Re: Calibrating output (digital)


Robert Barnes (bb333@earthlink.net)
Sat, 22 May 1999 22:05:02 +0100


Dear Dave Fokos.....
    It's nice to see your letter on the Alt newsgroup! I've been
wanting to ask you a couple of questions anyway, because I have been
testing my own digital negative process at work. I have been able to
download your paper and view it in Acrobat, but not your curves nor your
tablet. I have been scanning exhibition grade silver gelatin prints
(which is more difficult way to go than scanning negatives, but
expedient for me at this point). I have been making my scans at 100%
and 600 dpi. I have then been using a calibration set in a Harlequin RIP
that I have been testing (3000 DPI/300 LPI with a 45 degree elliptical
halftone, which looks to me to be a cleaner dot structure that the
"bitmap structures" that I tried in the Harlequin) and I have found that
my contrast (read highlight/shadow detail) is easier to control then
when I use the 3000 DPI/600 LPI set........I suspect that the the higher
resolution is loading too much information to image clearly with these
materials (using the analogy of it being easier to print a "clean"
halftone at 175 LPI on a lithographic press than it is at 300 LPI. I
have used a curve adjustment "press-gain bump" setting for my
calibration set which does help minimize that gain in contrast, but I do
get better highlight and shadow detail than at 600 LPI....do you think
that this is true? I would very much like to see your tablets and
curves that you post with your paper, but I cannot view them after my
browser downloads them. Could you please attach them to an email and
send them directly to my email address? I use Mac and Netscape.
Thanks Dave!

Appreciatively.......Bob Barnes

Beakman wrote:
> > <<and I can tell you that imagesetters are calibrated for PERCENTAGE
> > DOT but NOT for density.
> >
> > Hmmm.... the percent dot has a direct relation with density; if we are
> > talking about hard dot, theoretically as long as the Dmax is above 2.0 then
> > the difference should not make any difference. That is, 50% dot should
> > measure density of 0.30 (+base+fog) if the Dmax is above 2.0.
> >
> > Of course, if the Dmax is really as low as 2.0, it could mean that the dots
> > are not that hard because each dot actually has a gradation from Dmax to
> > Dmin, but imagesetter's Dmax usually is above 3.0.
> >
> > >> The density ranges all over the place, and as
> > the imagesetter runs, and the lasers heat up, it changes.
> >
> > I find this interesting and curious. It sounds like the dots are not hard.
> >
> > >> The change in
> > density will not effect the percentage dot values (as used in offset
> > printing) hence it is of no concern to printers. HOWEVER, the density IS
> > important for alt-photo work because the negs print as a hybrid. >>
>
> The Dmax of the film typically ran from 4.0 to 7.0 (!) -- incredible, but
> true. Suppose you ran film and made a box filled with 1% dot on the neg.
> and next to it a box with 0% dot (i.e. opaque film). Then let's say that
> you measured the optical density of the opaque square and it measured 4.0.
> Then you measured the % dot of the 1% box, and sure enough it is 1% dot.
> Next you measure the OPTICAL density of the 1% box and find that the
> reading is say, 1.8. O.K., now you crank up the lasers and run another
> neg. This time the optical density of the opaque square is 7.0, and the
> OPTICAL density of the 1% box is 3.1 EVEN THOUGH THE % DOT STILL MEASURES
> 1%! As it turned out, for me anyway, the OPTICAL density was important.
>
> > Do you use stochastic screening or traditional halftoning?
>
> I use 45-degree elliptical dot halftoning. If you like, you can read all
> about it in a paper I wrote which is posted on the Bostick & Sullivan web
> page. The only change I would make to the paper would be to suggest
> sacrificing some shades of gray in order to use a finer linescreen -- try
> 400 or 425 lpi -- you may not notice the loss in shades of gray, but the
> prints may be better.
>
> best regards,
> David Fokos



This archive was generated by hypermail 2.0b3 on Thu Oct 28 1999 - 21:39:35