In this work, we introduce a co-adaptive closed-loop interface driven by audio augmented with a parameterization of the dancer's heart-rate in near real-time. In our set-up, two salsa dancers perform their routine dance (previously choreographed and well-trained) and a spontaneously improvised piece lead by the male dancer. They firstly dance their pieces while listening to the original version of the song (baseline condition). Then, we ask them to dance while listening to the music, as altered by the heart rate extracted from the female dancer in near real-time. Salsa dancing is always led by the male. As such, their challenge is to adapt, their movements, as a dyad, to the real-time change induced by the female's heart activity. Our work offers a new co-adaptive set up for dancers, new data types and analytical methods to study two forms of dance: well-rehearsed choreography and improvisation. We show that the small variations in heart activity, despite its robustness for autonomic function, can distinguish well between these two modes of dance.
© Copyright 2020 Proceedings of the 7th International Conference on Movement and Computing. Published by ACM. All rights reserved.