Archived posting to the Leica Users Group, 1999/09/27

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

Subject: Re: [Leica] scanning Leica slides vs colorneg
From: Godfrey DiGiorgi <ramarren@bayarea.net>
Date: Mon, 27 Sep 1999 17:34:28 -0700 (PDT)

> It is my understanding that your typical scanner is really not up to
> snuff to slides and does better with ColorNeg while a Polaroid 4000
> can deal with slides fine. By the end of the year who knows the
> options but can anyone confirm this or know much about it?

Transparencies, being positive images, have more dynamic range than
negatives (the materials are not sensitive purely linearly, the response
curves are better characterized as partial differential equations of order
2 - the useful range of exposure tends to happen in the toe and linear
section of the curve, each transfer from negative to positive includes a
rise in contrast just like making slide duplicates includes a rise in
contrast). What limits scanning of transparencies is similar to what
limits making good prints from transparencies: the contrast gain of a
slide which has just too much dynamic range burns highs and loses lows.

Given this, the limiting factor in how well you can scan a transparency or
negative has to do with the scanner's bits per channel which relates to
its DMax capabilities (as well as all the other details of quality
alignment, good focus, etc). Shooting transparencies for scanning is much
like shooting them for printing: you have to be careful to keep the
significant portions of the slide to a tight lighting ratio not only to
keep within the transparency material's exposure tolerance but also to
minimize contrast gain. A 12 bit per channel scanner at 2700 dpi will do
quite a good job with slides where a 10 bit handles only the better
exposed slides well, and an 8 bit unit's capabilities are rather limited. 
Most negatives scan very well even with 8 bit units, although 10-12 bit
units do seem to wring more brilliance out if properly calibrated and
adjusted for the film. Part of this might be that the more expensive
scanners generally get better software as well, part of a professional
package. 

Added dpi resolution surely helps no matter what but the limitations here
are more related to how much detail your digital image will have at what
enlargement size rather than how well it will reproduce the color/tonal
scale. A 2700 dpi film scanner, calibrated properly, will return enough
pixels to make a quality 12x-16x enlargement from 35mm format (16x is
about 170 ppi, 12x is a healthy 225 ppi). While this doesn't beat the top
notch printing technology available today, it's adequate for most people's
home uses. Going to 4000 dpi gives you closer to the drum scanner
resolution capabilities if you want really large prints. 

Something must also be said for the fact that some scanners and their
drivers are better optimized for transparencies vs negatives and vice
versa. Part of the reason I chose the Polaroid Sprintscan 35E/S (10bit,
2700 dpi, built-in DSP) was that in testing it, it seemed to be much
better at rendering color negatives than the comparably price Nikon
scanner was in 1996, the Nikon at the same price was better at rendering
slides. 

Godfrey