Digital BW, The Print

Alessandro Pardi wrote:

> ...I first tried to modify the global RGB curve,
>but that caused banding, whereas changing the Red
>curve gave no banding

Jeff Randall wrote:

>Yes, the Red curve controls the Cyan position ink (which is the dark
>gray in the FS and Piezo systems) so adjusting it is the best way
>change the dark end.  You can't directly control the Black ink -- I
>think the driver automatically turns it on when each RGB values is
>below about 50%.  Paul Roark may want to step in to correct or add to
>this.

I haven't been following all this thread, so my comments may not be
responsive to the issues you were originally addressing.

>Using any combined RGB transfer function ... -- a meat axe ...

I agree with this.



A couple general thoughts I might add -- too radical slopes on curves
sometimes adds artifacts.  The magenta position ink is sufficiently light
that it might be run up into the highlights a little more, allowing the
slopes to be decreased a bit.

The black ink is turned on by the driver looking at the combined R, G, B
values.  The black ink is a potential source of dots as it starts to be
turned on.  And, when I measure the roughness of test strips, in general,
the worst areas are never the highlights, but the shadows.  The
concentration on highlight dots is a bit misplaced, in my view.

When the blue (yellow-position, lightest ink) and green (magenta-position,
next lightest ink) have both been run all the way down -- that is, turned on
to 100% flow -- the red curve (cyan-position, dark gray) ink becomes the
"switch" for the black.  If you look at the red curve of the PS_FS1160a
curve, you see a reduction of slope of the curve at about 55% (115 input on
the 256 scale).  That might be where the black ink starts to come on.

If there are any dots are roughness in the tests strips at that point (I
don't have one in front of me) it might be because the black ink was turned
on before the cyan ink was on full enough to hide the black dots.  What I
have done on some curves is pull the lightest ink back from being on 100%.
It is light enough that even a reverse slope can easily be overcome by the
cyan ink.  If the lightest ink curve is raised enough, then the cyan ink can
be more fully on and the lightest ink curve can be a "switch" for the black
ink.  It might be that this approach can reduce roughness in the shadows as
the black ink starts to be poured in.

Often in my VM curves all three curves are close to the bottom (full-on)
position, but not totally on.  Then all three can be used to control the
black.  The lightest gray will give the finer control than the cyan because
a one-unit increase in the light ink doesn't add as much density as a
one-unit increase in the cyan ink, the black ink, indirect increase being
equal for both.

Also, if curves have waviness down in the shadows (and my curves do
occasionally -- things I have to get back to as time permits), if you have 2
or 3 curves that are not totally on and able to  adjust the black, you can
stagger the control points and get finer control.  That is, we are limited
to 15 points per curve.  If there is a  bump between, for example, the "85%"
and "80%" points, keep the points of one curves on those points, but move
the points of another curve to midpoints, for example, the "82%" position.

Overall, I'm not sure that having the lighter ink curves full on early is an
advantage.  Their ability to add to the accuracy of control in the shadows
might be an advantage.

I run into this problem with the "warm" vm curves.  Because I'm holding back
the toner as much as possible, that curve becomes my sole control in the
shadows.  The result is sometimes persistent humps in the ramp that are
tough to control.  And, when I have them controlled on one printer, someone
else with the same printer type might report a problem there just due to the
normal production variations in these machines.  The more radical the curve
slope is, the more these production differences seem to surface.

And so it goes.  None of these things is perfect, and the ability of the
individual to adjust the curves to his/her own machine is a real advantage.

Paul
http://www.PaulRoark.com