Dune

Dune is Denis Villeneuve’s 2021 epic science fiction film for Warner Bros., adapting the first half of Frank Herbert’s 1965 novel. Written by Villeneuve, Jon Spaihts, and Eric Roth, the film follows Paul Atreides as the noble House Atreides is thrust into a war for the deadly desert planet Arrakis. The ensemble cast includes Timothée Chalamet, Rebecca Ferguson, Oscar Isaac, Josh Brolin, Stellan Skarsgård, Dave Bautista, Stephen McKinley Henderson, Zendaya, David Dastmalchian, Chang Chen, Sharon Duncan-Brewster, Charlotte Rampling, Jason Momoa, and Javier Bardem. Filming took place from March to July 2019 in Budapest, Jordan, Norway, and Abu Dhabi.

SCANable provided 3D scanning for the production, capturing performers as high-resolution body scans. Our photogrammetry data supported the film’s visual effects teams with digital doubles and face replacement work.

Why an ensemble this size gets scanned

A film with fourteen credited leads and a sustained visual effects program does not scan its cast as a formality. It scans them because every one of those performers is a candidate for a shot that cannot be achieved practically — a stunt too dangerous to stage, a crowd too large to hire, a plate where an actor needs to appear in an environment that does not exist.

Digital doubles are the answer to all of those, and they only work if the underlying capture is faithful. An audience will accept an impossible landscape without blinking. It will reject a human face that is a few percent wrong, instantly and without being able to say why.

That asymmetry is the entire reason body and face scanning exists as a discipline. The tolerance on a person is far tighter than the tolerance on anything else in frame.

Capturing a costumed cast

Wardrobe on a production of this scale is not incidental — it is sculptural, heavy, and specific to the character. Scanning a performer out of costume produces a body that then has to be re-dressed digitally, which is expensive and rarely matches.

Capturing them in costume produces the actual silhouette: how the garment falls on that body, where it creases at the shoulder, how the material sits when the performer stands naturally. Our Mobile CyberSCAN™ Studios were built around this — the array travels to the production, so talent can step out of a scene, scan in full wardrobe and makeup, and step back in.

That matters most when the schedule is international. A show shooting across Budapest, Jordan, Norway, and Abu Dhabi cannot route its cast back to a fixed facility between units. The capture has to go where the production is.

Simultaneous capture, and why it is non-negotiable

Our arrays fire every camera at the same instant. A scanning method that sweeps or takes multiple passes asks a human being to hold perfectly still for the duration, and human beings cannot. They breathe, they micro-shift, they settle.

The resulting drift shows up as soft or doubled geometry that has to be repaired by hand — and hand repair on a face is where likeness goes wrong. A single instantaneous capture removes the problem at the source rather than correcting for it downstream. This is the core of what cyberscanning actually is as a technique.

From scan to face replacement

Face replacement is the most demanding thing a scan gets used for. It requires geometry accurate enough to track against the performer’s real head, and surface detail fine enough to hold up at the scale a face occupies on a cinema screen.

That means capturing pore-level skin detail, not just an overall likeness. Cross-polarized photography separates diffuse color from specular response, giving a modeler the information needed to build skin that responds correctly to lighting — including lighting conditions the set never had. Our photometric and OLAT capture extends this into fully relightable assets for shows that need it.

For a film shot largely in extreme desert light, that relightability is not a luxury. An asset baked to one lighting condition cannot be moved.

Delivery into the VFX pipeline

What ships is not raw capture. Scan output goes through retopology, texturing, and render-ready asset builds before a vendor sees it, arriving scaled, oriented, and named to the show’s convention.

The goal is that a VFX artist opens the file and starts working, rather than spending a day figuring out what they were sent. On a film distributing work across multiple vendors, that consistency is the difference between a usable asset library and a collection of one-offs.

See more feature-film capture work, or read about set and location scanning for the other half of what a production like this needs captured.