HOME
TOPICS
ABOUT ME
I consider Picture Window essential for serious image editing and often keep it running all evening long.
Al Fasoldt's reviews and commentaries, continuously available online since 1983
Picture Window software can blend scans to reduce flaws
March 10, 2002
By Al Fasoldt
Copyright © 2002, Al Fasoldt
Copyright © 2002, The Post-Standard
You can get more than you asked for when you scan an image. Sometimes the scanned photo or slide is marred by thousands of tiny light and dark spots in dark areas of the picture. The mottled appearance is especially visible when you enlarge the picture and when you print it large enough for framing.
This visible noise is generated by the scanner's pickup element when it strains to capture dark scenes. It's sometimes called "dark noise." Inexpensive scanners usually show this kind of noise more than expensive ones do.
But there's a simple way to get rid of much of the noise in the dark areas of scans, even if your scanner is a low-end consumer model. I've used the technique I'm going to tell you about for many months, cleaning up scans I've made on my $160 PrimeFilm 1800 slide scanner.
I've compared them to scans made of the same slides on an expensive professional scanner I used last year, and all of them have less dark noise (and, as a result, seem clearer and brighter) than the pro scans. That's amazing, considering that the pro scanner costs more than 10 times as much as the PrimeFilm 1800.
Here's one small part of a high-resolution slide scan made by the PrimeFilm 1800. Three scans were blended in Picture Window to reduce "dark noise" in the darkest areas of the image. There's not much detail in the dark areas -- they're truly dark -- but there's almost no noise. That's how a good scan should look.
The technique is amazingly simple. To cancel out the dark noise, you combine two or more successive scans of the same image. Because the noise is random, noise speckles aren't in the same place in each image. When you combine the scans, noise is smoothed out.
Of course, you could do this with four or five scans instead of two. I've known photographers who use as many as 15 scans when they're trying to get noise levels as low as possible. Using more than three probably isn't worth it, since the smoothing effects show up right away, with one or two added scans. BR> Some scanners will perform multiple scans and combine successive images for you. But if your scanner is like mine, you'll have to do it in software. You can do it using layering functions in Photoshop and in many other image editors, but you'd be wasting your time.
Spend $50 instead and get Picture Window (www.dl-c.com), and you'll never regret the cost. Picture Window, available only for Windows PCs, can combine multiple scans in many different ways, and uses methods that far outstrip the functions in Photoshop. I consider Picture Window essential for serious image editing and often keep it running all evening long.
Picture Window puts its multiple-image blending options in its "Composite" menu, under "Transformations." There are 10 ways to combine images, and there are many ways to fine-tune the combinations.
The only drawback is that Picture Window can combine only two images at a time. If you have many successive scans of the same picture, you'll have to combine the first with the second, then combine the resulting image with the third, and so on.
This turns out to be a boon if you are manipulating scans to bring out detail, because one operation can reduce noise while a second or third emphasizes highlights, and so on.
Here are some of the basic image-blending methods Picture Window uses, with descriptions supplied by the software's help file:
Blend: Second image is blended with the input image.
Lighten or darken: The second image is blended with the input image, but only where it lightens or darkens the input image.
Additive Filter: The second image is applied to the input image as a filter, as if you were sandwiching two transparencies.
Subtractive Filter: The second image is used as a filter. This is similar to Additive Filtering except that light areas of the filter are opaque and dark areas transparent.
Negative: The two images are combined so as to reverse each other.
Add: The second image is added to the input image.
Subtract: The second image is subtracted from the input image.
Register: The second image is blended with the first in such a way that areas of the first image not covered by the second show up as black in the resulting image.