Good, that seems to have worked as intended.
You use a film based camera to shoot an image of the target (digital
camera also works, but scanning the image is bit hard ;-)), and try to
avoid reflections of bright objects and sky. The goal is to project
such a small version of the target on film, that the center detail is
smaller than the resolution limit of the camera lens + film.
Depending on the lens+film that will probably be the case when you
shoot the target at a distance of approx. 50x the focal length you use
(exact distance is not very critical). Maybe 25x times focal length is
enough, but around 50x will always work for film. Use a tripod to make
sure the image doesn't suffer from camera shake. If the camera is an
SLR and has a mirror lock-up, even better results can be expected if
you use that. A low ISO film (e.g. Provia or Velvia) will give the
sharpest results, but you can also use the film you normally use (that
won't stress test the scanner, but will give an impression of what
resolution limits you can normally expect).
My guess is that the resolution of your target on paper will be around
5-10 cycles/mm, so if you reduce that with a factor of 50, then the
finest detail will become 250+ cycles/mm, which is too fine for your
lens to resolve, let alone in combination with film. That will ensure
that the resulting image will be resolution limited by the lens and
film.
After processing you will now have a target for scanning, and you know
the scanner is unlikely to be confronted with finer image detail than
in your test film. If everything went as it should, it is now possible
to judge the resolving power of your scanner, by measuring the
diameter of the central (approx. circular) blur. The diameter in
pixels can be calculated into a cycles/mm value as follows:
60 / pi / (pixels x 25.4 / ppi) = cycles/mm (similar to line pairs per
millimetre).
The number you find then indicates the limiting resolution for your
image chain. If the target and the scanner are very good, you could
reach something like 70-80 cy/mm. A modest scanner would reach perhaps
30 cy/mm. If the scanner can be focused, you can test how consistent
that is by making several cropped scans and compare.
You will see that the contrast/modulation towards the center will
reduce from paper-white to ink-black sinusoidal waves to more or less
uniform gray (zero modulation). The decline could be measured and put
in a graph which would become an MTF curve, but that takes quite a
bit of work. Just the limiting resolution will suffice for your goal
of quantifying the scanner limits, and determine if it is out of specs
or not.
To give you a typical idea of what my imaging chain produces:
http://www.xs4all.nl/~bvdwolf/temp/100mmJTF20004.png (it was shot with
a telelens and from a bit closer because I didn't have enough room to
back up). My scanner scans with 5400 ppi, so your image will probably
be a lot smaller. You are looking at approx. 3.4mm of film, neither
sharpening nor noise reduction was applied.
Bart
