Store Closed until April

Assetizing your Sets

Sets are complicated

Especially when it comes to replicating a set or location digitally, so many things make up a scene.

Consider a scenario where we're shooting in a specially constructed village set. It's a bespoke design, built exclusively for the production, and doesn't have a 'real world' equivalent. The surfaces of the walls are unique, the ground tiles are asymmetrical, and the set dec is custom-built. This set is encased by green screens, and the concept art calls for large extensions of similar-looking buildings stretching down the road. So how is it currently captured?


How are they currently captured?

Conventionally, on-set VFX scanning companies capture reference images of these sets under existing lighting conditions, accompanied by a LiDAR pass and any HDRIs shot during the takes.

When this data reaches the VFX vendor, it has to be broken down and parsed for usable elements. The texture/surfacing departments will need to find reference photos that work as textures, then painstakingly de-light them, and potentially composite several images together.

Modeling departments must identify elements in the Lidar and reference imagery that can be used as a basis for new buildings, they cannot use the existing buildings as they are unique to what was seen in-camera.

This entire process is labor-intensive, costly, and consumes months of artist time.

The Scan Space Advantage

At Scan Space, we understand what the pipeline is from front to back, and as such we aim to eliminate as much of the tedious tasks involved in working with poor (or no) scan data.

Our approach is to look at sets and try understand exactly what the intention is for the extensions by looking at concept art, talking with production, art department and the VFX supervisor.

We then go around the set, looking at what surfaces, objects and forms repeat often.

After the set has been completely understood, we 'assetize' it, by capturing each element as its own scan, doors, walls, every floor surface type. the pillars, barrels, boxes and roof tiles.
We strive to capture the largest examples of each surface, to give the artists the best possible options to work with

Unless an item is a 'one-off' hero item, it will be included in this 'assetizing' process. Some sets could yield upwards of 100 individual scans!

We name, group, and track these assets collectively, providing productions with a live database that offers a general overview of this data.

Once processed, this data is delivered to production as a 'Set Data' package. This package contains all the sub-packages, along with preview renders.

Our approach allows VFX vendors to significantly accelerate their turnaround times, saving time, resources, and costs.