How to ensure a good depth perception on a light field display

Information captured by the eye allows the human brain to extract depth cues from the scene, to analyze them and to understand a complex 3D scene. In real life all these depth cues are naturally present at the same time and coherent. Rendering them on a display for several viewers is not straightforward. Today no technology can fully achieve this requirement. Per definition, a light field content is a collection of light rays that are corresponding to different viewpoints of the same scene. A light field display should be able to render these different viewpoints to a single viewer or to multiple viewers. The quality of a light field display will be measured by its ability to correctly render these different views and then the expected depth cues. In the paper we will define the technical requirements for a light field display to provide effective depth cues such as the binocular disparity, the motion parallax and the accommodation at a pixel resolution expected by the eye. These requirements will be evaluated with respect to existing technologies (e.g. integral imaging displays) or expected future ones (e.g. microLED displays). Simulation of light field displays will be proposed for different display size (smartphone, desktop or TV), resolution per view and pixel pitch. These simulations illustrate the main impact of the pixel pitch to reach the light field display goal in terms of depth cues requirements. The last part of the paper will focus on the data generation issues. How to generate such amount of views at the display side and what are the adapted data formats for such a purpose? The paper will be associated with supplementary videos that will illustrate the view generation topic using a dedicated data format that is optimized for current GPU processors. Video capture of early stage light field displays will be provided.