Re: [photo-3d] Re: Newbie question-near point distance

["Dr. George A. Themelis" <drt-3d@xxxxxxx>]

Mike Galazin wrote:

>George, you know perfectly well that the issue is
>whether or not it is *advantageous* to change the 
>stereo base when changing FL.  

It depends :)  It would be nice to have control over
the stereo base and the focal length independently,
but it would be a problem if the two were linked in
any predetermined way, using for example the formula
that Mike Davis gave earlier. 

The RBT X* family of stereo cameras come with either
65mm or 75mm stereo base and linked 35-70mm zoom
lenses.  There is a great variety of pictures that 
can be taken despite the limitation of the fixed
stereo base.  Problems arise when the near object
is too close to the camera (where the stereo base
should be reduced) or when the near object is too
far from the camera (when the stereo base should
be increased).

Question:  Would it be an advantage to link the 
stereo base to the focal length *and* the distances 
of the near and far away objects, using for example 
the formula for constant on-film deviation?

This might be an appealing idea for some, but I
still see it as a limitation.  Consider the case
of a portrait.  Quite a few photographers had
good success using twin cameras bottom-to-bottom
(this reduces the stereo base to about 3 inches)
and medium telephoto lenses (in the 80mm to 135mm
range).  Consider a tight portrait where the near
to the camera object is the nose of the model and
the far away object is the back of the head (or
ears).  By having the model turn the head, facing
into the camera or away from the camera, the depth
range can change quite a bit.  The constant on-film
deviation formula would then predict a different
stereo base for each turning of the head.  Since
the photographer is using the same lenses and is
positioned at a fixed distance from the model, it
makes no sense to me to change the stereo base
based on small rotations of the model's head.

So, let me repeat:  It is an advantage to have
independent control of the focal length and the
stereo base.  But it is not a good idea to link
the two in any predetermined way.  Let the
photographer decide/experiment, depending on the
situation.  Here are a few guidelines for starters:

- Portraits: stereo base 3" (twin cameras bottom-to
  bottom), lenses 80-135mm.
- Action shots: stereo base around 6" (twin cameras
  side-to-side), lenses around 135mm
- Animal close-ups photographed from far away using
  135mm and longer lenses and without infinity in
  the picture: Use McKay's principle (increase stereo 
  base proportionally to the increase in the FL)
- Landscapes which include infinity:  For normal lenses
  start from the 1/30 rule of thumb (stereo base = 1/30
  of distance of nearest objects).  If you use wide
  lenses then double this.  If you use long lenses then
  cut this in half (David Lee's approach, which is a 
  simplification of Mike Davis' "accurate" formula)
  
My only argument with Mike Davis is that he
used the word "accurate" as an answer to a question
regarding the stereo base.  In mathematics there
are accurate answers but not in photography (IMO).

George Themelis

 

Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/