Archived posting to the Leica Users Group, 1999/03/18

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

Subject: Silverfast, try a demo! [was Re: [Leica] Nikon LS-2000: Thanks!]
From: George Hartzell <hartzell@corp.webtv.net>
Date: Thu, 18 Mar 1999 10:18:44 -0800 (PST)

Nathan Wajsman writes:
 > Dear Joe, Tina, Robert, James, Tom:
 > 
 > Many thanks for your advice on the Silverfast. The opinions are strongly
 > divided, with some people saying it is basically not worth the price
 > while others sing its praises. I think that I will buy the scanner
 > without the Silverfast, learn to use it with the standard software, and
 > add the Silverfast later if I feel a need for it.

Silverfast used to have a demo version of their software available on
their web site.  It was full featured, except that it embossed a big
SilverFast logo in any image it saved.  It's obvious enough that you
can't really use the saved image, but not so overwhelming that you
can't evaluate the qualities of the image you got.

Demos still seem to be available at http://www.lasersoftint.com/, try
it and see if it "floats your boat".

There's a related argument about whether one should manipulate an
image in the scanning software, or downstream (e.g. in photoshop).
I've heard knowledgeable people espouse both views.

The most convincing argument that I've heard applies to scanners that
scan at a bit depth of more than 8 bits per pixel/channel.  Since
almost everyone works with 8 bit per pixel/channel images in
photoshop, the drivers for these scanners have to map the image into
an 8 bit per pixel space (discarding information in the process).  In
general, it's better to work in the richer data space, then map it to
the lower bit depth.

In reality, I think that it ends up being a contest between the
quality that you'll be using and the effectiveness of the available
tools.  Lower quality output devices or small changes to an image
might work well with either technique.  If the accuracy or usability
of the post-scan tools is much better than the pre-scan tools, then
one might get better results working post-scan.

Just how much this matters came home to me in a digital photography
workshop (http://www.sjphoto.com, highly recommended!), playing with
portraits with a monochrome medium format digital back that sampled at
14 bits per pixel.  There was so much data in the raw scan (both in
bit depth and raw Dmax) that amazingly different, yet high quality!,
images could be created by tuning the contrast, brightness, etc....
That rich information was thrown away in the 8 bit photoshop file.

That said, I use the supplied Nikon drivers for my LS-1000 and Epson
PhotoEX.  For high end projects I try to get involved in the scanning
step, since I'd like to make the decisions about what to throw away.
At a minimum, I like to talk to the person doing the scanning.  Think
of it as printing your own negs, there's a lot of room for
interpretation.

g.