Systems Over Spectacle

Most event visuals are finished files. Someone edits a timeline, exports a master, and the venue plays it back. It works, it is predictable, and it is over the moment the lights come up. The file cannot answer the room it is playing in.

We build the other thing. A system that produces the image while the event is happening, from rules we wrote and inputs the room provides. It costs more thinking up front and considerably less panic on the night.

Take the 100 Violins concert. Eight pillars in a circle around the audience, clad in P3.9 LED across two faces each, sixteen surfaces in total. The obvious approach is sixteen video files and a playback machine. Instead, one GLSL shader feeds all sixteen surfaces through TouchDesigner. Nothing is edited, nothing is exported. The image is computed frame by frame across nearly fifty minutes, and the inner faces stay warm while the outer faces stay cold because that is a parameter, not a render.

The difference shows up when something changes. A set runs four minutes long, a section repeats, the conductor holds a silence. A file keeps rolling and drifts. A system follows.

The same logic decides what the image is allowed to react to. For a cinema format in Dolby Atmos, we set one rule: the image follows the structure of the mix, not its amplitude. A visual volume appears when a sound travels across the ceiling. Lateral tension builds when the mix opens onto the surrounds. The image compresses on monophonic passages. Reacting to loudness is easy and looks like a screensaver. Reacting to spatial structure requires understanding the piece, and it reads as intent.

At Djerba Explore the principle went further. The island's cultural, sonic and environmental data were not illustrated, they were the material the image and the sound were generated from. Nine microphones spread at different heights listened to the room and the room drew itself. A jar of olives, wired through an ultrasonic sensor and an Arduino, triggered wall pixel mapping over DMX. The visitor was not watching a piece about Djerba, they were inside a system running on Djerba.

There is a commercial argument underneath the artistic one. A system can be re-run, re-scaled, re-pointed at a different venue or a different dataset. A file can only be played again. When a client returns for a second edition, we adjust parameters. That is the whole difference between a production cost and an asset.

Systems over spectacle is not a preference for restraint. It is a claim about where the work actually lives. Not in the output, in the conditions that produce it.