munish@s.ms.uky.edu (Munish Mehra) (01/28/91)
Does anyone have some tips on photographing computer screens. There was a discussion about a year or two ago, does anyone have a copy or pointers to it. I'd appreciate any help. I have a Cannon EOS Rebel with a 35-70 Zoom, so I guess that should be good enough. Thanks. -- Munish Mehra ARPA: munish@ms.uky.edu BITNET: munish@ukma.BITNET "Hardware and software, We are your slaves and masters, Our lives are programs." Geoffrey James, The Zen of Programming.
keeney@vixvax.mgi.com (Richard Keeney) (01/28/91)
In article <1991Jan27.232910.22575@ms.uky.edu>, munish@s.ms.uky.edu (Munish Mehra) writes: > Does anyone have some tips on photographing computer screens. > There was a discussion about a year or two ago, does anyone have > a copy or pointers to it. > I'd appreciate any help. > I have a Cannon EOS Rebel with a 35-70 Zoom, so I guess that should be good > enough. > Thanks. If you are going to be doing a lot of that sort of thing, I would recommend purchasing a high quality digital image recorder like those sold by Management Graphics (my employer). If you only need a couple-few images, I would consider sending screen dump files (in some form of well used image file format like PICT, TIFF, GIF, etc) to one of the many services that make a business of imaging other peoples graphics on film using high quality image recorders like the ones we sell. Many of these businesses charge as little as $5.00 per slide for two or three day turn around. I can supply the name of one of our customers near you who do this sort of thing if you are interested. In either case the quality of the image recorder produced slides will be far superior to that of a picture of the screen. If you insist on photographing the screen yourself, here are a few tips: Turn off the room lights to eliminate reflections off of the tube. It is easy to overlook such reflections with the result being a ghost image of your face, the room, etc on every slide. Use a daylight balanced film - most CRT's are calibrated to have a white of 5500 deg. K or so. Use a long focal length lens (75 to 100mm) that can focus close enough such that the screen fills the frame. The longer focal length helps overcome some of geometric distortion introduced by the curved CRT. Some macro lenses designed for copy work may be appropriate if you can find one of reasonable long focal length. This type of lens is design to not introduce any of its own geometric distortion when photographing a flat subject (like a pice of paper of a print) while lenses designed for general photography will compromise geometric distortion for other priorities (like aperture size, zoom range, color correction, etc). Use a tripod and a cable release since the exposure will be on the order of a few seconds. Be sure the camera is positioned on a line normal to the center of the image area of interest on the screen to avoid any "keystoning". Use an exposure of at least 2 to 5 seconds to eliminate the horizontal bar that can be present if a shorter exposure is used. Most cameras will not let you select an exposure time of say 10 seconds unless you use the "B" setting. When using the "B" setting, the built in exposure meter cannot give you any useful information. To get around this problem, set the shutter speed to 1 second (or whatever the slowest speed the camera can still meter at) and take a reading of the required f-stop. Then open the lens one stop for each time you double the exposure time. For example, if the meter reading was f-16 at 1 second, you could properly expose at f-11 for 2 seconds, f-8 for 4 seconds, f-5.6 for 8 seconds, or f-4 for 16 seconds, etc. The f-stops marked on most lenses follow the sequence: .. f1.4, f2, f2.8, f4, f5.6, f8, f11, f16, f22, f32, f44, f64 ... Descending from right to left one of these steps doubles the exposure (the numbers are proportional to the inverse of the diameter of the aperture, the exposure being proportial to the area of the aperture). Be sure and "bracket" the exposure for your first few attempts. "Bracketing" in photographic terms means to expose the same subject at a variety of exposures both above and below that recommended by the exposure meter. This can be a tremendous time saver (at the slight expense of the cost of the additional film used) since it can ensure that the best possible exposure of an exposure critical subject is obtained on the first shooting session, and thus avoiding the delay associated with trial and error when each try has to be sent out for processing. You will most likely need to increase the exposure 1/2 to 1 stop over the meter reading due to reciprocity and stroboscopic effects in the film (typically, film is not a perfect linear integrator of light and tends to be underexposed when longer exposure times or intermittent exposures like those of the flickering light of a CRT are made). This desire for an increase in exposure can be offset somewhat by the fact that most slide films produce more saturated colors when underexposed, but at the expense of overall brightness of the slide when projected. This may be important when the subject on the screen is a chart or graphic with large areas of saturated colors, and is of less importance when the subject more closely resembles a natural scene. Print films, on the other hand, have more lattitude for overexposure, and print quality will deteriorate rapidly with underexposure. When you find a setting that works for your CRT, Lens, and Film combination, be sure and write it down as a starting point the next time you have to do this. Also note the brightness and contrast settings on the monitor. I myself find it hard to put too much faith in the monitor settings, and I would be tempted to bracket on subsequent shooting sessions anyways Use a fairly small aperture (large f-number) of at least f11 or f16 to make sure that the depth of field is sufficient to ensure that entire screen is in focus. With larger aperture settings (smaller f-numbers), the corners of the screen may be slightly out of focus compared to the center. Adjust the exposure time accordingly, keeping in mind the requirement of having an exposure time on the order of at least a few seconds. Physics is on your side here since you can always use a smaller aperture to achieve a longer exposure time, or a longer exposure time to achieve a small aperture. The only way to run into difficulty with this is if you are using too fast of a film (ISO 400 or higher). The faster film is more expensive anyways, so I would stick with something like Ektachrome 100 HC or Fujichrome 100 for slides, or Kodacolor 100 for prints. Good luck. -- Richard A. Keeney Senior Software Engineer Internet: keeney@mgi.com Management Graphics, Inc. Phone: +1-612-851-6126 1401 East 79th Street, #6 Fax: +1-612-854-6913 Bloomington, MN, 55425 U.S.A.
