Perils of lenstesting

[Erwin Puts <imxputs@knoware.nl>]

LUGgers,

The extended discussion about the relative and absolute merits of the
Noctilux versions shows once again, that the (subjective) appraisal of a
lens and the objective (scientific) testing of a lens are worlds apart.
So it should be. Testing a lens gives a number of hard facts, which only
imperfectly correlate with a person's liking or disliking of that lens.
Comparing results taken with two versions of the Noctilux in different
circomstances is the best way to ensure  an inconclusive exchange of
opinions.
We must  first of all distinguish between the quest for high resolution
images (the domain of Leica lenses) and the quest for high impact or
emotional images. (also the domain of Leica lenses).
High resolution images are subject to the laws of image degradation and
these can be tested and confirmed quantitatively. The image degradation
equation is part of the optical design process and can be
tested/measured. If lens A or lens B is better in this respect can be
noted IF all the components of the image taking chain AND the image
taking conditions AND the test target (be it a landscape, a model or a
testchart) are exactly the same.
If any one of these conditions change the comparison is difficult. The
Noctilux 1.2 has a high contrast image at full aperture, but rather low
resolution. The MTF is optimized for the 5 to 10 cycles/mm, that is to
produce clear and sharp object outlines. Taking a picture with this lens
on a 400ISO film will give a very crisp image with good subject
outlines. Take the same picture on a 100ISO film and the low resolution
of the '1.2' fails to bring in very fine details, which will be
reproduced as unsharp. This cut-off of the MTF curve is not so marked in
the 1.0 version, which will reproduce very fine detail with good clarity
and also gives high contrast subject outlines. On a 400ISO film with
pronounced grain (like the Tri-X) the grain pattern begins to interfere
with the clear rendition of fine detail and the resulting image taken
with the Noctilux 1.0 could lack the bite of the 1,2 version.  If you
would use a fine graind 400ISO (like the Ilford 400Delta) the Noctilux
1.0 can give better image quality. If you develop your B&W films to a
different gamma (a C.I. value of 0.56 versus a C.I. value of 0.70) the
resulting negative contrast difference is quite substantial, but
difficult to see without a densitometer). This difference could give a
different impression and subsequent evaluation of full aperture contrast