Olympus-OM
[Top] [All Lists]

[OM] Taming the wild lens test

Subject: [OM] Taming the wild lens test
From: "Peter A. Klein" <pklein@xxxxxxxxxx>
Date: Sat, 12 Dec 1998 02:08:45 -0800
Wouldn't it be nice if we had a calibration point to compare the various
lens tests out there?  It's frustrating--everybody puts out tons of numbers
that can't be related to each other.  Lots of people say lens tests are
useless, but everybody reads them.  And usually end up comparing apples to
oranges.  Some sources claim contrast and resolution are different aspects
of the same thing.  Others treat them as completely different qualities.

Is there some sort of rough rule of thumb one can use on various lens
tests?  For instance, what is the rough equivalence between Gary Reese's
"ABCDF" ratings and, say, the Modern Photography tests one can find off the
List's Web site?  Between the ModPhoto tests and Popular Photography's
"percent contrast of a .01 mm slit?"  Between the pre- and post-1982
ModPhoto tests, and their new "subjective quality factor?" I always found
it amusing that they could discern *tenths* of a subjective percent.  :-)

I believe I have some "feel" for the resolution numbers and how they
translate into pictures.  As for evaluating contrast tests, I haven't got a
clue besides "higher is better."

I'd be interested in other people's thoughts on this.  Here's what I do:

 * * * * *

When I look at the Modern Photography tests on Web sites mentioned in the
Olympus List page or its close brothers, I tend to use 50 lines/mm as a
cutoff point for "reasonably good."  Much lower than that, and I may want
to use that f-stop with caution and not enlarge much; higher than that, and
there's nothing to worry about *if* the lens can handle the lighting and I
don't shake the camera.  When texture or sharp lines are important, I aim
for the best f-stop I've got.  The difference between 60 and 72 lines/mm
may be largely theoretical, but like chicken soup, it couldn't hurt. (I'm
referring to normal and moderate wide-angle lenses here--I relax the
standards a little for teles.

If it's a choice between handheld 1/15 at f/2.8, 1/30 at f/2, or 1/60 at
f/1.4, I'm going to opt for the wider openings unless they're really
terrible, because it's much more likely that camera shake will ruin the
shot than a little lens softness.

Here's what I do when I get a new lens.  I tape a 2-page spread of the
stock market pages from my local newspaper to the wall, flat as I can.
Load the camera up with slow slide film.  I used Kodak Elite 100 for my
last round, probably should have used K-64.  Put the camera on a tripod,
and set it horizontally, so that the beginning of the print on the left
page shows at the left edge of the frame, and the end of the print on the
right page shows at the right edge of the frame.  Shoot the pages at each
f-stop, using the self-timer to trip the shutter. (I'd lock the mirror up
if I could, but I've got an OM-2).

When I develop the slides, I look at them using my trusty Radio Shack 30x
microscope.  (Sorry, I don't own a projector).  I'm just looking for a
general idea of how good the center and corner performance is.  My grading
system goes like this:

A: Stock quotes clearly and easily readable.  Characters have distinct edges.
B: A little fuzziness, but still completely readable.
C: Moderate fuzziness, some characters are not quite readable.
D: Almost unreadable, severe smearing.
F: What characters?

Of course, this takes contrast into account as well as resolution. It just
so happens that the lower-case letters in a stock market quote are about
1mm high, and you can see three distinct lines in them (such as in the
letter "e"). Reduce a 68 cm -wide double-page spread down to 3.6 cm on
film, a factor of 18.9, and these three lines translate to about 56
lines/millimeter on film.  

Now all this is very rough and crude, mind you.  It assumes I've focused
the lens perfectly, and that the lens performs the same at medium and far
distances as it does close up. Quality above the 56 lines/mm level is going
to show up only in *very* subtle changes in the crispness of the image. So
the test tells me how bad the lens is wide open with more assurance than it
tells me how good it is.  And since I don't mix up the slides, I may be
influencing the results by what I expect to see.

Still, if I can read the stock quote characters, the lens is probably going
to look OK at that aperture. 

My 50mm f/1.4 SC Zuiko tests out thusly:

        Center  Corner
1.4     C-              D
2       C+              D+
2.8     B-              C-
4       B               B-
5.6     A-              B-
8       A-              B-
11      A               B+
16      A               B+

Note that my results are very similar to Gary Reese's in his recent "Lens
Test Part 2."  And the progression of relative image quality is similar to
published lens tests.  So this test tells me that my lens should to perform
about as well as I'd expect a 50/1.4 to perform.  No surprises.  Now I can
forget about the stupid tests, and go take some pictures!

Now, if only it would stop **raining** here!  It's Seattle in the winter,
and we just had the wettest November on record (sigh).
----
                                  :    -----==3==      ---      ---
       - Peter -                  :   |    |  |  |    |   |    |   |
                                  :  @|   @| @| @|   @|  @|   @|  @|

< This message was delivered via the Olympus Mailing List >
< For questions, mailto:owner-olympus@xxxxxxxxxxxxxxx >
< Web Page: http://Zuiko.sls.bc.ca/swright/olympuslist.html >


<Prev in Thread] Current Thread [Next in Thread>
Sponsored by Tako
Impressum | Datenschutz