srnelson@nelsun.Eng.Sun.COM (Scott R. Nelson) (01/29/91)
From article <1991Jan27.232910.22575@ms.uky.edu>, by munish@s.ms.uky.edu (Munish Mehra): > Does anyone have some tips on photographing computer screens. > There was a discussion about a year or two ago, does anyone have > a copy or pointers to it. > I'd appreciate any help. > I have a Cannon EOS Rebel with a 35-70 Zoom, so I guess that should > be good enough. > Thanks. First of all, I assume you are photographing from a color monitor. You will (obviously) need a tripod to hold your camera. A cable release is helpful in reducing camera vibrations, but it is not absolutely necessary. A good daylight film produces best results on a color monitor because it is sensative to the same colors as the phosphors on the monitor. For slides, Ektachrome 100 works well. For prints, Ektar 25 is probably the best you can get. The slower the film speed the better the results you can expect. You are more likely to get a good color match with slide film because the photo processing lab does not attempt to do any color adjustments on slide film. Nearly all prints have some color adjustments done on them which can change your results. Before you start shooting pictures, make sure the color CRT is totally clean. The darker the room the better. Be sure there are no reflections off of the CRT from any other light sources. It is also a good idea to make sure no other bright CRTs are nearby. Align the camera and screen so that the screen is centered both vertically and horizontally relative to the lens of the camera. I would recommend using the 70 mm setting on your zoom lens to minimize the distortion caused by the curvature of the CRT. The further back you place the camera from the CRT the less distortion there will be. I would also recommend turning off the autofocus on the camera and getting the picture properly focused once and leaving it there. Because the electron beam sweeps over the face of the CRT somewhere between 50 and 76 times per second, you want the exposure time to be very slow. I prefer to catch at least 10 sweeps of the electron beam to guarantee even light distribution. I never set the speed faster than 1/4 second. The light metering system on your camera will give you a good general idea of the correct F-stop to use. If you want the very best pictures from a CRT screen it is recommended that you bracket the exposure by shooting half an F-stop to either side of what your camera recommends, in addition to using the setting the light meter selects. The last good set of pictures I did were shot using Ektar 25 film at 1/2 second exposure with the F-stop set at 2.0 to 2.8. Your camera only has a maximum F-stop of 5.6 at the 70mm lens setting (I just got one for Christmas), so you would probably want to set the speed to 1 or 1.5 seconds if you are using Ektar 25. If you must have prints and want the best color reproduction, you will probably have to take something to the lab that they can use to match colors. The image called "lenna" is an excellent one for getting good color matching, since any good photo lab can produce a correct flesh tone. If you have a color digitizer, you could digitize a photograph containing the colors you are concerned about, photograph this from your screen, then have the lab match the colors. I haven't tried this yet, but plan to do it. Most good photo labs are willing to work with you to get the exact colors. Once you get it right, make sure they write down the settings they used so that you can reproduce the results. If you are shooting many different images, keep in mind that brighter images may need a different exposure setting than darker images. This should be obvious, but I ran into the problem once. That is everything that I can think of, but I would appreciate hearing other photography hints that I haven't mentioned. --- Scott R. Nelson srnelson@eng.sun.com Sun Microsystems "Proofread carefully to see if you any words out."
