Abstract

As a key element of music teaching, this paper focuses on the use of Bayesian network techniques with the characteristics of the three elements of sound in music for music teaching. The Bayesian harmonic model of audio is introduced in the analysis of music signals, and a Bayesian harmonic model is proposed on this basis, and then the fundamental frequency parameters and harmonic amplitudes in this model are estimated using a reversible hopping Monte Carlo sampling algorithm. The accuracy of the Bayesian harmonic model is studied and the error between the actual signal and the test signal is found to be less than 2%. The average score of the experimental class after the assessment of students' professional music ability with the help of Bayesian network technology and the tutorial teaching of the three elements of sound characteristics was 86, and the average score of the control class using traditional teaching was 79.8, with a score difference of more than 6 points. Therefore, the music signals synthesized by using Bayesian harmonic model analysis reproduced the original signals better, and the performance of the students subjected to model-assisted teaching improved significantly, which led to the improvement of the teaching quality of music majors.