Digital BW, The Print

I've been playing with transfer curves for "Nevins method" quadtone
printing (printing greyscale through the Epson driver, using the
transfer function to linearize the output).  Two approaches produce
rather different curves.  And neither matches the screen that well
(screen hardware calibrated with Photocal and MC7).

I noticed recently somebody pointing out that the histogram from a
step wedge of equal steps should be a nice comb of equally-spaced
spikes.  (In the real world the spikes are more like triangles, of
course.)  I looked at the histogram from a step wedge printed through
my previous transfer curve, and *eeewwwwwww*.  Bottom and top tones
all muddled together, and spacing not very even.  However, prints
through the old curve look pretty good, and match screen decently.

I built a new curve, evening out the tones.  The step wedge looks
better -- there was a big visual jump at 30% in the old one, for
example, that the new one cleans up.  And in a scan of the step wedge
I get much more evenly spaced peaks.  However, some pictures printed
through it look much lighter than the screen.  Others look okay.
[Update: they dry down some, and look more like the screen when dry]

What I've done with the new transfer curve is divided the dynamic
range of the ink+paper+printer up evenly and shared it out among the
input densities.   What I did before amounts to matching the high
densities in a more *absolute* way (densities that scan equal to a
reference photo print; from ofoto; as recommended by Prof. Nevins),
which results in the low densities piling together (since the MIS ink
on Aspen paper doesn't get as dense as glossy photo paper).  

Is there any argument for one approach being "better", or is this one
of those situations where it has to be an artistic choice on how to
render each print?  I *hate* artistic choices.  (Well, I hate artistic
choices on what I think ought to be technical issues.)

(One underlying assumption of the "evenly spaced comb" approach is
that the scanner does a decent job of representing tone relationships
on a given scan.  Is this nonsense?  Seems to me it *has* to do that
to be of any use as a scanner.  Obviously with a densitometer I could
be more precise and more accurate about this, but I don't have the
budget for a densitometer just now.)
-- 
David Dyer-Bennet, dd-b@...  /  Ghugle: the Fannish Ghod of Queries
 John Dyer-Bennet 1915-2002 Memorial Site http://john.dyer-bennet.net
        Book log: http://www.dd-b.net/dd-b/Ouroboros/booknotes/
                 Photos: http://dd-b.lighthunters.net/