[argyllcms] Re: ArgyllCMS: shadow details and blue color issues. What I'm doing incorrectly?
- From: "Xavi" <aruiz@xxxxxxxxxxx>
- To: <argyllcms@xxxxxxxxxxxxx>
- Date: Sat, 20 Oct 2012 00:16:36 +0200
Thank you for your reply Nikolay.
"Absolute cororimetric compiance between display and the paper is only
achievable in controlled condition when the paper brightness has the same
level as the display brightness because very vell controlled illumination.
Is this Your case?"
I have a home made lightbox that offers me D50/500lux/CRI>93. My main
monitor is a wide gamut with IPS panel, calibrated to D50, neutral black and
150cd/m2. Then I imagine that the white point brightness of the media and
the monitor will be very very similar.
" In such case it would be better to apply some kind of color mapping (for
example -ila) which actually can improove the image appearance but can't
gain colorimetric compliance.
"
I imagine that you want to write "-tla" as an option of colprof. It is
correct? But permits me explaining to you my problem, and then, please tell
me if I must to use the option "-ila"/"-tla" or read the patches with the
option "chartread -e" (that I'm not sure if it is supported by the
ColorMunki).
But my problem is not based to a visual taste, preference, appaerence or due
to illuminance.
I have used my subjectivity to evaluate my prints for a long time, but I
don't have the security that I print the same colors that I see on my
screen.
Only some days ago I have decided not to use the subjectivity to evaluate my
prints nor use the softproof. Then I have decided only to use the
ColorMunki to read the printed patches to compare with the source Lab
values.
Then it is not relevant the lightbox and also is not relevant my monitor
because my source are Lab values and the destination are these Lab values on
paper read by my Colormunki. Source=Lab values of patches. Destination= Lab
values on paper reads by Colormunki.
Here I don't use my eyes at all. I don't use the lightbox. I don't use my
subjectivity. Only objectivity. I use Photoshop to know the source Lab
values and the Colormunki to know the destination Lab values reading the
patches on the paper.
I use this method to try to discover, and to show clearly in this mail list
my results, why I see a more than desired variation of all the colors when
I print using the perceptual rendering intent, specially on blue colors and
why I loose a lot of shadow details when I use the relative colorimetric
rendering intent (seems to be due to a very non neutral black ink) with
profiles creatd with ArgyllCMS.
Once I discover the reason of this error it is more easy to find a solution,
could be using an option on colprof. I have exactly the same issues with 2
diferent printers, one of them with 2 different brand inks.
I attach an XLS file using as a source the Lab values of the 24 patches of
the ColorChecker chart and as a destination, the Lab values of these 24
patches printed on paper and beeing read by the ColorMunki.
On the first group you will see the delta E of each one of these 24 printed
patches. In this case, printed using the relative colorimetric rendering
intent without BPC, I obtain an average deltaE00 of 1,88 with a peak of 2,51
with the cyan color and a peak of 3,02 on the neutral 3.5.
Printing with perceptual rendering intent without BPC I obtain an average
deltaE00 of 4,12 with a color peak of 8,41 with blue color and 8,14 on
purplish blue color. With blue color the deltaE is greater more than 2 times
than the average deltaE and is more than 8 times greater that the same blue
color when printed with the relative colorimetric rendering intent.
I question if this difference on deltaE when using the perceptual rendering
intent, on all colors but specially visible with the blue color, is
"correct" compared with the deltaE of the printed blue color when I use the
relative colorimetric rendering intent?
The obtained values of each print on the differnent rendering intents are
correct?
Reading your previous reply and this, I have included a table with the
deltaE values when I print with the absolute colorimetric rendering intent.
The average deltaE00 is very similar to the printed patches with the
relative colorimetric intent but if you analize each color patch, one by
one, the differences are a lot greater.
A thing that I don't undestand and that make me question if I am doing all
perfectly is that the Lab values when I use any of the 2 methods to
softproof, aren't equals to the values on prints. Are different values.
But to not to send a very large email I prefer to wait if anybody can help
me with the formulated questions and after to try to discover what I'm doing
incorrectly or understand why the values of the softproof aren't the values
of the final prints.
I will be so grateful for any advice you can give.
Thank you !
Best regards,
Xavier Puente.
_____
De: argyllcms-bounce@xxxxxxxxxxxxx [mailto:argyllcms-bounce@xxxxxxxxxxxxx]
En nombre de Nikolay Pokhilchenko
Enviado el: viernes, 19 de octubre de 2012 16:48
Para: argyllcms@xxxxxxxxxxxxx
Asunto: [argyllcms] Re: ArgyllCMS: shadow details and blue color issues.
What I'm doing incorrectly?
Hello, Xavi.
Absolute cororimetric compiance between display and the paper is only
achievable in controlled condition when the paper brightness has the same
level as the display brightness because very vell controlled illumination.
Is this Your case?
If the brightness of media white is differ from the display white, You can't
achieve equal perception from the display and from the print even if
measured Lab values of the print (which are normalized to the print
illuminant) are qual to the values in PhotoShop. In such case it would be
better to apply some kind of color mapping (for example -ila) which actually
can improove the image appearance but can't gain colorimetric compliance.
If You want really same colors between display and Your lightbox, You can
perform the measurements of printed target with "chartread -e" while
calibrating white on Your display nor Your lightbox. I.e. You should start
the measurements, calibrate the instrument, calibrate the white (the light
source) right on Your display and continue to measure patches on lightbox.
With this workflow You'll obtain the same Lab on the print as on the
display. With the same basis. And after profiling You should be able achieve
equal Lab and equal perception of the colors in gamut. This is what I think
about Your problem.
Fri, 19 Oct 2012 13:41:06 +0200 "Xavi" wrote:
Hello Nikolay.
To print my photos I wish to use the intent that gives me similar results to
what I see on my calibrated screen or beeing more objetive, to the Lab
values that PS informs me that a color has.
I thought that the more simmilar results were offered by the relative
colorimetric intent. The other usually used intent, when speaking of
printing photographs, the perceptual intent, thought that was more visual
adaptative, than "real".
After receive your reply, I have printed, with absolute colorimetric intent
without BPC (is not posible to print with BPC option enabled), the same 7
patches that I use to do my tests , and reading them with an spectro, the
deltaE values are between the results obtained with the relative
colorimetric and the obtained when printed with perceptual intent. But very
very near from the relative colorimetric. Exactly only a difference on
deltaE2000 of 0,20 and 0,32 of deltaE76.
When I try to show my profile results, specially of the blue color using
perceptual intent, Graeme Gill, commented me that he don't see the results
on my prints and that he don't know how the application that I use do the
soft proof.
Then I have decided to not to use soft proof to compare results.
Then, the only way that I think that I can compare results is printing
patches that I know their Lab value(photoshop gives me the Lab value of a
color when the cursor is over it), read these patches with an spectro and
calculate the difference between the Lab value of the patch and the Lab
value of the same patch printed on paper. I use the deltaE76 and the
deltaE2000. I think that a correct profile must have deltaE values near zero
when speaking of inside gamut colors.
This method is correct? If yes, I must to print the patches using relative
colorimetric without BPC or absolute colorimetric intent?
If you have other ideas of how to verify if the workflow of creating a
printer profile, than use it and measure the results, please tell me it that
I will try it to check the results.
Any idea will be wellcome.
Thank you very much.
Best Regards,
Xavi.
Other related posts:
