When
Thematic Session 1: Design and Interaction (Monday, 14:15)
Abstract
Intuition is an innate part of performing, and yet it is seldom considered and has been under-theorized and under-researched in the context of performer-system interaction models. Intuition has significant relevance to the interaction and communication between an autonomous system and a performer. Interest in artificial musical agents continues to increase within interactive live performance, to produce works incorporating autonomous computer systems programmed to “listen” and contribute material. Interaction with such systems commonly focus on computers sensing the performer’s physical and/or sonic gesture. Traditional sense-response systems often focus on detecting the action and ignore the intention. However, intention fosters the development of performer trust, synchronization, and collaboration within interaction. And yet the more ‘AI-based’ systems that tend to be the preferred approach can similarly be perceived as black-box interaction, where the agent’s intention is hidden behind closed computation. I conducted three case studies on performers’ sense of intuition within interactive performance by designing and testing an interactive system that provides embodied cues of the system’s internal state and intention. The system’s intention to act and the quality of gesture was made perceivable to the human performer by simulating breath and a breathing pattern required by the system to play a synthesized flute. By altering the timbre and duration of the simulated breath, the system indicated the quality of intended gesture. The model was evaluated by collecting performer interview data, third person observation of performer interaction, and first-person accounts of the system designer testing the system as a part of the design process. The collected data provided insights into the human performer’s perception and social reading of the system’s (musical agent’s) presence, intentions, and actions as a co-performer. The results of this study can be used to further develop models of interaction with autonomous generative systems in live performance.
Bio
Greg Corness is a researcher working in the area of embodied interaction in media and performance environments. His work focuses on developing generative sound systems and autonomous agents for use in interactive environments such as games, audio installation and interactive performances. He publishes in the fields of electronic music and human-computer interaction and his work includes galleries installations, interactive museum exhibits and live performance in Canada and the US. Greg holds a MMus in Computer Composition from the University of Victoria and a PhD from the School of Interactive Arts + Technology at Simon Fraser University in Interactive Performance.