P3D Re: eidetic and Pulfrich?

[roberts@xxxxxxxxxxxxxxxxx (John W Roberts)]

>Date: Fri, 26 Dec 1997 08:22:19 -0700
>From: wier@xxxxxxxxxxxxxxxxx (Bob Wier)
>Subject: P3D eidetic and Prufritch?

>I got to thinking since there are a number of knowledgable people on the
>list regarding such things, I'd pose the following... Has prufritch been 
>adequately explained? I believe that the slowing of light thru the darker 
>lens is not significant enough to account for the stereo effect.

>I wonder if at least part of the mechanism might be in the latent eidetic
>image processing capabilities of most adults. Could it be that the darker lens
>improves (or perhaps degrades?) the eidetic memory coming from that eye, 
>causing a "projection" of one image onto the other producing the stereo
>effect? 

The problem with thinking of it as a memory model is that you could watch a
30-minute movie with continuous Pulfrich effect, and see stereo the whole
time. For the view from the eye with the dark lens, you would have to be
"storing" the view all that time without consciously seeing it, then
"recalling" it a fraction of a second later.

The traditional model for the Pulfrich effect at least seems plausible and
consistent to me. As an electrical engineer, I think of the visual system
as a pipelined processing system, where visual interpretation is done in
stages. Different parts of the brain do different parts of the processing,
and they all function at the same time - as each part performs some of the
job of processing, it passes the information to the next section, and accepts
data from the previous section. (The data path is not strictly linear -
multiple types of processing may be done in parallel at a particular stage,
and data can be tapped off at intermediate stages for use by the rest of the
brain. If a large carnivore tries to pounce on me, I can start trying to get
out of the way without waiting for the diagnosis of "siberian tiger".)

In a sense, a processing pipeline is a sort of "memory" (a transmission line
is too), but it wouldn't be the same as the "eidetic" memory I've heard
of - the images come out after a very short time, and in the same sequence
they went in. You couldn't use this sort of storage to recall an image an
arbitrary time later.

The model for Pulfrich is that the dimmer light caused by the dark lens
over one eye somehow slows down *something* (perhaps several things)
along the visual processing pipeline, somewhere between the optical input
to the retina, and the processing center where information from the two
eyes is combined to extract depth information from binocular disparity.
If the scene has been shifted by the right amount and in the right way
between the two points in time that are being compared, then a perception
of depth will be extracted.

I don't know of a physical mechanism that causes delayed processing in dimmer
light, but the results would be consistent with the electrical engineering
model. One of the important issues in an electronic design is "skew" - the
relative timing (or delay) among two or more signals. If you have two signal
paths with intermediate stages, with identical logical design, but one path
is built using (for instance) 74S series logic, and the other with 74LS
series logic, then the signals along the path using the 74LS logic will arrive
at the destination significantly later than the signals on the 74S logic
path. (Bandwidth, or the maximum input frequency that can be accurately
processed, will also be decreased. It would be interesting to find out
whether the number of "frames per second" that can be accurately distinguished
in human vision is affected by the overall lighting level (with or without
flicker detection).)

I have read that individual neurons function essentially as very sophisticated
majority gates (the output is a function of the number of input stimuli).
Since ultimately visual stimulus comes from photons striking the retina,
perhaps the reduced number/frequency of stimuli in dim light causes the
neurons to take longer to resolve their output states. The same issue
might apply to areas of visual processing outside of Pulfrich. Does it
take longer to recognize a person from a very bad photograph than from
a good photograph? Does it take longer to recognize depth in a stereo
pair with poor or ambiguous cues?

>Since this article mentioned that many if not most children 
>posses eidetic memory, it could follow that most adults have latent
>capabilities in that direction. Also it would be interesting if the period
>of eidetic memory in children might be coincident with that time that
>language acquisition skills are high, which I believe also tends to
>diminish later in life once the brain gets more "hard coded". 

That had occurred to me too, after reading your RDS post. Young children
seem to have a phenomenal ability to make correlations needed to learn
language.

Pulfrich question: if you make a traditional 3D movie with polarized or LCS
glasses (instead of Pulfrich glasses), but display one view much darker than
the other one, does the Pulfrich effect still work? It seems to me it would,
but it would be interesting to hear whether it's actually been tried. If
it does work, it would be interesting to find out what happens when 2-camera
stereo and the Pulfrich effect are used *in combination*.

John R


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