sloan@cs.washington.edu (Kenneth Sloan) (01/29/91)
In article <1991Jan28.010515.708@vixvax.mgi.com> keeney@vixvax.mgi.com (Richard Keeney) writes: [lots of good advice on shooting screens, except} >Use an exposure of at least 2 to 5 seconds to eliminate the horizontal bar >that can be present if a shorter exposure is used. > No. These exposures are unnecessarily long, and may lead to reciprocity failure. In my experience, exposure time should be between 1/2 second and 2 seconds. 1/4 second is OK in an emergency. Anything shorter than 1/8 second will probably be unsatisfactory. > >You will most likely need to increase the exposure 1/2 to 1 stop over the meter >reading due to reciprocity and stroboscopic effects in the film (typically, >film is not a perfect linear integrator of light and tends to be underexposed >when longer exposure times or intermittent exposures like those of the >flickering light of a CRT are made). Not necessary if you follow the advice above... >When you find a setting that works for your CRT, Lens, and Film combination, be >sure and write it down as a starting point the next time you have to do this. Even better: take the time now to design, produce, and save a test image. I like something which is predominantly 18% gray, with a few color bars and somm fine detail. Put your name in it, and your film has less of a chance of getting lost. Iterate until this image generates a dead-center meter reading for an exposure which provides results you like on "typical" images that you shoot. Then, bracket anyway. Give your co-author the "1-stop too dark" copies. -Ken Sloan
aipdc@castle.ed.ac.uk (Paul Crowley) (01/29/91)
keeney@vixvax.mgi.com (Richard Keeney) sez in <1991Jan28.010515.708@vixvax.mgi.com> >Then open the lens >one stop for each time you double the exposure time. For example, if the meter >reading was f-16 at 1 second, you could properly expose at f-11 for 2 seconds, >f-8 for 4 seconds, f-5.6 for 8 seconds, or f-4 for 16 seconds, etc. The >f-stops marked on most lenses follow the sequence: > >.. f1.4, f2, f2.8, f4, f5.6, f8, f11, f16, f22, f32, f44, f64 ... > >Descending from right to left one of these steps doubles the exposure (the >numbers are proportional to the inverse of the diameter of the aperture, >the exposure being proportial to the area of the aperture). This guy seems to know exactly what he's doing, but he's made a slight slip up there: he's right the second time, not the first. If the reading is f-4 for 1 second, you could expose at f-8 for 4 seconds, etc. I shall be using most of the advice he gave, since I'm making a film and I shall be doing some computer animation. (It will be brain-stompingly low quality anyway, our camera is thirty years old.) \/ o\ Paul Crowley aipdc@uk.ac.ed.castle /\__/ Trust me, I know what I'm doing.
gray@s5000.RSVL.UNISYS.COM (Bill Gray x2128) (01/30/91)
In article <1991Jan28.010515.708@vixvax.mgi.com> keeney@vixvax.mgi.com (Richard Keeney) writes: >In article <1991Jan27.232910.22575@ms.uky.edu>, munish@s.ms.uky.edu (Munish Mehra) writes: >> Does anyone have some tips on photographing computer screens. >> There was a discussion about a year or two ago, does anyone have >> a copy or pointers to it. >> I'd appreciate any help. >> I have a Cannon EOS Rebel with a 35-70 Zoom, so I guess that should be good >> enough. You ought to try to borrow something in the 105 or 135mm range, preferably a fixed focal length, not a zoom or varicofal. >> Thanks. > Mr. Keeney begins giving generally excellent technical advice for photographing CRT displays. He does fine until the following: >Most cameras will not let you select an exposure time of say 10 seconds unless >you use the "B" setting. When using the "B" setting, the built in exposure >meter cannot give you any useful information. To get around this problem, set >the shutter speed to 1 second (or whatever the slowest speed the camera can >still meter at) and take a reading of the required f-stop. Still ok, but now the error is introduced: > Then open the lens ^^^^ >one stop for each time you double the exposure time. For example, if the meter >reading was f-16 at 1 second, you could properly expose at f-11 for 2 seconds, >f-8 for 4 seconds, f-5.6 for 8 seconds, or f-4 for 16 seconds, etc. The >f-stops marked on most lenses follow the sequence: > >.. f1.4, f2, f2.8, f4, f5.6, f8, f11, f16, f22, f32, f44, f64 ... > >Descending from right to left one of these steps doubles the exposure (the >numbers