Technological advances in imaging and content production may render the current standard of 256 gray levels per color in most modern displays inadequate.

by Walter Allen

IN the January 2008 issue of Information Display magazine, Aris Silzars authored a thoughtful column on "Technology Asymptotes." At the end of the piece, he suggested that we may have reached the "good enough" point for home-entertainment displays with the current HDTV standard of 1920 x 1080 pixels at 256 levels per color. While I agree that there is little motivation at present to extend that resolution level (even 1280 x 720 pixels is adequate for most home applications), 256 levels per color (8 bits) may not be fine enough as displays and sources evolve and extend their contrast ranges.

Just how good is the 8-bit system? As evidenced by acceptance from the great majority of today's consumers, it appears to be good enough for present-day displays. But on many of these displays the steps between the lower individual discrete levels can be readily detected, and this can manifest itself as posterization in the imagery. Making matters worse, of the 256 levels currently available, in many cases, such as for DVDs, only 220 levels (16–235) are normally used, with the remainder serving as headroom. Thus, while 8 bits may be good enough for today, it is hovering near the "just good enough" end of the range, and new technology on the horizon may push it into the "not good enough" category. These new technological advances are leading to higher-contrast displays and the accompanying ability to display significantly more levels of luminance and color gamut. The big question then becomes how many of the great majority of users will appreciate and demand those additional shades – enough to increase the standard beyond the 8-bit level?

For a little background into the need for range in luminance and contrast, let's travel back to the time when I first became seriously interested in 35mm SLR photography. Initially, life in this image capture and display world seemed so good. I had the tools to compose a picture just the way I wanted, perfectly framed and exposed, and could select the film best suited for a particular photo session. When making an enlargement, I could further refine the process by dodging selected areas to enhance the exposure. But, even then, all too often it seemed there were times that I came upon awe-inspiring sights – spectacular sunsets or sunrises, scenes by a babbling brook in the forest, gorgeous shots at the beach, glorious fireworks, or a winter wonderland – when my camera would fail me. I would take a series of shots, carefully bracketing exposures and precisely composing the images, thinking about the incredible sights I was recording for posterity. After processing the film and viewing the results, either slides or prints or enlargements, the results were all too often a letdown – the impact and depth of the original scene would clearly be missing. The pictures would be good, maybe even very good, but would invariably fail to reproduce the drama of the original moment.

What was the most important ingredient missing from those reproductions? It turned out to be a restricted range of luminance levels and the related overall lack of contrast. The human visual system has evolved to a highly refined and truly amazing dynamic sense. We can perceive an incredible range of luminance levels, from seeing by starlight to viewing clouds backlighted by the sun. The range is well in excess of 1,000,000,000:1! Now, of course, there are caveats – all the range is not available at once, as the eye has adaptive behavior that needs time to be enabled, and within smaller areas of a scene we can discern luminance differences of only about 1%, and anything dimmer than about 1% of an adjacent bright area is seen as black. But the eye can easily detect luminance variations of well over 10,000:1 in non-adjacent areas of larger images.

When I would look at the prints (or slides) that were intended to recreate those original, awe-inspiring sights, I would be looking at a medium that was probably delivering less than 100:1 in contrast, and certainly well less than 1000:1. Also, in both the brighter areas and the darker areas of the image, the variations in luminance would be compressed and thus appear flat and lack depth. No wonder I was disappointed. The problem could be found in both the capture and display media – not enough luminance range can be captured in a single film exposure, and even if it could be, a conventional print, relying on the absorption of incident light, still could not display the necessary range.

Now let's return to the present, in which the situation is changing. In the world of photography, not only are sensors evolving to capture a greater luminance range, but a bracketed series of captured images can be further combined to create very-high-contrast source material. If we look to computer-generated images, there is virtually no limit to the potential "luminance" range of the content. On the display side, high-contrast-capable displays are showing up on many fronts. We have high-contrast plasma displays, LCDs backlit with arrays of controllable LEDs, LCOS projectors, DLP projectors with outstanding black levels, and the promise of OLED displays. All these displays can render wide contrast ratios and are capable of very-wide dynamic ranges, but in most cases either the display device or the content is limited to 8 bits per color. If more granularity of the luminance levels was provided, with added bits/luminance levels, the images can be significantly enhanced with additional realism and depth. Viewing such displays with current standard (8-bit) programming all too often results not only in posterization, but also in significant areas of the image appearing as completely black, without the "shadow detail" that would have been present in the original scene. As these displays make their way into the mainstream, the source limitations of only 8 bits per color will become more apparent, and just maybe will no longer be considered "good enough." At that point, users may clamor for a little bit (or two or ?) more.

The road to increased bit depth in displays is being paved by developments such as Deep Color in HDMI 1.3, by the Digital Cinema Initiatives, by PC graphics systems, and by capable digital cameras and image-processing software. While the road has potholes here and there, such as the 8-bit constraints of MPEG2 in DVDs and HD, not to mention the fact that most mainstream viewers do not currently perceive a problem with a lack of levels, I hope and expect that this new road to expanded luminance range and depth will be well-traveled in the coming years. Then I, and the overwhelming majority, can look forward to seeing dramatic sunsets and fireworks on our home displays. •