Hello Gerhard,
thanks for the clarification.
For calculation of XYZ and Lab for the tungsten illumination I had used the
formula provided by Bruce Lindbloom on his web page (Computing XYZ From
Spectral Data (Reflective and Transmissive Cases)). For test, I used my
spreadsheet and calculated the XYZ,Lab values for D50 and I recovered the
values in the original reference table. I think I should also use Argyll and
repeat these calculations there and compare the results.
Looking at Bruce's page, he also gives an explanation, why the reference
white is needed for transmission and reflective data. :-)
Hermann-Josef
--- Begin Message ---
- From: "Gerhard Fuernkranz" <nospam456@xxxxxx>
- To: <argyllcms@xxxxxxxxxxxxx>
- Date: Tue, 9 Jun 2020 12:19:48 +0200
CIELAB numbers are not fully specified, if you don't specify the reference
white point as well, which was used for the XYZ -> CIELAB conversion. That
like saying the length is 5, w/o specifying whether you mean 5 inches, or 5
feet, or 5 meter.
Usually, D50 is assumed if the reference white point is not specified, but
for numbers from an unknown origin you can never be sure how they are
supposed to be interpreted, if it is not explicitly specified.
I guess that the CIELAB values in your reference file were calculated for a
different reference white than colprof's CIELAB numbers (the latter are IMO
D50-based).
Regards,
Gerhard
Am 09.06.20 um 10:30 schrieb Hermann-Josef Röser:
Hello Graeme and Gerhard,this
thanks a lot for your replies. Unfortunately I do not really understand
(yet).0.465900
Here are the values for patch A1 in the reference file:
SAMPLE_ID XYZ_X XYZ_Y XYZ_Z LAB_L LAB_A LAB_B
LAB_C LAB_H D_RED D_GREEN D_BLUE D_VIS STDEV_X
STDEV_Y STDEV_Z MEAN_DE STDEV_DE
A1 2.622 1.970 0.758 15.331 10.511 2.674 8.89
15.75 1.47 1.88 1.93 1.70 0.06 0.05
And here is the corresponding line from the ti3-file:
SAMPLE_ID XYZ_X XYZ_Y XYZ_Z RGB_R RGB_G RGB_B STDEV_R STDEV_G STDEV_B
A01 2.622000 1.970000 0.758000 28.91538 19.94871 19.18301 0.720745
0.899324three
So the XYZ-data are identical. RGB_R ... are the measured values from the
target scan. This I understand.
And here is what profcheck tells me:
[1.146434] A01: 0.28915380 0.19948710 0.19183010 -> 14.802257 14.722387
13.324269 should be 15.329014 15.318528 12.124771
, but not mangled otherwiseWhat is the meaning of the number in [] at the beginning? What are the
numbers following after the A01: ?supplied
The values following "->" are quite different from the Lab-values in the
reference table. So is this due to the rendering intent?
My naive understanding of the ICC-profile was, that it is "just" the
transformation of measured scanner RGB-values to the XYZ/Lab values
by the reference table. So for any measure triple (R,G,B) interpolationvia
the profile gives me the absolute colour in (X,Y,Z) or (L,a,b). So still Ithe
do not see, why I need a rendering intent here :-( : Once I have the
absolute colours I then can render into a colour space of my choice (e.g.
sRGB) using one of the rendering intents.
I would be very happy for a reference so I can read and learn about all
these connections. What I have found up to now (e.g. the white papers on
color.org pages and several books from the Wiley series) was not of much
help regarding an understanding of these technical details.
Many thanks again and best wishes
Hermann-Josef
--- End Message ---