Happy New Year first of all!
(Although I assume most are not there, yet, at time of writing.)
I got some really strange results on images I've converted from sRGB to Lab
colour space, and then for output back to sRGB. So I've investigated a bit
deeper and set up a test.
I'm taking all 24 GretagMacbeth ColorChecker test patches with their
"official" colour values. I have set up two transforms, one sRGB -> Lab, and
one in reverse (perceptional intent). Doing that I am using 8 bit sRGB and
converting to double or 16 bit Lab (no difference, as internally it all works
on 16 bits AFAIK). So if I am correct, I should be returning after the round
trip to exactly the same 8 bit sRGB tuple. Maybe plus/minus a little due to
rounding differences. But for some patches I get *quite* some significant
differences, particularly on the C06 patch (turquoise), where I've got a
difference of 15 in the sRGB red channel. For completeness I've also compared
the differences in Lab according to the stated values. These are also quite
significant, although not nearly as extreme.
In the table below you can see the Euclidean vector differences of the
obtained colour tuples towards their "officially charted" values.
Any idea for the source of this? Or did I enter the task with some false
assumptions, that the transformation is not lossless and gamut mapping or
adaptation to the illuminant strikes hard on this problem? If so, what should
I do to obtain a "lossless" round trip conversion sRGB --> Lab --> sRGB?
Guy
diff Lab diff sRGB
A01: 0.380142474294 0.0
A02: 0.356133153789 0.0
A03: 0.459139562193 1.0
A04: 0.394233296707 0.0
A05: 0.18255860535 0.0
A06: 0.110803485303 1.0
B01: 0.28150167907 0.0
B02: 0.149504245982 1.0
B03: 0.179581391942 0.0
B04: 0.244501858579 0.0
B05: 0.132116260512 0.0
B06: 0.386083819914 2.0
C01: 0.195918729315 3.0
C02: 0.0980949761964 1.0
C03: 0.125333433035 1.0
C04: 0.270626755882 2.0
C05: 0.234848733492 0.0
C06: 0.343219124854 15.0
D01: 0.286984421446 0.0
D02: 0.149785097046 0.0
D03: 0.248573548852 0.0
D04: 0.404931550676 0.0
D05: 0.241744899875 0.0
D06: 0.561757483198 0.0
--
Guy K. Kloss
Institute of Information and Mathematical Sciences
Te Kura Pūtaiao o Mōhiohio me Pāngarau
Room 2.63, Quad Block A Building
Massey University, Auckland, Albany
Private Bag 102 904, North Shore Mail Centre
voice: +64 9 414-0800 ext. 9585 fax: +64 9 441-8181
eMail: g.kl...@massey.ac.nz http://iims.massey.ac.nz
------------------------------------------------------------------------------
_______________________________________________
Lcms-user mailing list
Lcms-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/lcms-user
