I've been pondering a novel (to me) possibility for scanning slides via a flatbed scanner. My reasoning was that the resolutions and enlargement possibilities using normal flatbed/slide adaptors are seriously limited by the small film size in relatation to the scanner's capture area. This is annoying because I have free acces to a flatbed but would have to shell out for a dedicated slide scanner.

However, if the slide were to be projected onto the flatbed (using a normal projector and either directly or via a 'rear' projection screen) the image could be fitted or enlarged to take full advantage of the flatbed's total capacity. Sort of like a very bright, digital enlarger.

I have no idea if I've thought of something that already exists, or even if it is possible. I forsee the need to switch off the flatbed's light (which I don't know how to do short of wire cutters!) and that a 'normal' projector may not be bright enough, though the image would be much smaller and so brighter than usual in this application. If it does work though, the scope for enlargement (being optical) could outdo even quite expensive slide scanners.

Before I decide to either reject or work more on this, I thought I'd open it up to the collective wisdom of the forum for comments/suggestions/ridicule.

Kevin

-- Kevin Purdy (k.j.purdy@derby.ac.uk), April 18, 2000

Answers

I don't think the direct projection method would work. The optical system of the scanner is much like a camera, and the result would be exactly the same as standing in front of a projector with a camera pointing back at it and trying to take a picture of the projected image. All you'd get is an out-of-focus image of the projector lens.

The other alternative of using a back-projection screen should work, but the quality of the result will be very dependent on the "grain" of the material and the evenness of the illumination. Flatbeds that take a transparency adapter automatically switch the internal lamp off when told to scan slides, so that may not be a problem. Give it a try. Let us know how you get on.

-- Pete Andrews (p.l.andrews@bham.ac.uk), April 18, 2000.

Hmmmm
I think this could be made to work - but by the time you completed all of the necessary modifications you could have purchased a Nikon LS....
You would have to remove the lamp from the scanner itself - it would only interfere with the process. The projected image would have to be carefully focussed on the expected focal place of the flatbed scanner - typically the top of the glass - but even better results could be achieved if the glass were removed (or removable). Placing a piece of frosted glass for focussing would be a nice idea - then remove it for the scan.
The analogy of the camera pointed back toward a projector isn't exactly accurate - it would work if you removed the lens from the camera and focussed the projector on the film plane. It's the same with the scanner.
Getting the exposure and color temperature right would also be the trickiest part - but some experimentation might prove fruitful.
But - as I say - by the time you get all this working you probably would have invested a bunch of money - and there's no guarentee the results would be any better.
If, however, you are the mad inventor type this kind of off-beat development can be exciting and fun - and occasionally rewarding.

Des

-- Dan Desjardins (dan.desjardins@avstarnews.com), April 18, 2000.

Well, frankly, I'm not sure... :-)

It seems to me that the rear projected version would work best, that you should probably leave the glass in the scanner and place the projected surface on top unless it's a really thick screen or you don't mind if you get dust or scratches in the scanning mechanism (SARCASM! Don't goof up the works!), that you'd probably have to simply remove or disconnect the scanning lamp or convince the scanner that a slide attachment with lamp has been attached, and that the projected image may not be what the scanner is looking for. I would think the color balance might be way off because you're scanning projected(additive) color not reflected(subtractive)? Or am I thinking of that the wrong way?

Best way to try it? Get a medium sized, 4" or so, magnified rear projection slide viewer or make similar with a cheap fresnel lens ($3-4 @ local drugstore), a diffused light source, and a translucent material for a projection screen. You might try a piece of thin white paper or tracing paper. A piece of rear projection inkjet film or Epson glossy film might be perfect. You'll probably want a dimmer for the light source.

If you get anything at all reasonable you can then get a better lens and do it right. For the diffuser try a piece of frosted glass or a piece of white translucent plexiglas. Bear in mind, that you'll probably need to put the projector components in at least a cardboard box to keep outside light sources from degrading the image by shining on the screen or through the lens.

This might be a fool's errand, but many said the same of the electric light bulb... :-) If you have fun doing it, and maybe gain something, it's worth the time and minimal cost.

Have fun!

-- Gerald M. Payne (gmp@francomm.com), April 18, 2000.

If any of the ideas previously suggested would work, then it must be possible to somehow attach a scanner to a normal projector. This would allow scanning of all the slides in a carousel, without unloading them, feeding them into the scanner, and reloading them. This is a task that we must do very frequently. Hmmmmmmm.

Dick.

-- Dick Monahan (dmonahan@physicianed.com), April 19, 2000.

Kevin, this WON'T work (adequately); no one is catching the glitch although Pete came closest. Let me try to explain simply; if you project a slide and stand in front of the projector while looking into its lens, you'll be able to see only a small part of the slide (if it's fuzzy, focus the image roughly at a very distant (infinite) screen; your relaxed eye will be able to focus on it as though it were very distant). The parts of the slide that your eye cannot see will also not be seen by the scanner lens. So the scanner will only see a small spot in the center of the slide.

To make this work, you would essentially need to place a large "field lens" at the scanner surface. This lens would need the proper focal length to image the projector lens aperture into the scanner lens. Unfortunately such a lens would introduce all sorts of new problems.

-- Bill C (bcarriel@cpicorp.com), April 19, 2000.

Bill C.,

Let me try to explain, SIMPLY. ;-) In a rear projection system the screen, in effect, becomes the field lens... It's the focal plane that the viewer focuses on. If the scanner is reasonably well focused on the same plane it should be able to scan the projected image as easily as a backlit slide or negative in a slide or negative scanning attachment. I would think that the rear projection screen should probably be as thin as possible in order to get the focal plane at the right distance from the scanning head, but don't think it would be incredibly touchy since the scanner's depth of field is likely to be substantial due to the size of the lens element.[similar to a digicam... :-)] Further, I'd suggest, that if in your example, you(the scanner head) simply held up a translucent piece of plastic and moved up and down or left and right you'd image a different section of the slide. We're trying to scan the magnified image PROJECTED onto a focal plane(the rear projection screen, or diffuser, or field lens), not focus on the original slide viewed through the "reverse magnification" of staring down a projection lens!

I think a simple bit of experimenting would prove this. Somebody, try it? I just don't have the equipment(projector) here(don't even have a slide...) and don't have time this week to construct a mock up slide projector from spare bits of shop debris. All you really need is lamp, a condenser/diffuser to even out the light, a slide, a cheap flat fresnel lens, and a translucent projection screen(sheet of white paper in a darkened room would do). Stuff it in a carboard shoe box or whatever is handy (may need to be longer depending on focal length)and vary the distance between the slide, lens, and screen until you get it right. Easily done if you hang the slide and lens from the "rails" formed by the sides of the shoebox and simply cut the end out and tape the paper over it for a screen. Gotta remember this one for my neice's science fairs... :-) You might need to mask the light source so it doesn't get 'round the edges of the slide and shine on the screen and wash out the magnified projected image.

Once you get the image on the "screen" you fix things in place and up end it on the scanner. Dollars to donuts, you end up building a better version and acting very self satisfied for months... :-)

-- Gerald M. Payne (gmp@francomm.com), April 19, 2000.

I think this would work in a crude fashion. The main difference I see is that the film (slide) laying on the glass bed is a flat plane lit from behind by light from many directions, both perpendicular and oblique. Whereas in projecting the image onto the bare glass, the image will only have a small portion perpendicular to the plane of scanning as Bill suggested. I suspect that the center of the image will be quite a bit brighter than the than the outer regions, much like a severely vignetted lens. However, there is some chance the idea of the backlit image plane could work. One of our local government offices uses a homemade imaging setup for aerial slides of farmland, I am often surprised how good those images look on their backlit screen. I`ll try to find out their source for the screen material and post back. By the way, your ideas don`t sound nearly as crazy as something I once did for a friends 50th wedding anniversary. The fellow had some older slides he wanted included in a computer slide show for the party. I improvised a way to get them scanned using my flatbed scanner, my wifes compact mirror from her purse, and sunlight coming thru my office window. The big limitation was that I could only scan betweeen 3:30 and 5:00 PM. (Don`t laugh) Actually they came out quite well for the intended use, and my friend was delighted.

