On Tue, 19 Jun 2001, Dan Honemann wrote:
> > >Take a look at the Leafscan 45 sample vs. the Nikon ED 4000 about halfway
> > >down the page at this site:
> > >
> > >
> >
> > One of us is hallucinating, or one of us is blind. I sure
> > don't see the "astonishing" difference you're talking about,
> > even when these two images are inspected under high magnification
> > in Photoshop.
>
> Really? Maybe it's my monitor (a 14" thoroughly uncalibrated notebook LCD).
> I don't know what accounts for the difference--maybe the one poster is right
> in saying it is contrast--but it is most apparent to me in the girl's face.
> The Leafscan image looks clear and _glossy_, while the Nikon image looks
> _flat_. To put it differently, the Nikon image looks like a scan, while the
> Leafscan image looks like a photograph (to my eyes).
>
> I don't have the vocabulary or the trained eye to articulate what the
> difference is or what causes it--but I sure can see it. That's why I was
> hoping someone here could tell me what it is, and if it could be addressed
> in Photo Shop so that the Nikon scan would end up looking as good as the
> Leafscan image after some tweaking. I ask because I'm leaning toward buying
> the 4000 now, so I'm hoping there's some way to get it to look as good as
> the Leafscan--cuz that's the sort of scan I'm aiming for (yep, I could
> always just get a Leafscan 45, but I don't know what I'm doing and figure a
> 4000 with ICE has a shorter learning curve--and scanning time per image).
First off, we're comparing two JPGs, each about 250K bytes,
which are intended to represent many megabytes of real image
data. So the comparison, as presented on the URL you gave,
is flawed from the start.
Secondly, I'm amused that you would cite an "astonishing"
difference between these two images based on how they
appear on a 14" LCD screen. FWIW, I looked at them earlier
using a 17" aperture-grille CRT, operating in 24-bit color
mode and 1600 x 1200 resolution. In the browser window,
there was almost no perceptible difference. Bringing the
two images into Photoshop, and observing them at 400%,
the differences became perceptible -- but hardly "astonishing."
The differences were in tonality, and (to my eyes) not at
all in sharpness or resolution.
Others have commented that the Leaf image had mildly blocked-
up shadows, whereas the Nikon scan did not. I agree somewhat
with that assesment; it's as if the Nikon image had a black
point that was somewhere around (10,10,10) or even higher.
The "fix" for the Nikon image might be to properly set the
black point. That would close up the shadow detail in the
Nikon image (making it look more like the Leaf image) and
improve its overall contrast.
When comparing scans from totally different scanners, I
tend to dismiss minor differences in tonality. It's very
hard to get these to match, and the differences don't
necessarily reflect on the scanner itself -- more often,
they reflect on the settings used to make the scan, and
the skill used in choosing those settings. If the
differences are minor, they're easily fixed in Photoshop.
Note also the subjective nature of the differences.
You clearly prefer the Leaf image, probably because of
its better contrast, while others point out that the
Leaf image lacks shadow detail. So, a "proper" scan
is often a matter of taste. In any case, color
correction is all about *where* you put the contrast,
and unfortunately, it can't be everywhere.
rafe b.
