CAPITAL CURRENTS: 3-D TV – A long way to go
Standards bodies have several years of work ahead to sort out 3-D TV interoperability issues.
3-D TV was everywhere at the Consumer Electronics Show this year, but it will be a while before cable companies deliver 3-D TV for home viewing. Numerous issues of standards and interoperability have to be worked out.
3-D programs can be created by using two cameras, or a camera with two lenses, to create a left (L) image and a right (R) image. And there are numerous technical solutions for displaying it, each with some advantages and some disadvantages.
At CES, there were numerous 3-D display technologies demonstrated, including a plasma screen with active shutter glasses, an LCD screen with passive polarized glasses, a projector with passive polarized glasses and a lenticular lens LCD with no glasses.
But this column is primarily about interoperability issues with coding methods for multiplexing the L and R images for delivery to the home, and then from the set-top box to the display.
I have come across two points of misunderstanding in reading press reports of 3-D TV. First, the picture format is not necessarily limited to 1080 lines 60 frames per second progressively scanned (1080p60). It can be 480p60, 720p60 or 1080i60. Second, the transmission coding methods (which I will discuss below) and the display technologies are independent of one another. The set-top box or display must use the proper demultiplexer to separate the L and R frames so the display can process them. But any transmission method should work with any display technology.
By transmission coding, I mean the multiplexing format used to deliver compressed (using MPEG or AVC compression) L and R images to the cable headend, and then to the home. Conceptually, the simplest transmission method is to send two full-resolution images (L and R). Or just like stereo audio, you could send a 2-D image plus a signal that represents the difference between the L and R images. But lossless methods require twice the data rate, so folks are examining lossy methods.
There are many approaches under consideration for lossy multiplexing that preserve the data rate. Most interest today seems to focus on sending separate L and R images at reduced resolution. For example, they could be sent as alternating separate frames, which reduces temporal resolution. Or they could be sent side-by-side in a single frame, which reduces horizontal resolution. Or L could occupy the top half of a frame, and R the bottom half, which reduces vertical resolution.
There is a form of multiplexing called checkerboard, which reduces diagonal resolution.
Finally, there is a method called anaglyph, which uses red-cyan or yellow-blue color filtering to carry two images within a single stream. For display, this multiplexing requires the use of glasses with colored lenses and was demonstrated during the Super Bowl telecast (to the discomfort of those without the glasses).
This approach degrades the color fidelity of the image.
So for delivery via satellite to the cable headend, and then to the set-top box in the home, there are numerous multiplexing methods possible, and there is no single standard preferred by the cable industry. In the absence of a standard, set-top boxes might have to support several of the multiplexing methods. But cable industry standards development activity has only just begun.
From the set-top box to the display, the HDMI interface is the most common (but not the only) interface in use. So while conceptually it’s possible for a STB to simply pass through the multiplexed L and R images over HDMI and let the display device deal with them, the STB will have to do some processing to produce 2-D analog outputs.
The HDMI interface was designed to carry uncompressed video. For 3-D, the STB might reconstruct full-resolution L and R images at twice the data rate of a 2-D program. Or the STB more likely might create uncompressed L and R images at the reduced resolution of the lossy multiplexing scheme used for transmission, maintaining the same data rate as 2-D. In either case, the uncompressed L and R signals will have to be multiplexed for carriage over this interface, which may or may not be the same kind of multiplexing used for transmission of the compressed signals.
The way HDMI works today, there is signaling that allows the STB and the display to discover the processing capabilities of one another and negotiate the best picture format that both the STB and display can support. This signaling is defined in a CEA standard – CEA-861. So it might be necessary to extend the signaling protocol to identify which uncompressed 3D multiplexing methods are supported by the STB and the display. It seems likely that a “baseline” 3-D multiplexing format will have to be defined for HDMI, and that all 3-D-capable STBs and 3-D-capable displays will have to support at least this baseline method.
In summary, while cable delivery of 3-D TV may seem inevitable, standards bodies like the SCTE and CEA have several years of work ahead to sort out these interoperability issues. Meanwhile, expect to see continued trials that are designed to winnow the options.