A multiple-F0 estimation algorithm for computer composed music based on supervised learning and harmonic selection

作者：Yi Guo^*; Xinyue Bai

来源：International Journal of Digital Content Technology and Its Applications, 2012, 6(11): 139-147.

DOI：10.4156/jdcta.vol6.issue11.18

• Estimation algorithm
• Estimation problem
• Good research
• Harmonic matching
• Harmonic structures
• Human auditory models
• PCA
• Training process

A multiple-F0 estimation algorithm is presented in this paper. It used harmonic selection and a mathematic thought for computer composed music specifically to multiple-F0 estimation. Computer composed music is almost free from similar instruments of the differences between different individuals, so it is a good research object at the beginning. This paper regards the multiple-F0 estimation problem as a typical pattern recognition task and solves it by first training and then identifies ways to deal with. The algorithm reduces dimension by PCA and selects candidates first by human auditory model and harmonic structures of notes. It does the estimation by combining a variety of methods. In training process, the harmonic structure of each note was obtained. And then, it changes the multiple-F0 estimation question into a mathematic problem and we can solve it in mathematic way. It can be shown in this paper that the experimental results indicate that this method has very good recognition results for the computer composed music.

• 出版日期2012
• 单位电子科技大学