Originally Posted by darinp2
This is AVScience. Sorry, but the premise that you can just watch some non-test pattern content you have and tell whether it is reproducing every pixel is not credible.
Other than that the images look really good to you do you have any actual proof that the 665 is reproducing every pixel properly? What is the basis for this claim that you keep repeating? Put another way, how would you possibly know whether it is reproducing every pixel properly?
If you are just going to say something like, "Because it wouldn't look this good", how would you know?
Do you think it is possible for an e-shift projector with 1080p chips to reproduce all those pixels properly too? Would you consider it credible if somebody said an e-shift projector reproduce all those pixels because they watched some content and the images wouldn't be that good if the e-shift projector wasn't displaying all those UHD pixels properly?
How can I say that?
1) I can SEE every pixel by walking up to the screen and looking at them.
2) I am an imaging systems engineer with 34 years of experience which included analyzing image quality of all types of film and electronic images on levels you can't imagine. This was one of the few companies in the world that assigned an image scientist to every significant program I worked on (professional and commercial products, not consumer products) had at least one image scientist assigned to the program... and nothing gets by those guys. They are very good at (and very patient) tutoring essential team members in all the specific issues with each product.
3) Because if the projector was high-frequency limited, I would be able to see blurred pixels and indistinct detail in native UHD/4K content and there ISN'T any. I can see single-pixel-level detail in UHD images. Period. And I mean I can follow a 1-pixel wide hair and follow the hair pixel by pixel in the (paused) frame.
4) Because no 1080p projector on the planet makes images that look this good. (there, I said it anyway because it is a valid observation)
5) Because I have had several e-shift models here and while they look better than "regular" 1080p, they produce softer looking images that are nice in their own way, but detail is not as sharply defined as the images produced by the 1000ES or the 665ES.
6) I sit close enough to the screen to see all the detail in UHD images and too close, really, for HD images, I see aliasing in HD images constantly because I sit so close to the screen (45 degree viewing angle for HD (THX recommends 36 degrees to see the most detail without artifacts like aliasing) and I move the screen (on legs) even closer to get a 52 degree viewing angle which is just a little more than 1.5 times the screen height. You can look at a page of text reproduced by a really sharp 1080p projector and see the raggedness on the edges of characters caused by the aliasing and if you get really close to the screen, you can see why you are seeing aliasing with a 45 degree viewing angle.
7) Because I can create a 3840x2160 or 4096x2160 computer image and make the entire screen all one color or 1 shade of gray except for 8 pixels that I can make some other color or black or white or some shade of gray, and in that frame I can find all 8 individual pixels when I get up close to the screen and look for them (I typically place them near the center, close to the 4 corners, and at the midpoints of each side of the screen with the pixels along the outer edge, spaced a couple of inches in (towards the center) from the very outer edge of the frame. A pattern like this doesn't violate Nyquist Sampling Theory and resolves perfectly, unlike alternating lines or checkerboards of black and white pixels that DO violate Nyquist Sampling Theory causing big problems for LCD imaging devices that rely on conversion of the digital signal to analog to actually drive all or part of the LCD imagers - but causing no problem for imaging technology like DLP that directly drives the chips with a digital signal. Alternatively, I can create other UHD or 4K images with a single color (or gray or black or white) background and draw a convoluted line around the frame with a 1-pixel-wide pencil tool in PhotoShop or Paint.Net and I can see that 1 pixel line clearly in every part of the screen from center to corners and points between.
I've got about 10 more points but it's early AM and I still have to shred the brisket I took out of the smoker an hour ago. This should be enough to assuage the science police.