[Leica] lens evaluations continued

[Erwin Puts <imxputs@knoware.nl>]

The ranking of CdI (from 5 to 3 stars) is too rough on two counts. The
equidistant steps of one star per rating category suggests an equal drop in
performance. But the performance of the Summilux-M  with 3 stars is much
closer to that of the Summicron-M with 5 stars and the difference between
the Summicron-m and R is much smaller too. Partly it depends on the
performance weightings. If you take an overall average of image quality
from let us say f/4 to f/8 all three lenses would get four and a half or
five stars. (including the new Summilux-R). If you only look at the full
aperture performance a jump of 3 to 5 stars is too big).
Even if we could agree on the rating being fair or accurately reflecting
the quality potential of imagery, then we do not have any clue on which
aspects the Summicron-M gets 5 stars and the Summilux-M gets 3 stars. If
the rating is comparative (that is compared to all other 50mm lenses ever
tested) we are in trouble. If the rating is absolute (that is reflecting
some type of MTF measurements) we are also in trouble.
As example: flare suppression is never measured with any MTF instrument.
But flare is very important for the overall clarity of the image and the
contrast of the very fine image details. Again: in my testing I give more
emphasis to flare and the image quality at 20lp/mm as I know that on these
accounts the qualitative discrimination starts. I am, relatively spoken, a
bit more lenient to corner performance and will never condemn a lens if the
far corners are out of synch with the outer zones. Now if CDI gives strong
emphasis on the cormers and less on flare the overall rating might still be
the same, but the *interpretation* of the rating and the evaluation of
performance would be vastly different.
I have taken this comparison to some depth (not so because I do consider
the CDI test as a reference point, but on the LUG the CDI results are often
referred to) to illustrate (and I hope to demonstrate) what I said in the
beginning of the earlier post. It makes no sense at all to start comparing
results all over the world and trying to average all conclusions. Nor does
it make sense to compare historical statements from many people about many
lenses in the hope that some golden common treadth will emerge on the
argument that so many people cannot be all wrong. My point is that unless
you can compare statements of persons who have an equivalent set of
criteria and the same categories and dimensions for evaluation, we are
searching for the Holy Grail at best. At worst we stay ignorant and
continue to make sweeping and inaccurate (or even wrong) comparisons.
Some last points:
Zeiss Sonnar lenses for the old Contax are often considered to be optically
superior to the equivalent Leica lenses. In the thirties this opinion had
some merit. Not because the Sonnar design inherently is better than the
Double Gauss design. Far from it! DG lenses theoretically and
computationally  outdistance the Sonnar types easily. But the higher
contrast (less lens elements) and the generally coarse grained emulsions
boosted sharpness impression.
Leica lenses are now designed to very low mounting tolerances. This assures
a lens performance close to theoretical maximum and for a very long period.
THis is not testable, but of the utmost importance. Some important lens
characteristics of Leica lenses (No, no not bokeh) that give the lenses its
fingerprint, can be analysed only if you know where to look and what to
look for. Standard testing procedures will miss these personality traits.

Erwin