Converging fields of view

[P3D Neil Harrington <nharrington@xxxxxxxxxxxxx>]

Larry Berlin writes:

>Thanks Neil for confirming the stereo camera geometry. It makes sense that
>such a feature would be designed into a stereo camera.

My pleasure, Larry.  

>I also appreciate the fact that with a two camera system you can image
>stereo of distant objects with a wider base than a fixed camera provides.
>This factor makes the two camera system as somewhat preferrable over a fixed
>base system, unless most of your photo subjects are at relatively close
>range. Even in the normal range it's fun to play with exagerated 3D which is
>difficult to do if the lenses are at a fixed separation. 

I haven't yet used a two-camera system.  I can see it would have advantages
all right, as you say, for hyperstereo.  With practically any 35mm cameras
on the market now it wouldn't really give you any _choice_ but to accept a
certain amount of hyperstereo, since the camera width is bound to be greater
than the usual lens separation on a stereo camera.  But that probably isn't
a problem in most cases.  I've done some single-camera stereo, shifting the
camera sideways for the second shot, and this always seems to work out fine
at moderate distances, even with no attempt to be really precise about the
distance of the shift.

I've been thinking of making up a two-camera system using a pair of Ricoh
R1s.  The R1 appeals to me for this because it has a dual-focal-length lens,
30mm full frame and 24mm masked for "panorama" shots.  This would avoid the
problem of trying to synchronize two zoom lenses, which (especially on
point-and-shoots) would be more of a job than I'd want to undertake.  I'm
assuming, of course, that the lenses would be close enough not to require
any special matching.  Also, the R1 has the option of using an infrared
remote release, which seems to me a nice quick-and-dirty means of getting
something _reasonably_ close to shutter synchronization.  Any thoughts on this?

>It's necessary to
>understand the factors involved for either camera system using terminology
>that can apply to both circumstances. I guess the ideal would be to have
>both a stereo camera and a pair of regular cameras. 

I think so too.

>Or better yet, with
>digital one could have a multiple lens (any number of lenses), continuously
>variable-base stereo camera with very little actual bulk, no chemical film
>processing, no waiting. Has one been built yet? :-)

Yikes!  I'd be more than satisfied if someone would just start making a new
Realist-format camera with up-to-date features like autofocus and
autoexposure.    :-)




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