Lenticular Thoughts

[P3D <LeRoyDDD@xxxxxxx>]

     There's been some bandwidth on lenticular prints recently...

     The major problem I see with lenticular prints is the reduced depth;
both as reduced percentage image parallax on the negatives and reduced
absolute image parallax in the typically viewed print.

     The reduced percentage image parallax problem:
     With the three and four lens 35mm lenticular variants, the separation
between lenses is about 18mm, 3/4in. Nominal interocular of 65mm is 3.6 times
that.
     So objects in the lenticular print scene at 50 feet, which still have
reasonable, pleasant 3-D effect with "normal" stereophotographic systems,
will have the relief of objects at 180 feet in the "normal" system. That
distance is at or near "stereo infinity." Don't take lenticular prints of
landscapes!

     The reduced absolute image parallax problem:
     This has to do with the size of the viewed print.
     The 3 1/2 x 5 inch lenticular print is only about 2 inches "deep."
     For background... in the Realist format systems, if one projects  with
infinity points separated by two to two and a half inches, then moves the
screen until the frame edges are superimposed(may take a larger screen than
most of us have), the result is(mostly) what the camera saw... and if you
view it from the camera natural window distance you've got ORTHO! At least if
*your* interocular is about the same as the lens spacing of the camera.
      And the view is as "deep" as it can get.
     In the above setup, if you move the screen away from you(and it's big
enough to keep the frame edges on the screen and refocusing of course) the
only result will be the stereo "window" will stay in place at the same
distance... but seem to "come off the screen."
     If you move the screen closer, the window will stay fixed in space, but
seem to be "sunk into the screen." As you bring the screen closer, the actual
size of the images will be smaller and the edges will separate.
     With typical home size screens of about 50 to 60 inches, we commonly
just superimpose the mask edges on the screen and accept an infinity
separation of one to two inches. With not much for comparison, the 3-D(kinda,
almost, maybe Ortho) illusion of reality still holds up fairly well since
convergence(as opposed to image parallax) is one of the weaker cues to 3-D.
      If you keep moving the screen closer and keep the edges coincident, by
the time the image is reduced to a width of three and a half inches with
superimposed edges, the illusion is broken. Infinity points are only
separated by a fraction of an inch, and apart from the fact you can't get
close enough for an ortho image size, you can see it's not very "deep." About
2 inches, just like the lenticular print.
     You can get more depth with a larger size lenticular enlargement, but
even these sacrifice displayed depth by mounting to the "window."

     Inherent loss of resolution is the other significant problem.

     Each print has only a given area on which to put the image(s). A
lenticular print using two images has to throw away half of each image to put
the stereo pair behind the lenticular screen. A four image print uses only a
quarter of the information from each image.
     If you decide to take lenticular prints, use the three lens system if
possible or at least have your four lens pictures printed only using three
negatives. This advice is probably less important in smaller print sizes
where you don't view close enough to see a lot of difference, but would
become more important the larger the print. 
     When the prints are made, the operator of the print machine typically
converges what they consider the primary subject at the plane of print
surface. This results in all the information from that subject plane in all
the images being used.
     On inspection, one can see that the objects behind(and in front) of that
point are much "softer." The best 3-D in lenticular prints have objects in
front of the main subject so that the "forward" depth in the system is used,
but those objects will appear softer even if in focus.

     These disadvantages can be overcome somewhat, but the lenticular print
in smaller sizes will never be more than what they are. Like color separated
anaglyphs, they are uniquely themselves, not a stereo transparency or print
viewed in somewhat ortho conditions.


LeRoy Barco
LeRoyDDD@xxxxxxx


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

End of PHOTO-3D Digest 1335
***************************
***************************
 Trouble? Send e-mail to 
wier@xxxxxxxxxxxxxxx 
***************************