Archived posting to the Leica Users Group, 1999/10/15

[Author Prev] [Author Next] [Thread Prev] [Thread Next] [Author Index] [Topic Index] [Home] [Search]

Subject: Re: [Leica] Nikon 4000 dpi scanner?
From: "Henning J. Wulff" <henningw@archiphoto.com>
Date: Fri, 15 Oct 1999 14:37:12 -0700

At 7:06 PM +0200 10/15/99, Pascal wrote:
>On 15-10-1999 00:20 Robert G. Stevens wrote:
>
>>Is the 3.6 with a multipass scan?  What is its single pass Dmax?
>
>Good question, and I honestly don't know if that 3.6 is only obtained
>when the 8x multisampling is used. I checked my Nikon manual and didn't
>find a clue at first sight.
>
>I normally use 1x sampling, unless slides have darker areas that need to
>be clearly "translated" into a scan. Then I am using the 8x, but it's
>very slow, unfortunately.

Again, the quoted figure of 3.6 for Dmax bears no relationship to reality.
Dmax for the Nikon is definitely less than 3.0, and probably around 2.6.
Advertised numbers for Dmax and dynamic range are nowhere near reality. The
dpi rating of scanners give you information on the size of file, but are
based on CCD sensor counts, not on achievable resolution. All scanners
achieve a lower resolution figure than advertised, with scanners that
acquire images through glass giving lower figures than glassless scanners,
as a general rule.

The Dmax capability does not change with multisampling. Only the CCD noise
level gets averaged out so that denser areas are cleaner. Sampling is
available at a 1x, 4x and 16x level, not 8x.

If you have a very dense slide, and it is somewhat underexposed, you can
raise the 'analog gain', ie, turn up the light, and get more information
that way. A good level to choose is one where the slide border turns out to
have a value of around 3-4 for all channels. Then you will probably get all
the densest information the slide can offer, and it will be clean with the
16x multisampling. If you have a slide with a lot of important information
both in the highlight and shadow areas, and you want the digital image to
show both (naturally compressing the midtones to some degree in the
process), the best way is to turn down the analog gain for one scan
(highlight), and do another scan with the gain turned up (shadow). In this
case you would probably have the border at levels of 8 to 10 for the shadow
scan, and clear film base at a level of 245 on all channels for the
highlight scan. The main problem with this is that sometimes the slide
moves a little bit between scans. Rotation can screw it up a lot.

   *            Henning J. Wulff
  /|\      Wulff Photography & Design
 /###\   mailto:henningw@archiphoto.com
 |[ ]|     http://www.archiphoto.com