T3D Re: Phanto-cam Lens system

[Larry Berlin <lberlin@xxxxxxxxx>]

>Date: Mon, 26 Apr 1999
>From: Brian Reynolds <reynolds@xxxxxxxxx>
>................
>Larry Berlin wrote:
>> Some of you may have read about the phanto camera on the SD-3D
>> list. .............
>
>Gee, I wish this were taking place on the tech list.

*****  :-)  Well, now this part of it is on this list!



>....................
>Pinholes will have problems because off axis the hole is no longer
>round.  This will cause significant light drop off (more than just
>that due to the pinhole to film corner distance being longer than the
>pinhole to film center distance).  You might want to check out the
>pinhole mailing list for more information.  The archives can be found
>at <URL:http://www.pinhole.com/discussion>.

*****  Thanks, that might prove interesting. In theory, a pinhole is merely
a very tiny hole and it serves as the focal point for all rays passing
through it. Ignoring some aspects of reality to follow the theory a bit, a
pinhole could literally be turned at 45 degrees to the axis (to favor the
off axis view) and it wouldn't affect the projection factors at all. All
rays still have to cross each other thru the pinhole. That would keep the
pinhole effectively round for capturing a phanto-view. It would be less
effective with on-axis viewing.

This changes when you substitute a lens for the pinhole. Tilting the lens
would NOT be the same as tilting a pinhole.


>......................
>Short focal length view camera lenses tend to have a large angle of
>coverage.  For example, a 90mm lens for a 4x5 camera would normally
>have at least a 100 degree angle of coverage (i.e., 50 degrees from
>the lens axis to the edge of the image circle).  The Schneider
>Super-Angulon XL series has an angle of coverage of 120 degrees.
>
>What size film will be used?  A 35mm or 6x6 frame could be place
>pretty far from the axis.  A 4x5 sheet wouldn't be able to moved that
>far off axis and still have the entire sheet remain within the image
>circle.

*****  This is good information. Since this process ignores the normal axis
view it would use that 50 or 60 degrees of off axis view, and not even all
of that. Maybe the bottom edge of the desired view starting about 5 to 10
degrees off axis and the top of the view being as far as possible to the
extreme.

I have no idea what size film might be used, but the obvious choices include
standard film sizes in sheet or roll. It really depends on what angle of
view you want in the results. If your viewing angle is mostly from above you
will need less film than if your angle is low and the projected view for
each eye may be extremely elongated. I think Bruce mentioned an historical
example where someone calculated the view of a horse that would extend
across a football field... It depends on where you put the vantage points.

The question resolves to, how to design a lens system (potentially including
a specialized aperture) that is optimized for the side view instead of the
axial view. That means in effect a lens with a view approaching 180 degrees
as close as practical, except that we can ignore half of that field and
ignore the on-axis view. The side 90 degree view would center on the 45
degree angle. Some phantograms could be done at this angle, but others might
center their view elsewhere.

Is such a lens merely a *wide angle* lens, or is there another way of
increasing the field of view without the familiar distortions of a wide
angle lens?

Larry Berlin

Email: lberlin@xxxxxxxxx
http://3dzine.simplenet.com/


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