are proportional to the inverse of the diameter of the aperture, >the exposure being proportial to the area of the aperture). > Unfortunately, Mr. Keeney has the f/stop progression backwards. Moving from right to left on his scale progressively HALVES, not doubles, the exposure. In other words, assuming his f/16 @ 1 sec. exposure example was what your meter recommended, the following f/stop and shutter combinations give equivalent exposures: Shutter 1/16 1/8 1/4 1/2 1 2 4 8 16 f/stop 4 5.6 8 11 16 22 32 45 64 Shutter speeds are in seconds. Many shutters use slightly different speeds; for example, 1/4 sec = .25 second, but most shutters use 1/30, 1/60, 1/125, 1/250, etc. Because of the reversing of his f/stop progression, Mr. Keeney's shutter speeds should be the reciprocals of what he listed. The rest of his advice is excellent. I would add that reciprocity failure does tend to get worse as shutter speed increases, as do certain lens flaws. Most lenses perform best in the f/8 to f/11 range. Using a lens in the 135mm length will make focussing easier, reducing the need to stop down excessively to gain depth of field. This will also reduce the need to use excessively slow shutters. Also remember that not all tripods are created equal: your camera must be *absolutely still* during the entire exposure or all you get is a carefully exposed fuzzygraph. I would extend his suggestions by adding one more: when you bracket, bracket in 1/3 f/stop increments, especially if you use transparency (slide) film. Transparency film has much less latitude (inherent ability to function well when incorrectly exposed) than either color or b/w negative film. The slower the emulsion, the better the image, but the less latitude the film has. For example, a slow Kodachrome (ASA/ISO 25, if I recall right) gives a fabulous image when handled correctly, but is less forgiving than the emulsions Mr. Keeney recommended. You can see the difference in 1/3 stop bracketing. >Good luck. You'll need it. CRT display shooting is a demanding commercial shoot, especially if you intend the pix for reproduction. Ask for a raise. :-) > >-- >Richard A. Keeney >Senior Software Engineer Internet: keeney@mgi.com >Management Graphics, Inc. Phone: +1-612-851-6126 >1401 East 79th Street, #6 Fax: +1-612-854-6913 >Bloomington, MN, 55425 >U.S.A. Bill -- : gray@rsvl.unisys.com : : : : My gun is safer than Ted : : Unisys has enough problems without being : Kennedy's car. : : blamed for my personal opinions. : :
clh@tfic.bc.ca (Chris Hermansen) (01/31/91)
In article <8130@castle.ed.ac.uk> aipdc@castle.ed.ac.uk (Paul Crowley) writes: >keeney@vixvax.mgi.com (Richard Keeney) sez in <1991Jan28.010515.708@vixvax.mgi.com> >>Then open the lens >>one stop for each time you double the exposure time. For example, if the meter [stuff deleted] > >This guy seems to know exactly what he's doing, but he's made a slight >slip up there: he's right the second time, not the first. If the >reading is f-4 for 1 second, you could expose at f-8 for 4 seconds, etc. or at f-5.6 for 2 seconds, or f-11 for 8 seconds, etc etc. >I shall be using most of the advice he gave, since I'm making a film and >I shall be doing some computer animation. (It will be brain-stompingly >low quality anyway, our camera is thirty years old.) Which brings me to my real point in responding. I've had an Olympus OM-1n for quite a few years; we regularly use it for taking photos of the Sun screens around here, just like the folks mention above. Unfortunately, I've never been that happy with the results until a month or so ago. The difficulty is that one normally needs to use a fairly long lens to eliminate screen curvature effects, and most long lenses (especially on older equipment) don't do a good job of focussing when the lens is wide open (eg f-4, f-2.8) at close range (actually, I believe that most lenses don't do a good job at close range, even many that are so-called macro lenses, but hey, I could be wrong). So, I popped into a local camera shop a month or so ago and explained my problem to them. The solution was a 100mm close-focussing lens from Tamron (model # SP 52 BB). The improvement in picture quality is nothing short of astonishing. Now all I need to do is learn how to fix the gamma, and I'll be off... :-) Chris Hermansen Timberline Forest Inventory Consultants Voice: 1 604 733 0731 302 - 958 West 8th Avenue FAX: 1 604 733 0634 Vancouver B.C. CANADA clh@tfic.bc.ca V5Z 1E5 C'est ma facon de parler.
