Re: parallel lens axes

[P3D Gregory J. Wageman <gjw@xxxxxxxxxx>]

John Bercovitz writes:

>You guys are saying the same thing, I think.  The axes of the 
>lenses themselves are parallel or you would indeed get keystoning.  
>However, lines drawn through the centers of the lenses and through
>the centers of their corresponding images are not parallel.

Yes I think you're right that we were saying the same thing, but I think
you expressed it more clearly.

Perhaps it's picking nits, but I think it's a bit confusing to use the
words "toe-in" or any such thing, because that's neither the intention nor
the result of the arrangement.

If I understand it right, the offset lenses simply cause the left image to
have more information from the right side of the scene, and the right image
from the left side of the scene.  This binocular disparity of the left and
right sides of the window border causes the window to have an apparent
location in the scene closer than infinity.  And, consequently, as someone
(sorry, forgot whom) stated earlier, the top and bottom of the window do
not have any inherent depth information.  The brain must "put" the top
and bottom of the window in the same place as the sides.  Hmm, I wonder
what would be the effect of having a *different* apparent depth for the
left and right sides of the window?  Another mind-bender, I'd assume. :-)

It strikes me that this would be much, much easier to demonstrate in
real-time, if one could simply dynamically mask a pair of stereo images
taken with the lenses centered on the aperture.  It should be possible
to move the window dynamically to any point in the scene, simply by
sliding a pair of masks (narrower than the images, obviously) inward
and outward over the images.  Since one isn't moving the film chips,
there should be no problem with convergence changes.  For that matter,
one should be able to do something like this in a computer, too.  This
could make a nice canned demonstration.

	-Greg


------------------------------