Hi--

Thank you for publishing your Nikon lens list which I found on searching
for more information about Nikon lenses for Leica thread mount cameras.
I found this information quite helpful to me.

I disagree with you on your assessment of the linear distortion in
Nikon's lenses, at least those wider than 35mm. I have this same
disagreement with the photo magazines' testers in rating various amounts
of distortion as "low." As long as testers consider 1% distortion to be
"low," manufacturers who produce 0.5% distortion lenses won't have the
competitive advantage they deserve. (1% distortion is a bend of 1/20
inch in a vertical line along each side (0.5% each side adds to 1% - I
assume that's how it's defined - and that's how I use it below) of an 8
x 10 print - that's pretty obvious and significant at the edge of a
print - you probably wouldn't accept that from your copy machine if you
were making comparison tests. The 2nd or 3rd generation copies would
look like they were printed on balloons.

Not from a copy machine - it is supposed to "copy" accurately, but 1% is
considered relatively low for SLR lenses, and most Nikkor non-zooms meet
this spec at the edges (though mebbe not further in...). BTW, I take 1%
to be a line displacement along *one* long side of the frame of 1/100th
of 36mm, or .36mm... In an 8x10, that would be 1/100th of 10 inches, or
.1 inch, which is quite low for 35mm SLR lenses, *relatively* (you
should
see what some do...! ;-).

It's like the tests of the digital cameras saying the autofocus response
is "fast" at 0.6 second. Give me a break! Lets rate all the cameras the
same, and just let people know they can choose between autofocus and
fast response, but can't yet get both.

But then whole classes of cameras would fail, yet they are useable for
some purposes...


You indicate that most Nikon lenses have relatively low linear
distortion. As one who up until about three years ago shot a lot of
urban and architectural scenes in my work and used my Nikons to prepare
original photos for use in computer visualization where the scene must
be simulated in a computer image, I have not found this to be the case.

I would not use SLR wide-angles for this work... The whole class of SLR
WAs and zooms is known to be inappropriate for much of architectural
work, for the reason you note (with a couple of exceptions, below).

In particular, I have found the linear distortion on the wide zooms to
be substantial, and that on the wide fixed focal length lenses to vary
considerably. I found the 18mm f/3.5 to be pretty good, the 20mm f/2.8
less so, the 28mm f/2.8 to be unacceptable for barrel distortion, and
the zooms to be gawdawful bad (though I didn't use any of the pro zooms,
and the images I have seen indicate that they may have distortion equal
to or less than the worst of the fixed-focus lenses I tested). Because
the wide zooms I have checked in the stores (24-120, 28-105, 18-35) have
barrel distortion or central barrel/outside pincushion obviously visible
in the finder when using a grid focusing screen at the wide end, I have
never purchased one.

Nor would I, for this purpose - which is a very particular one demanding
performance in one aspect beyond what is normally required... It is
simply
an inappropriate choice of gear for a particular purpose (for city-
scapes
more than for straight architecture, where the curves can often be
concealed
or accepted...).


In considering what was available and what I needed for accurate 3-d
visualization, I considered 0.5% distortion along the center 30mm along
a line 10mm from the top and bottom (combined variation between the two
lines bove and below the center to be excellent, but found no Nikon
wideangle lenses I observed except the 28mm PC-Nikkor (I did not
evaluate the 35mm PC) to meet this criterion. After looking at the
specifications for reproduction lenses, I would guess that most lenses
used for repro work have substantially better than this level of
distortion at the standard reproduction ratios for which they are
designed. However, it seems to be difficult to design a wide lens for
this level of distortion over the range of focal ratios used in 35mm
photography.

For SLRs, yes, but rangefinder lenses, often being symmetrical designs,
can often produce low levels of distortion (but framing and visually
leveling
these is a pain!).

Specifications for the Leitz 35mm Aspherical and 21mm Aspherical lenses
show they meet this specification at infinity, as does the 38mm Biogon
(just barely) on the SWC (displacement of measurements adjusted for the
format) and the 35mm Grandagon.

And so will most other 35mm rangefinder lenses... Try the Voightlander
12, 15, and 21mm lenses for "quality on a budget" (the 12mm is listed
in the Nikkor list - and it is for sale at
www.David-Ruether-Photography.com/fs.htm).

I considered 0.7% to be good but not observed in any other Nikkor lenses
I tested (I sold my 21mm f/4 long ago, alas), 0.85% to be barely
acceptable (the 18mm f/3.5 was about this level) and anything over 1% to
be unacceptable. The 28mm f/2.8 was about 1%. The one zoom I owned (28-
70mm f/3.5-4.5) had over 2% distortion at 28mm at this distance from the
center of the negative. I suspect that the retrofocus design used in
reflex wideangles has inherent barrel distortion. I found that
rangefinder wideangle lenses in general did not have this problem.

Yes. There are two very low distortion Nikkor WAs and one WA zoom
The 15mm f5.6, the 21mm f4, and the 28-105mm f3.5-4.5 (the 100-300mm
f5.6 is also very low in distortion). The 28mm f3.5 and f4 PC lenses
(I prefer the f4) are acceptable. The MUST OWN WA for Nikon SLR
architecture work, though, is the 15mm f5.6 (NOT the f3.5...) - it
has virtually no linear distortion, excellent-to-the-corner sharpness
by f11, even illumination, and fair "snap" (samples must be selected,
though...). The 28mm f4 is a very crisp lens. These two serve for
90% of architectural work for me, with an 85mm or some-such additional.


It would be great if somone would catalog the distortion specs on these
lenses from past tests in the magazines or other sources if available -
perhaps if I get back in this practice I'll do it.

One problem with this is that some lenses have very complex distortions
(like the 24mm f2.8 Nikkor, that shows VERY little frame-edge
distortion,
but just inside, it is very high); another problem is that some
distortion
types, of the same amount, look worse than others (pincushion is worst,
"wavy-line" is next, and I often like barrel distortion, even if it is
strong, for shooting people... You have particular needs, but these may
not
be common to most others...
DR

Pat Mann