ge@dbf.kun.nl (Ge' Weijers) (02/05/91)
munish@s.ms.uky.edu (Munish Mehra) writes: >Does anyone have some tips on photographing computer screens. >There was a discussion about a year or two ago, does anyone have >a copy or pointers to it. >I'd appreciate any help. >I have a Cannon EOS Rebel with a 35-70 Zoom, so I guess that should be good >enough. >Thanks. Just some extra observations: You can't connect a cable release to a Rebel (EOS 1000 to those this side of the Atlantic). Use the timer instead. The zoom should suffice I think, set to 70 mm. Unless you take pictures of a completely flat monitor some distortion will be inevitable, and this distortion will be worse than the slight distortion from the Canon lens. If you're not sure about exposure you can use the Tv setting on the exposure selection dial. This won't work for line drawings on a black background! As others pointed out you should use a long exposure time, 1/4 of a second or longer. Don't use too long a time though, as the so-called Schwarzschild effect gets very noticeable (long times make films less sensitive. You can get info from your film manufacturer.) By using the exposure compensation you can bracket your exposure. Write down the exposure times chosen. When you do it again you can switch the camera to manual and shoot away. Don't rush to the camera shop for 100mm macro lens. They're expensive and not justifiable for the low quality you will get anyway, as colour monitors are not so hot. What are you going to use the pictures for? Ge' Weijers -- Ge' Weijers Internet/UUCP: ge@cs.kun.nl Faculty of Mathematics and Computer Science, (uunet.uu.net!cs.kun.nl!ge) University of Nijmegen, Toernooiveld 1 6525 ED Nijmegen, the Netherlands tel. +3180652483 (UTC-2)
3ksnn64@cidmac.ecn.purdue.edu (Joe Cychosz) (02/06/91)
>In article <1991Jan27.232910.22575@ms.uky.edu>, munish@s.ms.uky.edu (Munish Mehra) writes: >> Does anyone have some tips on photographing computer screens. >> There was a discussion about a year or two ago, does anyone have >> a copy or pointers to it. >> I'd appreciate any help. >> I have a Cannon EOS Rebel with a 35-70 Zoom, so I guess that should be good >> enough. >> Thanks. > This is what I posted a few years ago. I have not tried Ecktar 25 or 100. Film: I use Kodak 100 print film for prints and Ektachrome 100 (Ektachrome 64 is no longer available) for slides. 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 at 70mm. This contradicts what one might think in that one would expect the flattest photo to appear by getting as far away as possible and shooting with as long of a lens as possible. The problem is in the distoration in the lens system at 210mm. Keep in mind that a macro lens works on a series of approximations. What results is concave pillowing which is oppisite of the convex pillowing caused by the curvature of the screen. The optimal flat picture occurs when the convexed pillowing of the screen is balanced with the concaved pillowing of the distortion in the lens system. Since I do not have a straight 200mm lens, I have not been able to study and measure the distortion in the macro. One final note: I tend to shoot images that are either 640x480 or 512x 512 off of the SGI. I have shot full screen images, but the curvature of the screen gets a bit difficult to control, especially with the 70-210 lens problems. 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, 1/60th. A wide black bar will appear on the screen. 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. I have shot 1 second exposures at f8 and have had similar results. Again, 100 speed film is pretty forgiving. I have found the exposue metering systems in cameras to unreliable when determining the exposue. This is especially true when shooting vector displays. Also, you want to avoid full white areas such as menu areas and the like. Change them to a 70 - 80% grey. The pictures will look much nicer when printed. 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 and fully saturated colors. Vectors are the hardest to process. I use a Noritsu and a Fuji machine for my processing. Usually one or two units of density must be added (i.e., +1, +2) for computer graphics shots. The process you are going through is to balance the quality of the color with the quality of black. If you don't do this you will end up with brown instead of black. The Noritsu and Fuji will will print on the back of the photo what settings were used. So if you do reprints bring the old photo in so the processor can see what settings they used last time. Examples follow below: Noritsu: 896 45 N N N 6 2 896 = Sequence number 45 = Film code (45 = Kodak 100) N N N = Yellow, Magenta, Cyan print levels. N = neutral, there will be numbers -1, +1, etc. for non-neutral settings. 6 = Density setting. 2 = Ignore. Fuji: 79 010005 28 +01 +0 -01+02 NN -5 0 79 = Sequence number. 010005= Film/vendor/speed code (Kodak 100) 010205 for Kodak 200 Gold 28 = Daily density setting, results from calibration. +01 = Density setting. +0 -01+02 = Cyan, Magenta, Yellow print leves. +0 = neutral. NN = Ignore. -5 0 = Ignore. 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. It 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 am in the process of working on a calibration image which will help the process people make there settings. I am also working on reducing the yellow edges which appear (especially at red and green boundries). I don't expect to have any of this done sometime soon though.