[argyllcms] Re: V1.1.0 Perseptual rending Qst
- From: "Wallden Magnus" <prn.scr@xxxxxxxx>
- To: <argyllcms@xxxxxxxxxxxxx>
- Date: Tue, 2 Feb 2010 06:26:55 -0800
Hi !
Exuse for the delay.
I have tried to doublechech almost everything, pherhaps I missed
somthing.
-----------
I have made thre scanner profiles:
xyz_cLUT
colprof ..... -qm -ax -u .........
Lab_cLUT.icm
colprof ..... -qm -al -u ..........
Matris Shape
colprof ........ -qm -as -u .........
(perhaps -u has no influence here)
---------------
The profiling target has some patches whit small valuse in some patches,
scanin delivery this result :
row 534 from the ti3-file
xyz cLUT:
534 0.0000 45.882 100.00 0.0000 52.245 41.091 -8.4961e-006
-small but negative value
lab cLUT
534 0.0000 45.882 100.00 0.0000 48.450 37.823 1.0219
matris/shape
534 0.0000 45.882 100.00 0.0000 60.461 47.425 6.0778
---------------
Perhaps it is possible to set small values below zero to zero, or to
some
small positive value, or abort processing, what now is the best.
---------------
The use of Matris/shape and LabcLut gives both good profiles, whit some
small
visible differences, in all rendering exept "abolute calorimetric",
where the white point is to low, I suppose.
I have problems to decide whitch I prefer.
best regads
Magnus
Other related posts:
