Apache-Talk @lexa.ru 

Inet-Admins @info.east.ru 

Filmscanners @halftone.co.uk 

Security-alerts @yandex-team.ru 

nginx-ru @sysoev.ru 




      :: Filmscanners
Filmscanners mailing list archive (filmscanners@halftone.co.uk)

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

filmscanners: Sprintscan 120 and new negative proile scheme (LONG)

OK, here's my 2 cents on this.

First, Polaroid are not alone in offering this. Silverfast ships with a
bunch of profiles, as does Vuescan.

Are they helpful? 


The first problem is that they don't keep up to date with the emulsions. It
is extremely confusing trying to work out which profile goes with which
emulsion. Vuescan for example has profiles for Tmax 400 in D76 at various
contrast indices, but no profiles for TriX or any Ilford films. It has
profiles for all six generations of Kodak Gold 400 but nothing for 400 VC.
And so on. Silverfast has profiles called Kodak 1, Kodak 2 and Kodak 3 but
no clue as to which emulsions they refer to. So, inevitably, they are out of
date as soon as shipped.

The second problem is that it's never clear what these 'profiles' are
supposed to do. I remember reading something in the Vuescan manual which
said something about 'making the image look as much like the original scene
as possible'. In other words, applying a inverse H&D curve, presumably, plus
a custom base removal mask in the case of color neg. This seems absurd (the
first part, I mean) since if successful it makes all emulsions look the
same. The Silverfast profiles, for their part, apply a custom 'color space
expansion', which means that they come with predefined min/max set points
for the individual color channels.

Does this *actually* work on color neg? No.

Do minilabs read the emulsion type before printing neg? No.

The third problem is that *even if the profiles were useful and worked
properly* film developing varies so much that you always have to tweak

I wish scanner manufacturers would stick to the knitting. What is ACTUALLY
useful (ok, to me) in a scanner driver?

-- faithful rendition of *actual image colors* in all cases. In other words,
proper calibration of the scanner CCD. Silverfast does this well on the

-- ability to handle color neg properly. This is SO simple and yet rarely
done well. In essence, remove the orange mask, invert and set the expansion
points for the individual channels. Vuescan does this well, but the GUI is
very confusing which negates the benefit.

-- ability to handle bw neg properly. As Austin says this means set
black/white points and tonal curve. This means we need a good, detailed
histogram and a curves box which functions as well as the industry standard,
Photoshop. Neither Silverfast nor Polascan are up to snuff on this.

-- the ability to output gamma-corrected high-bit scans. In the case of
neg, inverted gamma corrected high bit scans. In the case of color neg, mask
removal on high bit scans.

-- intuitive GUI which reflects standard interface guidelines. Silverfast,
Polascan and Vuescan all fail miserably on this one. Polascan not quite as
miserably as the others.

Now, once you can do all this you can add as many consumer-friendly bells
and whistles as you like. But UNTIL you can do it... yeah, well, you get it.

So, film profiles? Who cares? There's a lot of stuff to get right first.

John Brownlow



Copyright © Lexa Software, 1996-2009.