[rollei_list] Re: Contrast and Resolution
- From: "Richard Knoppow" <dickburk@xxxxxxxxxxxxx>
- To: <rollei_list@xxxxxxxxxxxxx>
- Date: Mon, 12 Nov 2007 12:50:11 -0800
----- Original Message -----
From: "Peter J Nebergall" <iusar4s@xxxxxxxx>
To: <rollei_list@xxxxxxxxxxxxx>
Sent: Sunday, November 11, 2007 8:52 PM
Subject: [rollei_list] Re: Contrast and Resolution
I started reading PP and Modern in 1964. I was ten. I
remember reading
of "resolution," and how distortions decreased as the lens
stopped down,
and I verified this through my old Praktiflex with manual
58mm Biotar. I
could see it improve.
What I didn't quite comprehend was "lines per millimeter."
I could get
the concept, and that a Summicron gave you more of them
than an old
Auto-Sekor or Xenon, but exactly how were they computed?
I know it's
like todays "DPI" for digitals... but I was taught the
Summicron and the
old Goerz Dagor were rulers of the roost. I still have
them, and using
them makes me wonder, every time I look at Mom's D2 work,
why it is
"unsharp." But how was a lens' resolution in "lines per
mm" computed?
Peter Nebergall
Resolution is mesured directly, not computed. To be
correct, the measurement is line pairs per millimeter or
inch for old measurements. A line pair is one black and one
white line. To make things more confusing scanned systems,
like television, are measured differently although
resolution is still stated in lines.
The usual method of measuring resolution is to use a
test chart. A number of these have been devised but the most
often used one is the United States Air Force chart. This is
composed of a number of patterns including a group of
graduated resolution patterns consisting of five black
separated by four white lines. In order to get the chart
image to be the right size it is photographed from a
distance of a certain multiple of focal lengths (forgot the
number at the moment). This distance is enough to get the
image into the field where a lens corrected for infinity
will not suffer from loss of correction from a close target.
A complete test requirs either several targets in an
array to cover the entire field of the lens or several
photographs with the target in different locaations in the
frame to accomplish the same thing. A series of tests should
be made at different f/stops.
The chart can also be used to provide a target for
examining the aerial image using a high quality microscope.
For this kind of test the lens, or entire camera where that
is more convenient, is mounted on a fixture of some sort,
ideally an optical bench. Target is viewed through the
microscope. The advantage of this method is that it
eliminates the variables that affect a photographic measure
where the entire camera and means of focus are actually
being tested along with the film characteristics. One may
want such a test but it is not an entirely valid one where
only the lens performance is to be compared.
A more modern test requires the use of an automatic
target and aerial image sensor. This set up can measure the
resolution at a given point on the image and automatically
plot it as an MTF chart. Since lens performance varies with
image angle measurements made at off-axis points of the
field are absolutely necessary to characterize a lens.
Because some aberrations are affected by both angle and
f/stop a complete series of tests must be made for both
different angles and different stops.
As a mentioned in another post MTF data can be plotted
in a couple of different ways. When complete plots are
supplied such charts can provide a good idea of the
performance of a lens, although they do not give any
information about some important characteristics such as
geometric distortion or illumination. Also, while resolution
is important the MTF charts do not, in general, separate the
effects of various aberrations on the image, they just
indicate whether resolution is good or poor. Chromatic
effects require a series of MTF measurements made with
monochromatic light.
Film resolution: MTF or resolution charts made by
photographing test targets are affected by the film
resolution and presense or absense of edge/border effects.
To know this it is necessary to test the film, which is a
non-trivial measurement. A suitable test target must be
provided and insurance that its image is sharp on the film.
Kodak measured resolution by means of a special camera,
which has been described in the literature. A lens with a
resolution of over 500 lp/mm was designed for this system.
Sometime around the early 1950's it was discovered that this
lens did not perform according to its design and another
lens was designed and made. All current products were
re-measured and the resolution of all Kodak film products
was changed accordingly. The effect was to increase the
resolution values in most cases. The camera, or fixture,
used was made of pretty much solid metal and very rigid. Its
been a very long time since I read the details of this
device and don't remember, for instance, how the film was
kept flat. I think they may have coated glass plates for
testing.
When a test is made photographically the resulting
images are a combination of lens and film characteristics.
While the rule of thumb is to calculate the average of the
two or to use the root-mean-square of the two resolution
numbers this is not really accurate. Two MTF curves must be
cmbined using a method called convolution where every point
on both plots is combined. Where the resolution of film and
lens are similar the resulting measure will be about half
the value of either. However, while the two can be combined
its difficult to separate them except approximately.
Both film and lens resolution measurements made using
the usual three or five line test chart suffer from an
effect similar to aliasing in scanned systems which produces
false resolution. This effect is present for both visual and
photographic measurements. It is sometimes difficult to
determine whether one is seeing actual or false resolution
tests made using more than one kind of chart will help
identify the effect. Where a chart shows a lens to have more
than the theoretical limit of resoution it should be
suspected that there is false resolution taking place.
There are applications where lens resolution is
extrememly important, for instance in the step and repeat
printers used for making semiconductors, however, for
general photographic purposes resolution is only one of
several characteristics which result in pleasing
performance. As an example, Mamiya lenses have a general
reputation for being very sharp. However, it may be that the
cameras they are used with have very good focus accuracy and
film flatness. Both may be true, but to make tests of
different lenses which are mounted in different cameras and
attribute the differences exclusively to the lenses is a
serious error.
---
Richard Knoppow
Los Angeles, CA, USA
dickburk@xxxxxxxxxxxxx
---
Rollei List
- Post to rollei_list@xxxxxxxxxxxxx
- Subscribe at rollei_list-request@xxxxxxxxxxxxx with 'subscribe'
in the subject field OR by logging into www.freelists.org
- Unsubscribe at rollei_list-request@xxxxxxxxxxxxx with
'unsubscribe' in the subject field OR by logging into www.freelists.org
- Online, searchable archives are available at
//www.freelists.org/archives/rollei_list
Other related posts:
