[argyllcms] Re: camera profiling with newest version of Argyll

  • From: Graeme Gill <graeme@xxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Tue, 01 Sep 2015 10:48:58 +1000

Chad Johnson wrote:

I'm attempting to build a color profile of my camera and I'm not able to get
a reasonable
result with v1.8.0. The last time I profiled my camera was with v1.5.1 and
the results
were good. The following is a simplified version of my workflow which I'm
using for this
particular test. I executed everything exactly the same in both versions of
the software:

scanin -v -dpin IMG_4147.tif ColorCheckerSG.cht ColorCheckerSG-xrite.cie

colprof -v -D"CanonEOS70D" -qh -am -u IMG_4147

Hi, the answer is that there was a bug in V1.5.1 in the colprof -am option -
even though "matrix only" is specified, the resulting profile has a curve,
and in new versions this bug is fixed, and you get a matrix only.

You can see all that in the verbose output and the fit errors.

Your data doesn't fit well to a matrix only - you don't appear to
have linear light (i.e. RAW) RGB from the camera.

To get a profile in V1.8 that is similar to what you were getting before,
use colprof -aG

Graeme Gill.

Other related posts: