Janne T. Morstol
As 3D technology continues to evolve, the amount of available content continues to expand and sales of 3DTVs increase at a steady pace, the worldwide 3D market will grow—though perhaps not at the radically exaggerated rate that was initially hyped. I don’t believe there is any doubt that, over time, 3D has the potential to become a key differentiator for broadcasters and programmers.
3DTV raises the bar for quality and precision and increases the demand for bandwidth. The creation of 3D video and particularly the need for the best possible picture quality has ramifications for how content is captured, produced and broadcast. Broadcasting live 3DTV brings even greater challenges than “canned” 3D content such as DVDs or movies for cinema. The TV broadcast infrastructure is complex, covers long distances and includes multiple technology components, editing phases and broadcast delivery mechanisms.
In the backhaul/contribution phase, left and right eye information must be transported along with other metadata required to generate the highest quality 3D images and must be compatible with all 3D compression and display solutions. To ensure a high-quality viewing experience, it’s critical that all the data required to display the left and right eye images correctly is conveyed throughout the transport chain. Whether producing live content or canned video, it’s imperative to keep the right eye and left eye images synchronized in order to preserve the 3D perspective.
When we look at real world objects, the differing images from the left and the right eye are processed in the visual cortex of the brain to produce a perception of a three dimensional composition called stereopsis. The goal of a 3DTV screen is to reproduce this stereopsis as closely as possible.
Full HD Contribution
For 3D video, the most important content is the left and right eye images. To ensure that 3DTV programming delivers the quality of experience the consumer expects, it is a key requirement that the 3D contribution/backhaul system supports both left and right images in full HD resolution as well as existing 2D HD.
Bringing 3D content from the venue to the editing studio, from there to the broadcast center, then to the playout primary distribution point, requires two HD quality links at full 4:2:2, 10-bit resolution. Careful consideration should be given to the choice of compression for the backhaul of 3D content, both in terms of how to best maintain content quality and the economic cost of the bandwidth.
The left and right eye images must be synchronized in order to ensure the best 3D viewing experience. Synchronization means that the left and right eye images, captured at the same time by the 3D camera, are delivered to the editing and processing facility as they were originally captured. This may sound simple, but in reality synchronization continues to be a challenge for video transport and sometimes even for audio signals. The SMPTE technical committee 10E is currently developing a standard to define the mapping of the left and right eye signal into a 3G HD SDI signal.
The images must contain sufficient information to facilitate the processing of 3D depth and parallax information, and camera metadata may also be required to simplify or improve 3D processing. This may include zoom, focal length, distance to the point of focus, sharp range, camera identification, position, direction, separation and alignment. Clearly the camera metadata has to be synchronous with the video images, and we believe this can best be achieved using a wrapping or container technology such as MXF (Media Exchange Format).
Optimal Compression for 3D
We believe that JPEG2000 compression is an excellent solution for the optimal transport of stereo images when used with dual encoders accepting HD inputs. The wavelet transformation of JPEG2000 compression provides the same processing for all pixels, and the frame-by-frame compression utilized delivers consistent high-quality for all images. JPEG2000 also ensures that the horizontal resolution, so important for the stringent quality requirements of 3D, is preserved. With JPEG2000 there is no error propagation between pictures, because JPEG2000 is a picture-by-picture compression technology. This ensures consistent, high-quality for all images, at very low latency.
The right and left eye images received from the contribution/backhaul link need to be processed to include logos subtitles and other graphics. At the end of the post-production process, the final product needs to be compressed and sent over an MPEG channel to the home.
To make this possible and simultaneously maintain the left and the right eye images, there are a number of possible format options that can be employed before compression. The most common format used today is side-by-side. The MPEG encoder and decoder are fed with a standard HD signal that can be handled the normal way.
High-quality 3D content is becoming increasingly available and more is being created—from live sports events to epic movies. With a growing number of both active and passive 3D screens now available in the stores, prices will inevitably fall. The third critical factor affecting how quickly 3DTV will be successful in the home is video transport infrastructure.
More 3D, More Bandwidth
From a programmer’s or broadcaster’s perspective, the provision of additional feeds of 3D content will mean the use of additional, costly bandwidth. Broadcasters can address this in a number of ways: by buying additional capacity, investing in new infrastructure, deploying IP-based technologies and by utilizing efficient compression technologies such as JPEG2000. Whichever alternatives are chosen, it’s worth remembering that a good quality viewing experience in the home starts with high-quality in the contribution/backhaul phase.
As more content becomes available in 3D including major sporting events and popular TV shows, sales of 3DTV sets will approach critical mass and the floodgates of consumer demand will open. Given the long lead times for capital projects and infrastructure investments, now is the time for broadcasters and programmers to start factoring 3D into their infrastructure planning.
The author is COO of T-VIPS.
For more on 3DTV infrastructure, read:
The Care and Feeding of 3D Signals, by Wes Simpson
Janne T. Morstol