Digital BW, The Print

Jim,

You wrote:

>... can I do something like what you
>describe below? Except print out the step wedge and bring
>it to a lab with a densitometer that reads CMYK?

I suppose.  I used to use a flatbed scanner to read the test strips.
However, they are not, in my experience, as good as a densitometer.  In
fact, about half of the ones out there with the people who wanted to help in
making curves were totally unusable.

For just getting even steps, however, a scanner, or even just using your
eyes can do a very decent job.

>I guess a question might be how to vary the cmy to get cool, neutral,
>or warm curves to taste- do you weight the midtones with more color,
>or is it more linear?

The paper sets the white point, and the black ink the black point.  I do
role off the gamuts to these points as I approach them.  So, the midtones
between 25% and 75% will be about the same tone/gamut, but outside those
points the curve approaches the black ink and paper white.

> I'd expect the k values read by the densitometer
>to be more straightforward...

The X-Rite makes the job easier and more accurate (I hope -- I just trust
the X-Rite to be giving me true values).

>Maybe one way is if I can print out some known good step wedges that
>work with my paper in PS6, then I can read the values and just try to
>duplicate them say ...

I originally just visually matched my Piezo output.  With tone is a little
more difficult, but with the density of the steps, it's really not a bad way
to go.

Paul
http://www.PaulRoark.com
_________________________________________________


--- In DigitalBlackandWhiteThePrint@y..., "Paul Roark"
<paul.roark@v...> wrote:

  It also looks like PS 7
is
> different than PS 6, though I still do not have a copy of PS 7 to
see what
> might be going on.
>
> With the Piezo driver now being separately available, it probably is
worth
> trying if you're having trouble.
>
> On the other hand, if you like a variable-tone inkset and are having
> trouble, learning how to tweak the curves makes a lot of sense.  If
you can
> identify the area of the problem with the eyedropper in the image
file, then
> make some blocks that are 1% different in that area and add them to
the
> 21-step test chart, the way the top and bottom 5% have been done
with the
> posted 21-step test file.  This will allow you to see whether you
are
> getting a smooth response in the area of concern.  Once you identify
the
> problem, you can go to the curves and correct it.
>
> The RGB curve is usually mostly free of points.  (The limited number
of
> points that Photoshop allows per curve is a problem.)  For very
small
> wiggles at the ends of the curve, I've been using this RGB curve to
correct
> the situation.  By putting enough points on the 45 degree line
(input =
> output) you can hold most of the curve as is while adding a bunch of
points
> in the area where the problem is.  By doing this you can often get
enough
> points in the critical area of the graph to do the job.
>
> Good luck.
>
> Paul
> http://www.PaulRoark.com