Archived posting to the Leica Users Group, 1998/11/28

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

Subject: [Leica] Dominique's comments about fairness (very long)
From: Erwin Puts <>
Date: Sat, 28 Nov 1998 14:58:56 +0100

Dominique wrote:
>According to Laney, the edge spread function method has been developed
>by... Leitz.(see. p.21 in the 2nd edition).One more time you gun down a
>research or a method which has been made by others than you.That's neither
>fair nor academic.
>The method seems to have "evaporated" (why ? I'm waiting for a rational
>explanation) and today Leica uses standard MTF tests.
>It is interesting to quote Lothar Koelsch, the father of recent Leica
>lenses, who wrote 4 papers in Leica foto (1996) on the photographic quality
>of lenses : "higher are the MTF curves, better is the lens" (1996/5, p. 45).

First of all, Dominique, I am well aware that the edge spread function
approach for image analysis, has been developed by Leitz, more to the point
by Mr. Thomas from the Leitz Rechenbüro (Optic Design department).
In fact I have the following articles about this function:
Delingat, Thomas, Vollrath: Uber die Ermittlung von Bildgütezahlen bei der
Messung der OTF, 1983
Thomas: Die Kante als Strukturelement zur Bildgütebeurteilung von
Fotoobjektiven, 1984
Marx: Umrechnumg der Thomas'schen Kantenbild-Breite in ein Bewertungsmaß
für die optische Ubertragungsfunktion, 1984
Vollrath: Berechnung und Bildgütebewertung von Foto-Objektiven, 1985.
What Laney did is extracting parts of these articles to compile/condense
them for the chapter in his book.

I am not aware that I gun down or condemn "once more" a research or a
method which has been made by others than you. That's neither fair nor
I have clearly stated in all my posts around this evaluation topic, that I
know and appreciate the methods used by PopPhoto (in fact I am quite alone
in defending the SQF as a valid approch), that I know the MTF measurements
by ColorPhoto, Photodo and Chasseurs.  I NEVER questioned the method or the
research around MTF analysis. Be it done mathematically on a computer,
generated by Ealing equipment or produced by microdensitometer. Yes, there
is much more around MTF theory that you dare to imagine. BTW I also use MTF
measurements, provided by Leica and others. So I do not question any method
not 'invented'  or used by me. My methods consists of techniques  that are
not different than the ones used by others. It is your accusation that is
unfair and unacademic. What I question is the way the results are
interpreted and presented to users of Leica equipment.
The weighting method by Photodo is the method I question, as is the one
used by Chasseurs. Again the micro-slit measurements by PopPhoto are OK,
but the use of these results in the complicated weighting used to generate
the SQF is at stake, not the Ealing equipment. Should Photodo or Chasseurs
give the MTFgraphs as presented by Zeiss and Leica and state clearly the
optical parameters to generate the MTF graph, I would have no quarrel with
anyone of them. Of course I know the motive of photodo/cdi et all. MTF
graphs are very difficult to interpret and no one wants to study these
graphs at length. Or better still: without sound optical theory and design
practise as a background  the connection between these graphs and real
photography is very difficult to make. Therefore these magazines 'help' the
reader to collaps all this information into one merit figure (4 stars or a
4,7). This approach I think is in our enlightened society  not viable, not
anymore. If you use the figures you present them in an intelligent way and
make sure the reader can correctly interpret them. If you are choosing
lenses because CdI gives 5 stars and you neglect the ones that get a 4 star
rating: be my guest. And if you prefer a leica lens because photodi gives
it a 4,7  where another one gets a 4,6: again be my guest. I still propose
that the choice of a lens based in this colllapsed merit number is at best

Why has the edge spread function evaporated?
Laney seems to miss the core of the Leitz argument. The image quality of a
lens is fixed by a function called the Optical Transfer Function. This is a
Fourier transform of quite complex form. It has two components: a value and
a phase. The value is representd by the wellknown MTF graph. The phase is
represented by the you guess Phase Transfer Function. The MTF values refer
to the contrast while the PTF refers to the spaial phase difference. The
MTF canbe produced quite easily graphically. The PTF not all all. The
approch of the Leitz engineers from the early 80's was to try to design a
graphical representation to evalute the phase difference of the PTF. They
tried it with the edge spread function.
It did not work: first of all their approach had no theoretical optical basis.
Secondly the idea to use the edge contrast can now quite easlily be done by
standard MTF measurements.
Thirdly the high speed of computers eliminate part of the need of assessing
the PTF component.
Fourthly: as soon as image quality does not vary that much over the image
field as is the case with modern Leica lenses the efect of phase
differences is minimized.
Fifthly: the MTF as a rule never gives all information to pertain to a
lens. Nor does the PTF.
Sixly the OTF is to be prefered and a large assortment of optical
measurements: point spread functions, lateral fans, and a host of others.
The MTF is a convenient shorthand result for knowledeable optical
engineers, it is NOT a goal in itself. After all the optical aberrations
are the problem to correct.

Summarizing: the edge spread function was not the solution to a real
problem and that problem partly went away and partly can be adressed by
other more exact menas and methods.

I am at a loss why you in the context of your argument refer to Mr Kolsch:
what he says does not contradict what I say : an MTF graph is a valuable
report about the image quality of a lens and as a rule the higher the
graphs the better the image quality.  It is not true that "today Leica uses
standard MTF tests" as this would imply  an exclsive reliance onn them.
Leica has one of the few design teams that use an avalnache of highly
sophisticated intrumnts and software programs to design lenses: the MTF is
a tiny part of them. Most of all Leice designers use quite amply a
diminishing asset in today's world: brain and creativity.

Now that was the explanation demanded by you. Eric does not want me to take
things personally . Still I feel a bit annoyed by your tone: "gun down"
"not fair or academic" especially as these accusations are made without any
knowledge of my way of work or background. Also your demand: "(why ? I'm
waiting for a rational
explanation)" sounds quite agressive to me. Why should I have to explain
indepth my statements to you where you never have given any explanation
about your statements or claims besides citing other sources.

I do presume that some Luggers might be interested in this matter and that
is the reason that I bring in into this forum.