Scanning Kodachrome Slides
It has taken me a little while to get confident with how to get the best results when scanning slides, especially Kodacrhrome slides.
These are the issues:
All slides have a much higher contrast then negative film. This leads to the risk of flare, especially where a dark and a light zone are close together.
The DigitalICE capability for removing dust doesn't always work well with Kodachrome slides. Artifacts are produced... where a dark and a light zone are close together.
When dust works its way into a Coolscan, the image is degraded, leading to blurring...... where a dark and a light zone are close together.Some of the earliest slides that I have started scanning with my Super Coolscan 5000 ED were a set of four films I took during a camping trip to Central Australia in 1984. One of the features of these photos is a lot of contrast, leading to pictures.........where lots of dark and light zones are close together.
As a result, I've had to wrestle with a lot of flare and artifacts in my images and work out ways to reduce this.
Digital ICE Artifacts with Kodachrome
With the first few generations of Digital ICE provided with Nikon (and other) film scanners, there was a simple rule: Digital ICE could not be used with Kodachrome or silver based Black and White films. The reason is that the infrared sensor used to detect dust (which picks up the dust, but which finds standard films to be invisible) cannot reliably shine through films containing silver. Although Kodachrome looks like a colour film, it is actually based upon a set of three black and white films. The colour is added during a very complicated processing process, but the slides remain at their heart a black and white film with colours added on later.
The version of Digital ICE can be used with Kodachrome (but not with silver based Black and White films).
There's no explanation given as to how this magic is accomplished. After all, if the infrared couldn't see through the film with previous versions, how does it see through the film now? Maybe they use a different wavelength of infrared. Anyhow: it mostly works, but they (honestly) warn that with some slides 'digital artifacts' will be produced. These aren't really explained in detail, so I'll give some examples of what they look like.
The artifacts appear on the border between dark and light parts of the slide. The light part of the border migrates into the dark part, looking either like a smear or a set of jagged indentations along the edge.
Here's a closeup of a photo with Digital ICE artifacts:
Here's a closeup of the same photo without Digital ICE being used:
A Simple way to remove Artifacts
This is simple, but it requires:
You take two scans of the slide, one using ICE and an identical one without ICE (a hassle, but I'm taking two scans as a matter of course on any half decent scan in case it turns out to have artifacts that I don't immediately notice.
You need to develop a little familiarity with a digital imaging program that can use layers (Paintshop Pro, Photoshop etc.)When you take the two scans (one with ICE, one without), be careful not to touch the slide between the two scans and don't change any setting other than ICE. This will ensure the two images are identical except for the presence of artifacts (and the absence of dust).
Open the scan that used Digital ICE. Select All, and copy it into clipboard.
Close that scan.
Open the scan without Digital ICE.
Paste the image in the clipboard directly on top of the other image 'as new layer'. The image with the artifacts (but no dust) will be the layer on the top.
Select the Eraser tool. Make sure it's a convenient size that's about the size of the artifacts. Somewhere between 5 and 20 pixels will probably work with an image scanned on the Coolscan at maximum resolution.
Enlarge the image, and go to a part of the image with artifacts. Drag the eraser along the edge, keeping mostly in the dark part of the edge where there is bluring or jagged parts of the light side of the line intruding. As you do this, you'll see the nice crisp dark part of the image showing through from the image below.
Merge the layers when you have finished to save a single layer image in the way that you have fixed it up.Bingo! Most of the image from the ICE scan is retained (the good parts that have no dust spots and no artifacts), but in the few places where digital artifacts had previously been, you've removed them.
Flare
Flare appears to be a significant issue with the Coolscan scanners when they scan slides.
Negatives don't seem to suffer the same issue.
Here's a scan that shows the flare:
You'll notice that the flare extends (in this case) up and down. If the slide had been a portrait aspect ratio, the flare would have gone left to right. The fact that the flare extends in a particular physical direction is the best indication that any flare you see is actually scanner flare rather than an imperfection in the original slide. To work it out, put your slide in sideways and see if the flare extends in a different direction. If so, that's scanner flare:
Compare the flare. In the first image of the bridge, it extends into the black borders at the top of the slide (and upwards from the bright spans into the shadows on the bridge). In the second pictue, the flare still reaches upwards, but because the slide has been rotated it appears in different parts of the image. You'll see the flare extending up, but in this image it is in the horizontal direction of the bridge instead of vertical.
Almost anything seems to exacerbate the flare: using DEE, using SIE, increasing the Gamma, using ICE. The slide above has a gamma of 1, a DEE of 20 and it uses ICE.
The slide below minimises flare: it doesn't use DEE or SIE, it has no ICE and it has a Gamma of 0.6. There's still some flare visible, but much less.
The other thing that causes flare is dust inside the scanner. I've seen somebody recommend keeping your scanner inside a plastic bag when it's not in use. Since I use my scanner most days, that's impractical. At some stage I'll need to open my scanner to remove the dust. There are instructions on the web on how to do this. It's a fiddly job, but it can make quite a difference to the amount of flare.
Table of Contents