Digital BW, The Print

Pete Bergstrom <petebergstrom@...> wrote:
> ... I found is that with a flatbed scanner
> (Epson 4870 in my case) I could calibrate a new set of EB6 inks (my own
> dilution ratios) in about 4 prints, ...

> ..., the first step is to choose an ink limit for the black.
> You can efficiently do this with a 21-step image and QTR's
> calibration mode. ...

Yes.

> I ended up with about 25-35%.

I made a profile for Red River Polar Matte and found the K ink limit to be 35.

The new QTR profile as well as an ICC for this paper are now in
http://www.paulroark.com/BW-Info/1400-Eb6-Profiles.zip .


> Next, using this overall ink limit, I printed the calibration mode image
> that QTR provides. I let it dry, then scanned it (RGB). Using (in my case)
> Photoshop Elements, I then measured the "relative density" reading of each
> patch (green channel, 5x5 pixel sample) ...

I usually find that the lighter inks can handle a higher ink limit
than the denser black ink.  As such, once you find the K ink limit,
you can set that as the default and leave all the other ink limit
boxes blank so that they use the default/K ink limit.  Then scanning
the entire second Calibration Mode print that was made at that default
ink limit makes finding the relative densities in Photoshop, using the
eyedropper, a very easy approach.  Hopefully a scanner is going to be
consistent enough that you could simply measure the 100% patch for
each lighter ink with the eyedropper and then see where that density
occurs on the black ink 21-step.  That would give you the relative
densities for that same default ink limit.

The QTR profile for Red River Polar Matte used this approach and
linearized quite well.  It did need linearization, but this is normal.

For the ICC I used the general NCA correction curve that is in the Zip
file.  With newer software, ICCs always use the driver set to ICM/OFF
(No Color Adjustment).  With that setting, the 21-step test print
makes a curve that has too steep a step between 95% and 100%.  That is
why the NCA correction curve, which is generic, helps with the
process.  It is embedded in the ICC by dropping it into Create ICC-RGB
along with the Lab L data.

Paul
www.PaulRoark.com