Re: scanner profiling + w/b
Re: scanner profiling + w/b
- Subject: Re: scanner profiling + w/b
- From: neil snape <email@hidden>
- Date: Mon, 09 Jul 2001 14:17:11 +0200
on 9/07/01 12:58, Andre Sch|tzenhofer at email@hidden
wrote:
>
1. there's the target with a number of fields who are designed to describe
>
the gamut of the scanner.
Yes sort of. The patches are in fact a reflection of the pure and mixed
colors produced by the dyes/colorants of the films. They only describe the
film and don't know anything about your scanner.
>
2. this target is to be measured by a spectro creating data which is
>
dedicated to be the PCS.
Yes these are the measured values referenced to Lab or XYZ.
>
Maybe it's not that simple in particular, but basically I think that's it.
>
So there's just this one direction (RGB->LAB), just this one LUT and there's
>
no mapping, it's just called perceptual bacause it is the rendering default.
>
Choosing any other intend in the convert to profile dialog makes no
>
difference, because the other intents refer to perceptual/this one table.
That's a better way of saying default perceptual as in the case of input
profiles rendering intent doesn't have a meaning.
I lead this back
>
to the CMM who interpolates the missing steps. Now the file is ready to
>
color finetuning and composing. The results are ok, but theres still much
>
interpolation involved. What about setting w/b in the raw scan and/or in the
>
profiling process? Due to the fact that we scan individual Images, don't I
>
compare apples and oranges if I do so?
In fact as many have said before you profile a process and not the film
itself. So if you profile using the B/W points as experienced scan ops do,
then you can correctly use this as a workflow. In any case if the software
doesn't use ICC correctly then remapping to the B/W points wouldn't be much
worse than pushing the gamut around with curves.
Neil Snape email@hidden
http://mapage.noos.fr/nsnape