[argyllcms] Re: Gutenprint and ink limits.

  • From: "Lars Tore Gustavsen" <lars.tore@xxxxxxxxxxxxxx>
  • To: argyllcms@xxxxxxxxxxxxx
  • Date: Wed, 8 Aug 2007 11:25:57 +0200

On 8/1/07, Kai-Uwe Behrmann <> wrote:
> If you have choosen a Premium or Lustre paper type, thats a setting you
> could start with. It is calibrated already by Robert L. Krawitz.

That's fine, but maybe it is not black enough.

Short version of this question:

I have printed a grayscale wedge. Is my understanding right that the
darkest patch cannot be printed blacker with profiling, but only by
tweaking in the driver?

Longer version:

Since I have read a lot about linearity I found many references to a
rip called quadtonerip.  It is noting I consider running on my system
since I'm mainly on linux, but it was nice reading anyway.  I also
read about a small tool included called qtr-create-icc.exe. It is
mentioned in these pages:


Using that tool involved printing a 21-patch grey file. Instead of
using that I tried to make one with argyll like this:

$ targen -v  -d1 -g 11 -m 0 -e 0 -f 0 -s 21 greytest
$ printtarg -v -ii1 -r  -pA4 greytest
$ gs   -sDEVICE=pnggray -dNOPAUSE -r300    -sOutputFile=greytest.png
 -c quit

The resulting tif file looks very similar to the Step-21-gray.tif
provided with qtr, but as a bonus I now had a ti2 file.  I opened this
file in photoprint, since cinepaint have problems printing grayscale
images, printed it (with option "uncorrected") and measured it.

Since the qtr-create-icc.exe tool requires lab values I did a:

$ spec2cie -f   greytest.ti3 greytestlab.ti3

I have no idea if this is the right way to do this.  The lab values
looked more correct when I enabled fwa compensation, so I did so. If I
use spotread with an fwa measurement on the paper, I got about 1 lab
value lower on the deepest black patch. I guess there is something
going on here that is above my head.

After that I exported the gray and the lab values like this:

$ grep ^[0-9] greytestlab.ti3| cut -f 1,2,42,43,44 -d " "> greytestlab.txt

I edited the file with the appropriate header and footer.  I then run

$ wine QTR-Create-ICC   greytestlab.txt

I ended up with two files, one graph and one grey profile.  I could of
course have used gnuplot for the graph more easily and freely. But
here is an output from this program, converted to pdf.
http://www.mulebakken.net/div/qtr-output.pdf  (6 kb)

And here are the highlights

                        Step    Dens    Lab     A       B
gutenprint              100.00  1.827   12.54   -3.70   1.19
gutenprint density+10 100.00    1.893   11.12   -2.68   -0.53
windows abw             100.00  2.421   3.42    -0.11   -0.22
windows "no color"      100.00  2.122   6.83    -0.36   -0.27

I try to show you here the high lab value for the black patch, when
printed with gutenprint, compared to the two windows driver print.
The reason can be my printing method in linux, or that a grayscale
file is not very well supported in gutenprint.

Am I right in guessing that I should lower this value before I try to
profile my printer?  I guess it most be at least possible to match the
normal windows driver with lab 6.83.

As a simple test I applied the generated grey profile to the
testtarget like this:

$ tifficc -v  -i
-o greytestlab.icc greytestlab.tif greytestlab-curve.tif

When I measured this, the L channel was quite linear. But as I
expected the Lab value for the dark patch was not any lower.

Maybe this topic about deep blacks is a b&w issue? Or is a general
printing issue?

Sorry for this long post, I included all that qtr stuff since maybe
other find it useful if it is the correct workflow.


 Lars Tore Gustavsen

Other related posts: