Chuck, I agree with your observations IF it was a case where I'm determining
the resolution of a sensor. In my color-checker test fixture, I have a
converging line test in both red-white and black-white. The red-white
tester maxes out the resolution at least 50% less resolution.
But in this case, where I'm comparing lens to lens, I believe a standard
black-white line-pair chart is adequate for telling me what I needed to know
about the lenses. Also, in this case, I believe it to be more accurate as it
is using all four cells in the bayer quadrant whereas a test of, say,
red-black or blue-black, would reduce maximum resolving ability by around
2/3 and be prone to excessive aliasing. Since time immortal, lens tests
have been done using high-contrast monochrome targets (typically
transmittion targets, not reflective targets) on monochrome film, so i don't
think (although I could be wrong) that I'm out-of-line for using this
methodology in my testing. I could care less about the sensor as long as it
reveals enough detail (which the E-1 didn't) to reach an extinction point
beyond that of the lens itself.
I have an issue with Gary's lens tests in that it is subjective and not as
scientifically measurable as is a standard line-pair chart. Also, his test
is subject the vagaries of lens contrast. For example, one reason why
digital images look sharper than comparable film images is the ability of
the digital image to hold edge definition. The 35-80 images LOOK sharper
than the 35/2.8 images at, say F8 where both match each other on the
line-pair chart. With the survey map, this increase in contrast would have
made the 35-80 look possibly an entire grade better than the 35/2.8. A
change of films (specifically to one that that combines with the developer
to migrate grain for local-contrast-enhancement) would have possibly
revealed other results other than what he saw with the Ektachome 64T. His
tests are good (and good enough that I republished them on my website), but
represent one method of testing--one which is not strictly calibrated and
repeatable from one observer to another.
As to why your exposures were way off at f/22 I can't say. It seems
> highly unlikely that all of your lenses should be that much off true
> f/22 but I can't argue with your observation. But I don't think it
> matters. With a 4/3 sensor you should be avoiding f/22 like the plague
> as your real world resolution limit there is only about 1MP.
>
In principle, I would agree with you. Diffraction "typically" reveals itself
as a form of gaussian blur. As such, it is possible to counter some of that
blur with USM sharpening or a deconvulation sharpener. What I have seen with
some of the F11-F22 images is a noted softening or reduced contrast, but
subject-dependant. I wouldn't use F22 on a foilage-rich photograph, but
wouldn't hesitate to use it on a high-tech image where I have expanses of
smooth with sharp edges where I needed the DoF. I can fix the edges, but I
can't fix chaos/fractal-like subjects.
AG
--
_________________________________________________________________
Options: http://lists.thomasclausen.net/mailman/listinfo/olympus
Archives: http://lists.thomasclausen.net/mailman/private/olympus/
Themed Olympus Photo Exhibition: http://www.tope.nl/
|