The Piano Lesson

The Piano Lesson is Netflix’s 2024 drama adapting August Wilson’s 1987 play, directed by Malcolm Washington and starring Samuel L. Jackson. Washington co-wrote the screenplay with Virgil Williams, and the cast includes John David Washington, Ray Fisher, Michael Potts, Erykah Badu, Skylar Aleece Smith, Danielle Deadwyler, and Corey Hawkins. Set in 1936 Pittsburgh in the aftermath of the Great Depression, the film follows the Charles family of the Doaker Charles household and their heirloom piano, decorated with designs carved by an enslaved ancestor.

SCANable provided 3D scans of several sets and actors in support of the film’s visual effects, delivering terrestrial LiDAR and photogrammetric data for the VFX team’s work.

An object the story is about

Most props can be approximated. A prop that the title refers to, that the camera returns to, and that characters argue over for two hours cannot.

Carved relief is specifically difficult. The detail that matters is shallow — a few millimeters of depth carrying the whole meaning of the surface — and it is exactly the kind of information that a modeled approximation smooths away without anyone noticing until the shot is in front of a supervisor.

Photogrammetric capture records that relief as measured geometry, at a density where the carving survives into the final asset rather than becoming a texture pretending to be one.

Why the capture has to happen in one instant

People cannot hold still. Not performers, not crew, not anyone. Breathing alone moves the chest and shoulders enough to matter at the resolution character work requires.

A capture method that sweeps across a subject, or takes several passes, is therefore recording several slightly different subjects and averaging them. The result is soft where it should be sharp, and the softness lands on the parts of a face that carry likeness.

Firing every camera in an array at the same moment removes the problem instead of correcting for it afterward. That single-instant capture is the core of CyberSCAN™ as a method, and it is why the data needs so little hand cleanup.

Interiors are harder than exteriors

There is an assumption that a large exterior is the demanding scan and a domestic room is the easy one. In practice the room is usually harder.

An open exterior gives a scanner long, clear sightlines. A furnished 1930s interior gives it doorways, furniture, stair rails, and dozens of small occlusions, each of which hides some part of the space from any given position.

Covering it properly means many more scan positions, planned so that each one sees enough of the previous to register reliably. Get that wrong and errors compound through the chain until rooms at the far end of a house no longer line up with rooms at the near end. Careful set and location scanning is mostly a planning discipline, and the deliverables go out after cleanup and render-ready preparation.

More feature-film capture work is in the portfolio, or see everything we capture.