[Leica] lens evaluations

[Erwin Puts <imxputs@knoware.nl>]

- --============_-1308258681==_ma============
Content-Type: text/plain; charset="us-ascii"

Recently a long quote by Jim Williams (Contax G fame) hs been circulated on
the LUG. I at first resisted temptation to react. Still some comment is
appropriate I assume. First of all his text is full of statements, not a
single one prooved or explained. He is also demagogially clever ("superb"
Zeiss lenses versus "computer designed" Summicrons). It is however not my
goal to refute Mr Williams. He seems convinced of his opinion and so be it.
At stake is the reputation of Leica or its "reputational myth-making". And
the large group of Leica users who "because of their own fixed
beliefs in Leica superiority -- certainly weren't about to disabuse the
masses of their illusions".
So are leica lenses superb or just front runners or me-too products. Were
leica lenses unsurpassed in the past or not?
From my postings (and test result) you will know that I am not a person to
give super high marks to older Leica lensgenerations. But I am also not
inclined to give older Zeiss or Japanese very high marks. The Planar 2/50
for the Contarex, once described as of unsurpassable optical quality, is
very good but not state of the art anymore. And so is the seven element
Summicron.
Many Leica lenses are excellent and often the number one for its fovcal
length and specs. But leica also had and has lesser designs, quite apt for
its task, but not state of the art: the 1,4/35 Summilux and the Summicron
2/90.
The problem with all these generalisations is this: they assume one fixed
evaluation standard along which all lenses can be measured and given a
simple merit figure (be it a number, a word or a range of stars).
This is quite wrong and even more importantly it misrepresents the reality.
You still owe me the Summilux-R report. Why? I have all bench results,
hundreds of comparison transparancies and many MTF graphs. So it should be
easy to make a comparison and a clear ranking? Ok, let us start:
I compared the Summicron-M and R and the Summilux-R (old) and new and the
Summilux-M. That gives a bit body to an evaluation and it brings out small
differences, which may not be important for a star ranking, but are
important for real life shooting.
Well overall the Summicron-M is slightly ahead of its R-sibling. At full
aperture the M has a bit more contrast and its ability to record extremely
fine details with outstanding clarity is greater. Also the performance of
the M extends over a larger image circle than does the R, which is a bit
weaker in the corners.
The Summilux-R (old) is of low contrast and the recording of fine detail in
the center is very good, but drops off fairly quick (after an image circle
of 6mm radius). Stopping down the quality improves slowly until at f/5,6 to
f8,0 where we can find excellent quality.
The new Summilux-R has a high contrast at full aperture with fine to very
fine detail excellently rendered over the whole image field. Stopping down
one stop brings image quality better than its predecessor at f/4,0 and on
stopping down image quality rapidly improves to a new optimum level at
f/5,6.
The Summilux-M is about equal to the new R in the center (6mm radius) but
stays behind in the outer regions. Stopping down improves the contrast and
the clear rendering of very fine detail. At f/5,6 we have excellent
quality. The Summicron-M at the same aperture could be qualified as of
exceedingly high imagery.

Now I do hope that in the context of these descriptions designations like
"very fine details" and "outstanding clarity" make sense and can be related
to photographic practise and expectations of image quality. To be brave and
reckless for once I would rank the Planar 1.4/50 for the Contax RTS as
above the old Summilux-R but below the new one.

Now how should we rank these five lenses based on the text above?
Summicron-M is still the best, followed at a little distance by the
Summicron-R, which is closely followed by the Summilux-R new. This lens in
some important areas inches ahead of the  Summilux-M.
The Summilux-R old is a bit lagging in all areas and outdistanced by the
new Summilux-R.
This would be Ok as far as my *current* set of evaluation criteria will
allow. If you put different weightings to the several components of image
quality (as defined by me, mind you) the ranking and the evaluation could
be different. The clear and contrasty rendition of very fine detail in the
zonal region at the wider apertures is for me a very important componant as
it will define the capacity  of big enlargements from negatives (or
trannies) taken in adverse luminance conditions. Most testers I know of
(yes also CdI) are looking for different aspects of image quality.
The overall ranking might or might not be the same. But a simple star
rating system (the Summicron-M gets 5 stars from CDI, the Summicron-R gets
4 and the Summilux-M gets 3) does not give insight into the finer points of
the differences in optical performance. It will also not tell you how the 5
or 4 or 3 stars have been put together.
If we now would try to compare lenses from several reputable companies over
long ranges of lenses and over several decades with different and often
non-reproducable test paramaters from many different persons who also
change their evaluation criteria (or even more worse did not change their
criteria) we would be insane. At least I would be insane.

Why then did I start with telling you that I have not yet made my final
report. Because I do conduct tests at several distances ( 1 meter, 3
meters, 5 meters and 10 meters or infinity). It is quite logical that at 1
meter the capacity of rendering fine details is different from the 10 meter
distance if you are looking for the *same fine details*!!. At one meter we
would need let us say 10 lp/mm of good contrast to record these details. At
5 meter we need 50lp/mm to record the same details with the same contrast
and clarity. As 50lp/mm represent unbelievably  fine details that can not
be seen in a transparancy projected to a 4 meter image and viewed at close
range we must carefully reflect how to evaluate performence differences at
this demanding level of image quality. Now the 5 lenses mentioned above
have different characteristics in this respect and I just need time and
some study to phrase my conclusions.


Erwin
- --============_-1308258681==_ma============
Content-Type: text/enriched; charset="us-ascii"

Recently a long quote by Jim Williams (Contax G fame) hs been
circulated on the LUG. I at first resisted temptation to react. Still
some comment is appropriate I assume. First of all his text is full of
statements, not a single one prooved or explained. He is also
demagogially clever ("superb" Zeiss lenses versus "computer designed"
Summicrons). It is however not my goal to refute Mr Williams. He seems
convinced of his opinion and so be it.

At stake is the reputation of Leica or its "reputational myth-making".
And the large group of Leica users who "because of their own fixed

beliefs in Leica superiority -- certainly weren't about to disabuse
the

masses of their illusions".

So are leica lenses superb or just front runners or me-too products.
Were leica lenses unsurpassed in the past or not?

From my postings (and test result) you will know that I am not a person
to give super high marks to older Leica lensgenerations. But I am also
not inclined to give older Zeiss or Japanese very high marks. The
Planar 2/50 for the Contarex, once described as of unsurpassable
optical quality, is very good but not state of the art anymore. And so
is the seven element Summicron. 

Many Leica lenses are excellent and often the number one for its fovcal
length and specs. But leica also had and has lesser designs, quite apt
for its task, but not state of the art: the 1,4/35 Summilux and the
Summicron 2/90. 

The problem with all these generalisations is this: they assume one
fixed evaluation standard along which all lenses can be measured and
given a simple merit figure (be it a number, a word or a range of
stars). 

This is quite wrong and even more importantly it misrepresents the
reality.

You still owe me the Summilux-R report. Why? I have all bench results,
hundreds of comparison transparancies and many MTF graphs. So it should
be easy to make a comparison and a clear ranking? Ok, let us start:

I compared the Summicron-M and R and the Summilux-R (old) and new and
the Summilux-M. That gives a bit body to an evaluation and it brings
out small differences, which may not be important for a star ranking,
but are important for real life shooting. 

Well overall the Summicron-M is slightly ahead of its R-sibling. At
full aperture the M has a bit more contrast and its ability to record
extremely fine details with outstanding clarity is greater. Also the
performance of the M extends over a larger image circle than does the
R, which is a bit weaker in the corners. 

The Summilux-R (old) is of low contrast and the recording of fine
detail in the center is very good, but drops off fairly quick (after an
image circle of 6mm radius). Stopping down the quality improves slowly
until at f/5,6 to f8,0 where we can find excellent quality.

The new Summilux-R has a high contrast at full aperture with fine to
very fine detail excellently rendered over the whole image field.
Stopping down one stop brings image quality better than its predecessor
at f/4,0 and on stopping down image quality rapidly improves to a new
optimum level at f/5,6.

The Summilux-M is about equal to the new R in the center (6mm radius)
but stays behind in the outer regions. Stopping down improves the
contrast and the clear rendering of very fine detail. At f/5,6 we have
excellent quality. The Summicron-M at the same aperture could be
qualified as of exceedingly high imagery.


Now I do hope that in the context of these descriptions designations
like "very fine details" and "outstanding clarity" make sense and can
be related to photographic practise and expectations of image quality.
To be brave and reckless for once I would rank the Planar 1.4/50 for
the Contax RTS as above the old Summilux-R but below the new one. 


Now how should we rank these five lenses based on the text above?

Summicron-M is still the best, followed at a little distance by the
Summicron-R, which is closely followed by the Summilux-R new. This lens
in some important areas inches ahead of the  Summilux-M. 

The Summilux-R old is a bit lagging in all areas and outdistanced by
the new Summilux-R.

This would be Ok as far as my *current* set of evaluation criteria will
allow. If you put different weightings to the several components of
image quality (as defined by me, mind you) the ranking and the
evaluation could be different. The clear and contrasty rendition of
very fine detail in the zonal region at the wider apertures is for me a
very important componant as it will define the capacity  of big
enlargements from negatives (or trannies) taken in adverse luminance
conditions. Most testers I know of (yes also CdI) are looking for
different aspects of image quality.

The overall ranking might or might not be the same. But a simple star
rating system (the Summicron-M gets 5 stars from CDI, the Summicron-R
gets 4 and the Summilux-M gets 3) does not give insight into the finer
points of the differences in optical performance. It will also not tell
you how the 5 or 4 or 3 stars have been put together. 

If we now would try to compare lenses from several reputable companies
over long ranges of lenses and over several decades with different and
often non-reproducable test paramaters from many different persons who
also change their evaluation criteria (or even more worse did not
change their criteria) we would be insane. At least I would be insane.



Why then did I start with telling you that I have not yet made my final
report. Because I do conduct tests at several distances ( 1 meter, 3
meters, 5 meters and 10 meters or infinity). It is quite logical that
at 1 meter the capacity of rendering fine details is different from the
10 meter distance if you are looking for the *same fine details*!!. At
one meter we would need let us say 10 lp/mm of good contrast to record
these details. At 5 meter we need 50lp/mm to record the same details
with the same contrast and clarity. As 50lp/mm represent unbelievably 
fine details that can not be seen in a transparancy projected to a 4
meter image and viewed at close range we must carefully reflect how to
evaluate performence differences at this demanding level of image
quality. Now the 5 lenses mentioned above have different
characteristics in this respect and I just need time and some study to
phrase my conclusions.



Erwin    

- --============_-1308258681==_ma============--