About SMART and Contrast

SMART expects the light levels of a properly calibrated projector to follow the theoretical gamma=2.2 curve. If the contrast on your projector is set too high then the projector may not be able to produce all of the light that is required to follow this curve at the highest IRE levels, and the light output will essentially flatten out. So you really haven't raised the actual contrast ratio by raising the contrast past its optimal point. You will get more light out at the lower IRE levels and this may make the picture appear to have more contrast or to be brighter. The problem with doing so is that you will lose details and/or have color shifts in the brighter parts of the image. You won't notice this in many scenes, but in scenes where there are white clouds or white shirts, etc, you may notice that the brightest past of the image lack detail or are slightly off color. So setting the contrast properly is an important part of getting the most accurate picture.

Setting the contrast has two goals. The first is to raise the contrast until you have the maximum light output from the projector at IRE 100. The second is to assure that levels below IRE 100, e.g. IRE 90, have less light intensity than IRE 100. Ideally, the light intensity increases with increasing IRE level right up to IRE 100, and for the most accurate picture, it should follow a curve that represents a gamma value of 2.2. SMART II checks for this and gives a "lower contrast" message if it sees that the light output for any color at IRE 100 is significantly lower than one would expect from the gamma curve.

With most projectors I have reviewed, this logic is correct, as setting the contrast too high is the most common cause of this problem. I have seen several cases recently however where two other problems have cause a similar deviation from the theoretical gamma=2.2 curve, and where the SMART message to lower the contrast is not appropriate.

The two other things that can cause intensity deviations at high IRE levels. The first is that the detector itself can become non-linear or reach its limits at the brighter light levels, especially if it is too close to the projector. The second possibility is that the projector's gamma tracking is not quite right.

So this section reviews what to do if you are getting this message, and suggests some tests to distinguish between these three possibilities,

1. The contrast is too high,
2. A non-linear detector,
3. Poor gamma tracking in the projector

How to test what's causing the poor tracking

Lower the contrast a few points and do another run with everything else the same as the run that gave the message, including the distance to detector.

If at the lower contrast setting, the level at IRE 100 remains at or close to the value in the previous run, but the level at IRE 90 and below fall, that means that the contrast was a bit too high. You can tell either by looking at the charts as you switch between data sets, or looking at the data in the table in the analysis area. In this case, lowering the contrast is the right thing to do. Since he level at IRE 100 didn't drop, you can be assured that you projector is producing its optimal contrast, even at the lower contrast settings. If, on the other hand, the when you lower the contrast, the measured light level at IRE 100 falls significantly, more than experimental error, then you are probably being asked to lower the contrast below it optimum point. This likely means that one of the other factors is causing the apparent tracking problems. In this case, you should raise contrast back to the value where the IRE 100 level is at its maximum level. You can do this with the meter and the IRE 100 window without doing whole runs - sort of the other end of the bias settings.

If in the above comparison, the whole curve moves down without changing shape, then the detector is probably not the problem either and it is likely the projectors tracking that is off.

If the whole curve moves down and starts to straighten out, then detector linearity may be an issue. The sure way of eliminating the detector linearity concerns is to do another run, with every thing the same, but with the detector at a slightly greater distance, e,g, an additional 20 cm or a foot. If you still get exactly the same curve, but at a lower level, then the detector is fine and it is likely the projector that is off its gamma.
If the shape of the curve does change with the greater distance, then obviously detector linearity was the issue, and you should find a distance where the detector is operating below it maximum level. (Most photocells should not have a problem at 6 to 8 feet, ~ 2 Meters distance, but may if they are half that distance to the projector.)

So what can you do if it seem that the projector tracking is not quite right? You have two choices - live with it and don't worry about it, or attempt to fix the tracking problem. If you want to try and improve tracking at the higher IRE levels, you can do so by adjusting the 3D gamma settings, most likely only level 7. I had to do this on my projector, as my Red was much too high at IRE 90 and 100. Sounds bad, but it only takes about 10 to 15 minutes. Obviously going into the factory mode is not for the faint of heart, so you will have to make your own decision here. If you are interested in seeing what is involved in adjusting the 3D gamma correction matrix, click here.

One more note.

Some people may prefer a higher contrast setting that those indicated by SMART. I believe that what most people will find is that as you raise the contrast, you may not get any more light out at IRE 100, it already at its maximum. So you really haven't raised the actual contrast ratio by raising the contrast past its optimal point. You will get more light out at the lower IRE levels and this may make the picture appear to have more contrast or to be brighter. The problem with doing so is that you will lose details and/or have color shifts in the brighter parts of the image. You won't notice this in many scenes, but in scenes where there are white clouds or white shirts, etc, you may notice that the brightest past of the image lack detail or are slightly off color.
The trouble with using the Avia test to set contrast is that you can still see the almost white bars long after their color has changed. So simply seeing the bars is not a good enough test of not having contrast too high. You should look for a color shift in the "white" background or the bars as you raise the contrast value at stop and go back when it occurs. Neither the bars nor the background should ever show any color or color change. The use of measurements and SMART saves you from the need to make this somewhat subjective judgement, but I'll admit it errs on the side of accuracy rather than maximum apparent contrast or brightness.

SMART II v1.2c

SMART Version 1.2b use IRE 50, 70 and 100 to test for the contrast being too high. SMART Version 1.2c. only uses IRE levels above 70 and is therefore hopefully a more reliable test for contrast as it will be less affected by detector linearity.