Olympus-OM
[Top] [All Lists]

[OM] Re: Profile for FS4000

Subject: [OM] Re: Profile for FS4000
From: Nils Frohberg <nilsf@xxxxxxx>
Date: Tue, 12 Jul 2005 15:10:27 +0200
Thanks for your (as always) great explanation.

On Mon, Jul 11, 2005 at 03:11:02PM -0700, Moose wrote:
> It's a little more complicated than I had hoped. Profiling a flatbed for 
> reflective material is really simple.Scan the target, match the target 
> outline to the scan and make a profile.
> 
> With transmissive material, it seems like it should be very similar, but 
> a two step process. First profile the scanner with a film target, then 
> profile inidvidual films.

I thought it would be possible to do things in such a way: You take
transmissive material with a clear base. This allows you to register
the color of the light source. Further you add various "real" color
fields. (Whatever that may be.) With these you can calculate the way
the scanner interprets these "real" colors. Obviously, the film you
want to use later should contain a subset of the colors of your
calibration strip. (Unless you want to start extrapolating the
color rendition of colors outside the spectrum).

After you did this, you know how the scanner handles "white", "red",
"blue", "green", and so on. Thus, it should be possible to render
the scanned film with all its characteristics into your graphics
file.

After you did this, profiling the film will return the scene into
its "true colors", no matter what the films looks like. With this
method, Velvia will scan the same as Sensia. (In theory at least.)
And the scans will look like the scene in real life.

> If course, I didn't have a target and went ahead and profiled a few 
> films. Works great, but theoretically weak. Now each film profile is 
> like a combo scanner and film profile. That's fine for the moment, but 
> means aging of the light source or use of a different scanner make them 
> less than accurate.

Yes. Plus, you might want a Velvia scan to keep that Velvia look. But
you don't want a red cast in the green trees. (This is why you need to
profile the scanner by itself!)

> When I've tried it and have some idea of how it works, I'll post what I 
> find.

I'm looking forward to this! :)

> > (I hope I can prevent buying an expensive color chart for a one-time 
> > action.)
> >
> Well, in theory, it shouldn't be a one time action, as light sources age.

True, this is something I did not consider up to now.

I wrote quite a bit now without adding any value, I think :)

Nils

-- 
The Moon is Waxing Crescent (30% of Full)

==============================================
List usage info:     http://www.zuikoholic.com
List nannies:        olympusadmin@xxxxxxxxxx
==============================================

<Prev in Thread] Current Thread [Next in Thread>
Sponsored by Tako
Impressum | Datenschutz