A probabilistic expert system for automatic musical accompaniment

This article presents methodology that allows a computer to play the role o f musical accompanist in a nonimprovised musical composition for soloist an d accompaniment. The modeling of the accompaniment incorporates a number of distinct knowledge sources including timing information extracted in real- time from the soloist's acoustic signal, an understanding of the soloist's interpretation learned from rehearsals, and prior knowledge that guides the accompaniment toward musically plausible renditions. The solo and accompan iment parts are represented collectively as a large number of Gaussian rand om variables with a specified conditional independence structure-a Bayesian belief network. Within this framework a principled and computationally fea sible method for generating real-time accompaniment is presented that incor porates the relevant knowledge sources. The EM algorithm is used to adapt t he accompaniment to the soloist's interpretation through a series of rehear sals. A demonstration is provided from J.S. Bach's Cantata 12.