P3D Re: Matched lenses in stereo

[abram klooswyk <abram.klooswyk@xxxxxx>]

George Themelis has quoted part of a discussion in
New.3D.Film.Gear@xxxxxxxxxxxxx (which list I don't
read) regarding
>the tolerances in matching lens focal length (or image
>size) of R and L lenses.
(Photo-3D Digest 3537, 8 Oct 1999)

Greg Erker added (Photo-3D Digest 3538, 8 Oct 1999):
>There are two issues here (IMO) one is
>the size missmatch your eyes can handle
>and the second is problems from the slide
>mount window cutting off the two images
>unequally.
(...)
>So I'd say 1% is way too much due to
>the unequal image showing through the mount.
To which George agreed.

I should like to comment on the "focal length (or
image size)" and on the tolerances of the eyes,
and quote an interesting experiment from Wheatstone.

1. It probably has been said before and maybe also
in the discussion quoted by George, but it might be
interesting for others to point out (which George
and Greg of course know) that image size does not
depend on camera lens focal length.
It only depends on the distance from the lens to
the film. (Of course, you are not free to choose any
lens to film distance when you want sharp images...)

When e.g. the left focal length is 50 mm and the
right 50.5 mm (1% more), image size is equal when
both are set at 50.5 mm from the film plane.
(For simplicity ignoring optical finesses like
principal planes).
But, the right lens is focussed for infinity and
the left for about 5 meter, in this example. This
is no problem when the lenses are stopped down
to get a large depth of field.
(The "Wray Stereographic" exploited this scheme)

Greg is right of course in pointing out the mounting
difficulties of unequal image size, which seem to
contradict the large tolerance of the eyes.
The contradiction comes from mounting unequal image
sizes in exactly equal mount openings.
When you would mount 2% larger left images consequently
in 2% larger left mounts, only the projectionist would
complain. (Skip this unpractical nonsense :-))

2. The tolerance of the eyes for unequal images is
often pointed out by quoting the different diopters
which some people have in their glasses.
There is however a more simple experiment anyone can do
now, and Wheatstone was the first to do it, as is
clear from his 1838 reading.

To demonstrate that both eyes see a slightly different
image, newbies often are told to take a pencil, hold
it vertical at some distance for the eyes, and look at
some distant object - the pencil doubles.

Now repeat this, and without moving your head, move the
pencil and the viewing direction some 45 degrees to one
side. Still looking in the distance, the doubled pencil
will show considerable size difference, up to 10% or more,
depending on the various distances and angles involved
in the experiment (in using a ruler instead of a pencil
some measurements can be estimated).
But in looking at the pencil (or the ruler) itself,
you have no problem to see it single.

Wheatstone described a similar experiment with a coin
and looking at a closer point. This experiment was one
of his arguments to show that images falling on
non-corresponding retinal points could be fused, which
Brewster attacked so vigorously.

------
(To be explicit: I also agree that matching lenses for
stereo to near zero % should be the goal.)

Abram Klooswyk