Abstract
It is very difficult in the music harmonization system of the modern popular music that the complicated music, that have the chord progression and the key change reflected in the music theories and arranger's personality, are composed by using the techniques of artificial intelligence and neural network.
In this paper, we propose a musical arrangement system using fuzzy reasoning by max-min operation in two fuzzy sets. There are the parts consisting of some bars in the melody of the music. Each part has tonality very related to images of music. Then, we make a chord progression fit to the tonality of each part by taking notice of the importance of each note constructing a melody. The two sets are classified with (1)“fitting a chord to a melody” and (2)“propriety of a chord progression according to considering music theory. We get the parameters of the sets from the shapes of the membership functions, which are changed by arranger's feeling and experience. We have the good natural chord progression without inharmoniousness by using the reasoning. It is shown in the experimental results that the tonality takes advantage of the reasoning to propose the final chords.
