[argyllcms] Re: Profiling for translucent materials
- From: Graeme Gill <graeme@xxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Fri, 05 Aug 2005 11:57:29 +1000
Andrej Javorsek wrote:
I have question concerning measuring and profiling for translucent
materials, that can be viewed as normal (reflecting) materials or as
transparencies (on a viewing box).
I suspect that each mode of viewing will need own profile (with
different -c and -d parameters)...?
But the real problem for me it seems, how to measure testchart...
- as transparent or as reflective...
- if as reflective, what backing to use for measuring (black, white...).
I will be happy to get any other suggestion in the mater of profiling
for translucent materials.
I don't have any well researched answers for you, just some thoughts.
You are really looking for a compromise profile, so you can expect
that in neither mode will it look perfect. One approach might be
to characterize it as reflective and transmissive separately,
and then blend the two results in some proportion.
Another approach that I've read about is to do a reflective
profile, and tweak it to have several darkened versions (effectively
a gamma adjustment), choosing which profile to use simply by trying
them out.
One thing I have noticed about backlit material is that with some
inksets (presumably pigment ones), there is often a reversal in
dark areas, particularly if the balance of CMY to K goes through a
transition (perhaps due to hitting the ink limit). There is
no such effect when illuminated reflectively, but backlit the
K has less density than the CMY, and some of the dark areas
aren't dark enough. Some careful manipulation of the black
generation may be needed to minimize this effect.
Presumably dye based inksets suffer from this effects
somewhat less.
As for how to measure it, the usual answer is to measure it
as close to how it is set up for viewing. If under reflective
viewing it has a white backing, measure with a white backing.
Graeme Gill.
Other related posts:
