[argyllcms] Re: GCR Profiles, Gray Balance

From: Graeme Gill <graeme@xxxxxxxxxxxxx>
To: argyllcms@xxxxxxxxxxxxx
Date: Wed, 10 Jul 2019 13:17:28 +1000

graxx@xxxxxxxxxxxx wrote:

Hi,

I enclosed a 4 pages PDF which detail my analysis.

Page 1 show the gray balance curves from this profile.

using the default black generation.

Page 2 show the device values and the measurements.

Page 3 show the a* / b* for all those values.

Doesn't look right for relative colorimetric, so it's not
clear what you are showing. Expected Relative colorimetric
value for white is (by definition) neutral:

xicclu -fb -ir Argyll_Enhance_2019_07_09.icm
100.000000 0.000000 0.000000 [Lab] -> Lut -> 0.000007 0.000006 0.000004
0.000019 [CMYK]

xicclu -ff -ir Argyll_Enhance_2019_07_09.icm
0.000007 0.000006 0.000004 0.000019 [CMYK] -> Lut -> 99.998594 -0.000040
-0.000118 [Lab]

I have only one question.

The original IT8.7/4 characterization target show that the 400% CMYK build
has a CIE
Lab value of 24.94 1.36 0.62 which seems rather “neutral” to me.

The profile (which strives for smoothness/consistency as well as fit)
gives the value:

xicclu -ff -ia Argyll_Enhance_2019_07_09.icm
1.000000 1.000000 1.000000 1.000000 [CMYK] -> Lut -> 24.960532 1.345977
0.648357 [Lab]

which is consistent with the measurement sample.

What I wonder is the determined Total Ink Limit so far of the 400%?

The colorimetric table priorities maximum accuracy (minimum delta E)
over every other aspect except ink limit. And the chosen black value (according
to the
profile) has a lower L* than CMYK 1 1 1 1:

xicclu -fb -ir Argyll_Enhance_2019_07_09.icm
0.000000 0.000000 0.000000 [Lab] -> Lut -> 1.000000 0.000000 0.729204 1.000000
[CMYK]

xicclu -ff -ir Argyll_Enhance_2019_07_09.icm
1.000000 0.000000 0.729204 1.000000 [CMYK] -> Lut -> 23.794386 1.488683
1.866661 [Lab]
1.000000 1.000000 1.000000 1.000000 [CMYK] -> Lut -> 26.294332 0.801951
3.300543 [Lab]

So the "blackest neutral" happens to have a total ink of 273%. Because the set
(default
because you or the .ti3 haven't overridden it) ink limit is 390%, other colors
are free
to use more ink.

Note that the black point is a compromise between low L* value and neutrality.
The
compromise is reduced as you increase L*:

xicclu -fb -ir Argyll_Enhance_2019_07_09.icm
25.000000 0.000000 0.000000 [Lab] -> Lut -> 0.993247 0.261571 0.623036 0.991476
[CMYK]

xicclu -ff -ir Argyll_Enhance_2019_07_09.icm
0.993247 0.261571 0.623036 0.991476 [CMYK] -> Lut -> 24.999838 0.000975
0.000714 [Lab]

As you will see on Page 2, Argyll determine TIL is 285%, made up of 99c 24m
63y 99k
which corresponds to a measured CIE Lab value of 23.85 -0.46 -3.87.

No it didn't. See above.

As you will see on Page 3, using RelCol, I would expect a “monotonic”
transition from
the substrate color to the “darkest neutral”, in this case 285% TIL.

The Colorimetric table doesn't compromise minimum delta E for "continuity".
There are no
such guarantees - they depend on the device characteristic and black generation.

Is there an option, perhaps, I am missing? That would force the darkest
neutral to be
“neutral”? It seems that this would be the cherry on the icing…

It is neutral - you haven't shown otherwise as far as I can tell.

P.S. Is it possible to use HTML for posting to the group? I’ve recently seen
inline
graphics?

You can include attachments as long as they aren't too large.

Cheers,
Graeme.

References:
- [argyllcms] GCR Profiles, Gray Balance
  - From: graxx

[argyllcms] Re: GCR Profiles, Gray Balance

Other related posts: