Originally Posted by krichter1
It must have been a looong day debugging Cobol code (my brain… she is mush!), but I literally read this post like five times when I was left with…
Sorry. Not sure if this will help.
Here is an article by SoulNight from here where it shows graphs where on/off CR is on the left edge, ANSI CR is on the right edge, and in between are a bunch of intra-image CRs for different patterns:
The DI doesn't help ANSI CR because the ANSI CR test is a test of a 4x4 checkerboard of white and black, where the DI doesn't close.
The DI does help the CRs between say a 5% window and black (the intra-image CR between those 2) on a test pattern with just those 2 because the DI can close and the projector can modify the signal.
As a simple example, if a 5% window would be 0.1 nits and black 0.01 nits with the DI off (for 10:1 intra-image CR) the DI system could close the iris down so that half the light was blocked. This would take both to half their values, but the projector could just tell the chips to put out twice as much light for the 5% window so it was still at .1 nits off the screen, but the black would now be about 0.005 nits off the screen, for 20:1 intra-image CR.
The CR for a 5% window on black is most affected by the on/off CR with the projectors we normally discuss since 5% video should be about 1/500th of white. So, even if you had a great 500:1 ANSI CR, if that was also the on/off CR then the black floor would be where you would like to put the 5% video level. So, very little CR between 5% video level and black.
The JVCs can do 20:1 or higher with a checkerboard of 5% video and black because they have high on/off CR, not because they have high ANSI CR.
I measured a 5%/0% 4x4 checkerboard on a JVC RS400 tonight and with the iris open and DI disabled I got 23:1. Enabling the DI increased that to 67:1. Starting with the iris at -15 resulted in 33:1 and that wouldn't change with the DI enabled because the DI doesn't close any further for that pattern if the manual position is set to -15.
Did I confuse you even more?