Analysis of Music Rhythm Based on Bayesian Theory
Xiaolan Lin, Chuanzhen Li, Hui Wang, Qin Zhang · 2009
Automatically extracting rhythmic information from musical recordings is inarguably one of the most critical subtasks in many systems of music information retrieval. This paper presents a system for automatically extracting rhythm feature of audio music signal in the WAV format by using a new approach based on metric structure and Bayesian theory. In this system, an detected method is applied in the first step to extract the onset data, which will be used as input data to track tempo by a dynamic Kalman filter in the second step. Then a metric-based method is used to infer meter, which together with tempo will represent rhythm. Experimental results show that the accuracy of our approach ranges from 43.2% to 68.2% according to the music genre.