Re: Flatbed scanners vs Pro Photo CDs
Re: Flatbed scanners vs Pro Photo CDs
- Subject: Re: Flatbed scanners vs Pro Photo CDs
- From: Jan Steinman <email@hidden>
- Date: Mon, 4 Dec 2000 20:21:42 -0800
From: tflash <email@hidden>
I did a comparison on the Epson site of these two units and from the specs
the relevant differences I see are these:
Perfection 1640SU Photo
3.2 Max
42-bits per pixel internal, 42-bits per pixel external (selectable)
Expression 1600 Pro
3.3 Dmax
36-bit per pixel internal/external 24-bit per pixel external (selectable)
It seems that the only spec the 1600Pro has over the 1640SU is transparency
size (4x5 meets my needs) and Dmax 3.3 vs 3.2. Is the Dmax a significant
enough issue to tip the table?
Something is rotten here. Bits-per-pixel and Dmax should roughly
correspond, unless those are "marketing bits" instead "real bits." A
"real" 14 bit scanner should achieve Dmax circa 3.6. If it's
quantizing to 14 bits, but for whatever reason (noise, whatever), it
can only achieve 3.2 Dmax, then a number of those bits are simply
returning noise.
Also, these scanners cannot actually resolve 1600 spi. That's more
marketing crap. They use two sensor arrays that are offset 1/2 cell
from each other, which (unless you are playing tricks with phase
differences, which I seriously doubt these scanner are) simply
doesn't work. Shannon says you cannot resolve finer that the sampling
aperture, which in the case of offset sensors, is still 1/800th of an
inch. You're probably no better off than if you bought an 800 spi
scanner and then up-sampled in Photoshop, because staggered sensors
(as well as micro-stepping) is simply doing mechanical interpolation,
NOT resolving more real data.
(My qualifications: I'm an EE who has designed sampling systems, and
have worked on synthetic aperture sonar systems. I don't believe
Epson can mount the sensors within the tolerance of a portion of a
wavelength of light in order to use synthetic aperture theory.)
<rant>This situation reminds me of the early transistor radios. They
often loudly touted the number of transistors inside, never
mentioning why more transistors should make it perform better than a
basic five-tube radio, and also never mentioning that many of those
transistors were not fully utilized, sometimes only used as diodes,
sometimes simply soldered to isolated pads on the printed circuit
board!</rant>
Another question to ask: why are the specs of a $400 scanner so
similar to those of an $1,100 one? Definitely something fishy here.
Unless you part with $400-$1100 on a whim, I'd recommend taking some
difficult material down and having it scanned on each. That's what I
did when I bought a film scanner, and the side-by-side results were
enlightening. For a distant second best evaluation, find some good
magazine reviews.
IMHO, the situation has gotten so bad that scanner specs are nearly
meaningless -- just like printer specs.
--
: Jan Steinman <
mailto:email@hidden>
: Bytesmiths <
http://www.bytesmiths.com>