ProfilerPlus experiences

["Paul D. DeRocco" <pderocco@ix.netcom.com>]

My printer's an Epson 2000P, and my scanner's an Epson 1640. I was trying to
calibrate the printer using Epson photo paper, which, though it's not listed
as one of the standard media types in the 2000P driver, I've found produces
reasonably good prints if I select Watercolor Paper.

The results were absolutely horrible. I followed the instructions very
carefully, twice, once using Photoshop 6 and the second time using Photoshop
5 (under Win2K). For printing, I tried both method 1 (associate profile with
printer in Windows, and then select ICM in the print driver) and method 2
(turn off color mangement in the driver and select the profile in the Print
dialog box. The two results were vastly different, which implies that one of
them is wrong. According to the docs, method 1 is buggy, so I would expect
that the results from method 2 are what the program was "supposed" to
produce. However, the results from method 1 were only a bit duller than the
screen, while the results from method 2 were very over-saturated.

However, the method 1 print had a serious problem in that a gradient with a
mild purple tint turned heavily saturated purple at the dark end, with a
fairly sharp boundary between the dark, saturated purple and the more nearly
correct purplish-gray around it. I suspect that somehow the profile has a
discontinuity in it, but the ICCInspect utility that I use to look at
profiles doesn't know how to display 3D curves graphically.

They boast about how their package doesn't require a calibrated scanner, and
doesn't need an IT8 target to be scanned along with the printout. Well,
thanks a lot. I'd much rather get an accurate reference target to scan,
since the profile built from a scanner can't possibly be any better than the
calibration of the scanner. They're obviously operating under the assumption
that any decent, modern scanner will produce fairly accurate color out of
the box, without user calibration, and that the resulting printer profile
will be roughly as close to the screen as an original scan was. In my case,
my Epson 1640 has always produced okay color, but not so good that I
wouldn't feel the need to tweak it in Photoshop. Therefore, I'd expect to
have to tweak colors in Photoshop roughly the same amount, prior to printing
via a profile built with that scanner. But that doesn't leave me any better
off than using the canned profile. And indeed, the ProfilerPlus generated
profile is worse, due to the nasty purple discontinuity in it.

Their profile builder includes sliders for tweaking the color, saturation
and contrast of the profile, and they suggest making iterative test prints
in order to get the profiles right. But that defeats the purpose of
profiling. I'd sooner start with the canned Epson printer profile, and tweak
that using the same tools, and entirely skip the part about printing a
target and running it through the scanner.

So, two questions. First, can anyone explain why turning on ICM in the 2000P
driver and associating the profile with the printer in Win2K (method 1)
produces such very different results from assigning the profile in Photoshop
and turning off color control in the print driver (method 2)? And second,
can anyone give me any ideas on how I might coax some better profiles out of
this thing?

--

Ciao,               Paul D. DeRocco
Paul                mailto:pderocco@ix.netcom.com

-
Turn off HTML mail features. Keep quoted material short. Use accurate
subject lines. http://www.leben.com/lists for list instructions.