Hunter Brown


Heretic: AI-Driven Improvisation Software

Heretic is an artificially intelligent computer music system to be used within the context of human-machine free improvisation. Heretic is written in the SuperCollider programming language with the machine learning aspects of the system implemented in Wekinator. The motivation behind Heretic’s inception was to create an autonomous system that uses my own improvisational methodology as a computational and conceptual framework for machine improvisation. To achieve this, Heretic’s architecture is divided into three interdependent modules: Interpretive Listening, Contextual Decision Making, and Musical Synthesis. Each of these modules serves an important musical function for the spontaneous creation of novel improvised music. When working collectively, these modules are a complete computational model of my improvisational methodology that interacts with human improvisers in real-time. Heretic is trained on my approach to improvisation, but through its interactions with a human performer, Heretic’s own improvisational voice and modes of musical expression emerge.

Github Repository

Machine Listening Demo

On the top-right corner of the screen in the above video, you will see a list of 10 improvisational "Language Types" accompanied by green sliders. These sliders are a visual representation of Heretic's ability to detect which language type I am playing in real-time. When I start, the sliders indicate "Silence" as I am not playing. As I begin to play short, staccato drum attacks, the sliders indicate "Sparse Formings." This process continues as I traverse all 10 language types in this video.