[argyllcms] Re: help with camera profile

  • From: Stephen T <stwebvanuatu@xxxxxxxxxxxx>
  • To: "argyllcms@xxxxxxxxxxxxx" <argyllcms@xxxxxxxxxxxxx>
  • Date: Thu, 19 Jul 2012 04:14:41 -0700 (PDT)


I have used ufraw, I have an IT8.7/2 and, after several attempts at camera 
profiling, I don't use them anymore.

1. Ufraw is based on dcraw but it does some weird scaling of the linear TIFF 
Output the same raw photo of the camera target as a linear TIFF in ufraw, 
dcraw, DarkTable and in Raw Therapee.
Compare the scanin results: ufraw is anomalous.
Ufraw also lacks tools like RGB tone curves that are most useful for rendering 
the output.

2. It is really difficult to control glare (flare) and get good uniformity of 
REFLECTANCE when shooting in sunlight. An incident light meter is not helpful. 
Check the average RGB around the edges of the target in ufraw. If the variation 
is more than 2 or 3% then the profiling will likely have problems. Secondly, 
noon sunlight is not D50.

3. The IT8.7/2 target has a good spread off patches but the Kodak paper has 
flourescent whitening additives. The target white point will not be neutral in 
direct sunlight (lots of UV).

4. Digital cameras have good system linearity and a simple matrix profile 
should be satisfactory. The ColorChecker Classic 24 has good spectral 
properties (more natural than IT8.7/2), is adequate for matrix profiles, 
reasonably priced and well-supported by Argyll CMS and other profiling software.

What you get in the end is a rough and ready ICC matrix profile that should 
deliver realistic colour for sunlight and similar daylight illuminants.


Other related posts: