watson@ames.arc.nasa.gov (John S. Watson) (11/16/88)
[this discussion started in comp.graphics] Here's how I get 24 bit photographs of my images from my 8 bit Sun screen (The basic idea is to triple expose the film, once with each R,G,B channels). First I set up my 35mm camera in front of the our Sun's 8-bit color monitor, on a tripod. I level the Sun screen and make sure the camera is pointed at the center of the Sun screen. A good sturdy tripod is best, but I use my ultra-cheap ultra-lite backpacking tripod. I also connect my remote shutter controller, and turn on the 2 second delay to minimize vibrations. I use a telephoto lens of about 100mm to minimize the "wide-angle" distortion. Once everything is set up I run a program I wrote that take the three red, green and blue image files and make a grey scale "intensity" file. Then I display this intensity image. Next, I turn out all the lights in the room and take a meter reading of the intensity image. I have a fancy camera that has all sorts of automatic modes, but I just use it's light meter to tell me the right exposure. I use as slow a film as 100 ISO or less, usually Kodak. I want a long shutter speed to avoid flicker, on the order of 10 seconds, so I set the aperture to 8 or more. Next I set my camera to expose the same frame 3 times, and set the exposure to 1/3 of the meter reading I got from above (so for the 10 second example I'd expose with three 3.33 second exposures). I have 3 little programs with display each of the red, green and blue file separately, each with it's respective "red scale", "blue scale" and "green scale" colormap (i.e., for the "red scale" the red values in the colormap are from 0 to 255, and the green and blue are 0). So then I: Display the red channel. Expose the film. Display the green channel. Expose the film. Display the blue channel. Expose the film. That's it. I also redo the whole process at with a slightly longer and at a slightly shorter exposure (bracketing), just in case. Turn on the lights, go home, have a beer. I'm sure there are lots of nitty gritty details that I ignore; things like gamma corrections of the monitor and all that sort of stuff, but what the heck, the results are fine. Have fun, -- John "Crash" Watson, Civil Servent from Hell ARPA: watson@ames.arc.nasa.gov NASA Ames Research Center UUCP: ...!ames!watson Any opinions expressed herein are, like, solely the responsibility of, like, the author and do not, like, represent the opinions of NASA or the U.S. Government.
kurtk@tekcae.CAX.TEK.COM (Kurt Krueger) (11/18/88)
I've seen some good advice here, but thought I would add some more info. Notice that camera systems designed for this work are set up to photograph a black and white screen through 3 different color filters (i.e. the red part is displayed and photographed through a red filter, etc.) This solves a few problems. 1. The mismatch of the color phosphor's output spectrum vs. the film's spectral response is fixed by the color filters. 2. Gamma correction can be done, as well as differing sensitivity to the three primary colors. 3. If you take a magnifying glass to the color screen, you will notice that if you have an entire green image (for example) that only 1/3 of the screen is actually lit up (the red and blue dots are black in this case). A well focussed photograph will be 1/3 green and 2/3 black. The results will be a muddy green. By photographing a b&w screen through a green filter you can get 100% green. (Note: you can help this situation by overexposing or defocusing).
good@pixar.uucp ("It's kind of fun to do the impossible." -- Walt Disney) (11/21/88)
In article <18196@ames.arc.nasa.gov> watson@ames.arc.nasa.gov (John S. Watson) writes:
:
:[this discussion started in comp.graphics]
:
:Here's how I get 24 bit photographs of my images from my 8 bit Sun screen
:(The basic idea is to triple expose the film, once with each R,G,B channels).
You can also get pretty good results just shooting off the monitor. I
routinely make slides here with Ektachrome 100 and a 200mm Macro Nikor.
Off our Barco and Sony RGB monitors I wind up with 1 second at f5.6.
Back in 1984 I shot "The Adventures of Andre & Wally B" by setting a
Mitchell with an animation motor right in front of a 512 x 488 Barco.
It turned out better than it should have.
Better systems will get you better pictures, but you'll be amazed
how well simple setups work.
