3D Video: From Capture to Diffusion

Chapter 4

Shooting and Viewing Geometries in 3DTV

4.1. Introduction

A three-dimensional (3D) perception induced by a 3D display, operated according to various modalities (optics, colorimetrics and alternately shutters) through spatial and/or temporal, generally planar, mixing of colocalized 2D images in front of viewers, is essentially only an illusion. These mixed images are separated before being received by viewers’ eyes so that, through stereopsis, their minds are tricked into seeing a deceptive 3D scene instead of two superposed flat images. This generic viewing geometry must be taken into account when capturing media for 3D television (3DTV) because the relationship between shooting and viewing geometries directly affects the quality of the viewer’s experience, as well as depth distortion of the perceived scenes.

In this chapter, we will describe and characterize the viewing geometry and then present compatible shooting geometries. We will then study the potential distortions in perceived scenes that a combination of these shooting and viewing geometries may cause. The relations between these distortions and the parameters of the geometries used will allow us to propose a specification methodology for the shooting geometry, which will ensure that scenes are perceived with a set of arbitrarily selected possible distortion on the 3DTV device used. Lastly, we will also provide practical details on how to use this methodology in order to place and configure virtual cameras when ...