[argyllcms] Re: Printing using an ICC profile
- From: "Alastair M. Robinson" <profiling@xxxxxxxxxxxxxxxxxxxxxxx>
- To: argyllcms@xxxxxxxxxxxxx
- Date: Thu, 21 Jan 2010 18:56:06 +0000
Hi :)
Pedro Côrte-Real wrote:
I was just trying not to waste too much paper. 600 was just a random
guess on my part. How many should I aim for for a good profile?
600 should be plenty to give you satisfying results for a first try -
but I think you actually used 3 runs of 190 - so no individual profile
had more than 190 patches?
I can get pretty good results on inkjet-printable CDs using a single run
of just 144 patches - but that's with a customized patchset biased
towards neutral colours. A lot depends, too, on how well-behaved the
printer is.
Speaking of which, when colprof finishes, it prints an error summary
which gives you an indication of how well the profile fits the source
data. What did it report for your profile?
You can get that reading after the fact, using the profcheck utility,
like this:
profcheck file.ti3 file.icc
I took the -d3 from the docs. I assumed it had to be that because the
profile was for RGB not CMYK even though eventually gutenprint is
going to convert it into device space. Is this not the case? If it is
the printer space then shouldn't it be one of 4-11 considering the
R2880 has CMYK + LC + LM + LK + LLK?
You're right - in your workflow gutenprint handles the messy business of
converting from RGB to Device space. While it's theoretically possible
to get better results profiling directly in Device space, it would
require a much larger number of patches to be read, and there's no
open-source solution as yet for doing the separation to more than
4-channels, so for the time being the RGB workflow is probably your best
bet.
I'm not sure gutenprint+GPLin(?) are up to it already but I right now
I only want a reasonable profile that allows me to print something
usable.
With a little perseverance you should certainly be able to get usable
results - hopefully even very good results.
As for GPLin, it's still very much experimental, and isn't yet very
useful. Argyll's new calibration tool looks very promising, but I've
not had a chance to play with it in depth.
All the best
--
Alastair M. Robinson
Other related posts:
