Steve said:
None of my slides are perfectly flat in their mounts and I've noticed,
during preview, when I move the focusing crosshairs from the edge out to the
center of the slide the numbers change. The biggest change, so far, was from
117 to 138(in the manual focus pallette). I'm wondering what the range of
good focus is, if anybody knows?
Maybe I should set the focus in between the min and max numbers? Or the
set the cross hairs on the subject? Or is this a non-issue?
No, it certainly isn't a "non-issue"; it is probably the most serious
limitations of the Nikon scanner range! However your films don't look
too curved at all - I have seen some a lot (maybe 3-5x) worse than that.
How objectionable you find the problem depends on many things including
the image size that you print at, the original image sharpness, the film
granularity, how much post processing you do and, typically, just how
tolerant you are of slight defocus.
For the largest prints I make, typically 18-20x enlargement of the
original 35mm frame (ie. cropped from a 16x24" print), I find that my
range of good focus is around +/-5 of the optimum, with a tolerable
level of about +/-10. Anything further than 10 Nikon focus units from
the optimum level makes the grain on Velvia and Provia film I use
visibly softer. Using GEM to reduce the visibility of the grain
increases this range of acceptability out to about +/-12 because the
image on the film can never be as sharp as the grain itself. So a total
range of about 25 from one extreme to the other is probably the most
that I could tolerate with the film I use and the largest print size,
but I would prefer that to be closer to a maximum range of 20 or even
less.
Your data ranging the optimum focus from centre to edge of around 21
units seems borderline but likely to give acceptable results if you set
the focus point somewhere between the edge and the centre of the frame.
Since the film is tilted more off the horizontal plane towards the edge
than the centre, the rate of focus degradation gets worse towards the
edge, so some have recommend a default auto focus point around 2/3rds of
the way towards the edge. Doing it manually, you should probably pick a
number a little more towards the edge value than the centre. However
this really depends on the mounts, the film and just how curled they
actually are, so its best to experiment. As Bart mentioned - the
subject matter will also influence your choice of optimum focus point.
There really isn't an ideal solution to this. Some folks have
experimented with multiple scanning at optimum focus for different parts
of the image and then stitching the results together via masks in
Photoshop - seems like hard work to me! The only alternative is to
transfer the film to glass mounts, but that introduces potential Newton
rings issues and then, if you use anti-Newton ring glass, the image is
somewhat diffused - maybe worse than the original defocus.
Someone did suggest a proper drum scanner oil mount arrangement once,
but I would not recommend that due to concerns about the oil getting
into the scanner mechanism. Although it should evaporate off quickly,
who knows what non-volatile contaminants it will have picked up getting
there.
Your films don't look too bad, but if you have to scan some worse
examples then you might have to experiment with some of those
sub-optimal solutions.
