Re: Paper white and the seven dwarves
Re: Paper white and the seven dwarves
- Subject: Re: Paper white and the seven dwarves
- From: "Andre Schützenhofer" <email@hidden>
- Date: Mon, 04 Feb 2002 17:44:23 +0100
> >By altering the data in the patch text file for the
> >white patch to numbers which provided a better visual match and
rebuilding
> >the profiles I was able to achieve a much better end result.
> I've been commenting on such issues here for some time now, without
> generating much positive response... Andrew says it isn't a problem,
Henrik
> says to use good Euro paper without whiteners, and I keep on correcting
as
> needed...
Some month ago I made some tests with an approach dealing with the UV
problem, but I still have not found someone to verify this. It depends on
the way how paper white is calculated out of the measured data.
Normally, the aim is to have a neutral white point in the profile. But I
think that's not all, because - caused by the optical brighteners - not only
the paper appears blue to the measuring device, the colors include this blue
cast too, mainly the lights and quatertones. In order to generate D50 LUT's,
paper color has to be calculated out. So the D50 graybalance within the
profile is based on blue paper, seen through the measuring device.
Therefore, to neutralize the blue cast, the graybalance becomes "warmer"
than it would be on neutral paper.
When we edit the white point tag, the graybalance is still based on blue
paper. What we have now is a profile with a neutral white point but there's
still the graybalance not matching the real visual appearance of the paper
to the human eye.
If the white patches of the target files are responsible for this
calculations, there is - in theory - a solution. What we want is to have a
neutral profile with a neutral graybalance without shifting of colors and
with a neutral white point on neutral paper, which is impossible caused by
optical brighteners. To edit the white patches in the measured data file to
lets say L=95 a=0 b=0 will not have the desired result, because it does not
neutralize the graybalance.
The trick is to edit the white patches to a value twice as high as the
actual distance to the gray axis.
Example: measured L=X a=1 b= -4 should be edited to L=X a=2 b= -8.
What I want is to make the profile creator think the paper is "twice as
blue" as it is - but not the rest of the measured colors. Therefore, by
building LUT's refering to D50 we should be able to calculate papercolor out
of the measured data because we calculate with a papervalue refering on
measured colors half the distance to neutral, so the result should be 0. Not
to mention to not forget to edit the white point tag afterwards.
It looks like a practicable method and some tests showed good results, but
it should be confirmed mathematical, and perhaps its nonsense. If so, sorry.
Regards, Andre Schuetzenhofer
_________________________________________________________________
Werden Sie Mitglied bei MSN Hotmail, dem grv_ten E-Mail-Service der Welt.
http://www.hotmail.com/de