Re: Dan / Would Perfect Lens have Bokeh?

[Paul Schliesser <paulsc@eos.net>]

>There is a difference, Bokeh or whatever, and I have difficulty in
>understanding that it cannot be defined and/or measured. Once you know what
>it is, the optical designer should be able to engineer it back into future
>designs.

Gerard,

It's tempting to make comparisons with audio equipment design. Especially 
during the "monster amp" days of the 1970s, a lot of companies made 
equipment whose main design goal was to test well. The component with the 
lowest distortion figures isn't necessarily the one that reproduces music 
the best, but if you make your decision based only on the test report 
charts, this is the one you will buy. It's been a long time since I have 
paid attention to the audiophile world, but if I remember, you can't hear 
some kinds of distortion when levels are below about 2 percent. 
Manufacturers were fighting over levels in the 0.005 range. Very low 
distorition figures became a fetish with people who were into the tech 
end of the hobby.

One way to dramatically reduce distortion is to use large ammounts of 
negative feedback. This uses a circuit which feeds an amplifier's own 
distortion components back into the amplifier with reversed polarity. 
This negative signal cancels out and conceals most of the amp's 
distortion artifacts. This can seriously degrade performance, but boy, 
will it look good in certain tests--which is more important than 
reproducing music well.

In Dennis Laney's _Leica Lens Practice_ book, he describes the methods 
that Leica uses to test their own lenses. He claims in his book that the 
tests done by most magazines don't do the  battery of tests necessary to 
get the whole picture (pun intended), because they are expensive, time 
consuming and difficult. Although I don't know much about optical design, 
I found the way that the graphs show the _balance_ of lens 
characteristics to be very interesting. The emphasis was on balance, not 
an obsession with doing well in a few, essy-to-measure areas. There are a 
lot of tradeoffs that can be made in different areas.

It would, of course, be possible to design the lens so that it excelled 
in magazine lens tests, but at the expense of more subtle characteristics 
which are difficult to measure.

If you are taking a test in school, and you know the questions in 
advance, and you memorize the answers, the results of the test don't 
accurately reflect what you know.

- - Paul