Digital BW, The Print

For newbies like myself, who are struggling to tame B/W color casts on 
their 2200 by using Photoshop curve adjustment layers, I'd like to suggest 
a trick that has made a world of difference for me, and which I've not seen
mentioned before. (And for you experts, if you see a problem in this 
process, please correct me before I spread a bad idea.) 

(1) Find or generate a simple grayscale step wedge that you like, oriented 
horizontally, so it prints out about 1 inch high and 5-6 inches wide. You 
are going to print this many times on the same sheet of paper, so to be most
useful it should be less than an inch high when printed. 

(2) Use Print Options to set the image's vertical position on the page 0.5 
inches from the top (uncheck the "Center Image" check box.) Print the step 
wedge without any color adjustment, using whatever printing method you 
wish to calibrate (I'm using the Schofield method as posted on The 
Luminous Landscape site, which gives me okay results, but a magenta cast at
20% and below.) This is your baseline. Write "baseline" next to the step 
wedge when it comes off the printer. 

(4) Create a new curves adjustment layer. Make the smallest change you 
possibly can, such as (say) adjusting ONLY the green curve at the position 
corresponding to the 20% step so it outputs 1 unit more green. Name this 
layer "Curves 1". Lock it. You will never change it again. 

(5) Select File>Print Options. Increment the vertical position of the step 
wedge by 1 inch. Put the same sheet of paper into the printer in the same 
orientation, and print. When it comes out, write the name of the new layer 
next to the new step wedge, which will now appear just below the last wedge 
you printed. 

(6) Compare the results from (5) to your baseline and any other step wedges 
printed so far. If you are like me, you've fixed some problems and created
some others.

(7) Duplicate the curve adjustment layer from the last step, and increment 
the number in its label, e.g. Curves 2. DO NOT change the last adjustment 
layer--it is now part of your "trail." In the layers palette, disable the 
old curve, enable the new one, and edit the new layer's curve with another 
small change. Lock the new layer and go back to (5), making sure that only 
the background and the new curve are enabled in the layers palette. 

After a while you'll have 9 or 10 step wedges all lined up on a single 
page, which makes it quite a bit easier than any other method I've used 
(including Gray Balancer) to hunt down color casts. For example, it is 
simple to compare how all the variations you've tried react under 
different lighting conditions, since they are all on one piece of paper. 
You learn just how distantly even miniscule curve adjustments propagate 
through the step wedge, etc. Each iteration is relatively quick. And if you 
overshoot your corrections, it is easy to go back to the adjustment layer 
that produced the best results so far and start over from there. 

One other note: I also print out a second baseline step wedge and 
carefully trim it so I can hold it directly against any of the step wedges 
on the page to see the difference side-by-side. This has significantly 
improved my ability to tell if I'm making a global improvement at each 
step or whether I am slowly color-shifting my way into, for example, a 
uniformly pea-green tonal scale. 

I fully intend to buy/build custom profiles, use ImagePrint, wait for 
Epson to improve their drivers, or some combination of these 
things... eventually. But as a newbie to color wrestling, I feel 
I am learning something by doing it the crude, manual way before I buy a 
canned solution. I'm definitely gaining an appreciation for the complexity 
of the process, in any event. I don't have a curve I am satisfied with 
yet, but I'm getting close, and anyway I'm not expecting this process to 
yield perfectly neutral output, just an inexpensive & relatively quick 
improvement over what I get with Schofield's method. 

The only question I have is, can it do any harm to run the paper through 
the printer with ink already on it from the prior trials? Will that gunk 
anything up if I'm always printing on a clean part of the paper?

Thanks,
David Wroblewski