I m testing a workflow where a Mac camera recognizes a hand or body pose locally and turns it into OSC, Art-Net/sACN, or FreeD data for the rest of a live-production system.
The most useful mappings so far seem to be simple and reversible: launching a spare clip, toggling an effect bypass, changing a dashboard control, or triggering a safe lighting look. For reliability I m using a restricted recognition zone, separate ON/OFF thresholds, a short hold time, and a cooldown so one pose cannot fire repeatedly.
I m curious how other VJs, lighting operators, and virtual-production teams would use this kind of control. Would you prefer momentary cues, toggles, or continuous values? And what safeguards would you consider essential before using gestures during a real show?
Runs super smooth on my Mac Studio and the multi-camera support actually holds up well, no dropped frames even with two angles running pose tracking at the same time. The custom reference poses are a nice touch for triggering lighting cues on cue.