Digital BW, The Print

Hi Roger,

> ...your 21 step grayscale wedge (which is VERY useful!)

I use 21-step files for most tests.  QTR can handle 51 steps also, 
and I wish I had a better one.  I made a vertical one that sometimes 
helps, but I found the horizontal, 3-row one in QTR appeared to be 
affected by vertically running artifacts.  Even on matte papers the 
rollers and perhaps other things do leave their marks, even if 
invisible in most situations.  The 51 steps are so close, however, 
that they can be affected by the artifacts -- or so my tests suggest.

So, I generally use 21-steps and keep in mind that the curves must 
have smooth transitions to avoid a wiggle that is smaller than the 
relatively wide 21-step spacing would "see."

 
> You seem to be using it to both linearize and get a
> neutral grayscale simultaneously. 

Yes, I look at and adjust for L, A & B and the same time.

> Do you have any references on how to do this?  I 
> don't have a background in manual curve creation.

I'm not sure I've written up much detail.

First, I use Excel to open and graph the PFP text files.  So, open in 
Excel, highlight the L column, then go to the top and Insert a 
Chart.  Use the Line type.

For L, just take a straight edge and draw a line from the first to 
the last dot.  This is what a linear rip will print when linearized.  
RGB (like Gray Gamma 2.2) will compress the shadows, so in fact, draw 
a line from the paper white to a 95% that would be barely above where 
the 100% is.  This line is your guide.

Where a dot of above the line, you'll want to adjust R, G, & B curves 
accordingly.

If the points in the Lab A & B measures are neutral, adjust all three 
equally.  Nore on the graph how much, so you learn at the next 
iteration how much a move by X points changes the output.

You need to be familiar with Lab A & B.  If the print is too yellow, 
you'll adjust the blue curve accordingly.  If the point has a perfect 
Lab L but has a high Lab A, you'll reduce M and increase C by the 
same amount.

You'll get the hang of this after a few iterations.  

So, it's a process of printing the 21-step file with the last curve 
(which is saved with the iteration number in the name), measuring the 
print (after a hair drier is used on it), graphing the results, and 
making appropriate adjustments in the curve.


> Say if step 17 (80%) is L26.47

I don't think in L numbers.  I just graph and draw a line for 
reference, keeping in mind the final should have the compressed toe.

> a= 1.98 b= 6.61

A=1 is OK.  See where the paper white is and use that for a 
reference.  "Neutral" is defined by the paper for Lab A the way I do 
things.  These spectros are not very accurate at these very low 
readings, so the paper white serves as a calibration point, and I 
think you'll see most papers aim for a slightly positive Lab A.

 
> I'd want to counter the postive b by increasing the blue channel 

Yes, but confusion can arrise is what is meant by "increasing."

The curves have 2 modes.  The 0 - 100% ink mode is easier to deal 
with, but less accurate.  You'll find you're limitted even by the 256 
steps of the 0 - 255 luminance mode.

Just keep in mind that moving the blue curve toward the darker side 
increases the yellow ink.  So, how people think in complementary 
colors probably differs, but I mentally label the curves by theie 
complement.  That way "more" ink in the 0 - 100% ink mode means more 
yellow is injected when the blue curve (which I think of as yellow) 
is moved to a more positive number. 



> somewhat and reduce the a by increasing the green channel,

No, green is not opposite of blue.  I make the Lab L, A and B moves 
separately just to keep it simple.  To keep L equal I'd increase Blue 
2 units and reduce G and R one  unit each.

> ...  Is there an equation to assist with this?  

There is trial and error, but you'll find you learn how much to move 
each point.  By keeping track of your moves on the graph you've 
printed you can see how much to move the next time.  The iterative 
process works well, and I think you'll find you learn how much to 
move things quickly.


> ----------------------------
> 
> Also, on an unrelated note, I'm going back and reprofiling my R220 
> using UTR2 with glossy media.  By using RPM and checking Microweave 
> I'm getting much higher dMaxes- increasing from 2.13 to 2.32 on 
> Kirkland Glossy, getting 2.25 on Silver Rag, 2.15 to 2.36 on Harmon 
> FB Al, etc.  It also has no microbanding under a loupe which is
> nice. Downside is that prints take forever.

Yes, I do the same thing.  I'm using an old 2200 with Carbon-6 (with 
cheap SuperJet CIS -- working perfectly) and have it on 2880 slow for 
ultimate smoothness.  If I were a service bureau I'd probably crank 
it up to 1440 HS and hope the customers didn't use a loupe on the 
print.  But, given my low volume, I feel better with the perfect 
prints, and the speed is very secondary. 

Good luck with the profiling.  Once you get the hang of it, you'll 
seldom have to do more than a half dozen iterations.  Fortunately, 
most papers of the same basic type use the same basic curves.  I have 
only one layer for the 260 and use is with all my matte papers.  The 
opacity slider on the layer is usually all I need to adjust to 
account for paper differences.  I also find I usually want the 
saturation increased 25% with the 260.  The MIS pigs simply are not 
as high gamut as the Claria inks.

Paul
www.PaulRoark.com