[argyllcms] profile <1 million dollar question!>
- From: Roger Breton <graxx@xxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Fri, 29 Feb 2008 14:08:57 -0500
Well, with the help of Graeme and Alastair, I have the feeling I'm getting
somewhere with Argyll -- I never doubt it. But I'm all excited to find
interesting questions as I go along, little nuggets of color management
knowledge.
Here's the comments from a profile session:
>>profile -v test_with_stats
> Total ink limit being used is 390%
> No black ink limit being used
> No of test patches = 182
> Estimating white point
> Approximate White point XYZ = 0.814167 0.842132 0.744550, Lab = 93.543000
> 0.423000 -4.416000
>
> Creating optimised per channel curves
> Initial White Point XYZ 0.814167 0.842132 0.744550, Lab 93.543000 0.423000
> -4.41 6000
>
> About to optimise temporary matrix
> .....................
> About to optimise input curves and matrix
> ..............................................................................
> ..............................................................................
> ..............................................................................
> ..........................................................
>
> About to optimise output curves and matrix
> ......................................
> About to optimise input curves and matrix again
> About to optimise input, matrix and output together
> .....................
> About to adjust a and b output curves for white point
> About to create grid position input curves
>
> Create final clut from scattered data
>
> ******************************************************************************
> ******************************************************************************
>
> Doing White point fine tune:
>
> Before fine tune, WP = XYZ 0.000000 0.000000 0.000000, Lab 99.786657 -0.259466
> 0.309300
>
> Creating fast inverse input lookups
> Compensate scattered data for input curves
> White point XYZ = 0.808475 0.837511 0.743901, Lab = 93.342285 0.184253
> -4.705911
>
> Find black point
> K only value (Lab) = 8.856753 1.517002 0.394401
> Black point XYZ = 0.007579 0.007461 0.005880, Lab = 6.739928 1.551594 0.519903
>
> Done A to B table creation
>
> Creating B to A tables
> 100%
>
> Done B to A tables
>
> Creating gamut boundary table
> 100%
>
> Done gamut boundary table
> profile check complete, peak err = 4.363131, avg err = 0.961138
OK. How I just love this stuff.
My question comes from what I was doing with the cctiff converted to Lab
file. You see, I want my students to verify for themselves that when they
point to a patch in the Lab version IT8.7/3 Basic they can compare with the
actual EyeOnePro measurement, since I ordered cctiff to convert AbsCol.
A quick comparison turns out a very good fitting of the device data by
Argyll. Very good. For example, for the 100C patch, the EyeOnePro measured
51.499 -34.625 -57.789 whereas the TIFF file opened up in Photoshop shows
52 -35 -57. I understand Photoshop rounds off the data for the Info palette.
To me, that's an excellent result. So I continue with the 100M patch. This
time the EyeOnePro reported 40.705 77.134 -7.750 while Photoshop showed this
one at 41 79 -10. Not quite as close as on the Cyan. I continue my
comparison and, in most cases, the fitting is excellent -- good math and
modeling, Graeme!
But then I turn to the comments by profile and I think to myself, maybe
there are some profile generation statistics I missed that would tell me how
good was the fitting of the colorimetric data by the profile? And the only
thing I can think of is the very last line:
> profile check complete, peak err = 4.363131, avg err = 0.961138
My thinking is, I would like to demonstrate to the students the effect of
using a larger device sample in modeling its behavior. So, if I use
something like an IT8.7/3 928 sample set (as opposed to my lowly 182 sample
set) or an ECI2002 or an IT8.7/4 or, something like 3000 patches by Argyll,
can I reasonably expect those two statistics to go down?
Or is that a sign of the device ill-behavior? This is from an Epson
printer, by the way. Linearized by a CMYK RIP.
Just curious...
Roger
Other related posts:
