Body-Driven Interface, MoveNet, Gesture
A dancer stands ten feet from the computer, mid-phrase, and wants the skeleton a different color. In most software, that means walking back to the mouse. On the movement wall, it means clapping. What TouchDesigner does with a node graph, this does with a webcam — the body is the remote, and the interface has disappeared.
TLDR
Three gestures replaced the mouse. The body is the interface.
For the Curious
The movement wall has a lot of settings — skeleton color, skeleton mode, background, blend mode, color cycle, triangles, keypoints, mirror. The first version put all of them in a side panel, which is fine if you are sitting at the desk. A dancer is not sitting at the desk. A dancer is ten feet away, moving, and the last thing they want to do is break the phrase to walk over and click.
So the panel became a gesture. Clap, and the menu advances to the next setting. Raise both hands, and it retreats. A high beep on the way forward, a lower beep on the way back. The feedback is a large on-screen HUD plus a chime, so it works even facing away from the screen.
The point is not that clapping is faster than clicking. The point is that the body never has to leave the dance. The interface stops being a thing you walk to and becomes a thing you are already doing.
For the Curious
Cycling through discrete values is fine for a color. It is clunky for a thickness. There is no satisfying way to “increase the opacity” by clapping — you want to feel the value move, the way you feel the weight of a fabric between your fingers.
So the slider gesture was born. One hand rests on your hip. The other hand rises and falls. As it rises, the line thickens, the opacity deepens, the keypoints grow. As it falls, everything goes lighter and more sheer. The HUD shows a live progress bar, so the value tracks your hand in real time.
It is the closest thing to touching the garment that a screen can offer. The hand is a fader, and the body is the control surface.
For the Curious
The last piece was the photograph. The movement wall already had a Style Snapshots strip — a row of thumbnails of captured looks — but taking one meant pressing a button. A dancer mid-phrase does not want to press a button.
So the shutter became a pose. Put both hands on your hips — the natural “take my picture” stance — and a 3-2-1 countdown begins. A short shutter beep, and the current webcam and skeleton view is captured and dropped into the strip, right alongside the other thumbnails.
The countdown is the gift. It gives you three seconds to move from the trigger pose into the pose you actually want photographed. The body signals the moment; the machine waits for it; the photograph lands where the others live.
Every feature on the movement wall is the same sentence told differently: the system reads the body and responds. The clap menu, the slider, the shutter — they are not three features. They are one idea, and the body is the proof. We don’t put clothes on a body. We let the body tell the clothes what to do.
The Next Move
The movement wall is at /movenet. Clap, raise, slide, and pose — the body is the remote.