Digital profiling of a lens's optical characteristics—distortion, focus marks, focal length. Essential for matching CG elements and 3D tracking accuracy.
You need accurate lens data if you want to seamlessly integrate digital objects into live-action footage. Lens mapping systematically captures how your physical camera distorts the world—and it's significant. Every lens has optical peculiarities: chromatic aberration, distortion (barrel or pincushion), focus shift between color channels. A VFX supervisor who ignores this data will later see in compositing that the 3D geometry doesn't fit, the perspective vanishes incorrectly, the CG character floats instead of standing.
The capture usually happens with checkerboard patterns or special calibration rigs in front of the lens—multiple shots from different distances, focus positions, and zoom levels. Modern software (internally developed or commercial) then calculates the exact lens characteristics from the pattern's deformations. The result is a lens distortion model and a 3D camera matrix, which you later integrate into tracking, matchmove, and compositing software. The 3D reconstructor then knows how to transform spatial coordinates back into the correct image space.
In practice, you particularly need this data for large-format VFX shots: when you have to composite greenscreen characters, when camera tracking is critical, or when you're processing stereoscopic footage. With a flat production lens (standard prime, calibrated by the manufacturer), you can often work with approximations. But with old anamorphic lenses, fisheyes, or experimental optics, things quickly become hair-splitting. We've seen shots where a barely visible focus shift between camera test and principal photography made the entire match plate impossible—a solid calibration would have revealed this immediately.
The data is then available either as distortion curves, as a LUT file, or as structured parameter sets (focal length, principal point, sensor format, aberration coefficients). The matchmove artist loads this directly into their software—whether it's Nuke, 3DEqualizer, or Maya. This saves iterations in compositing and guarantees that your CG integration sits stably in the end and doesn't need to be readjusted with every render pass.