Archived posting to the Leica Users Group, 1999/07/06

[Author Prev] [Author Next] [Thread Prev] [Thread Next] [Author Index] [Topic Index] [Home] [Search]

Subject: [Leica] Why do I not give figures in a lensreport
From: Erwin Puts <>
Date: Tue, 6 Jul 1999 12:38:10 +0200

Some of you might assume that I do not have figures to provide you with.
The contrary situation exists. I have thousands of data about every lens I
tested.  But I will not give them and I will now explain why not. First of
figures acquire quickly a status of their own and will be used out of
context. The wellknown Mr Crawley also refuses to give figures for this
reason. Assume I have a lens with an on axis resolution of 125 lp/mm (this
is a real figure). Then I have another one with 100 lp/mm. Our cultural
bias will assume that the first lens is 'better'. Now I measure both lenses
again on an image location 3 mm from the center (the axis position). The
first lens has a resolution of 77 and the second one keeps its 100lp/mm.
Now which lens is better? I my view the second one as this covers a larger
area of the image with a very high resolution, while the first one is a bit
higer on one center spot but rapidly drops when in the field.
But there is much more. Resolution figures are not that important today,
but let us continue. The resolution figures change depending on the line
orientation of the testpattern. (aberrations etc). So the first lens (A)
has a resolution for the horizontal pattern of 77 and for the vertical one
of 50. Which one do you prefer. Now you rotate the lens and in its worst
position you get 50 and 33 resp. Again which one do you want. Lens B has 60
in all directions and does not change while rotating.
But not enough yet. Resolution patterns are visually analysed. As long as
two lines are distinguishable you have the figure. But how to make a
distinction between barely noticeable and clearly visible. For the
resulotion method both are valid. But yet more. The black line is greyish
because of flare. The resolution is the same as with a fully black line,
but the image is degraded. The resolution figure does not tell you this.
Two lenses have both good solid black lines, but one of them has color
fringes next to the black border. Chromatic aberrations not being related
to the resolution figure. One lens has fuzzy black borders and another one
very sharp lines.
How do these patterns change from center to corner? How do they change when
stopping down.

So the idea is clear. One figure is most misleading and at least inadequate
to get a proper evaluation of a lens. Or you need all figures.  My
testmethod evaluates Leica lenses at 12 different positions from center to
corner. The testpattern goes from 1 to 200lp/mm, is sagitally and
tangentially oriented. Is anyone interested in  24 values (12 horizontal.
12 vertical) per aperture, where every individual figure must be qualified
with flare, and other aberration information.
Why did Leitz close the famous testlab in Wetzlar?  Becuase their computers
generated so much figures that any  evalution of a lens became impossible.
So they shifted to the wellknown MTF graphs, which stop at a resolution of
40 lp/mm. My MTF graphs of Leica lenses show me that the limiting
resolution of a Summicron is far above 300lp/mm. Does this figure give you
a performance indication of the lens. Why should Leica limit their
presentations at 40lp/mm when much more resolution is available?
Because figures do not tell you the story about a lens.
That is why I do not give figures, even if I have more figures that you can
dream about.