P3D Re: The Imagination Effect

[roberts@xxxxxxxxxxxxxxxxx (John W Roberts)]

>Date: Sun, 1 Nov 1998 22:51:32 -0700
>From: wier@xxxxxxxxxxxxxxxxx (Bob Wier)
>Subject: P3D Re: The Imagination Effect

>>Sounds like the emperor's new photography to me.
>>
>>David Lee

>Yeah - in my case, as I indicated, I wouldn't have believed it if I hadn't
>experienced it. To add a bit more detail - just as there is a different sense 
>of reality (for most people) between waking reality and dreaming, there is a 
>textural differance between "seeing" 3D and analyzing clues in 2D images to 
>understand 3D spatial relationships.

There are many different perceptual mechanisms for depth perception, some
based on logical analysis and some based on automatic systems. "Seeing" 3D
takes place in the brain - it sounds from the context that you're referring
to one or more of the automatic depth cues working to produce a consistent
perception of a 3D world.

I agree with that. But I think it's still important to protect the terminology
to some extent. "3D" has already been taken to allow inclusion as a reference
for depth cues other than stereopsis. It appears from some of the previous
posts that there's now a tendency to do the same to the term "stereo".
Since that would make "stereo" just a synonym for "3D", and since it's useful
to maintain a distince term for the effect, I think it would be a good idea
to reserve "stereo" (for imagery) just for applications that involve the
two eyes (or the two inputs to the visual centers of the brain) receiving
different images.

>I've experienced another dimension (slight pun) in stereo vision where there 
>is actually a hyper effect - I saw this most strongly in driving down the
>Interstate south of Salt Lake City. The highway parallels the Wasatch 
>mountain range for a considerable number of miles. As there aren't a lot of 
>trees, you get a good view of the mountains off to the East. Moving south 
>at about 60 mph, I could look (briefly!) over at the mountains with both 
>eyes open, and get a sense of the rounded character of the shapes which is 
>absent with regular stereo vision because of the short baseline. I'm assuming 
>again it's something like Prufritch, where there is stereo processing going 
>on in the brain in discrete quanta, so that one image is processed after 
>another, but since these are separated by more than normal spacing (because 
>of the vehicular motion) it results in a hyper effect. 

When people look at a moderately distant scene, it's fairly common for them
to move their heads from side to side, as though to obtain a longer baseline
for the judgement of depth. So it would seem that humans have the processing
capability to handle this sort of depth information. But I suspect that
what's at work here is primarily a depth perception mechanism involving
*motion detection* - the motion detectors in the visual system pick up the
differential rate of motion of nearby and distant objects, and place them
in distance by the relative rate of motion. (As evidence, if you move your
head side to side to better judge the distance of a distant object, the
efficiency of the distance perception appears to be drastically reduced if
you shut your eyes between the extreme left and right views. If it were just
a matter of collecting two images, separated in time, and comparing them,
closing the eyes between the two views should not cause such a large
degradation of distant depth perception.)

Judging from the descriptions on P3D in the past, the Pulfrich Effect is *not*
based on perception of motion by the visual centers, but by the left and
right inputs to the visual centers receiving *different images* at any given
time (the difference in the images being due to the lateral motion of the
scene and the difference in processing delays due to viewing through lenses
of different darkness).

Which brings up and interesting point. When Pulfrich 3D is viewed, there is
usually a *combination* of the claimed time difference and motion detection,
providing two cues to reinforce the perception of depth. Has there been
experimentation on ways to isolate these two effects? Possible examples:
 - viewing of a scene illuminated by a flashing strobe light, to "freeze"
   motion, with the rate of motion and the rate of flashing adjustable to test
   various combinations. 
 - two slides flashed a fraction of a second apart, one a left view of a 
   stereo image, the other a right view of the same stereo image, and with 
   the subject wearing Pulfrich glasses. If the interval between the flashed
   views is close to the nominal Pulfrich delay for the particular viewing
   conditions, the brain should pervieve *three* flashed views, the first and
   last with just one eye, but the middle one a stereo image. (Not sure how
   to deal with persistence of vision - maybe by flashing a dark image on
   an already-light screen.)

>In fact, the whole 
>upper part of the tree seems to rotate a bit as I drive past, which I assume 
>is a perceptial conflict between the nearer branches "moving past" more 
>rapidly than those on the other side of the tree. 

I actually wrote a short story on that phenomenon when I was in high school!
:-)

John R


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