How can Technologies like Deep Compositing help to start the Look Development Process as early on as possible?

Abstract

Deep data opens up the possibility to save not only 2D, but 3D pixels, leading to  a rendered image which can be edited in 3D, giving the option to insert additional objects and effects without the need to create holdout mattes, as this method is not layer-based but relies on the depth values per pixel. 

This allows for a much more flexible, non-destructive and procedural workflow.  The rendering process in a production can start as early as the camera position and lighting is set, meaning that additional assets can always be added later on without any problems and worrying about holdout mattes. Also, volumetric objects can be rendered separately without having to consider the intersecting objects, and changes can be made after rendering.  This workflow is handy for small teams with tight deadlines, as the options for rendering on the last second are limited. Only the storage needs to be more capable, as deep data image files take up substantially more space than traditional 2D images, which also leads to a slower workflow which comes with working with large files. Deep Compositing is not a faster rendering approach, but a more flexible one.