P3D Re: Alignment of seperate 35mm cameras...

["Dr. George A. Themelis" <DrT-3d@xxxxxxxxxxxxxxxxxxxxxxxxxxx>]

Stuart, let me clarify the situation.  There are 3 ways to align the
cameras:  Diverging axes, converging axes, parallel axes.

- Diverging:  Have never heard anyone doing this way.  I addition
  to potential distortions, you also lose part of the image.

- Converging:  There is a keystone distortion which can be a problem
  or can be unnoticed (depending on the degree).  The advantage of this
  technique is that it sets the stereo window at the point of convergence 
  of the axes so it is useful when there is no other way to set the stereo 
  window or if the entire frame must be used without any cropping, like in 
  the case of video, etc.  

- Parallel axes:  This is the "politically correct" method, recommended
  if you want no keystone distortion.  The problem is that the stereo
  window is set at infinity.  This is a problem... It can be corrected
  by masking part of the image (as you have found).  It is easily done
  in slides or prints mounted in a fixed mount (Q-V type).  Prints without
  mount need to be trimmed (as you have done).  More masking/trimming is
  required in the case of long focal length lenses and infinity or 
  close-up/macro photography.  

  [Interesting note:  At the extreme (v. long FL lenses or stereo
  microscopy at high magnifications) you cannot use parallel axes but
  you must converge the camera(s).  The interesting thing is that after
  some point this does not matter any more because you are entering
  the region where linear perspective is almost non-existent, hence no
  keystone distortion (this is the case of Scanning Electron Microscopy,
  etc.)]

Do you understand the concept of stereo window?  Somehow print users,
mounting stereo prints without a fixed mask, do not understand this
concept easily.  Proof for this is the decades of commercial stereo print 
photography in the 19th and early 20th century where the lack of stereo
window is noticeable and annoying (for me)...  My guess is that it
is much easier to experiment shifting the pairs over a fixed mask as
slide film users have been doing and also the concept of the stereo 
window is more pronounced in stereo projection than viewing with a viewer.

There are many good books out there, explaining the concept.  I liked
the "3L" rule that says:  "The Left eye sees Less in the Left side"
When I mount slides, I follow a variation of this:  I am trying to
make the right eye see more in the right side.  Does this make sense?
If the window is set far back then objects in front and intercepting
the window appear to have "floating edges".  These edges must be masked
or cropped out for the window to be repositioned.  When mounting with
a fixed mount this means that the two pairs (prints or film chips)
must be placed further apart from each other.  When cropping prints
that means that the two outer edges must be cropped.

Let's see what you are saying:

>       A while back, there was some discussion
>    on the proper alignment of two single 35mm cameras
>    to yield the best results.  Until I heard this discussion
>    I had simply made sure that "infinity" did not come 
>    together (let's say for the sake of example that the
>    exact center of the focusing screens did NOT come
>    together anywhere in the image, and did NOT cross
>    either -But that there was always some attempt at a
>    "space" between them.)

Hmmm... You are using a strange terminology for me.  "Infinity did not
come together".  What does that mean?  I assume that you were trying
to converge the lenses of the cameras instead of having them parallel.
    
>       But I heard some people say to bring the 
>    centers together. ....

Converge the axes of the lenses? 

>    I tried this on several rolls and
>    the result I got was that it appeared to strain the 
>    eyes somewhat when I put together the prints.  

I wonder why?  There is no strain when the axes are parallel.

>    In addition,  (even though all I did was put the object
>    of interest that was the furthest away in the same
>    place of both lenses) I noticed that some of the right
>    -hand camera had as much, or more of the left-hand
>    edge image information.  

That must be the "floating edges" I mentioned earlier, which means
that the stereo window is set far back, possibly at infinity.

>    Hmmmmmm. Well, I then
>    simply "cropped" that "extra" bit of picture info on
>    the right-handed image (left edge) to make sure 
>    that the right image had LESS information on the
>    left-edge of it then the left camera phot had on its
>    left side.  

Actually, as I said, the right eye (or image) must see MORE in the
left side, in order for the stereo window to be set properly.  You
appear to be saying the opposite.  You are confusing me again! :-)

>    Presto!  Everything seemed closer to "normal" then.

That's an interesting way of putting it. 

>       The conclusion I have now come to is 
>    that it appears to be best to make sure that if
>    using two seperate cameras for instantaneous
>    shots, that to make certain that the exact center 
>    of each lens does not meet the other in the taking
>    of the photograph.  -Basically, simply put a little
>    seperation through-out infinity.  Has anyone else
>    encountered this and found this to be correct?

Still puzzled by your empirical terminology here.  It is much easier
for me to understand parallel or converging camera (lens) axes.
    
>       By the way, I shoot print films (25 asa)
>    and develop 3-1/2 x 5 prints. Then I crop them
>    to 3" x 4" tall each mounted on 4" x 7" cards.
>    Then I use them in any of the older viewers.

This concept is universal.  Applies for prints or slides equally.
But the ways of cropping the image might be different in each case.

>                       Stuart Johnson
>                       Sacramento Ca.

You asked about a stereo clubs in the Sacramento Ca. area.  Stereo clubs
are a precious commodity.  Very few in the entire USA (maybe a dozen?)
Check Alex Klein's page for more info on existing clubs:

http://www.stereoscopy.com/clubs/

Best regards,

George Themelis
Cleveland OH, home of the Ohio Stereo Photographic Society - One year one
              month and one week old :-)


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

End of PHOTO-3D Digest 2767
***************************