--Craig
...{ucbvax,sun}!pixar!goodflip@pixar.UUCP (Flip Phillips) (11/22/88)
In article <2709@pixar.UUCP> good@pixar.uucp ("It's kind of fun to do the impossible." -- Walt Disney) writes: >In article <18196@ames.arc.nasa.gov> watson@ames.arc.nasa.gov (John S. Watson) writes: [...] > [...] >Back in 1984 I shot "The Adventures of Andre & Wally B" by setting a >Mitchell with an animation motor right in front of a 512 x 488 Barco. When I was at Ohio State CGRG we shot all of the films using an Ariflex 16mm with an animation motor, shooting directly off of the monitors. We would use Ektachrome News stock, as I recall, cant recall the stock number, maybe someone who remembers can pass it along. >It turned out better than it should have. So did ours. -- Flip Phillips {sun | ucbvax}!pixar!flip Pixar - Marin County, California
dave@onfcanim.UUCP (Dave Martindale) (11/23/88)
In article <18196@ames.arc.nasa.gov> watson@ames.arc.nasa.gov (John S. Watson) writes: > >Once everything is set up I run a program I wrote that take the three >red, green and blue image files and make a grey scale "intensity" file. >Then I display this intensity image. Next, I turn out all the lights in >the room and take a meter reading of the intensity image. A much simpler method: fill the screen with full-intensity white. Take a light meter reading. Increase exposure by 2.5 stops. This puts the brightest whites near the upper end of the linear portion of the response curve for slide film. With negative film, you might give just a bit more exposure if you want more shadow detail. (Note that 2**(-2.5) = 0.18, and thus that this method is also equivalent to displaying and reading an 18% grey patch, except that by using white you can get away with a less-sensitive light meter, and also don't have to worry about how well your gamma correction has been done - i.e. whether your 18% grey is really 18% of white). >I want a long shutter speed to avoid flicker, >on the order of 10 seconds, so I set the aperture to 8 or more. Most films recommend using colour filters to correct for colour shift due to reciprocity failure with such long exposures. An exposure time of 1 second seems a good compromise - long enough that you don't see a dark bar due to recording a non-integral number of video fields, but short enough not to worry about reciprocity problems. >Next I set my camera to expose the same frame 3 times, and set the >exposure to 1/3 of the meter reading I got from above (so for the 10 second >example I'd expose with three 3.33 second exposures). This is wrong - you should use the same exposure, not 1/3 of it. Your light meter tells you to expose the film to the full-colour image for, say, 1 second. If you had a 24-bit frame buffer, you would do exactly that. Since you have just 8 bits, you do each colour separately, but you still need 1 full second of red, 1 second of green, and 1 second of blue to give the same exposure as 1 second of white.
3ksnn64@pur-ee.UUCP (Joe Cychosz) (11/26/88)
In article <2709@pixar.UUCP> good@pixar.uucp ("It's kind of fun to do the impossible." -- Walt Disney) writes: >In article <18196@ames.arc.nasa.gov> watson@ames.arc.nasa.gov (John S. Watson) writes: >: >:[this discussion started in comp.graphics] >: >:Here's how I get 24 bit photographs of my images from my 8 bit Sun screen >:(The basic idea is to triple expose the film, once with each R,G,B channels). > >You can also get pretty good results just shooting off the monitor. I >routinely make slides here with Ektachrome 100 and a 200mm Macro Nikor. >Off our Barco and Sony RGB monitors I wind up with 1 second at f5.6. I agree, however I use a few different settings. I also shoot Ektachrome 100 (Ektachrome 64 is no longer available). For prints I use standard Kodak 100 print film. Setup: This is the longest and most crucial step. It is important to get the lens as perpendicular as possible to the screen. Things to watch for are monitors which tilt (like Suns or SGIs). I also try to get the lens as close as possible to the screen and still be able to focus. This helps reduce the effects of the curvature of the screen. For a 70-210mm Macro Nikon this is about 2 ft. Exposure: I use a 1/2 second exposure at f8 with half a stop added in. To bracket the exposure I also shoot a f5.6 and a f11 exposure. With 100 speed film the exposure can be fairly forgiving. That is to say, you'll get a reasonable picture as long as the exposure and f-stop are within reason. Never shoot faster than the refresh rate of the monitor. This even goes for taking pictures of monitors sitting in rooms. If you do you will endup with black bars on the screen. For room shots I use 1/8 second or longer. Processing: For prints I usually tell the people that process my film that the roll contains computer graphics. Many film printers are computered controlled and are setup for printing people, sky, grass, trees and other things which you would find in normal photos. These printers usually lose it when the get a picture with lots of black. Vectors are the hardest to process. One other problem the film processing machines have is alignment. Lets say you are shooting pictures of some square 512x512 images. The film processor will more than likely assume that the left edge of the image is the left edge of the picture. I will not center it for you. This will also happen with slides. To solve this problem two things can be done. One: shoot a few fully exposed pictures at the beginning of the roll so the machine can set where the left edge is for the rest of the roll. Or two: us an alignment background pattern like I do. This looks something like this: ---------------------------------- ! ! ! ! ! ! ! ----------------- ! ! ! ! ! ! ! ! ! !-------! Image ! ! ! ! ! ! ! ! !--------! ! ! ! ! ! ----------------- ! ! ! ! ---------------------------------- I'll have to try a 1 second exposure at f5.6 in the next roll I shoot. Joe Cychosz
billd@celerity.UUCP (Bill Davidson) (12/02/88)
In article <9900@pur-ee.UUCP> 3ksnn64@pur-ee.UUCP (Joe Cychosz) writes: ..... >monitors which tilt (like Suns or SGIs). I also try to get the lens as >close as possible to the screen and still be able to focus. This helps >reduce the effects of the curvature of the screen. For a 70-210mm Macro >Nikon this is about 2 ft. Er, that's not what I learned in photography school. You can't totaly eliminate curvature but it is definitely INCREASED by getting closer. Your best bet should be to get as far away as possible. Since you will still want to nearly fill your frame, the distance will be dictated by the length of lenses you own (and the amount of space available in front of your screen :-) Shooting from far away with a long lens flattens any picture. --Bill Davidson -- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ....!{ucsd|sdcsvax}!celerity!billd