At 02:56 AM 6/11/01 EDT, Ed Hamrick wrote:
>Unsharp masking isn't a reasonable way to compare the scans, since
>this doesn't get to the root of why there's a difference between the
>results from the two scanners.
I disagree here, Ed. Here's why.
It seems some scanner vendors (maybe all) implement internal
trade-offs between noise and resolution.
Sharpness and resolution can't be considered separately from
noise, since the are inter-related.
I did a round of tests on my Epson 1640SU (a flatbed touted
by Epson as a film scanner also) and was thoroughly convinced
that Epson's 1600 dpi claim was a sham.
However... with enough USM, the Epson's output can be shown to
contain much more detail than you might think. With the extra
detail comes lots of noise, of course.
I'm guessing that Epson traded off resolution for lower noise
in this model. PS: the results of these tests (and scans from
several medium-format film scanners) may be seen at:
You can see for yourself what I'm talking about -- download one
of the JPGs from the 1640, and apply USM in Photoshop. The
results may surprise you.
rafe b.
