All code and the open dataset are published with the article. We also provide a dynamic dashboard for data exploration: thomasgnuttall.github.io/KarnatakGest...

The approach we take can be applied to any multimodally ensembled signals in both human and nonhuman communication, to determine co-structuring profiles and explore any reduction in degrees of freedom.

As individuals differ in their gesturing styles, we provide a computational approach to characterizing individual vocalists’ unique multimodal vocal-gesture co-structuring profiles.

We find that in vocal performance (South Indian, karnatak singing), vocal motifs that are more alike co-occur with gestures that are more alike, suggesting that gestures are constrained by co-occurring vocal motifs.

We go beyond looking at stereotypical gesture-motif pairs, and instead base our analyses on a wide range of freely varying vocal motifs (595 in total) and co-occurring hand and head kinematics, from 3D motion capture.