With the discussion going on the other site on figuring out coverage I've got to ask. How to tell coverage from just ilumination? I guess it depends on the final product. Contact prints can get by with a lot less on the film then something getting enlarged quite a bit. But in general how do you do it?
Very simple: You can't.
Some lenses are bulit to physically vignette at the edge of the sharp image (e.g. Zeiss Planar 135mm/3.4), most are not (e.g. Schneider Tele-Xenar 360nn/5.5).
I'm afraid you have to test, or trust manufacturer's data if available.
Certain lenses like Angulons have a circle of illumination that goes significantly beyond the circle of good definition, and no amount of stopping down will fix that. Interestingly, Dagors are of similar design to the Angulon, but they don't have such a dramatic falloff of image quality at the corners.
Ok is there a standard for good definition or is it something that is up to personal taste? There are stories of some lens manufacturers reducing the published coverage figures over the years. The only change supposedly being how they defined good definition.
I guess what I'm asking is. Just how much can I get away with when looking at older lens? One of the sellers on Ebay uses coverage figures that far exceed anything I can find published. He seems very well respected so I'm guessing he isn't puffing things up unfairly.
I'm guessing this is one of those questions that can't really be answered since it's too dependent on each different lens.
The "standard" could be offered in line pairs per mm or MTF, but it would depend on other factors like enlargement and viewing distance.
Much really depends on the design of the lens. Some lenses like the Angulon I mentioned above (not the Super-Angulon or later versions) or meniscus lenses have a very rapid drop in resolution and really can't be used beyond the circle of good definition unless you are going for an effect (like Sally Mann, for instance). Others, like some early ultra-wide designs might show a smoother and less severe decline in definition at the corners that can be quite tolerable.
This image, for instance, was made on an 8x10" camera with a pre-war 120mm/f:14 Berthiot Perigraphe at f:44:
This lens just covers (illuminates) 8x10". If you look at a print or the negative with a loupe, you'll notice visible coma in the corners, but it's not such a distraction really. I can work with it.
On the other hand, I was looking at some 5x7" negs that a friend made with a 90mm Angulon, and the sudden loss of sharpness at the edges was just unacceptable. It was as if the photograph had been made through a Coke bottle that was drilled out in the center.
See if you can try it before buying it, or make sure they'll refund it after a few days if you don't like it.
Then, take a few pictures with even illumination for the subject in the corner(s) with detail you can judge with your eye. For the first picture, don't use any movements, and for the following ones, add some rise/fall or shift. After printing straight from the negatives, compare the corners where you're getting closer to the edge of the circle, and hopefully you'll see how much movement you can get while still getting acceptable results.