Next-Generation Video Coding and Streaming

CHAPTER 73D VIDEO CODING

3D movies, especially animated movies, have enjoyed some measure of success. Autostereoscopic (glass-free) displays may be ready to move 3D entertainment viewing up a new level. Multiview 3D video bitstreams contain much redundancy yet they are sometimes coded separately (e.g., left-eye and right-eye views) to improve parallelism (i.e., simultaneous coding of multiple views with no degradation in spatial or temporal resolution). This increases the transmission bandwidth and storage requirements, especially for supporting wide-angle coverage involving many simultaneous views. This chapter focuses on the coding of multiple 3D views for efficient storage and network transport.

7.1 INTRODUCTION

The simplest form of 3D TV is a stereoscopic 3D (S3D) TV, where the display is capable of rendering only two views, one for each eye, so that the scene is perceived as 3D when wearing appropriate eye glasses. Thus, such displays project stereoscopic image pairs to the viewer. The source video captures stereo pairs in a two-view setup, with cameras mounted side by side and separated by the same distance as between a person's eyes. The active eye glasses turn dark over one eye for about 8 ms while showing the other eye an image and then vice versa for the complementary image. The S3D displays normally employ infrared signaling to activate and synchronize to the glasses. More sophisticated autostereoscopic displays are able to render and display more views in such a ...