[argyllcms] Re: Characterizing a scanner using Argyll

  • From: Klaus Karcher <lists@xxxxxxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Wed, 04 Feb 2009 09:30:49 +0100

Gerhard Fuernkranz wrote:
One limitation of a matrix model is certainly that it artificially
limits the gamut, i.e. you'll lose those valid RGB numbers which turn
into negative XYZ numbers after applying the matrix (and which are then
clipped to zero by the CMM). No subsequent transformation can bring the
clipped numbers back.

In my case, this limitation seems to be an advantage: the results of the matrix profile stay largely within reasonable limits while the "colors" returned by the CLUT profile fall outside the human visual subspace near the bottom of the device space (e.g. RGB 0/0/0 -> Lab 0/-8/-4)

The lowest value in the ti3 file (i.e. the light trap) is RGB 18 15 15 -> XYZ 0 0 0 (RGB values scaled from 0 to 255). I think it's reasonable to map device values below a plane trough this point to black.

Klaus

Other related posts: