Originally Posted by darinp
Fortunately, you don't have to buy it for it to be true.
Fortunately, your proclamation of it being true does not mean it actually is.
The fact that the DLP doesn't have persistence doesn't matter since the persistence is in our eyes, just like it is with primary colors where the single chip DLP puts them up at different times, but our eyes see them as put up at the same time (except when people see the RBE artifact).
It would matter if you wanted to capture motion between subframes/frames that could organically occur in LCD/LCOS, as you mentioned "long exposure" in your previous post on this issue. That "long exposure" you referred to is a function of the way a sensor captures light over time vs shutter speed, not a function of the lens' resolving power. And it is the only time exposure would be relevant to this case.
With 2716x1528 only pixels it is the size of the pixels that needs to be resolved.
This is correct. And regardless of pixel shift or not, the physical size of pixels being projected through the lens does not change given the same DMD size/native res.
With the two 2716x1528 it is the smaller "pixels" that your eye sees that matter.
This is correct. It is your eyes/brain perceiving it is seeing the smaller pixels, not the lens actually physically projecting smaller pixels. It physically projects larger pixels at double the rate to virtually create smaller pixels in your mind's eye (not in the lens).
Interesting that you push these as 4k over and over, then seem to think they don't need to be able to resolve those 2k sub-pixels (which is what 4k pixels can be thought of as) they are trying to display.
The subpixels are resolved by your eye/brain. They are not resolved by the analog lens. The lens in an 2716x1528 XPR projector needs to be able to resolve 4M pixels twice as fast as a lens in a 2716x1528 projector. Lenses have no image persistence issues, nor do they combine data like a camera sensor or your brain, nor do they operate on any digital pixel grid - they simply project whatever light is projected through them in analog fashion. In this case, that would be 4M pixels in both cases. And if it can fully resolve those 4M pixels (no blurring, loss of resolution, light spilling into adjacent pixels, etc), then that is all there is to it.
You seem to agree that these XPR pixels are trying to show detail smaller than a single flash of 2716x1528 is capable of.
Nope, I have always argued they are projecting 2716x1528x2, creating a composite higher virtual resolution that your eyes/brain perceive on the screen - and that is exactly what is happening. There is not a higher physical native resolution. Your brain perceives more detail from the pixel offset, which your eye/brain perceives as a higher resolution image than is actually being projected at any one moment.
It is what ends up on the screen that your eyes are supposed to see that matters.
What ends up on the screen is 2716x1528 projected twice as fast as a non-XPR projector. Your eye/brain perceives this as a higher physical resolution, but it is not actually a higher physical native resolution.
I demonstrated this in post #14
of the thread I linked to. That is here
I'll include the 3 images from there to make this easier.
Would you say that all 3 of these pictures require the same lens resolution to show by themselves?
Poor example as this implies image persistence in a lens, and lenses have no persistence. The same resolution lens could display all 3 images if the 3rd image is rapidly alternated. Your eye/brain then combines these lower resolution images for a perceived higher resolution in the 3rd image.
It doesn't matter if the 3rd image is displayed all at once or displayed by putting up the 1st image and then the 2nd image, as long as it happens fast enough that a human doesn't know it is happening, just like with red, then green, then blue.
It definitely matters. The color wheel is an excellent example to disprove your theory, actually, so it is good your brought it up. If a lens existed that could only display one single color at a time, it would have absolutely no problem displaying a full color picture from DLP. Because it is not the lens that combines the colors, it is your eye/brain that combines the colors. Just like pixel-shifted resolution.
According to your argument you'd also need a lens that was capable of reproducing all colors simultaneously for DLP color wheel to work in reproducing a color picture, and that is also false. You'd just need a lens that could display 3 different colors, 1 color at a time, in rapid succession.
If anybody thinks that a lens only has to be good enough to show the E and then show the 3 because they are put up at different times (although fast enough that humans see one image) in the following example, they should rethink their position. A bad lens will show an 8 to humans when the image they were supposed to see is E3. That lens could have been perfectly fine for showing a large E and perfectly fine for showing a 3, but not fine for showing an E that close to a 3.
With a good lens that E and 3 in fast sequence should look like the following after persistence of vision, not like a blocky 8:
That means the lens needs to be able to resolve the black between the E and the 3, even when that fine detail only came from low resolution images that were displayed at different times.
This is also incorrect. If the lens showed a blocky 8 it would not be fully resolving the E or the 3 at their original resolutions. If there was a blocky 8 it means that the lens was unable to resolve the E or 3 without erroneously spilling light into the next row/column of pixels. Lenses are analog and thus are not bound by pixel grids like sensors or DMDs, so it can leak light by 0%, 0.5%, 1.5%, 2.5%, 10%, 25%, etc into adjacent rows/columns of pixels. You seem to be thinking in a manner where a lens is digital or bound by some pixel grid, and it is not as lenses are fully analog devices.
If the lens can fully resolve the E and the 3, in rapid succession it could also fully resolve the E3. If it can't fully resolve the E3 in rapid succession, the lens is not fully resolving the E or 3 separately.