Extended Reality can put remote musicians in the same virtual room, but should that room be entirely invented, or just an augmented version of wherever each musician actually is? A new study — “Where do we Jam?”, accepted at the IEEE International Symposium on the Internet of Sounds (IS2) 2026’s Musical Metaverse workshop, held in Cannes in late October — puts that question to an actual test rather than a design assumption.
The setup
Researchers from the University of Augsburg and University of Trento ran a laboratory study with 28 musicians, comparing a simulated networked music performance in Virtual Reality against the same scenario in Augmented Reality. In both conditions, each musician saw their remote partner rendered in real time as a 3D point cloud, using an optimized Gaussian Splatting shader — the same underlying rendering technique this site has tracked as a creative medium elsewhere, here applied to a live performer’s body instead of a scanned object. The difference was the space around that point cloud: in VR, participants could select and customize a fully virtual 3D environment, while in AR they saw a pass-through view of the actual lab room they were sitting in.
What held steady, and what split down the middle
Across both conditions, positive affect, flow, and creativity support were strongly correlated — when one rose, the others tended to rise with it, regardless of which reality the musician was in. But involvement behaved differently: it correlated with flow and creativity support only in VR, and participants reported feeling significantly more involved in the fully immersive virtual condition than in the augmented, pass-through one.
The twist is in the qualitative interviews layered on top of those numbers. Preferences for AR versus VR, and for how a creative virtual environment should even be designed, varied substantially from musician to musician — enough that the researchers’ real conclusion isn’t “VR wins,” but that the surrounding space itself is a design variable, not just scenery behind the people you’re playing with. Their recommendation for future XR networked-performance tools is to support a broad span of the reality-virtuality continuum rather than picking one point on it, because no single setting serves every player’s sense of presence and creative flow equally well.
Part of a longer research line
This builds directly on earlier work from the same Augsburg lab — a CHI 2023 study, “Wish You Were Here,” that first tested point-cloud-rendered remote musicians against audio-only collaboration and found the point-cloud condition measurably improved presence. “Where do we Jam?” keeps that rendering approach (now upgraded to Gaussian Splatting) and turns the next dial: not whether to show your bandmate’s body, but where you both appear to be standing while you do it.