[filmscanners] Re: umax film scanners

["Peter Klein" <pklein@2alpha.net>]

I used a Nikon LS-2000 (2700 dpi) up until recently, when I replaced it
with a Canon FS4000.  I rarely print larger than letter size, but it was a
very worthwhile upgrade nonetheless.

To expand on what Karl mentioned, there's a big difference in the grain
aliasing (or artifacting) issue, particularly with fast "real" B&W
film.  It's not just that the prints from the 4000 dpi scanner are sharper
or more detailed.  They are, but you don't really notice that on an 8x10 or
letter size matte print, and you only begin to notice it on an 11x14.  What
you do notice is that 4000 dipi prints look smoother and less grainy. A
little grain never hurt anyone--it's part of the nature of faster film. But
there's no need to have more of it than is really there.

The 4000 dpi scanner shows Tri-X film grains pretty much as they look when
observing the negative with a high-powered loupe or microscope.  With the
2700 dpi scanner, Tri-X scans looked more speckled or blotchy.  At 4000
dpi, many film grains scanned as a few pixels or less.  At 2700 dpi, they
scanned at a half dozen to a dozen pixels or more. And remember, the pixels
cover more film area at 2700 dpi.  With Neopan 1600, the effect was even
worse.  Essentially, it made these films look even grainier than they
actually are.

Call it aliasing, artifacting, interference patterns or what you like.  The
point is that straight 2700 dpi scans *looked* decidedly grainier.  Partial
solutions were to use VueScan's grain reduction feature, or run the TIFF
through Neat Image after scanning.  But even applied very sparingly, these
measures cut down on the detail just enough to make the prints have a
little less edge.  With 4000 dpi, these measures are not necessary.

All this matters less with the better fast color print films or chromagenic
film (T400CN), and is not noticeable with Provia slide film. With these
films, the smaller grains or dye clouds didn't hit 2700 dpi's "not so sweet
spot."

I enjoy available light work.  I find real B&W film is just plain better
for stuff with dark shadows.  So for me, 4000 dpi is better.  If I was
primarily a color photographer, I'd probably have not bothered to
upgrade--I'd just send the negative or slide to a lab when I wanted the
occasional bigger print.

Now if only the files weren't so *big*!   :-)

--Peter Klein
Seattle, WA

>From: "Karl Schulmeisters" <karlsch@earthlink.net>
>----------------------------------------
>
>4000dpi is not just '20% better'.  In point of fact it is more like 50%
>better resolution.  4000x6000 resolution scales to 307x315dpi - just barely
>enough to get a decent print.  Which makes sense, since 13x19 is pushing the
>limits of grain when you print optically - indicating that's close to the
>information boundary for 35mm film.  Other digicam vs. film sites have put
>the data limit of 35mm film at between 16 an 24 Mpixels whereas 2700dpi only
>pulls out 10Mpixels of data..
>
>The worst part is that sampling at about 1/2-2/3 grain resolution is likely
>to give you the worst grain artifacting possible because of how the
>'sampling grid' interacts with the 'grain grid' - especially on tabular
>grained films like TMAX.


----------------------------------------------------------------------------------------
Unsubscribe by mail to listserver@halftone.co.uk, with 'unsubscribe 
filmscanners'
or 'unsubscribe filmscanners_digest' (as appropriate) in the message title or 
body