[rollei_list] Re: Lens Tests
- From: Richard Knoppow <dickburk@xxxxxxxxxxxxx>
- To: rollei_list@xxxxxxxxxxxxx
- Date: Sun, 27 Mar 2005 03:52:32 -0800 (GMT-08:00)
-----Original Message-----
From: Fred Fichter <ffichter@xxxxxxx>
Sent: Mar 27, 2005 12:56 AM
To: rollei_list@xxxxxxxxxxxxx
Cc: rolleiusers@xxxxxxxxxxxxxxx
Subject: [rollei_list] Lens Tests
Rollei vs Hassy vs Mamiya...
http://www.hevanet.com/cperez/test/fourcameras.html
Chris Perez does careful testing and is aware of the pitfalls of in-cam=
era testing. I would like to see comparisons of the lenses where some of th=
e unknown factors were in better control. I am not convinced that film flat=
ness is the only variable. Aerial image testing would give a better idea of=
what actual lens resolution is although it is not so good a method for mea=
suring either edge contrast or overall contrast.=20
The compromise made in lens design is between resolution and edge cont=
rast by choosing the way in which the energy is concentrated in the image. =
One measure is Strell ratio, which measures the ratio of the light energy i=
n the central diffraction peak to the overall energy, i.e. how much is in t=
he peak and how much is in the surrounding rings. The narrower the central=
spot the higher the resolution, everything else being equal. However, this=
results in a relatively large amount of energy in the rings. By making som=
e changes that broaden out the centeral spot the energy in the rings is red=
uced improving the contrast but reducing the resolution. This is sometimes =
known as apodising (removing the feet). A similar choice is made in the des=
ign of directional radio antennas such as satellite dishes. Since lens reso=
lution and film resolution are likely to be nearly the same for practical M=
F lenses the resolution on film may be misleading. Resolution must also be =
measured at a stated contrast between the resolved lines and the background=
. The lack of such a specification is one reason there is sometimes disagre=
ement about the resolution of lenses.=20
If Chris's values for the Mamiya of 120 lp/mm at the center are for an i=
mage on film it is phenominal since it is close to the resolution limit of =
T-Max and would indicate the lens resolution is so high that it is not redu=
cing the film resolution by much. If its for an aerial image it is still ou=
tstanding but more reasonable.=20
I am always suspicious of resolution numbers that are higher at some mid=
way portion of the image than at the center. This IS possible given the ef=
fects of higher order aberrations but is usually a good reason to make sure=
that there isn't some defocusing going on.=20
Lens testing is not trivial and it is asking much to control all the var=
iables but they should be kept in mind when interpreting tests like these. =
It is clear that Mamiya is doing something right because nearly everyone w=
ho uses the cameras remarks on the outstanding quality of the lenses.
As far as compromises made by Zeiss vs: Mamiya or Schneider who actually=
knows. This is guess work unless the word comes from an actual lens design=
er involved with the lens design. Much design done in the last 30 years, or=
even longer, has been computer assisted. This significantly reduces the ch=
ance for a bad design since it is possible to characterise lens performance=
in a very much more complete way than was done in the era before electroni=
c computers. A program like OSLO or Zemax can make thousands of ray traces =
in a fraction of second and analyse them where without the computer a singl=
e ray trace might take half an hour. Many classic lenses were designed with=
only a few ray traces and a lot of hand adjusting of actual models. Modern=
lenses can be made with much less guesswork and with a much better budgeti=
ng of tollerances so as to insure that production lenses will come up to ex=
pectations.=20
A note on Schneider quality. My impression is that previous to WW-2 Schn=
eider was a maker of mostly economy lenses of somewhat loose quality. Somet=
hing happened during or right after WW-2 which resulted in Schneider becomi=
ng a first quality manufacturer. The Xenars found in Rolleicords and MX Rol=
leiflexes are an example and certainly the innovative Xenotar design is one=
that would be expected of a quality manufacturer. Pre-war Zeiss seems to h=
ave had much better quality control than most other optical manufacturers, =
although I have encountered a couple of dogs. The material in the Zeiss Ind=
ex suggests that they kept experimenting with improved designs of existing =
lenses as well as new designs. Perhaps the splitting up of the company into=
two different companies had some effect, as one would expect. In any case=
, Schneider lenses of 1946 and later are not the same stuff as up to 1941.=
=20
--
Richard Knoppow
dickburk@xxxxxxxxxxxxx
Los Angeles, CA, USA
Other related posts:
