Scanning slides on a flatbed scanner does not produce the best possible images, but it's usually good enough for displaying them on a computer monitor.  After initial disappointment, I worked out the following method that produces reasonable results with no additional expense, minimal equipment, and little fuss.  (I've used a Microtek ScanMaker E6 600 dot per inch (optical resolution) flatbed scanner and PhotoImpact, the software that came with it; but other scanners and software ought to provide similar capabilities and results.)
 

A.  PHYSICAL ADJUSTMENTS

1. I remove the scanner's lid.  The Microtek E6's lid just pops off when you pull the hinged end up gently.
2. On the scanner's glass I place a sheet of black construction paper, with a cutout the size of a 35 mm slide holder.  This keeps stray light from influencing the scan, and it also makes it easy to get subsequent slides in the same place on the glass, saving a little time.  Thicker paper would work better, but construction paper seems to do fine.
3. I lay a piece of translucent whitish plastic on top of the slide.  This is to diffuse the backlighting that I'll add (in step A3) so the slide will be backlit evenly and the scanner won't "see" the lamp.   The plastic I use is actually part of a slide copying gizmo I've used for years on a 35 mm camera. but any relatively thin translucent plastic ought to work.  The important things are that it transmits light, but not the image of the light source, and that its color is neutral, so it doesn't affect the color of the image on the slide.  Plain white paper works, but it isn't uniform in density, so it produces a speckled effect in the resulting image, especially in the light areas.

 
4. Then I add light.  I've found that a compact fluorescent desk lamp works pretty well.  The lamp I use has a 13 watt U-shaped fluorescent tube about 5.5 inches (14 cm) long, so it concentrates the light nicely above the slide.  It swings out of the way when I want to change slides, and the height is easy to adjust.  Placing the lamp about 2 or 3 inches above the translucent plastic usually works well.  On very high resolution scans the fluorescent lamp sometimes induces striation in the image--narrow horizontal brighter and dimmer stripes, caused, I think, by a slight phase-shift interference between the two light sources.  If that's a problem, I've used a photoflood lamp (whiter light than ordinary incandescent) or a halogen lamp (only slightly less reddish than ordinary incandescent).  These lamps get very hot, however, so to avoid melting everything, I almost always use the compact fluorescent.  I rarely get striation with Exposure settings between 200 and 500.  (There's more on exposure in "Scanner Settings," below.)
5. If you'd like to try my earlier method, just place the slide on the scanner and lay a sheet of white paper on top of it, then close the lid and follow the "Scanner Settings" instructions, below.  This method works, but dark areas in the resulting images are about twice as dark as in the original slide.  That's because the light will have to pass through them two times--from the scanner's lamp through to the white paper and then back to the scanner's sensor.  Lighter areas in the slide, on the other hand, allow proportionally more light to pass through to the paper and back again.  This reduces the range, or ratio, of lightness-to-darkness in the scanned image.  Adjusting the brightness and contrast in software can help with this somewhat, but at the cost of reducing the image quality.  The improved method supplies at least as much light from above (behind) the slide as the scanner' lamp produces, so the resulting image looks a lot more like the original.

    The following is specific to Microtek's ScanWizard, the Twain-complaint scanner controller that came with the scanner.  It's likely, however, that other scanners' software provides comparable features, although they may have different names and be hidden away in different menus.  ScanWizard's View menu contains an item for the Settings Window, which must be in view because that's where adjustments to the incoming scan are made.  (After the scan, adjustments are made in PhotoImpact or some other software designed to enhance or manipulate images.)

1. First I reset all of the adjustments to their defaults, and then--this is the important part--set the Exposure to around 400. Less for generally light slides, more for very dark slides.  If the backlighting is very strong, it may not be necessary to increase the exposure at all.
2. Then I do a Preview and set the scan area to coincide with the slide image. Increasing the exposure above normal dramatically increases the duration of the scan, so I usually set the scan resolution to 300 dpi initially, and then bump it up to 600 or higher when I'm comfortable with any other adjustments I'll make.
3. Ideally, the resolution should match the capabilities of the output device--meaning that for display on a computer monitor, the scanner's maximum optical resolution will probably produce the size image you want from the slide.  I've found, however, that it's sometimes advantageous to use a higher, interpolated resolution because the resulting image is less pixelated.  This is especially useful if I want to display only a small portion of the slide.  Also, I have a lot of half-frame 35 mm and 16 mm slides taken with a number of early pocket-sized cameras (before full-frame 35 mm pocket-sized cameras were available).  I've gotten reasonable results with these slides by scanning at 2400 dpi.  This doesn't add any more detail, because the maximum optical scan is still 600 dpi, but it reduces pixelation substantially, so the image usually appears sharper.  It's also larger, of course--which can be either good or bad, depending on what dimensions you need to display.  For example, there's no sense in making an interpolated 4800 dpi scan and then reducing the image to 640 pixels across.  You're better off scanning at 600 dpi optical.  You can see a few examples of half-frame slides here.
4. I've found that several things affect the color balance in slide scans.  The color of the backlight is the most significant.  One reason I like the compace fluorescent lamp is that it's closer to natural daylight than a standard fluorescent, but still much less red than any household incandescent lamp.  Sometimes I find it necessary to adjust the color balance of an image after scanning, but so far I haven't found it useful to mess with the color balance during scanning.  However, I have often found it helpful to adjust the Shadows and Highlight tool, in Microtek's ScanWizard, following  Wayne Fulton's instructions.  (See an example of mine at D.3., below.)

C.  EXAMPLES--compare the two methods

Here are two 300 dpi scans of a rather dark 35 mm slide, first using the white paper method and second, this improved method.  The difference in brightness is obvious, but you'll also see that the colors are different and the lint is much more obvious in "paper" method .  The slide is an old one from the 1950s, of farmers planting rice in Korea, and has deteriorated quite a lot.  Fortunately, it'll be easy to perk up the colors, now that the image is scanned.
 

Early "white paper" method (no backlight), 300 dpi, Exposure 600.
 

Improved method (compact fluorescent backlight), 300 dpi, Exposure 402.

I chose this particular slide because it's a hard one to work with.  The shot is badly framed, but the subject may be interesting.  If you'd care to see more detail, here's a slightly enhanced 600 dpi scan of this slide.
 

D.  RELATED PAGES

1. Other comparisons--a "difficult" slide.  The white paper and the improved methods used on a slide of an interior scene with strong lighting contrasts.
2. The same "difficult" slide enhanced in three different ways.
3. My own experiments with resolution in scanning slides.  The "difficult" slide scanned at resolutions varying from 300 to 4800 dpi, using the "white paper" method.
4.  Be sure to see Wayne Fulton's excellent adviceon improving the quality of scans.  That's his main page.  He has specific info on scanning slides here.

My home page.

Send me email

(Please type it into the "To:" box in an email message.) 

Don Maxwell