| |
| |
|
The quality everybody wants first from a projected image is brightness. And certainly some amount of projected light is always needed if we're to see the image. But how much brightness do we really require? If we have lots and lots of brightness, because we've got a very powerful projector, aimed at a modestly sized screen, then we might suppose that we're going to enjoy quite a good picture. Let's find out if that's necessarily so.
On the first screen the letters were exactly as bright as the surrounding area. On the second they are many times darker. Thus there is a large perceptible difference between the brightest portions of the second screen (the background) and its darkest elements (the characters). This differentiation between light and dark is the essence of contrast. From Figures 1 and 2 we understand that the absolute value of the measured luminance of any display is no indication of its contrast. A screen reflecting 15 foot candles doesn't have to have better (or worse) contrast than one showing 200. Contrast depends exclusively on the ratio between the maximum and minimum light levels within any image.
We determine this ratio according to the formula: ------------ Min Let us suppose that we have a slide image of a picket fence (Figure 3) - really just a series of alternating black and white bars. If we use a photometer to read the amount of light reflected off a screen where there is a white bar, we could get a number like 1200. When we point it at a black bar we get a number like 3. Plugging those values into our formula we find this display to have a contrast ratio of 399:1. We can believe this extremely high ratio because when we take the slide out of the projector and hold it up to the light, we see that its white bar areas appear virtually transparent and its black bar stripes are nearly opaque. So the difference in this slide's transmission density can actually be four hundred to one. Can video projection provide a contrast ratio of that magnitude? Regrettably not. CRT projectors generally produce contrast that is less than 100:1 (and often much less). Why this is so has much more to do with these devices' capability to project black (the Min value) than their ability to project peak white. Blacks emanating from video projectors are really closer to shades of dark gray than true black. Thus if we take that same picket fence image and put it through a CRT projector we will observe the effects of light intended only for the white bars leaking over into the dark bars and thereby significantly reducing the contrast ratio. The primary cause of this phenomenon is that the phosphors on the surface of the CRTs emit light in a pattern that is much broader than the tightly focused beam of electrons which excites them. Thus their action is similar to a diffuser and the light energy they transmit spreads in a pattern that is roughly Lambertian. This scattering makes it extremely difficult for CRTs to keep dark areas separated from light ones. Notice that the challenge of maintaining a high contrast ratio occurs in the control of the Min value and not, as might otherwise be supposed, the Max. The CRT theoretically can be made to produce every bit as much luminance as a slide projector, so the Max need not decrease. It is the increase to the Min which debases the contrast. And, as is evident from the formula above, even a small increase in the Min term will have a dramatic effect on the ratio. Another way of thinking about contrast is to observe that it's the only attribute of a display system that can suffer from too much light. As we have seen unwanted light from the projector is bad enough; unwanted light from other sources can be calamitous. If the houselights were suddenly turned on while we are in a theater watching a movie we would immediately notice that the screen which just a moment before had seemed so lustrous and bright has now become hopelessly "washed out." Of course the screen itself hasn't changed. It continues to do an excellent job reflecting light incident to its surface. But our eyes have changed, quickly adapting to the suddenly increased average light level before them. Now when we look at the screen the light coming off it containing imagery (the movie) has to compete with the reflected room light. Both the screen and our eyes add the two kinds of illumination together because there is no way to distinguish between them. We are left looking at all light and no dark. And without that dark, there can be no contrast. And without contrast there can be no imagery. Another way contrast can affect image quality concerns the placement of a display within its audience's field-of-view. Since a screen rarely fills up our entire visual field, the experience of watching it is going to include some peripheral awareness of the surrounding environment. A screen placed before a dark wall will appear to be brighter than one in front of a white background. A screen with a black border will seem more attractive than one that's white all the way to its edges. We often confuse these impressions with questions of brightness. But they are actually perceptions of contrast. In order to ensure a real impression of brightness the contrast between a screen and its surrounding field must be at least 5:1. If this minimum ratio is not achieved, human eyes will not judge the screen to be "bright", no matter how great its actual luminance. Brightness, then, can properly be understood as a comparative term. An image will be perceived as "bright" only when it is seen to be brighter than something else. And if the "something else" is a totally dark room, very little actual luminance off the screen will be required to produce a strong sensation of "brightness." Unfortunately total darkness is not practicable in the majority of applications which employ projection screens. Some amount of extraneous light is almost always present and the question becomes, what can be done to minimize its impact on the screen's contrast ratio? If our screen is a front projection display, our options are limited to trying our best to keep energy from all the light sources other than the projector from striking the screen. Carefully recessed ceiling lighting, or properly shaded task lights, for example, will not excessively diminish an image's contrast except insofar as some portion of their illumination is directly incident to the screen's surface. With rear projection screens the available options are less limited. Since all rear projection screens are designed to transmit light rather than reflect it, the majority of the light striking a rear screen's front surface is not reflected back at the audience. Thus it does not compete with the light projected at the audience. Better yet, rear projection screens actually can improve the contrast of a display by the inclusion of darkening agents in their diffusion. Da-Lite has been a leader in the development of these High Contrast tints and now offers them throughout its Polacoat line of rear projection screens. Here's why they help. If we determine that the Max from a conventional rear screen is, say, 100 units of brightness and the Min is measured at 5 of the same units, we know from our formula that the Contrast Ratio is 19:1. Now suppose we put a little bit of colorant into the diffusion; just enough to reduce the transmission by 2 brightness units. Our new tinted screen has a Max that has been reduced to 98 and a Min that's been lowered to 3. It's true the new screen is a little bit (-2%) less bright, but the benefit of that cost is a contrast ratio that has jumped (72%) to nearly 33:1. If we compared the two screens side by side, the brightness differential wouldn't even be noticeable while the improved contrast would be recognizable instantly.
Across the entire range of visual displays, brightness is not the element which most influences image quality. It is contrast, the degree of separation and distinction between the light and the dark elements of an image, which most strongly affects our perception and ability for visual discrimination.
M. K. Milliken, Jr.
CONTINUE: ANGLES OF VIEW |
Figure 1 is an accurate representation of an illuminated projection screen flooded with light. Right across the center in big, bold, sharply resolved characters are the words HIGH GAIN. There are no other light sources impinging on the screen. How do we like this very bright, very uniform, high resolution picture? Would it be better if we made the screen a rear projection display? How about if we changed its aspect ratio? Neither alteration would help at all.
So whatever could be the problem? With all that light, how come we can't see anything?
Now look at Figure 2. It's the same screen, displaying the same message, but this time we can clearly make out the text. What's changed? Now we can read it not because we added more brightness, but because we took some away.
Figure 3 shows the picket fence we discussed above. Scanning from left to right we can see the contrast degrade from the first picket to the last. The rightmost picket appears foggy and difficult to make out not because there isn't enough light illuminating it, but because there's too much. What is missing is something to bring its darkness (the Min) back to the level exhibited by the leftmost picket.