Digital BW, The Print

Roy/Paul

This is very well put.  Anyone who argues abandoning the "well profiled
monitor" route is talking through a whole in their head.  There will always
be measurement error in such profiling but systematic measurement and the
application of that measured data (where possible) and an understanding of
what it means has to be the starting point.  The use of profiling is
pervasive through  the colour world (and I don't just mean photography) for
a reason - within measurement limits, it works.

Roy highlighted the difference between Gray Gamma 2.2 and Lab which was why
in my earlier response I noted my workspace values.  But whether you choose
to work in Lab, Lab Grey or GG2.2 is not the key question - in all these
spaces a particular shade of grey, visually, still looks the same.  The key
question is how you map those values to the print space and the shape of the
print space.  

Previously, in the RIP world, while we had a linear L print space there was
still no sensible mechanism for mapping (in a Same as Source workflow) from
a workspace of A with a gamma of x to a print space of B with a gamma of y.
We linearised the print space but it wasn't profiled and so we could not
draw on the wealth of knowledge and technique available to the colour world
for the transition.  (In a long discussion a while ago I was arguing that
the default mapping was inappropriate and causing "flat" prints and people
who didn't understand why it occurred.)  While I won't say it is perfect (I
would like to see some sort of perceptual rendering that was image
dependent) Roy's advent of (ICC) profiling the print space has been a major
step forward.

This advent was useful for two principal reasons.  Firstly, it allowed us to
use a much more sophisticated technique for the mapping of image values to
the print space (rendering rather than a point to point mapping).  But
secondly, by using an ICC profile it allows us to use PS' built in soft
proofing features to edit an image as it will be printed.  We got the use of
the tools the colour world has at its daily disposal.

Changing the monitor to meet the output is a fudge (and screws any colour
managed colour work). Better to measure the output - BOTH as it is displayed
and as it is printed - and have a translation mechanism for the two.  (BTW
this is why I asked my question re the PS setting used by people for their
generic rendering which includes the translation from file value to display
value.  Many use Perceptual for file-to-print but have relcol in their PS
preferences for file-to-display.)  At least this way if you don't like the
results of the translation mechanism (ie the effects of Perceptual
rendering) you can see them and adjust them  -  and see the affects of the
adjustments as they would be printed.

Arguably one should measure the print space for each printer, ink, and paper
situation (as we do for colour), ie go through the process Roy did when
creating his "generic" profiles for each combo.  This should, I believe (I
am a little puzzled by Roy's last sentence), entail a measurement of actual
ink black and paper white - as well as deal with any printer specific blips
in linearity between.

Steve


> From: Roy Harrington <roy@...>
> Reply-To: <DigitalBlackandWhiteThePrint@yahoogroups.com>
> Date: Mon, 04 Apr 2005 22:24:04 -0000
> To: <DigitalBlackandWhiteThePrint@yahoogroups.com>
> Subject: [Digital BW] Re: Matching Monitor and Print
> 
> 
> 
> 
> Hi Paul,
> 
> I was similarly bothered by the compressed shadow values that are
> typically seen.  I thought it was strange that I could have a high-end
> display and a high-end profiling device and software and not get decent
> separation.  What I finally realized was that this has nothing do to with
> my system being off -- it was entirely the definition of Gamma 2.2.
> The gamma curve is just a simple exponential: output = input ^ gamma,
> and the slope at 0 (i.e. black) is 0 (i.e no contrast) and you have to get
> a ways out before you get decent contrast.
> 
> Each of the possible gray spaces has it's own shape -- i.e what the internal
> numbers "mean" in gray value is different for each space.  In other words
> K=95 means something different in each space.  It's instructive to take
> a 21step and Assign different gray spaces to it and see the changes.  You'll
> see it different on the display and on the PS info pallette (setup with K and
> Lab).
> For instance, K=95 in GG2.2 is only Lab=1.
> 
> This is all the reason for my making a Gray Lab space, it makes for evenly
> spaced steps.  With this gray space I have no problem seeing nice even
> steps with a 51 step wedge (2% increments).
> 
> Photoshop together will color management have the capability of either
> approach in getting the monitor and print to match.  Your "first" is basically
> softproofing where you check "Preserve Color Numbers" and you see on
> the screen what the print will come out like.   It's done by measuring actual
> output and the softproof simulates that on the screen.  The "second" is
> by making a print profile to be used as the Print Space (rather than
> Same As Source).  Here the profile is used as a correction for the output.
> 
> I think the second is easier to use since don't have to setup softproofing
> each time and you don't need to target your data file to a specific output.
> But the first is probably more accurate since it can simulate ink and paper
> color as well as gray value.
> 
> Roy
>