Digital BW, The Print

Traditional compensating development techniques for negatives are not 
required to produce a good print when the negative is going to be 
scanned and not printed traditionally.  It is, in most circumstances, 
not material to the success of the print whether you use 'normal' or 
compensated development (N+ or N-).

How's that for an iconoclastic statement!

My premise was that as long as the negative was exposed correctly for 
the shadows (ie the density range of the negative was up off the toe) 
it would not matter whether you developed it for a 'normal time' or N- 
time or N+ time.

This makes sense to me because of how I believe the scanner and 
software work.  When you scan a negative and apply set points you are 
mapping a density range on the film to a set of numbers (say between 0 
and 16383 for a 14 bit scanner), and then mapping those numbers to a 
grey scale.  For example if the negative's optical density range is 
from .4 to 1.8 then if no data is clipped by set-points then the 
density of .4 produces a 100% black in the file and the density of 1.8 
produces a 0% black in the file.  If that same negative had been 
developed differently ( say N- ignoring the 1/3 stop more exposure) and 
its density range was .3 to 1.5 it would produce the same file.  .3 
gets mapped to 100% and 1.5 gets mapped to 0%.  So the scene brightness 
that produced the shadow density value of .3 or .4 produces the 100% 
black in the both cases  and the scene brightness that produced the 
highlight densities of 1.5 or 1.8 produces 0% black in both cases.  The 
absolute value of the densities produced by the development doesn't 
make any difference except perhaps in extreme cases with very high 
(3.3) or very low (.5) maximum densities.

I made a simple test yesterday that seems, to me, to validate that 
premise. Here was the set-up: I exposed took two rolls of film with 
each frame containing the same image.  I used the new plusx.  The first 
roll was at my 'normal' EI 64 ( exposed 1/60s @ 5.6) developed in D76 
1:1 for 7:15.  The second roll was exposed at EI 40 (exposed 1/60s @ 
f4) and developed for 5:00 D76 1:1.

As you would expect the 'normal' negative was much denser in the 
highlights.  I scanned one frame from each film strip on a sprintscan 
120 using vuescan.  The little density meter in vuescan (for whatever 
it is worth) gave a density range of ~.35 to ~1.85 for the 'normal' 
negative  and ~.4 to ~1.5 for the N- negative.  I set vuescan to scan 
B&W, 16 bit, Neutral color balance, with black and white points set to 
0%.

The two raw scans are  different as you would expect.  The N- histogram 
is offset a little further from the left  and was shorter in length 
than the 'normal'.  The set pointed scans were virtually identical-!!  
I made a levels adjustment to each image to bring the white point in to 
clip of the specular highlights in each image and the two images are 
pretty much indistinguishable.  Their histogram data is:

                    Normal            N-

Mean             178             182
Std Dev           52               51.9
Median           198             204

My new zone mantra is: expose for the shadows and let the highlights 
fall wherever they want.

It is arguable that a more dense negative is better because the scanner 
will extract more data into the raw scan file but that does not seem to 
make any practical difference.  Perhaps it would if you were to apply 
aggressive curves.