At the time we didn't have a slanted edge device, but do now. Maybe when the opportunity arrives, we'll do another - and even include the DF for Eric's sake! I am 100% serious about the time investment. No, it does not take days and days. But to do it right, it does take time and at the expense of something else, returning a phone or an email, supplying someone with some researched information, etc. All of the normal things that come with service and that never ends and always takes priority.
Hi Steve,
I agree, and the testing is usually not a goal in itself. However, by adopting a repeatable(!) methodology one can hugely reduce the testing error component, and focus on the phenomenon to be evaluated itself. It also becomes easier to repeat the procedure later, without introducing new variables.
My test target is insensitive to distance variations (as long as shot from more than 25x focal length). So a given lens will always produce the same resolution (on the star blur diameter as well as on the edges), regardless of smallish differences in shooting distance. The star makes it easy to focus accurately, especially with a loupe or when tethered, just aim for the smallest diameter of the central blur area. When shot reasonably square, when the star is in focus, the nearby edges will also be in focus. It also allows to judge a single specimen, amidst others. When a given lens/sensor combination is known to be able and achieve a 98 pixel blur diameter, then any test that produces more blur means that the lens is of lesser quality, or that focusing needs to be done better.
The horizontal slanted edge is used to measure vertical resolution, which is where most of the mirror slap effect is expected to manifest itself. When the target is shot approximately level, the slant will allow to super sample the resolution at 1/10th of a pixel. That's accurate enough to detect even sub-pixel differences.
Since vibrations peter out over time, it will even allow to measure the quality of any vibration dampening effect since we introduce a time-weighted average during the exposure time. If the shake is vigorous, but only lasts a fraction of the effective exposure time, it will be of lesser importance.
It also becomes easier to compare tests performed by others.
Cheers,
Bart