[UFRaw-Devel] WB temperature calculation

[Udi F. <udi...@gm...>]

Hi,

I just committed a new calculation for the WB temperature. The old
calculation was a reminiscent that I never knew what it was based on.

The current calculation is based on data from
http://www.brucelindbloom.com/ which seems trustworthy. I did
generalize the calculation to the range 2000K to 4000K so don't trust
me on this range. Still I'm most certain that it is correct.

I'm using the color matrices to convert from RGB with a reference
white of D65 (6500K) to the camera color space. The linear algebra,
was a bit more complicated than I thought, but I finally got it right.
There are some comments in ufraw_set_wb() that explain the
calculation.

Still, you will see that UFRaw's temperature settings does not match
Nikon's temperature. I thank all involved in adding the D2X presets,
which give a very good base for comparison. At low temperature the
mismatch is small and can be ignored, but as the temperature rises the
differences become bigger.

There are two things that I don't trust, which could lead to this
mismatch. One is Nikon and the other is Adobe. Nikon does not require
explanation. Adobe are responsible to the color matrices in dcraw. It
could be interesting to try and find a color matrices that reproduces
Nikon temperatures. If it is possible, then I would wander if such a
matrix also reproduces better Nikon colors.

Any new insights are welcome,
Udi