OK Bob,
I see clearly what you're alking about now, although that is a pretty
wacky colour... I know the stuff you shoot may very well have these
kinds of colours in them, although my intuition would be to stay away
from the colorimetric intents when using these "far out of gamut"
colours. Actually I would probably reduce the gamut with all kinds of
trickery in photoshop and then attempt to convert.. (I expect this is
what you're doing anyway)
None the less, you are right, this particular colour is rendered
poorly in the ECI profiles when using colorimetric intents, BUT if
you take the test one step further and instead of creating a fill of
the colour create a gradient fill to white, you will see that the
gradient remains somewhat smooth with the ECI profiles, and differs
between perceptual and colorimetric in Hue only. More what I would
expect.
The Adobe profiles however do not have the same "smoothness" and
introduce visible bands of shifting hues in the gradients... not so
desirable in many cases.
Neither is perfect, and it's a question of using the most appropriate
for the content!
-Andy Psarianos
F E Burman Limited
On 10 Nov 2005, at 16:25, Bob Marchant wrote:
On 10 Nov 2005, at 11:15, Andy Psarianos wrote:
I have tried to reproduce this blue/purple
problem a few times now
and have been unable to reproduce
Hi Andy.
Make Adobe RGB file (or sRGB if you prefer) .Fill with LAB
reference 30,68,-112.
Preview with ISO coated in relative and perceptual modes. See the
difference in hue (hopefully!)
Prior mails may have mentioned colours that are not real. We shoot
flowers that fluoresce into ultra violet and infra red, products
with spot colour , fluorescent backgrounds etc etc. We know that we
don't stand a chance of matching these colours in CMYK , but a good
profie will give a good first stab at it without shjfting the hue
too much .Ir's mostlly at the cost of saturation.
The ECI profiles seem to lose saturation with perceptual
colorimetric rendering intent , and shift hue in saturated blues
with relative colorimetric intent. Building profiles with the FOGRA
data using Profiemaker 5 is giving us better results, and building
profiles with the same software using our own data does likewise.
BTW we proof to ISO standards and get the same results , so it's
apparently not a screen thing.
Hope this helps.
Regards,
Bob Marchant.
_______________________________________________
ECI-EN mailing list
ECI-EN(a)lists.transmedia.de
http://lists.transmedia.de/mailman/listinfo/eci-en