Hello again Paul. I would suggest several experiments to try. Scan an image
with Digital ICE off. Scan it again with Digital ICE on. With an image
editor, subtract the two images (I would suggest this experiment for other
settings as well, like multi sample scans, analog gain, ...). The results
may surprise you. There will probably be a lot more surface defects than you
expected. With a VERY clean negative, you may notice some difference because
the two scans won't be perfectly registered. If you look at the very highest
frequencies (where you would expect to loose detail due to blurring) you
will probably see very little difference. One reason is because our software
has improved since Digital ICE was first reviewed several years ago. Another
reason is because images without surface defects APPEAR to be less sharp.
Defects tend to be very high frequency. Without this high frequency detail,
you eyes get tricked and the image APPEARS less sharp. But when you examine
actual targets or other metrics, you will find very little difference.
Another problem with Digital ICE is its reputation with Kodachrome type
films. The problem with Kodachrome is that the cyan dyes interfere with the
functioning of Digital ICE. In images with a lot of cyan (which includes the
darker parts of images since they are usually near 100 percent cyan) you
tend to loose image detail. I had this problem with a picture that was quite
light (cyan wasn't saturated in many places), it was very damaged, and there
was a man wearing a dark blue baseball cap with yellow lettering. Of course,
what I wanted to show was the lettering on the ball cap which was quite
blurred. What I ended up doing (since I had already removed the film from
the scanner) was to rescan twice using the technique described in:
http://www.luminous-landscape.com/ice.htm. Only I reversed the technique. I
used the Digital ICE image
on bottom and the original on top. I then "burned" through the Digital ICE
image to the original to show the detail just around the lettering on the
hat. It worked perfectly because the defects in the lettering on the hat
weren't that noticeable and I didn't have to do any cloning!
I hope that helps. Good Scanning!
Applied Science Fiction
From: Wilson, Paul [mailto:PWilson@gomez.com]
I owned the Nikon LS8000 for a little while and now I have the Minolta Scan
Multi Pro. I've read various things about how ICE can soften an image and
http://www.luminous-landscape.com/ice.htm even has an article about how to
lessen the problem.
With both scanners, I wasn't able to detect any softening due to Digital
ICE. Is there something I'm missing? Is there a particular situation where
this would be more apparent?
With the first versions of Digital ICE on the LS2000, there were three
settings, Normal, Sharpen and off. In the "normal" mode compared to the
compared to the "off" the image appears to be some softer. But in the
"sharpen" mode compared to the "off" mo