-- Bob T. (bternes@ix.netcom.com), April 20, 2000.

I just had to try this out, so I lugged my flatbed over to my enlarger and sat it on the baseboard, shoved a slide in the negative carrier and fired up the computer.

Without any interposing screen, as I suspected, all I got was a small blob of light from the lens (Sorry Des!).

The only piece of translucent material I had to hand was a negative filing sheet, so I tried it with that on the scanner. Hurrah! I got an image! Not a very good image, I must admit, but a genuine recognisable image. It was a bit gritty from the texture of the glassine paper, and quite red from the Tungsten lamp in the enlarger, but it was an image. I could probably have "tuned out" the red cast using the colour filters on the enlarger, but I had other things to do.

If someone wants to persue this further, then good luck to them. I'll stick to my 35mm film scanner and the tranny adapter on my flatbed thanks.

-- Pete Andrews (p.l.andrews@bham.ac.uk), April 20, 2000.

Good job, Pete! (Seldom do I hear of someone actually carrying out a test and reporting results).

Gerald, thanks for the explanation. I concur that a translucent screen WOULD work if it were perfectly diffuse. However, perfect diffusion implies that the light is scattered in all directions, including sideways in the screen, filling up dark areas and killing your image contrast. Also, as far as I know, such a diffuser is not available. I think (not sure) that real rear projection systems would have probably a fresnel lens built in to the back. Actually this situation is identical to viewing the ground glass of a view camera; if the appropriate fresnel lens (and eye position) is used it equalizes brightness at sides.

I still maintain that the only (reasonably feasible) way to get a good image with this general method is to have the appropriate field lens at the image plane. I'm not very familiar with the scanners, but if they used a fixed lens, the field lens would be conventional style. If it scans (is that why they call them scanners?), you would need a cylindrical lens. Again, the focal length would have to be appropriate to project an image of the the projector (or enlarger) lens through the aperture of the scanner lens. I am confident this would work well (provided projector condenser and field lens are adequate quality and not producing dark spots). The problems you would get from the field lens are: if it has a thicker center (like normal lenses do) it will not be able to simultaneously be focused at both the center and sides (field curvature), or, if a flat lens (fresnel type), you will pick up the rings (or steps) on the lens. If you move the fresnel above the image plane it will be doing some work on the image, but you should be able to refocus it. I don't know if anamorphic fresnels are readily available.

PS; if anyone is going to experiment further, the crude formula for field lens focal length is: 1/focal_length = 1/scanner_lens_to_plate + 1/plate_to_enlarger_lens. This means (I think) the focal length must be less than scanner lens to plate distance; due to f/number limitations a physically wide enough lens is probably not available.

-- Bill C (bcarriel@cpicorp.com), April 20, 2000.

I swear this is my last post on this, unless I build a working model.

Once your image is focussed on the rear projection screen, which can simply be any reasonably regular translucent material(doesn't have to be a bloody lens), the job is pretty much solved. You don't need a lens to refocus the entire image onto the scanning head, you need a surface to project the image onto so that the scanning head can move in two dimensions and scan the projected image a bit at a time, not all at once as though it were a camera. Phew!

Now, on to serious concerns. The real problems with this approach are going to be the resolution of the lens used to project the image and the lack of brightness at the projection screen due to the imperfect transmission through the screen material. And there is going to be some tendency for the image to be brighter at the center as long as a point source with a diameter smaller than the image is used to illuminate it. That's probably why the scanner works the way it does taking the light source and point focussed lens with the sensor as it goes...

