[argyllcms] Re: Flat shaper

  • From: Graeme Gill <graeme@xxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Sun, 22 Mar 2009 16:38:00 +1100

Pascal de Bruijn wrote:
Can anybody explain to me why I'm getting a nearly linear
shaper?

You're characterizing a linear light device ?

If you take a look at some of the .ti3 grey patches. ie:

SAMPLE_ID XYZ_X XYZ_Y XYZ_Z RGB_R RGB_G RGB_B

O05  45.860 51.520 38.130  50.245 55.401 49.783

You'll notice that (roughly) 50% device gives you 50% Y. This
means it's roughly a linear light device, so the shaper
should be almost a straight line.

Are you shooting this with a camera using RAW ?
If so, this is to be expected.

You'll get different shaped curves with a CLUT
profile, because the aim of the per channel curves
in that case is to make the CLUT grid spacing perceptually even...

Previously with my IT8 target I noticed pretty big corrections...

Please note that device profiles don't make corrections, they
characterize (ie. model) the device. Because in a matrix profile
the matrix is a linear light transform, the per channel curves
must convert from device to linear light, so their shape
is determined by the device characteristic. In a CLUT profile
both the per channel curves and CLUT table can represent
non-linear transforms, so the shape of the curves is not
determined solely by the devices characteristic.

Graeme Gill.

Other related posts: