Re: Discounting the joy of Pulfrich 3D

[P3D John W Roberts <roberts@xxxxxxxxxxxxxxxxx>]

>Date: Sun, 8 Jun 1997 17:17:59 -0500
>From: P3D Scott Langill  <slangill@xxxxxxxxx>
>Subject: Discounting the joy of Pulfrich 3D

>> Am I the only one that did not "discover the joy" of Pulfrich 3d...

>     I wholeheartedly agree. In a sense you were seeing an illusion of
>depth, but you were not seeing depth or 3D.

I don't know what the previous poster was seeing, but I disagree with the use 
of the term "illusion" in this context.

>     The Pulfrich Effect utilizes a colored or shaded lens over one eye
>to delay the signal from that eye to the brain. The delay is falsely
>interpreted by the brain as representing two simultaneous views of the
>same object displaced from one another spatially due to the presence of
>depth in the object. The brain recognizes the signal as disparity
>information of the sort used to resolve depth (i.e. in stereopsis). The
>effect is wholly dependent upon movement (e.g. a pendulum was used in
>the original experiments). Factors such as relative true depth or
>apparent velocity (i.e. motion parallax) would perhaps interact with the
>effect (as would other two dimensional cues to depth), but real movement
>across the retina is necessary for the brains false resolution of
>"apparent" disparity. Without movement there would be no sensation of
>three dimensional depth perception. In other words, the faster the
>movement the more the depth.

You seem to have a pretty good idea of the working principles 
behind Pulfrich 3D.

>     Moving objects have depth, static objects do not, faster objects
>have more depth, objects moving perpendicular to the observer's line of
>sight have more depth than those moving angularly at the same real
>velocity, and moving objects in the distance have depth even though
>stereopsis is normally functional out to about twenty feet. This does
>not bear much resemblance to the real world (despite what the 3D
>Frequently Asked Questions File would lead you to believe).

Perhaps you mean that depth cues from *focus* (accommodation) are functional
out to about 20 feet? Unless you have really bad depth perception, you should
be able to see depth (without motion) much further out than that. Also, humans
*do* automatically move their heads to get parallax to judge the distance of
very distant objects. And Pulfrich works in real-world applications - I
experimented with it about 5-6 years ago looking out of car and airplane
windows, and others have also done this. The brain already uses the relative
motion to reconstruct depth - the Pulfrich effect allows another part of the
brain to pitch in and reinforce the depth perception.

I don't understand this apparent degree of hostility to depth cues that are
not purely stereoscopic perception of a pair of images. Humans use dozens of
depth cues (watch out, or I may post my updated list of cues to depth
perception), and if you stick to just one cue and ignore or discourage use
of all others, you're never going to have really convincing 3D.

>     Two dimensional cues to depth can be very powerful, and as is the
>case with most illusions of three dimensional depth perception, they
>will interact with and can overpower three dimensional cues (such as
>with distorted rooms of the mystery spot variety). 

In the distorted room illusion, there is no "three dimensional cue" - you're
only allowed to see it with one eye, from a single point of view, so the only
cue it has to work against is the accommodation cue (and if the room is kept
bright, that can be held to a minimum). A much better example is a pseudoscopic
picture of a human face - even when the stereoscopic cues strongly indicate
that you should see the face as "hollow", it's extremely difficult to do so.

>It is important to
>differentiate between perceived depth due to the Pulfrich Effect and
>perceived depth due to two dimensional cues or stereopsis. 

Why? If your premise is that both Pulfrich and 2D cues are false 3D,
what is the benefit of differentiating between them?

>Any
>coincidence between perceived depth due to the Pulfrich effect and real
>depth is (despite the ingenuity of television writers) accidental at
>best. 

I don't think that's a fair statement. If by "coincidence" you mean "exact
numerical match of apparent depth", then I would agree that Pulfrich is hard
to "calibrate", but it sounds more like you're saying that it's logically
impossible for television producers to consciously utilize Pulfrich to
create an appearance of depth where they want to create an appearance of depth.

You mention analogies - I'll attempt one. In the realm of "conventional"
3D, there are two groups of people that are particularly prone to hostility
toward the methods used: those who are "stereo blind", who are sometimes
antagonistic to the whole idea of 3D, and those who are unable to decouple
their perception of size from degree of convergence - these people sometimes
object strenuously to the use of hyperstereo because "it makes everything
look small". By analogy, is it possible that you are "Pulfrich blind"? If you
ride in a fast moving car during the daytime, looking directly out a side
window, with a very dark lens in front of the eye nearest to the rear of the
car, does the depth of the scene not look any different to you? Or is
it possible that you are one of those people who strongly prefers that the
degree of convergence in a stereo image be "just right", or things don't look
the right size to you? If either is the case, then I sympathize with your
situation, but I can assure you that Pulfrich does produce useful perception
of depth to many people.

>     The theatrical presentation of 3D began in 1915, the first
>anaglyphic feature films (utilizing red green glasses) were created in
>the 1920s and the first vectographic feature films (utilizing polarized
>glasses) in the 1930s. 
>...Commercially available 3d
>cinematic processes suitable for two dimensional viewing (in both
>anaglyphic and vectographic formats) have been available since the
>1970s,...

You appear to be using the term "vectographic" for all stereo films to be
viewed using polarized glasses. The discussion to date on P3D appears to
indicate that "Vectographs" are made by a very specialized process which
puts both views overlapping one another on the same frame. I believe most
or all polarized films put the two views on different frames. Did you intend
to refer to polarized-view stereo in general?

>Enter the Pulfrich Effect, no special cameras, inexpensive
>dark and clear glasses, a suitable 2D appearance (with no "fringes" at
>all), a promotable (albeit unreal and non-stereoscopic) "3D" effect, and
>an accepting audience.

>From a certain point of view you could regard Pulfrich as "not true stereo",
but to be consistent I think you would also have to disqualify print cards
for Holmes viewers, 5p and 7p slide pairs, anaglyph, sequential field,
and so on, as they are currently implemented. *Nothing* that currently exists
provides even a reasonably convincing reproduction of visual reality. I don't
really expect such a technology to be available for 10 years from now,
possibly 20 or more. Every 3D technique that exists makes certain compromises
to get the desired effects, and these compromises sacrifice some degree of
realism.

But I don't consider that a useful point of view to hold (for now, anyway).
Given the current technology, I think it's more useful to define "true
stereo" as something which utilizes the stereoscopic perception of the
human visual system to create an appearance of depth, with the apparent
depth being a function of what might be perceived in a real scene. Given
that definition, a Pulfrich film or broadcast can be made which exhibits
true stereo. If you try to narrow down the definition of true stereo (in a
logical manner) to exclude Pulfrich, you're going to start knocking out 
the cherished traditional methods of depicting stereo as well.

>     It is unfortunate that a group as influential as the readers of and
>contributors to the photo-3d mailing list and 3D Frequently Asked
>Questions file are so enthusiastic about the use of the Pulfrich Effect
>on television. It would seem apparent that bandwidth would better be
>spent attempting to convince television networks and producers to use
>the available and true stereoscopic techniques.

Those being...?

Speaking of "illusions", I have heard a respected expert in the field of
scientific visualization refer to 3D itself as an illusion! [Scientific
visualization is the practice of creating a visual representation of
information (classic examples being charts and graphs), so that the observer
can more easily absorb, interpret, and correlate the data. Advanced
techniques can depict a surprisingly large number of parameters at once -
for instance, an arrow on a weather map can show location, direction, and
magnitude of an air flow at the same time.] When asked whether 3D could be
useful in visualization, this person discouraged the use of 3D, pointing out
that your eyes see in 2D, and the brain creates the illusion of 3D, so what
good is an illusion? (I didn't ask the person whether he was stereo blind. :-)
Date: Mon, 9 Jun 1997 00:20:19 -0500
Errors-To: 3d-moderators@xxxxxxxxxxxxxxx
Reply-To: photo-3d@xxxxxxxxxxxxxxx
Originator: photo-3d@xxxxxxxxxxxxxxx
Sender: photo-3d@xxxxxxxxxxxxxxx
Precedence: bulk
From: photo-3d@xxxxxxxxxxxxxxx
To: Multiple recipients of list <photo-3d@xxxxxxxxxxxxxxx>
Subject: PHOTO-3D digest 2101
X-Listprocessor-Version: 6.0c -- ListProcessor by Anastasios Kotsikonas
X-Comment:   The Stereoscopic Image (Photo-3D) Mailing List  
Status:   


So if 3D itself is an "illusion", I guess we could accept the idea that
Pulfrich is too. :-)

John R


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