Re: Optics question (not alt)

From: Richard Knoppow ^lt;>
Date: 03/10/05-03:50:30 PM Z
Message-id: <030901c525bb$42db2f60$17f55142@VALUED20606295>

----- Original Message -----
From: "Joe Smigiel" <>
To: <>
Sent: Thursday, March 10, 2005 12:39 PM
Subject: Re: Optics question (not alt)

> Is it the circle of illumination or the circle of
> definition (or a
> different circle) that increases when you stop the lens
> down to the
> prime aperture?
> Looking at some lenses I have which barely cover certain
> formats, I can
> see mechanical vignetting at the widest apertures occuring
> as I peer
> through the back of the camera through the cut groundglass
> corners
> towards the lens. The aperture has a truncated elliptical
> shape (much
> like the profile of an American football if appear cut in
> half
> vertically) at this point. I assume this causes a
> significant decrease
> in illumination at the corners of the image, but since I
> can still see
> the central portion of the lens, the circle of
> illumination must still
> be there from that central portion.
> Relatively, the corners are receiving much less light
> since the central
> portion of the image is accumulating exposure from all
> portions of the
> lens while the corners only get it from the center of the
> lens. As the
> aperture is reduced, this imbalance disappears until the
> film is getting
> exposure only from the central part of the lens at a
> smaller aperture
> (like f/22 as quoted for comparison of many lenses by the
> manufacturer).
> At this point the entire aperture is now visible from the
> back of the
> extended camera without vignetting, and the aperture is
> once again an
> approximate circle.
> As the camera bellows extension is increased further away
> from the
> infinity position, the film receives a greater proportion
> of light from
> the center of the lens while the formerly vignetted areas
> now lay
> outside the diagonal of the film. Isn't that what is
> really going on
> with the circle of illumination, infinity focus, and so on
> ?
> So really, isn't only part of the circle of illumination
> getting
> brighter and the entire field getting more even
> illumination as the lens
> is stopped down rather than the outer margins of the
> circular field
> actually increasing?
> Joe
   The "optimum" stop is something of a misnomer. Optimum
for the center is not necessarily optimum for the edges. The
reason is that several aberrations vary with the stop. The
most important ones are spherical aberration, coma, and
oblique spherical. Spherical affects the entire image, coma
and oblique spherical are proportional to image angle. So,
at some point the improvement at the center of the image due
to reduction of spherical aberration will be balanced by the
increase in diffusion from diffraction at the stop. This is
optimum for the center. At the margins the same things
happens for coma and oblique spherical so the optimum stop
varies with image angle. For a Dagor, for instance, optimum
for the center to coverage of around 60 degrees is f/22. For
the maximum coverage of 87 degrees is it f/45. Some very
well corrected lenses are best nearly wide open because
diffraction is least there.
   Vignetting is from something else. In a lens designed to
reproduce a flat surface onto a flat surface there are there
are a couple of contributors, one is the inverse square law
from the lens to the film. Another is the obliquity of the
stop which is partially obscured as one moves away from the
center. Note that the stop becomes cat's eye shaped
(eliptical) at the margins.
   By designing the lens to have a smaller exit angle, i.e.,
using a retrofocus design, the inverse square law fall off
can be reduced. By introducing coma into the stop the stop
magnification can be made to vary with the image angle. This
is the principle of the Roosinov lens used for most modern
wide angle lenses. This is also known as a tilting entrance
pupil. If you look into such a lens and move it from side to
side the image of the iris will appear to follow you. By
using special designs its possible to completely eliminate
the usual cos^theta fall off and even to make the edges
brighter than the center.
Richard Knoppow
Los Angeles, CA, USA
Received on Thu Mar 10 15:51:23 2005

This archive was generated by hypermail 2.1.8 : 04/08/05-09:31:00 AM Z CST