T3D Re: Panum's limit

[Larry Berlin <lberlin@xxxxxxxxx>]

>Date: Mon, 6 Apr 1998
>From: Peter Homer writes:
>.....................
>I have a photocopy from a book on pschology refering to the "Panum phenomena"
>I an not sure whether this is the same as John Bercovitz was refering to as
>there are some differences ." On a homogeneous background Panum (1858)
>presented  a vertical line to one eye and two vertical lines to the other.
>He found that if one of the double vertical lines is fused with the single
>line in binocular presentation the subject percieves the unmatched line to
>be further away than the fused line.(As exploited in stereo mounting grids
>such as the one manafactured by Reel 3D to set the near and far points in a
>slide?).This depth effect was found to occur only when the distances
>between the double lines was very small(aproximately 6' of arc) ". Beyond
>that is refered to as Panums limiting case.
> Julez (1968) had subjects look at one line with the right eye and another
>with the left ."The two lines were seen as a single fused line when they
>were registered within 6' of arc. If any greater two lines were seen. But
>Julez found that once fusion occured the amount of area seperating the two
>lines could be increased markedly without loss of fusion .In fact distances
>of over 2 degrees (120' of arc) could be reached before fusion was lost.
>Reducing the distance between the lines after loss of fusion did not
>produce a single image untill Panums limits were reached again".
>  There is a big difference between 10 or 20 and 6 minutes for Panums limit
>but the work by Julez implies that 2 degrees would be possible provided
>fusion was achieved at 6' first.      P.J.Homer
>

Well, this gives something to think about. I've long observed what may be
this phenomena while working on digital images. My understanding of the two
line scenario is a bit different though. When you fuse the two vertical
lines seen by one eye, with a single vertical line seen in the other eye,
you are actually fusing both lines at one time with the same single line.
Since the two lines are separate and separated by a small distance, OF
COURSE they will appear as lines at different depths. It is this parallax
factor that provides ALL depth cues in stereo imaging. (The supposedly
*unmatched line* is not really unmatched. Calling it *unmatched* is a
misunderstanding of the most fundamental aspects of this situation. As to
whether it is further away or closer would depend on which one is assumed to
be *unmatched* and whether the binocular method is parallel or crossed viewing.)

I understand the limit has to do with a separation distance beyond which
getting both of them to fuse at the same time becomes very difficult. The
mind will have trouble associating both of the two lines with the single
line at one moment, so will pick one or the other. When this happens, the
one left out becomes a floating object of nebulous depth since it's
seemingly by itself and only seen with one eye. This is *somewhat
equivalent* to having too great a depth in a stereo scene so that foreground
and background are so dissimilar as to need a very different postioning of
the eyes to obtain fusion of each part.

Any person familiar enough with freeviewing would be able to fuse the two
lines in the  Julez experiment at the greater separation or close to it from
the start and not require starting at the 6' point before increasing the
separation.

I routinely identify and correct repetitive elements in images when they
induce this kind of situation by accident. Or in working with star fields in
3D, one can make use of the effect to create more stars without having to
place it's matching star in the other side. Or, if such a circumstance
exists and is in a confusing borderline situation, moving one of them
vertically out of alignment fixes the problem, though you have to provide
it's eqivalent in both sides.

Larry Berlin

Email: lberlin@xxxxxxxxx
http://www.sonic.net/~lberlin/
http://3dzine.simplenet.com/


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