Last update: 2011/08/21

Riding the media bits

Leaving flat video for good?

After three quarters of a century of flat television, it is time to add a 3rd dimension

At the 58th MPEG meeting in December 2001 the Japan Electronics and Information Industries Association (JEITA) made a proposal to MPEG to investigate standardisation of "3D Video". The proposal was accepted and one of the outcomes of the investigation was the development of requirements for Multiview Video Coding (MVC). Very much like in the MPEG-2 MVC profile the idea was to give the encoder the possibility to exploit the correlation that exists between the output of an array of video cameras. A Call for Proposals was issued and this work item, too, was entrusted to the JVT. The development of the MVC technologies for AVC gave the opportunity to make yet another edition of AVC.

Currently MPEG is running the  3DV project with the goal to define a 3D Video format that enables both advanced stereoscopic display processing and improved support for auto-stereoscopic N-view displays as depicted in Fig. 5.

Fig. 1 – Expected functionality of 3DV format

 The figure depicts two specific application scenarios targeted by the 3DV format:

• Stereo devices able to cope with various display types and sizes, and different viewing preferences, including the ability to adjust the depth perception to mitigate fatigue and other viewing discomforts;
• High-quality auto-stereoscopic displays where the new 3DV format enables the generation of many high-quality views from a limited amount of input data, e.g. stereoscopic video and respective depth maps.

MPEG has recently issued a CfP in which some requirements are identified:

1. Compression bitrate not exceeding twice the bit rate of state-of-the-art compressed single video
2. Minimal visual distortion introduced by compression on synthesised views
3. Compatibility modes with AVC and HEVC supporting stereo and mono video
4. Possibility to extract bitstreams for stereo and mono output in a simple way and to reconstruct with high-fidelity samples from the left and right views of the stereo video
5. Extraction of bitstreams that represent sub-sets of the views available in the compressed domain
6. Improved rendering capability and quality compared to state-of-the-art representations
7. Real-time decoding and synthesis of views, required by any N-view display, with computational and memory requirements available to devices at the consumer electronics level
8. Independence from display type
9. Rendering of stereo views with a variable baseline
10. Support of an appropriate depth range determined by [-W/2;+W] or less (W = display width)
11. Support of display-specific shift of depth location, i.e., whether the perceived 3D scene (or parts of it) are behind or in front of the screen.