In the Display Daily of Sept. 21, 2007, I wrote on the lofty figures for on/off contrast published by manufacturers of the latest projectors for home theater. Complaints are that these figures, ranging from 10,000:1 to over 100,000:1, are meaningless, or worse, misleading. The article suggested that additional measurements, across a range of features sizes are needed to better characterize image quality and information content. Apparently striking a nerve, it was interesting to receive more responses than to any previous article, some quite impassioned. Here’s what the readers had to say:

John DiLoreto
Analyst and Editor for
Insight Media

Most, about 90%, were in agreement that overuse of the sequential contrast figure was a detriment to the industry.

"Hear, hear!," from one reader, and "I completely agree with you. The on/off contrast is not the value of interest," said another. "You’re preaching to the choir, as far as I’m concerned," wrote Alfred Poor at the HDTV Resource Center. "Full field contrast measurements rely on an image that carries precisely zero bits of information, which is hardly a useful metric."

Our associate Pete Putnam writes, "Good article. Contrast "buzz" has replaced the lumens nonsense from the late 1990s." He wrote an article in 2003 on contrast in monitors, and it supports the point: http://www.hdtvexpert.com/pages/shmontrast.htm.

Almost in violent agreement was Martin Euredjian of eCinema Systems who wrote, "I don’t even know why we use sequential measurements besides to confuse people with over-stated marketing specs. When was the last time you sat in a dark room watching a movie and had to look at a full frame of white or black for any given time?"

But on the other side was Matt Cowan Chief Scientific Officer of REAL D who wrote, "I think you downplay the importance of sequential contrast…for movies it is large," adding that "Movies are dark media. A typical movie frame has an average picture level of 5% or less."

He further maintains that, "The human eye can manage a simultaneous luminance range of around 100:1… This might tell us that any contrast over 100:1 is overkill."

His point is not widely held, however. In paper 3.2 SID 2007 digest by G Damberg et al state that the human eye "will adapt and can observe about 5 orders of magnitude of luminance simultaneously." Others reported a similar ability for the human eye’s recognition of simultaneous dynamic contrast. References were also made to movie film having intra-image contrast of 4 to 5 orders of magnitude, that’s 10,000 to 100,000:1.

But how well can you get this contrast out a projection lens and onto a screen is the question. Fundamentally, sequential contrast is incomplete. So where do we go from here?

The most elaborate scheme was proposed by Candice Elliott, Founder & CTO of Clairvoyante. "I like the idea of MTF(a), the area under the MTF curve, but feel that an MTFR(a) would be better. This is Modulation Transfer Function Ratio, which is also known as the Michaelson Contrast. The area under the curve for MTFR is the MTFR(a). Michaelson Contrast comes closer to the experience of contrast by the viewer than does MTF or conventional contrast ratio."

I must agree. This approach is very similar to one that I proposed at the Flat Information Display Conference in 1999 that used as a lower bound the Contrast Threshold Function of the human eye. In either case, the area between the curves gives a measure of the information that can be contained in the image.

How practical is this? Until we get into the testing business ourselves we are reliant on others. As Alfred Poor explains, "Your ‘area under the curve’ suggestion is a good one, but far too costly to implement. And don’t rely on the magazines to test all the projectors on the market; they don’t have the resources. The ideal answer would be an independent testing agency that is industry-supported, and the manufacturers send production units for testing (no lab queens, please) in order to get a certified measurement. And we could have them measure lumens while they’re at it."

Whether or not such testing might happen depends on a number of things. Most critically, in our estimation, is that we, as an industry, need to universally understand these concepts and agree that it is important to measure. Magazines, such as Widescreen Review and The Perfect Vision, painstakingly measure something as obscure as "color temperature vs. IRE" (before and after recalibration, using their ISF-trained experts.

What were asking to be done with measuring information content is no more laborious. We just have to think it’s important enough to ask for it.

Reply to the author