[argyllcms] Re: how many patches, profiling Epson printers

> See also Graeme's paper
> http://www.imaging.org/store/epub.cfm?abstrid=22190
> which describes the principles of the implemented algorithm.
> (Actually I'm wondering, whether the other professional profilers are
> probably based on Graeme's algorithm as well?)

I ordered the conferences proceedings. Looking forward to read Graeme's
presentation. Because, in my humble experience, with all the instruments I
experimented with and all the software that I have used, that have some
specific settings to deal with FWA compensation, I can't say the results
spoke for themselves.

GMB has a fixed setting for dealing with FWA when it detects their presence
in the spectral measurements. PrintOpen v5 has a variable setting (from 0 to
10) to compensate for FWA, which is not triggered by any analysis of the
data but entirely under user control. Those are what I call software
solution and I'm curious to see how argyll's approach differs. At one point,
I was so fed up with this issue that I purchased a 530 with a separate,
installable UV-cut filter to experiment with making profiles *with* the
filtration ON and *withtout* filtration, and I can't say I am convinced.

Now, it's very possible that all my efforts over the years have always been
defeated by Epson's inks (Ultrachrome or not) but I can't say for sure. In
my office, on my Epson 4000 printer, I use a proofing paper sold by Fuji
that is not too fluorescent, 93 0 -4 to 94 0 -5 (depending what I use to
measure it) hopefully. Still, I realize that b* of this magnitude is by all
counts fluorescent. For comparison, the same paper measured with a DTP41UV
shows 95 0 -0.75. I once tried a paper called Saphira by Heidelberg. On my
Iccolor it measured 96 -0.16 -0.99. Now that's very *unfluorescent* but the
%$!@# UC inks don't adhere well onto this substrate :(

So back with Fuji paper which a number of my prepress clients use.

My experience with FWA is that it invariably translates to shifts into the
grayscale, from L=100 to L=TIL. It's as though the calculated gray axis,
when I measure it after converting from Lab to Paper profile, is twisted
towards the midtones, with b* values as high as -3. In fact, all the
profiler's I tried so far (including argyll?) have never given me a complete
and uniform neutral scale. Because of printer's non-linearity never
completely under control of linearization? I don't know.

But that's why I am skeptical of any FWA compensation approach, hardware
(UV-cut filter) or software (spectral sharpening, I'll call it). Maybe it
works in theory but in practice, under all the lamps I have used for viewing
the results, it does not work to my taste.

Regards,

Roger Breton  |  Laval, Canada  |  graxx@xxxxxxxxxxxx
http://pages.infinit.net/graxx

Other related posts: