Billy Lynn’s Long Halftime Walk

Billy Lynn’s Long Halftime Walk is Ang Lee’s 2016 drama following a 19-year-old soldier’s victory tour after a harrowing Iraq battle. Based on the widely acclaimed, bestselling novel, the film is told from the point of view of Private Billy Lynn (newcomer Joe Alwyn), who along with his fellow soldiers in Bravo Squad becomes a hero and is brought home temporarily for a victory tour. Through flashbacks culminating at the spectacular halftime show of the Thanksgiving Day football game, the film contrasts the realities of war with America’s perceptions. Kristen Stewart, Chris Tucker, Garrett Hedlund, Vin Diesel and Steve Martin also star. Lee shot at an ultra-high frame rate — a first in film history — to create an immersive digital experience that dramatizes war in a way never seen before.

SCANable performed a detailed 3D scan of the Georgia Dome in Atlanta, Georgia. The resulting 3D mesh drove camera tracking and computer-generated crowd population for the stadium set seen throughout a large portion of the film.

An extreme frame rate removes the places you hide

Shooting at very high frame rates and high resolution strips away the motion blur and grain that conventional capture supplies for free. Those artefacts have always done quiet work for visual effects, softening edges and hiding small inconsistencies between the real and the added.

Without them, everything is legible. Geometry that would have passed at 24 frames a second is examined at leisure, and any element sitting at the wrong scale or the wrong distance announces itself. The response is not more polish downstream; it is more accurate measurement at the source.

A digital crowd is a placement problem before it is a rendering problem

Populating a stadium is usually described as a rendering challenge, but what audiences judge is arrangement. Figures have to sit on real seats, at real seat pitch, on tiers that rise at the real angle, receding correctly toward the far side of the bowl.

That arrangement can only come from the venue’s true geometry. Given a measured model, a crowd system places agents against surfaces that exist; given an approximation, every figure inherits the error, and the aggregate reads as wallpaper. Our set and location scanning supplies the seating and structure the placement runs on.

Lighting reference and geometry are read together

An HDRI describes the light in a space. Geometry describes the space itself. Neither is much use alone.

Together they let a lighting artist reproduce a real environment’s response — falloff, bounce, and which surfaces sit in shadow. Capturing both in the same coordinate frame is what makes them agree once a CG element is dropped into the plate.

Large structures fail in specific ways

Registration error that is trivial across a room is not trivial across a stadium bowl. Repeating architecture gives alignment algorithms too many plausible answers, and long sightlines turn small angular deviations into large positional ones.

Handling that is closer to survey practice than to set scanning: controlled positions, verified registration, and a single datum the whole capture ties back to. It is the rigour our enterprise survey work applies to industrial sites, and it is what lets the resulting mesh be trusted for tracking.

More feature film scanning is in the portfolio, along with the services list.