|
Michael,
Lets say we want to simulate what a print of some
image would look like on screen. A print that would have been made with a strait
PDN or RNP negative of course. One simplistic way to do this would be to
use a strait line "curve" where you change point(0,0) to say point(0,15) and
point(255,255) to say point(255,233). Now, I would believe not many of you would
like this new on screen image.
The question that come to my mind is why a print
that transformes your original in a very similar fashion as above become
something that everyone seems happy about? Do you all accept these
limitations as part of doing alt-prints?
Regards,
Yves
----- Original Message -----
Sent: Monday, March 10, 2008 11:51
PM
Subject: Re: curves and gum and
Christopher James book
OK but with making the digital negative for monochrome prints
"linear" is what I want. If I want to do any tone-mapping using advanced
algorithms in 32-bit HDR I'll apply that to my RAW negative in Camera Raw or
perhaps the camera manufacturer will build the algorithm into the next model
of Leica camera (the M10 due out in 2062) ;^) and I'll select it as an effect
from a menu. That's what the technology outlined earlier in that paper is
destined to be used for. It's interesting but well beyond the need of a
reflected print. A transmissive positive...perhaps. (Now you've got me
thinking about that Jeff Wall guy again. Maybe he's onto something. Urrrg.!) I
really think it's an apples an oranges argument.
~m
On Mon, Mar 10, 2008 at 8:29 PM, Yves Gauvreau < gauvreau-yves@cgocable.ca>
wrote:
Michael,
this idea of custom ICC profile is excellent,
it would solve quite a few problems except one. Because prints in general
have less dynamic range then your originals, someone as to decide which part
or parts of the original range will have to be sacrificed or more
positively which will need to be enhanced somewhere in the workflow.
Basically it's tone mapping again. Suppose this profile maker exist, then
with experience someone could learn how to prepare each original to get the
best out of it when printed. I suspect that those using current methods do
exactly that knowingly or intuitively but they attribute this to the
method instead of their experience unfortunatly.
Regards
Yves
-----
Original Message -----
Sent:
Monday, March 10, 2008 5:00 PM
Subject:
Re: curves and gum and Christopher James book
Yes. It should be inverted. But you have to note which
"mode" the curve window is in. If you give a handful of curves a quick
glance they may look the same. You have to REALLY pay attention to the
curve mode being used to display the curve. A curve calculated for graphic
mode (applied to a positive) will, on the surface, look exactly the same
as a curve applied to a negative -- BUT since PDN users generally work in
binary mode it is this change of mode which functions to invert the
curve. In the end the effect is generally the same. Ink (read:
density) is diverted via the curve toward the dark-mid tones and print
shadows in the negative to compensate for printer's ink profile. When
someone figures out how to take a curve and programmatically graft it onto
a custom ICC profile life will becomes much simpler for digital negative
"curvers" because it'll become as simple as selecting a paper in the print
dialogue. ~m
On Mon, Mar 10, 2008 at 1:09 PM, Henry Rattle
< henry.rattle@ntlworld.com> wrote:
Michael – you are right, of course. But my
question was more about – shouldn't a curve that you apply to a
positive, curve in the opposite direction from one you app ly to a
negative?
David's curve, like mine, has the shadows in the
bottom left corner.
H.
On 10/3/08 18:44, "Michael
Koch-Schulte" <michael@mondotrasho.ca>
wrote:
Henry, it all depends how the curve data was
gathered in the first place. Neither is right or wrong it's more a
matter of workflow. I choose to apply my curv es to the positive
because I'm taking my readings from a developed positive. I also like
to work in PS with the curve putting my highlights in the bottom left
corner using the graphic scale of 0-100 rather than the binary scale
of 0-255. The important t hing is that you apply the curve to the
image at the correct stage of the procedure. Dan Burkholder started
doing it this, I also do it this way. It intuitive for me. Others
apply the curve to the negative.
However one thing puzzles me - there's a
note alongside the curve which says
"Note from David: The curve
is applied before inversion to a negative and
the image should be
RGB."
Henry
|