AI-assisted Piano Music Composition: Innovative Approaches and Practical Explorations

Abstract

AI-assisted can save the cost and time used in piano music composition, and AI composition techniques are categorized into deconstruction, classification, and reconstruction, as well as three common models. In this paper, Markov model is chosen to compose piano music, describing pitch and duration, level, note duration, converting key and duration, and establishing piano melody, harmony and rhythm Markov modeling process. Setting up the test server and experimental software to train the Markov model so that it generates new piano music and completes the creation of piano music. The analysis shows that the piano music created by AI is more in line with the user's needs, the melody of AI piano music is more attractive can score about 8.3 points, and 86% of the users think that the piano melody created by AI is more fluent, at about 8.3 points. It has a high degree of emotional expressiveness and recognizability, and can meet the public's music aesthetic, so that the listener has a sense of immersion.