[argyllcms] Re: Generated camera profile too bright

  • From: Graeme Gill <graeme@xxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Tue, 03 Apr 2012 17:00:22 +1000

Andreas F.X. Siegert wrote:
> Nope definitely not. That is the original one before  started to push
> the exposure in the app.

Well, photoshop shows the 16 bit samples having a 16% value for
the white patch, and that's what the .ti3 file shows too:

      X      Y      Z      R      G      B
> D01 87.813 91.316 73.949 15.561 15.898 15.735

> Maybe I am misunderstanding something here. Exposure in camera oder
> pushing later on in linear mode?

That depends on your workflow, and what you want the profile to
do for you. If you want it to correct for non-linearities,
then you can't change the RGB values. If all you are after is
chromatic characterisation (you are assuming perfect linearity),
them linear scaling will be fine, but you need the white
patch to be at 100% to avoid the matrix scaling everything up.

> When I got close to overexposure on the white patch in profiled mode
> with the stock profile and then switched over to linear to generate the
> calibration TIFF, I still had a profile that resulted in in blown patches.

The profile is just doing what you're (implicitly) telling it to - it makes
the white map to PCS white. So if you've got the white patch
at 16% of the range, the profile will scale everything by 100/16
to make the white patch == PCS white.

Maybe there's something about how your workflow converts
from RAW to .tiff that is a problem ? Maybe it's not scaling
the RAW numbers to the full 16 bit .tiff range ?

Graeme Gill.

Other related posts: