Printer Profile

[HG]

I'm having trouble creating a new profile. Printer is an Epson 9890, paper is
Hahnemuhle Photo Rag Pearl (no OBAs), measuring instrument is a Colormunki,
operating system is OS X 10.9.3.

I started by building a simple profile to pre-condition. . .

targen -v -d2 -g18 -f426
printtarg -v -iCM -h -T 360 -C -m8.7 -p11x17
chartread -v -c1
colprof -v

. . . followed by. . .

targen -v -d2 -G -g86 -f1273 -c precondition.icc
printtarg -v -iCM -h -T 360 -C -m8.7 -p11x17
chartread -v -c1 -B -T0.4 HPR-Pearl
colprof -v -qh -S AdobeRGB1998.icc -cmt -dpp

I’m using this test chart:
http://www.northlight-images.co.uk/downloadable_2/frontier_color57s_2.zip
(I converted it to AdobeRGB since that’s my typical working space).

Compared to the manufacturer provided profile (also compared to my profiled
monitor and prints on Epson Premium Luster made both with the manufacturer’s
profile and with the print driver) the one I produced with Argyll was a bit warm
in the skin tones and with a considerable shift in the blues toward s a darker
more purple hue. I'm inspecting these prints under photofloods (2850K) so I
tried changing the illuminant in colprof with the -iA command. This completely
fixed the blue to dark purple issue, but now the reds are too dark and too
saturated.

Any suggestions as to how to move forward would be greatly appreciated.

Thanks,
HG

[Graeme Gill]

Post by HG
I'm having trouble creating a new profile. Printer is an Epson 9890, paper is
Hahnemuhle Photo Rag Pearl (no OBAs), measuring instrument is a Colormunki,
operating system is OS X 10.9.3.
http://www.northlight-images.co.uk/downloadable_2/frontier_color57s_2.zip
(I converted it to AdobeRGB since that’s my typical working space).
profile and with the print driver) the one I produced with Argyll was a bit warm
in the skin tones and with a considerable shift in the blues toward s a darker
more purple hue.

Hi,
that test chart seems suspect. It has no embedded profile, and assuming
either sRGB or AdobeRGB space, the patches that are expected to be blue (ie. the
colorChecker blue) is very purple, and it looks somewhat dark.

So it's possible that the Argyll print is representing it more accurately than
the other prints. I'd suggest using some different images, and doing a subjective
comparison to the images appearance on a profiled display, even though that
approach has it's limitations (ie. different viewing conditions and white points
typically making side by side comparisons impossible).

Graeme Gill.