Originally Posted by darinp2
This is the part I'm not seeing. The assumption of softening the image. Now I could be wrong and if so you or somebody else will probably show me the error of my ways so I will finally get it, but at this point I don't.
Are you assuming that the image will be softer and if so, based on the edge of the lens having less quality mostly? As I mentioned, somebody who actually measured MTF found higher MTF at the larger image end of the zoom, not the longer throw end. At least for 1 projector. I'm guessing they were measuring near the center of the image, but even there, why would the MTF go up at the end of the zoom that is generally thought of around here as the worse end for sharpness?
By anomoly size I mean things like bumps on the lens or cracks or divits in the surface of the lens that can be seen when it is put under a strong enough microscope. A lens may feel like it is flat, but isn't really down at the small scale that is relevant here.
To me this is related to why a 4K projector requires a better lens to resolve its pixels than a 2K (or 1080p) projector does. The anomoly size relative to pixel size seems relevant in both cases.
If the same lens was used for a 4K projector and a 2K projector, but the 2K projector was put at the long throw so that the center inch was used for 2K pixel width and the 4K projector used the center 2 inches, then in both cases the center inch would be used for 2K pixels. If the 4K chip was twice as wide with same fill ratio and related parameters, then would the sharpness on screen differ for the center 2K pixels for those two cases? If so, why?
If they don't differ, then for those center pixels zooming the 4K image to have a bigger exit image improved the quality for those pixels. I'll save the logic of that one for later in case that doesn't make sense, but it is related to the lens not being high enough quality for 4K if only 1" of it was used, but being high enough quality (at least for the center 2K pixels) if 2" of it was used.
Thanks. I appreciate it.
I definitely agree that 4K requires better lenses. A lens that resolves 1920 lines may not resolve 3840 lines. An analogy is todays mega pixel DSLR's easily show weakness in some older lens designs that are mitigated to some extent with lenses specifically designed with those mega pixel cameras in mind (CA is one of the biggies that I've noticed).
As to the assumption of softening the image as you move out from the center of the lens....
First off, I am not an optical expert but I have some experience with 35mm lenses and I think similar principles are involved in projector lenses. I've attached two Nikon MTF charts for their well regarded 24-70 f/2.8 professional zoom lens (I've also attached Nikon's key to reading these charts). Notice how resolution and contrast vary (worsen) as you move away from the lens center and also notice that there is also considerable variation with zoom position. Of course some lenses perform better than others but generally speaking contrast and resolution degrade as you move from lens center (prime lenses being much better than zooms, etc). With todays 36 megapixel DSLR (Nikon D800) I can certainly see that the image softens and contrast lowers along the outer portions of the frame. I have never had (or seen) any lens where this did not happen (the charts shown are made at f/2.8.....stopping down the lens improves performance across the board but the outer portion will always remain lower than the central portion).
I'll also note that most of Nikon previous [tele] zoom lenses performed best at minimum zoom (sharpness and contrast fall off at higher zoom levels) but recently Nikon has introduced zoom lenses that are better at the zoomed end. So a lens can be designed to work at best either end of the zoom range....it may well be that the projector lens in the unit where MTF works best at short throw vs long throw.
One other comment.....The nikon D800 has 36 Megapixels in a 24x36mm area.....this far exceeds any 4K projector density of 8.3 Mega pixels. And with w DSLR lens the exit image is going to be a circle that is larger in diameter then the diagonal of the 24x36mm frame (the diagonal is ~43mm). In this case there is significant pixel density along the edges of the frame.
Many projectors have shift lenses......Nikon and Canon make these too.....a shift lens has to output a significantly larger image circle. Consider the 35mm system wherein the image circle must be at least 43mm for a non shift lens to avoid vignetting -- it needs to be much larger for a shift lens (the 24x36 mm rectangle must lie within the image circle to avoid vignetting and as the image is shifted the position of that rectangle relative to center of the lens circle changes). Sharpness, contrast, CA are all affected when the lens is shifted (as most projector owners have likely noticed when they use lens shift).
It would be interesting to measure MTF at various positions on the screen.....I may be full of hogwash but I do believe that MTF would vary more on the outer portions of the image relative to the center of the image (although I acknowledge the central portion is more critical that the outer portion).
One final thought for now...there are always exceptions to rules and generalities....I suspect that the unit with the higher MTF at the larger image end of the zoom is the exception.....
To clarify, the horizontal axis on charts below go from 0-22...however, the edge of the image occurs at 18mm - the reason the chart goes to 22 (or thereabouts) is to show corner performance (half the diagonal is roughly 22mm).Edited by Geof - 9/18/13 at 7:32am