I'd still bet that with a good even light source, and a high resolution lens you could get away with a conservative magnification of 3 to 4 times which should increase your effective scanning resolution by the same factor. I just wonder if the image manipulation that would probably need to be done afterwards would make it worth the trouble? We'll see.

-- Gerald M. Payne (gmp@francomm.com), April 20, 2000.

I am the owner of an aerial photo company in Denver, CO. I have been considering the subject concept for many years. My first experiment consisted of hauling a small enlarger to a dealers showroom, setting it up above a scanner and focusing a B&W image onto the plane of the glass. Scanning that "image" resulted in nothing. Once I thought about it a little more and studied some basic optics principles, it became obvious to me why my experiment didn't work, and the reason is reflected in the preceding discussion. It is true that one could scan an image projected to a ground glass on the scanner's glass plane, but this would not satisfy my own objective. The image resolution would be greatly reduced. My objective is to combine the quality and flexibility of analog enlargement using an image projected through a quality enlarger lens, and the many advantages of digital image processing. The idea is to project the portion of the image needed to the size of the scanner glass. By enlargeing the small area in an analog optical process, then digitizing the enlarged image on a relatively low-resolution scanner, one can achieve the digital output of a very expensive, high-resolution scanner. This is my objective.

To get back to the practical problems, I believe the issue is this. A scanner has a row of CCD's recording light from a reflected image, or a backlit transparency. There must be a long fresnel-type lens placed over the CCD array to take the reflected light and focus the image to the array. I intend to experiment with the concept of removing the glass plate, removing the lens focused on the array, and projecting the desired image onto a plane which is coincident to the plane of the CCD array. I believe that by projecting the image to the plane swept by the CCD array, one would effectively combine the two forms of image projection and capture. The problem I forsee is one of light intensity. The array may not be sensitive enough to read the relatively faint image projected from an enlarger.

In any case, this is a suggested experiment and the reasoning behind it. I would appreciate comments and if anyone has experimental results, or related ideas, I would be very interested. I have held back from actually doing the experiment out of fear of damaging my scanner, but I think I will get the courage to proceed and will report back the results.

-- John Young (capsden@ix.netcom.com), March 27, 2001.

You just put the slide on the scanner bed and backlight it with difuse light. A florecent flashlight or battery powered camping lantern will work better due to the switching frequency being ususaly higher than the 60hz. of the scanners own light source. The florecent lamp has a wider spectrum than an incandecent lamp, but you might try a quartz halogen lamp also. Another way is to make a little collector /difuser out of foam core poster board. This way you can use the scanners own light source. This is how the "boughten" adapters work. The difuser would be triangle shaped with the hypotenuse three to three and a half times the size of the slide. As wide and as high as the slide, with the sides looking like gable ends of a house. Place the collector difuser on one end of the scanner and cut out white poster board to "white out" the rest of the scanner bed. Place the slide under one end of the little roof contraption and try it. If you slides or transparencies are larger than 35mm, you will have to experiment with a larger back lighter/difuser. Progecting directly onto the bed will not work. Drawing mylar an old clear /translucent tupperware lids make good difusers. Try looking at www.abstractconcretedesigns.com , I hope the URL is correct.

-- Bob Dare (sorry@nospam.net), November 18, 2002.

RE. Collector/difuser: The correct URL is; www.abstractconcreteworks.com/essays/scanning/scanning.html Another idea comes to mind. If you can disable the lamp in the scanner and use a projector above the bed. This will let you project the image to a larger area of the scanner bed. You will have to use a piece of drawing mylar or something for the translucent screen on the scanner bed and block out all other ambient light. The mylar screen will have to be flat and near flawless. The type and texture of the mylar may affect the potential sharpness of the scan. If you think you need to reduce the brightness of the projected image, do not use a dimmer on the projector lamp, it will dim ,,but it will also shift colors to the red end. You may be able to get decent results without disabling the scanners own lamp.

-- Bob Dare (dare@custom.net), November 18, 2002.