Anthony, see my site with a few samples that show the LS40 with Nikon Scan
3.1 with "difficult" slides (Provia 100 F RDP3, Velvia). This combination
never clips highlights and gets a lot out of the shadows with little noise.
It isn't perfect in highlights because I think it doesn't have quite enough
contrast in this range - effectively there is an S-shaped curve in the
scanning process which places more emphasis on the middle tonal values at
the slight expense of tonality in the shadows and highlights.
Unfortunately NS 3.1 blows the highlights in negatives, if those negatives
cover a very very wide range (I'm guessing 10 stops+) - NS behaves a bit
like a digital camera in its exposure, i.e. it sacrifices the highlights in
the name of creating a "punchy" image - meaning that you don't get such
great control. Luckily most negatives (I've shot!) don't contain the range
for this problem to arise.
Vuescan totally avoids this problem.
I omitted to mention that I had Auto Exposure turned off whilst scanning
with Nikon Scan. It turns out, in my opinion, that Auto Exposure OFF is
better than on, when scanning slides.
The LS40 is not noisy in the shadows. If you use slide film, grain aliasing
problems are ameliorated, too, as far as I can tell (by comparison Supra 400
negative film shows enough grain aliasing in shadows that one is almost
obliged to chop the bottom off using Levels in Photoshop).
I dare say Nikon tuned the LS40 (and NS 3.1) for slide scanning.
I've learned that when using Vuescan one should use a higher gamma and
greater brightness than I was using. I dare say that as I don't have my own
slides to scan I am not interested in optimising this process. I originally
posted this site to compare multi-scanning with single-scanning (something
that Vuescan supports with the LS40 - but Nikon Scan cannot do
multi-scanning). I spent about 10 hours all told creating that site and I
think it would have taken me another 10 hours to optimise Vuescan's
settings - so I apologise for the "introductory" nature of the tests.
Jawed
> -----Original Message-----
> From: owner-filmscanners@halftone.co.uk
> [mailto:owner-filmscanners@halftone.co.uk]On Behalf Of Anthony Atkielski
> Sent: 26 August 2001 10:30
> To: filmscanners@halftone.co.uk
> Subject: Re: filmscanners: Best film scanner, period!!!
>
>
> Austin writes:
>
> > Why do you consider 3.4 "too low", and for what
> > is it too low?
>
> I scan slides.
>
> > I don't mean this to come across snide, but do
> > you actually know what a density range of 0-3.4
> > means?
>
> Yes. It means 12-bit output, which gives a _theoretical_ dynamic range of
> 4096:1, or log(4096)=3.6, for density range. (A range of 3.4 actually
> corresponds to about 2500:1.)
>
> Unfortunately, actually achieving this with the scanner is much
> more difficult.
> My LS-2000 is just barely able to extract most of the detail from
> a slide, but
> it is very noisy in the shadows and tends to blow out in the highlights.
>
>
