Application of Bayesian Probability Network to Music Scene Analysis

We propose a process model for hierarchical perceptual sound organization, which recognizes perceptual sounds included in incoming sound signals. We consider perceptual sound organization as a scene analysis problem in the auditory domain. Our current application is a music scene analysis system, which recognizes rhythm, chords, and source-separated musical notes included in incoming music signals. Our process model consists of multiple processing modules and a probability network for information integration. The structure of our model is conceptually based on the blackboard architecture. However, employment of a Bayesian probability network has facilitated integration of multiple sources of information provided by autonomous modules without global control knowledge.