Hi all,
All of this information is very informative, especially about the light
source being more stable with the L.E.D. light. After looking at the
specs. on both the Dimage Scan Elite 5400 II and the Nikon V ED (which
are comparably priced) the Dimage has a 3 line color CCD with 5340
pixels per line, primary color filter whereas, the Nikon has a 3964
pixel linear CCD image sensor, the Dimage has 4.8 dynamic range and the
Nikon has 4.2. The Light source on the Dimage has only one L.E.D.
light source, and the Nikon does have R,G,B & IR L.E.D.'s. I wonder
then which specs. play the most importance.
The approach that Nikon use is to measure the density of the film dyes
at the very specific wavelengths of the LEDs, and these are selected so
that there is minimum crosstalk from the dyes in the other film layers.
This results in maximum colour purity at the time of the scan itself.
The "white light" scan that almost every other desktop scanner uses is
based on averaging the density of the film dye across the filter
response of each of the three colours. If you look at the spectral dye
density of most (all?) films you will see that while the red dye has
most of its density in the red are of the spectrum, there is also a
significant density in areas that are passed by the green filter and the
other dyes have some density in the other filter areas too. In
addition, if you look at the spectral response of the filters used on
CCDs (which are only dyed etch resist!) you will find that their
responses spread into areas of the spectrum corresponding to the other
film dyes as well. In simple terms, the colours smear into each other
and saturation is lost. This is no different from when you copy images
from one type of film to another under white light - each generation
results in reduced colour saturation because of the cross contamination
of the dyes and responses. What the scanner then does, knowing the
spectral response of the colour filters reasonably well (but not
accurately) is to increase the saturation by matrix manipulation to
compensate for the loss.
Don't read too much into the dynamic range figures, since both scanners
actually deliver less than they claim. The manufacturers compute their
dynamic range from the number of bits that the ADC produces, which
ignores the noise in the analogue circuit, the intrinsic ADC noise, and
any noise gain as a result of any necessary colour correction.
Effectively, each bit corresponds to 0.3 in dynamic range. So, with a
14-bit ADC, Nikon claim a dynamic range of 14x0.3 = 4.2, while Minolta
use a 16-bit ADC and claim 16x0.3=4.8. Both are unrealistic and, whilst
the Minolta may actually have a better dynamic range at the ADC output
(as a consequence of quieter analogue and better ADC circuits) it is
questionable whether this is retained in practice at the scanner output.
Sorry to complicate this
further, but I am not an expert in all of this and just want to get the
best for my money. Although everyone, after reading many reviews, seem
to really like the original Dimage Scan Elite 5400, and it seems that
there are a few complaints about the newer version, although it doesn't
seem to come from actual owners.
The complaints about the MkII mainly come from owners of the Mk1 and are
almost invariably about the removal of the Grain Dissolver - but nobody
seems to have ascertained if it is a loss of capability or if it is an
option that has become unnecessary as a consequence of the new light
source. Most of the complainers are the same people that were
complaining about the fact that the Mk1 took so long to scan when the
grain dissolver was used.
A diffuse light source might have actually been built in permanently,
but Mk1 owners are complaining because the words "grain dissolver" are
missing from the specification list. In other words, the complaints
appear to be based on nothing more than assumption.
I am not aware of anyone having evaluated a "Scanhancer" plate (the
original independent version of the "grain dissolver") with the MkII
scanner to see if it actually delivers the same benefit as in the MkI.
The complaints would be quite ironic if Minolta have actually delivered
a shorter scan time and a diffuse light source in the MkII ! ;-)
The other major change is the loss of Firewire interface. Well, if you
have several scanners running at the same time and use XP then that
could force a change of workflow to overcome some Micro$oft limitation,
but that is hardly Minolta's fault. One interface on the scanner is
certainly cheaper than two, and two interfaces is one more than will be
used.
