Interactive composition and performance framework with evolutionary computing

Choi, I ORCID: 2017, Interactive composition and performance framework with evolutionary computing , in: 43rd International Computer Music Conference, 15-20 October 2017, Shanghai Conservatory of Music, Shanghai, China.

PDF - Published Version
Download (599kB) | Preview


Evolutionary models such as genetic algorithms and cellular automata have been well adopted by composers. A challenge still remains in practice, how to convey the dynamics of evolution that can be perceivable through performance realization. This challenge becomes more explicit when adopting an agent model such as swarms in which temporality is implicit in their behavioral patterns governed by self-organizational and social dynamics. In this work a compositional approach and system architecture undertake some paradigmatic shift by situating the evolutionary dynamics at the locus of tone production through an ongoing engagement with a performer. Essentially, swarm agents are given sounds to play with, and performers play with these agents. By articulating a kinesthetic framework in performance, Performance Gestural Articulation Unit is introduced as a generalizable action repertoire with implicit and relative duration. The kinesthetic framework of performance factors is proposed in 6 categories as an abstraction applicable as design requirements for interactive performance. This articulation aspires to AI modelling for human-machine performance with kinesthetic evolution. The paper concludes by summarizing how the abstraction is applied to the composition, Human Voice, as a use case context.

Item Type: Conference or Workshop Item (Paper)
Schools: Schools > School of Arts & Media
Journal or Publication Title: Proceedings of the 43rd International Computer Music Conference
Publisher: ICMA
ISBN: 9780984527465
Related URLs:
Depositing User: I Choi
Date Deposited: 14 Nov 2017 12:37
Last Modified: 15 Feb 2022 22:38

Actions (login required)

Edit record (repository staff only) Edit record (repository staff only)


Downloads per month over past year