Originally Posted by darinp2
I just noticed something that seems a little odd. I noticed that you said the on/off CR was about 2000:1 with the DB off. But your chart of intra-image CRs as gray is swept from 0 to 100%stim shows less than 1000:1. I had been thinking that you were using just a little bit of non-black in that image, so that with a high ANSI CR compared to native on/off CR projector like this you would get close to the native on/off CR for your measurements in that image when you used 100%stim. Do the images for that one contain just a little bit of non-black? Do you do your measurements for black near those non-black parts? I'm trying to understand how you got so much less intra-image CR than on/off CR there without DB on and then what that means for the almost 2k:1 intra-image CR you got with DB on.
This is a very good question. The test pattern uses perfect black (RGB=0) rather than video black everywhere except for 4 small rectangles that surround the probe area being measured. The four small rectangles contain the white stimulus that is ramped from 1% of video black to video white (235) and they are relatively close proximity-wise to the probe area being measured. They are close enough to where even a high ANSI contrast projector like the 8200 can have a measurably raised black level compared to full field black. In this particular case though the 8200 has such a high black level floor (DB=off) that it looks like it's capturing the floor throughout the range (this can be seen from the graph of the black readings). So it should end up yielding very close to the numbers that I measured for on/off.
Looking through the data I can see that the white reading is low throughout and it's not something that I'm surprised at. As I mentioned, sometimes I would see situations where the lamp brightness would be inconsistent even though nothing had changed. For example, I once measured ANSI contrast at 720:1 with a white reading of 1390 and black at 1.93, later the lamp output inexplicably changed so that the white reading dropped to 760 which gave a black reading of 1.37 and a contrast of 553. This was independent of any modes or adjustments being set and happened after the lamp had cycled (flickered and gone dark for a brief instant) due to the bug that I mentioned. I had reported this to Optoma and I think it's related to test pattern bug which they have confirmed. In fact, I've been calling it lamp flicker but what is really going on is the lamp steps down quickly in brightness and then goes black briefly before coming back up and I think something is going on with the ramp down where it doesn't come back up to full brightness.
Because of the lamp and iris inconsistencies I had to reduce my expectations a little as far as the data that I took and rather than comparing it to other projectors like I had done with the static low APL test patterns. Instead I focused on the dynamic data only and in particular only relative measurements between modes, that way if the lamp was reading low or there were other issues at least it was doing so consistently between modes. It was a disappointment though to not be able to add it in the mix of projectors measured in the contrast thread as far as static contrast vs luminance. The 8200 is an interesting combination of high ANSI and average on/off and it would be interesting to see where the crossover is at similar to what we had done with the Sharp 20K, RS1 and Sony VW50.
To ensure that the readings were reproducible, I took multiple readings at each point to ensure that it was repeatable and then also went back and repeated the entire suite of measurements a second time to ensure that the measurements were consistent and that the peaks and dips weren't due to something that had changed between readings. Both the white and black readings were consistent the second time around.
Given all of the issues though, I really wondered how much data to post and if it would be better to post just ANSI and on/off numbers, but I thought that the results showing the large amount of intra-image contrast benefits and also black reduction were worthwhile. All in all though I'd really like to repeat all of these results once Optoma fixes the lamp/test pattern bug.