How to determine the resolution actually reached in a photograph a practial example
Let us assume we are about to take a photo of
a landscape. Somewhere there is a farm. And a wooden fence. Its white
painted vertical slats, each about two inches wide (which we will
convert to 50 mm) and spaced with about two
inches distance to the next slat (we may have to go there and measure
these values), stand clear against some darker background. This wooden
fence represents a structure of dark and bright lines similar to those
in resolution test
targets. Let us further assume that we are using a Carl Zeiss 50 mm Planar standard lens and a Contax SLR camera. The fence may be a quarter of a mile away (as a map can tell us or a binocular with a distance measuring device).
Now, let's make a rough calculation of the resolving task facing our imaging system: From a quarter mile distance, which is about 400 meters, the reduction ratio, let's call it "r", for the fence to be imaged on film with the 50 mm lens, is as follows: r = focal length divided by object distance, expressed in the same unit. This gives: 50 millimeters /400 x 1000 for a ratio of 1/8000.
So the original subject, the farm and the fence, and every detail will appear at 1/8000th of its original size in the negative (or slide, whichever you prefer). One slat plus one space, both representing one line pair in the original subject, are 4 inches wide. This equals roughly 100 millimeters. Reduced by the reduction ratio of 1/8000 it should appear on film with this width: 100 millimeters / 8000 = 1/80 of a millimeter. So 80 fit into 1 millimeter, which means, we are
about to resolve 80 line pairs per millimeter.
Can our imaging system handle this task? For a Zeiss Planar in a Contax camera and any modern 100 ISO film, 80 line pairs per mm is an easy challenge, if we focus precisely and avoid unwanted vibrations (by using a good tripod). So let's go ahead and take the picture. The fence should be well resolved in our photo.
Or, the other way round: If we took the photo and later found the fence resolved, then we knew we did achieve at least 80 line pairs per millimeter.
Camera Lens News No. 3, winter 1997/1998