C.A.L.V.I.N.N.™

Camera Array for Lightfields, Volumetrics, Imaging, and Neural Networks.

A 150+ camera 4K capture system built for lightfields, volumetrics, and neural reconstruction — capturing how a subject looks and moves from every angle at once, instead of freezing a single pose. We built it because nothing off the shelf did what our pipelines needed.

[ 150+ CAM / 4K / LIGHTFIELD + VOLUMETRIC ]

Talent standing inside the C.A.L.V.I.N.N. camera array, surrounded by its ring of camera modules

WHAT IT IS

150+ cameras, one subject, every angle.

C.A.L.V.I.N.N. surrounds the performer with 150+ synchronized 4K cameras. Every camera sees the subject at the same instant from its own angle, so what the system records is not one photograph but the light leaving a person in every direction at once — the raw material for lightfield, volumetric, and neural reconstruction.

The name is the spec sheet: Lightfields (dense multi-view coverage of how a subject reflects light), Volumetrics (full-body geometry in the round), Imaging (color-calibrated RAW, not compressed reference), and Neural Networks (data structured for the neural reconstruction pipelines VFX vendors now build on).

SINCE 2017

A borrowed name, from a show we scanned.

The acronym is honest engineering. The name it spells is a reference. In Watchmen — which we scanned for HBO in 2017 — the one character who experiences every moment simultaneously spends the season hiding in plain sight under an ordinary name: Cal.

C.A.L.V.I.N.N. traces back to the genlocked 4D array we built for that show, and its capture drove the series’ hologram effects. Every camera on one clock, one instant held from every direction — the same requirement 4D Gaussian splatting has today. We were meeting it six years before the technique had a name.

The hologram story — 2017 to 4DGS →

TWO DIFFERENT SYSTEMS

Not the same rig as our scan studios.

It is worth being precise, because the two get conflated. Our Mobile CyberSCAN™ Studios house a 150+ camera static scanning system — cross-polarized RAW, three-point beauty, and eight-direction OLAT passes, about 60 seconds with the actor. That is the system behind our digital doubles.

C.A.L.V.I.N.N. is a separate 150+ camera 4K system. Where the scan studios freeze a pose in perfect detail, C.A.L.V.I.N.N. is built to capture a subject across time and viewing angle — the lightfield and volumetric data that 4D reconstruction and neural pipelines are trained on.

Untextured scan geometry surrounded by the solved positions of the capture array

[ 150+ CAM 4K / LIGHTFIELD + VOLUMETRIC / NEURAL-READY ]

WHY IT MATTERS

Built for what comes after the scan.

A single scan describes a person at one instant. A lightfield array describes how that person returns light across many directions — which is what volumetric playback, 4D sequences, and neural reconstruction actually need to look convincing instead of merely accurate.

That is the reason the system exists under its own name: the pipelines it feeds are the ones moving fastest right now, and the capture has to be good enough that a dataset shot today still trains well against whatever method arrives next.

WHAT IT FEEDS

One capture, many pipelines.

4K multi-view coverage150+ synchronized 4K angles of the subject, in the round.
Lightfield & volumetric dataDense multi-view coverage in the round, not a single-angle capture.
Neural-ready datasetsSynchronized, calibrated imagery structured for radiance-field and neural reconstruction.
ML training dataStructured capture for machine-learning and neural rendering pipelines.

START HERE

Capture more than a pose.

Tell us the performance and the pipeline it feeds. We’ll scope the capture around it.

Request a Quote