations based on music theory by which presented music works are chosen. Furthermore, visual interface is designed by which users can follow the presented musical works. From the experimental results to confirm the validity of the present system, it is found that the subjects evaluation values of composed musical works becomes higher with the progress of generations. Furthermore, the subjects answer that the present system is better than the previous system from the viewpoint of understanding musical works visually.