Early Development of a Flexible Procedural Approach to Automatic Jazz Improvisation

Daniel Field Part of CW17

This paper describes early work on an approach to automatic improvisation in the jazz idiom, derived from analysis of human practice, with algorithm development in the Jython Environment for Music (JEM). From the outset this project sought to be inspired by the human process of jazz improvisation and to apply as directly as possible the logic and processes of a human engaged in the activity of jazz improvisation – without necessarily using any specific computing technique or algorithm class. This paper describes the thinking and early experimentation leading to the current prototype version which improvises essentially in real-time and demonstrates how the use of variable parameters can greatly increase the flexibility of procedural programming.