SHOW CONTROL
Video, light and motion each carry their own delay. Making them land together is the job; everything else is decoration.
Every element in a show has a delay between being told to do something and doing it. A media server has one. A lighting fixture has another. A motor has a much larger one, and unlike the others it is not even constant, because a mass takes time to accelerate and that time depends on what it is carrying.
None of this is a problem while the elements are separate. It becomes the entire problem the moment they are supposed to happen at once. An audience will forgive almost anything about a show except a bang that arrives after the flash. Human perception is far more tolerant of low resolution, poor colour and thin sound than it is of things that should be simultaneous and are not.
It is worth being precise about why. Human perception fuses events that arrive close together into a single occurrence, and the window in which it does so is narrow and unforgiving, especially between sound and vision. Inside it, everything reads as one event. Outside it, the audience does not think the timing is slightly off — they perceive two separate things, and the illusion the whole room was built to produce is simply not there.
So the useful mental model is not a timeline of commands. It is a timeline of arrivals, worked backwards. You decide when each thing should be perceived, then subtract what you know about each path, then issue the commands at whatever ugly staggered times fall out of that arithmetic. A cue list that looks tidy is usually a cue list that has not been measured.
Measurement is the part people skip. The specification sheet gives you a figure that was true in a laboratory for one configuration. The real number depends on the network, the media file, the fixture's own processing, and how much else is happening in the same frame. It has to be measured on the installed system, and re-measured after anything changes.
Motion deserves separate treatment because it is the one element whose delay is not a constant. A media server's latency is stable enough to measure once. A motor's depends on mass, on load, on temperature, and on where in its travel it starts. So the cue that fires it cannot be a fixed offset copied from another show; it has to be derived from the profile of that axis, and re-derived when the mechanical team changes anything.
The reward for getting it right is that nobody mentions it. Synchronisation is only ever visible when it is absent, which makes it a difficult thing to be proud of in a meeting and the first thing an audience feels.
Build the measurement into the commissioning plan rather than treating it as troubleshooting. A camera at the guest's position, recording sound and picture together, played back frame by frame, will settle in an hour arguments that otherwise run for a week. It also produces the only artefact that survives a handover: a record of what was actually happening, not what the cue sheet said should be.