What we learn building a visual instrument, written down. No fluff, no funnel — the boring problems that decide a show, explained properly.
We are building a visual instrument for the booth. The things we learn on the way get written down here.
What happens at 2 a.m. when the software dies with a full room watching. Measured, not promised.
Guessing the tempo by listening is a hard problem. Here is why it drifts worst exactly when you need it most.
A cut that lands perfectly in the booth can look late forty metres away. It is not your timing: it is physics.
That recurring stutter is usually not your GPU. It is 50 Hz meeting 60 Hz.