Survey tracking coverage before choosing the shot

Define coverage as a tested volume

Tracking coverage is not the outline of the room and not a provider’s maximum range. It is the part of the camera’s planned position, orientation, height, and lens use in which the whole system produces an acceptable image. A coverage map must include obstructions, operator paths, practical set pieces, LED structure, and the actual tracking body or marker arrangement. Moving the camera body, adding a monitor, or changing the lens can change what the system can see or what the shot reveals.

Unreal’s ICVFX quick start describes Live Link as the integration path for camera tracking and calls for an on-set test. That is an important boundary: engine connectivity is necessary, but the survey must prove the intended physical working volume.

Walk extents and occlusion deliberately

Put a floor grid or measured reference into the survey plan. At each planned extreme, hold long enough to observe stable data, then pan, tilt, roll, and walk the intended arc. Test behind the tallest practical obstruction, at the edge of the LED or set, and at any point where crew will naturally stand. Repeat with the real camera build; a lightweight survey wand can be useful for a preliminary map but is not evidence for a fully dressed camera.

Judge the recorded image, not only tracker health

A tracker can report a valid pose while calibration, latency, or lens data produces an unacceptable composite. Pair the coverage path with a physical/virtual alignment reference visible to the production camera. Epic’s calibration overview states that virtual position/orientation and tracking timing must match the physical camera/video for alignment. The stage survey should therefore mark both data dropouts and visible drift.

Illustrative example: a dolly move is technically covered until the operator passes a scenic column. The data remains present but a known 80 mm lens exposes a small alignment error at the far end. Mark the path “usable to 1.8 m at 50 mm; conditional beyond that; do not use 80 mm past column line until recalibrated.” This is more actionable than a green/red room diagram.

Publish a production-facing map

• Draw tested, conditional, and untested zones; do not colour untested space as safe.
• List camera build, tracker configuration, lens range, and scene/calibration revision.
• Give each restriction a consequence: reblock, re-survey, or technical escalation.
• Re-run affected cells after physical stage changes or sensor/marker relocation.

This Atlas method does not replace the tracking vendor’s calibration procedure. It turns that system’s output into shot-specific production evidence.