If this is your first visit, be sure to check out the FAQ by clicking the link above. You may have to register before you can post: click the register link above to proceed. To start viewing messages, select the forum that you want to visit from the selection below. |
|
|
Thread Tools | Display Modes |
#1
|
|||
|
|||
Digital enlarger DeVere 504 DS
The DeVere 504DS is a unique application, but piezo movement technology is
being used in digital camera backs to allow a relatively small CCD to create a large high resolution image file. If you do a Google search on "digital camera piezo movement" you'll find quite a lot of information. "Jorge Omar" wrote in message ... Well, here we have something unusual. Jorge |
#2
|
|||
|
|||
Digital enlarger DeVere 504 DS
"David Strip" wrote
You appear to be assuming that as the LCD is shifted, the image is constant, and that the resulting printed image is just a blurred image of the LCD. No, I wasn't, though I may have given that impression with references to 'etched glass' and 'vibrating enlargers' producing much the same effect. I was referring to elimination of the jaggies caused by imaging sharply defined pixels at 300 DPI. To achieve a smooth gradation between pixels you would, of course, have to print at intermediate pixel intensities at intermediate positions to achieve a smooth gradation between pixels. DeVere clearly states that this is what they are doing: Lots of discrete exposures at different display intensities with a filter and each exposure really being three exposures, one through each of the positions on the filter wheel. The same effect, though, can be achieved by with a motion that varies velocity with position, much the same as burning a print: the burning hole spends a lot of time in the middle and less and less time as it is moved farther from the circle. While I have no idea what DeVere is doing, other systems have been built around this concept in which the LCD is stepped to discrete positions and a different image is projected in each position. Since LCD's often have a fill factor of ~25%, you can image 4 distinct pixels/dots by moving the LCD to four distinct positions. Agreed, if an LCD has less than a 100% fill factor then a step and repeat puts down a higher resolution image with more information. The net effect is to create virtual pixels. Think of it as 100% fill. The resolution is not increased "apparently", it's increased for real, assuming the imaging optics aren't the limiting factor. I wouldn't say it is increased. The pixel size is the same, and so the minimum feature size is the same. It is just that without motion 3/4 of the pixels are missing. So, lets theorize a system with pixels with a 100% fill factor, as that is what has been accomplished by DeVere. Now we are now where we have started ... 300 sharply defined spots/inch on the paper. And the problem is how to blend the boundaries between the spots/dots/pix to make the whole pleasing to the eye. To do this one can move the screen in a smoothly in a 2-D Gaussian manner (shudder), approximate same with discrete exposures, or approximate the same by utilizing an optical system that produces a Gaussian blur of just the right sort. My guess was they already had the step and repeat action just to achieve pixel fill. And frosted glass or a coke-bottle lens just didn't cut it, so the solution was extra exposures at intermediate pixel positions. Now, back to this idea that moving (packed) pixels around will create resolution that is greater than achieved by stationary pixels. I propose the following experiment: Use a 1 pixel system in one dimension as the machine/printer/enlarger/ latest-term-thought-up-by-marketing. And further simplify the system by constraining the pixel to horizontal moves - sort of like moving to "Lineland" from "Flatland". If it works for one pixel and one dimension it should be generalizable to many pixels in 2 dimensions. This one pixel can print any shade from white to grey in an additive manner: 2 white exposures = white, 1 white + 1 grey = grey, 2 grey exposures = black ... It is, obviously, possible to make a black one pix spot next to a white one pix spot. A repeating pattern of white/black pixels is your basic MTF input, though really one should use a sinus density profile a square wave profile works about as well. Now, with multiple exposures, is it possible to make a periodic image that is of a higher frequency than that above? I.E.: one where the distance from the center of one black area to the center of the next black area is less than two pixels. If we allow the pixel to print with opaque ink then black and white then it is easy to see that zebra patterns of any frequency can be made can be made. But only at integer multiples of the base frequency. One leading cycle, starting from a blank stretch, can be made at any frequency, but the pixels will make a mess of the image to the right (assuming the pixel moves left to right) so that an arbitrary frequency can not be repeated smoothly. The ability to image frequency multiples is the same effect as size of the lens's bokeh (the size of the pixel) causing apparent lens resolution in an MTF test to suddenly spring to life again after having dissolved into pure grey. The ability to image at discrete frequencies is called an artifact. Rather than producing more detail it produces moire patterns in the output. The moving pixel problem is a variation on 'box car averaging' filtering, a well known technique in digital signal processing. It produces a low-pass filter with a comb-like pass band after the first filter zero. The mathematical literature for box car functions is hundreds of years old. Just search for 'box car filter' with google. In summation the effect of moving pixels around is a blur, no increase in information transfer is possible. Sorry, no free lunch. And again, there is nothing new under the Sun but that thinking makes it so. -- Nicholas O. Lindan, Cleveland, Ohio Consulting Engineer: Electronics; Informatics; Photonics. |
#3
|
|||
|
|||
Digital enlarger DeVere 504 DS
The proof is in the printing. Be nice and share. I look forward to receiving my complimentary DeVere 504DS. When's it arriving? |
#4
|
|||
|
|||
Digital enlarger DeVere 504 DS
"Claudio Bonavolta"
I've seen this enlarger [DeVere 504DS]... The basic technology is from MuellerSOHN, DeVere is incorporating it into its own product, as are several firms in the photolab business. http://www.muellersohn.de/LCD_E.html Looks like the digital LCD mini-lab printer is on its way to a drug-store near you. And the name of the game here is _speed_. Resolution only has to be consumer grade. On another front, Intel has developed high resolution LCOS (Liquid Crystal On Silicon), slated initially for projection TV, but going who knows where.... a minilab near you, maybe. http://www.reed-electronics.com/elec...668?text=intel -- Nicholas O. Lindan, Cleveland, Ohio Consulting Engineer: Electronics; Informatics; Photonics. |
Thread Tools | |
Display Modes | |
|
|