Originally Posted by Noah
I remember reading a thread here about a test that was done to attempt to lower ANSI contrast (while keeping other conditions identical) and determine to what extent a viewer could perceive the difference in A/B testing. In a batcave, a screen or white sheet was raised and lowered behind the viewer without their knowledge. I don't believe anyone was even able to tell whether the reflective surface was present or not.
I did that with a couple of friends. We were at Kris Deering's house and I used two portable Da-Lite High Power screens that I have. I figured since they were retro-reflective the back screen should wash the front one out pretty well. I don't remember how much the ANSI CR went down when I put the back screen up, but I think it was quite a bit. We didn't spend a lot of time doing the testing, so it is possible that with more testing we could have found out a lot more.
I recall that we tried some different scenes from Hellboy as well as some intro kind of image where most of the screen outside the center was black. I think we were using a Sharp 20k mostly (or just the Sharp 20k), which had high ANSI CR, but not the best on/off CR. If we had something with really high on/off CR too maybe the system ANSI CR differences would have been easier to see.
If I did it again I might look hard for some scenes with shadow detail that might disappear. I just don't recall how much time we spent on that. With nobody could pick out the difference between the back screen being up or down I think we gave up fairly quickly.
Originally Posted by Seegs108
Could you extrapolate on why DLP looks the way it does during brighter scenes compared to LCOS? Does it have nothing to do to with ANSI contrast? As you've said incremental differences can't be perceived by the human eye and for it to be noticeable it needs to be magnitudes higher for it to be.
That could be one factor, but I think a big possibility is that DLPs had (and may still have) a tendency to magnify noise. If the source had a spike of noise in it the DLP may show it 3 times as much as it really existed in the source. At least with a certain generation of DLP chip. I haven't looked at that much in years.
Back when I had a Planar 8150 I think I had an RS40 at the same time (although I could be mixing my JVC models up since I've had many of them). My memory is that if I used a little bit of sharpening on the RS40 (for some reason 10 or 20 in the menu comes to mind) then the sharpness between the 2 looked very similar to my eyes.
I do recall a bright desert scene with a dark jacket in one of the James Bond movies where a DLP did a better job with the black area to my eyes, but it seemed like I had to struggle to find scenes where I thought that and didn't have to struggle to find scenes where I thought the on/off CR advantage of the JVC showed up. Same kind of thing when I had the Samsung 900AB.
Getting colors equalized was often a problem, but when we would shoot 2 projectors side-by-side on the same screen with half of each image covered and one image digital swapped (so both projectors were showing the same half of the current frame), with an attempt at getting the same ft-lamberts from each, much of the claimed sharpness advantages of the DLPs over the JVCs didn't show up to Kris, myself, or some of my other friends.
I recall a Kris's JVC looking sharper than my Sharp 20k to all of us, but some of that may have been related to color differences that we hadn't been able to get rid of and at the time weren't willing to spend whatever effort (and possibly other equipment) it would have taken to equalize the colors.
High resolution MTF and related areas could also be a difference, but that one gets a little bit complicated too. There are the MTF numbers that may not tell the whole story, such as how square the pixels look, etc.
I know I'm not giving you solid answers, but like I said I don't claim to know every aspect of it, and most of the people I've run into who claim to understand everything about this subject matter are pretty clueless IMO. I could make a list of the ignorant statements I've gotten on the CR subject matter from people who are supposed to be expects, but I think I'll skip that (for now).
--DarinEdited by darinp2 - 9/20/13 at 5:08pm