Re: [photo-3d] Re: again those figures!!!

[Owen Pearn <owenp@xxxxxxxxxxxxxx>]

for the purposes of talking about the limits to fusion, our eyes do 2
things:

1) they accommodate (focus) so we see things clearly, and
2) they converge so we see things singly (ie. both eyes see the same
thing)

when we look at the real world, accomodation and convergence happen at
the same distance.

when we look at a coffee cup on the desk, our eyes accomodate at 18
inches (46 cm) to see the cup clearly, and they converge at 18 inches
(46 cm) so that both eyes are seeing the coffee cup. when we look at the
distant mountains, our eyes accomodate at 5 miles (8 km) to see the
mountains clearly, and they converge at 5 miles (8 km) so that both eyes
are seeing the mountains.

when we look at a stereoscopic picture, accomodation and convergence
happen at different distances.

when we cross-eye freeview a stereoscopic picture held at arm's length,
our eyes accomodate at 18 inches (46 cm) to see the picture clearly.
this accomodation distance will not change for as long as we look at the
picture. it will always be 18 inches (46 cm). as we "look around" the
picture, our eyes will converge at different distances as we look at the
objects at various depths in the scene. in only one case will the
convergence distance equal the accomodation distance. this is when we
are looking at an object whose depth is at the plane of the picture -
"the plane of zero parallax". the rest of the time, the convergence
distance will not be the same as the accomodation distance.

this disassociation of the accomodation and convergence ocular reflex is
a cause of stereoscopic viewing discomfort. there is an abundance of
human factors literature which quantifies this. numbers vary, and people
vary, but a maximum of "1.5 degrees" is common.

this number means "the difference between the accomodation angle and the
convergence angle must not exceed 1.5 degrees anywhere in the picture".
to work this out, you have to know the final viewing conditions.

one reference is:

- Y. Yeh and L. Silverstein (1990), Limits of Fusion and Depth Judgement
in Stereoscopic Colour Displays, Human Factors, 32(1): 45-60
- http://www.ergoweb.com/Pub/Info/Ref/HumFac/h90v32n1.html
- "maximum 1.57 degrees uncrossed disparity and 4.9 degrees crossed
disparity"

terms to websearch with are:

 accomodation convergence fusion stereoscopic

-o-

Oleg Vorobyoff wrote:
> 
> Thank you Sergio Baldissara and Abram Klooswyk for your detailed
> explanations.  I understand now that 2° convergence is a recommended (but
> not absolute) maximum for viewing stereo pairs, especially under projection.
> I still do not understand why greater convergence is difficult to view,
> considering that we routinely look at things with 12° convergence and more,
> for example, while reading.  My guess is that since a normal photograph is
> actually a rectilinearized projection it does not recreate precisely enough
> the curvature of objects viewed with a lot of parallax.  If that is true,
> perhaps the center portion of a fisheye image, enlarged to fill the frame,
> could be comfortably viewed despite greater than recommended convergence.
> Has anyone actually tried that?
> 
> Oleg Vorobyoff
> 
> My original question:
> >...Is there some definite physical or physiological
> >basis for the 2° maximum, or is that just a guideline?
> >After all, most of us can easily fuse the printing in a
> >book a foot away - that is a full 12° of crosseyedness.