Re: Pulfrich Take 2

[P3D Gabriel Jacob <jacob@xxxxxxxxxxxxxx>]

I wrote

>> To the brain, it is seeing very real disparity. But in a sense it is
>> "apparent", will explain in next paragraph.

Scott W. Langill replies

>The common perceptual reference for disparity would involve non-
>corresponding spatial points on the left and right retina. The
>Pulfrich Effect is dependent upon the slower transmission and processing
>of neural signals from the rod receptors in the dark adapted eye (the
>eye with the darkened lens covering it) versus the signals from the cone
>receptors in the eye which is not dark adapted. Light falling
>simultaneously on corresponding spatial points on the left and right
>retina give rise to an inaccurate representation to the brain or an
>"apparent disparity".

I am not sure how you came to that conclusion. The light is not falling
simultaneously on the corresponding spatial points.

Example, a point moving left to right on a monitor; with a dark lens on the
right eye. At the initial point, no 3D is observed. Now the point moves
slightly to the right. At this moment, the left sees the instantaneous
point, but the left eye sees the original point before moving. Thus the
spatial points on the left and right retina are not on the corresponding
spatial points and the disparity manifests itself as the point in front of
the monitor, or 3D!  

>I should perhaps have been more precise and indicated that by illusion I
>meant an incorrect representation or distortion. In the classic Pulfrich
>experiment a pendulum moving in a straight arc is seen as moving in an
>elliptical orbit in depth. Usually a stereophotograph (depending on
>viewing distance and assuming it is not hyperstereo, etc.) portrays an
>accurate representation of the reality of the subject. This is not the
>case with the Pulfrich Effect

Using the pendulum effect to demonstrate how Pulfrich distorts is not a
fair example. But before I explain why it's not a fair example, I would
like to say that no conventional stereo portrays an accurate representation
of reality. They all distort, in one form or another. Even if it's not
hyper, hypo, streched, squashed, or whatever; "reality" requires not only
stereoposis but also accomodation among other things; to not "distort".

Now as to why the pendulum example you gave is not a fair example;
yes it's true that it "distorts", but this is because the pendulum effect
lacks visual cues for the eye/brain to make any correct judgement.
The pendulum effect serves to demonstrate the Pulfrich effect and only
that. Even traditional stereopairs suffer from this kind of limitation
and not only Pulfrich, which I don't know why deserves this unfair
reputation. 

An example of stereopairs introducing distortions or as I prefer to call
it, limitations, is images that lack visual cues and erroneous judgments
being made (as in the Pulfrich), is as we see below in the ascii stereo pairs.

In this example, viewing it wide-eyed, we see an x
in front of the window.
 ________ ________
|        |        |
|        |        |
|    x   |   x    |
|        |        |
|________|________|

In the next stereopair, the x is behind the window.
 ________ ________
|        |        |
|        |        |
|   x    |    x   |
|        |        |
|________|________|
 
Now, in the last example if we combine the above two examples,
into one stereopair, we would expect to see an x in front and
another x behind the window. But alas that is not the case.
What we see is neither x's in front or behind the window but
rather on the same plane.
 ________ ________
|        |        |
|        |        |
|   xx   |   xx   |
|        |        |
|________|________|


Another example of lack of visual cues giving erroneous results in
the pendulum effect is an example using the computer mouse. If wearing
the dark lens on the right eye and you move your mouse pointer to the
right, it will seem to be in front of the monitor window. Steady motion
to the right will not give any distortion as in the pendulum effect;
due to constant motion. This is unlike the pendulum, due to simple
physics describing the motion, slowing down as it nears the end of
it's travel. This slowing down accounts for the elliptical orbit, but
without the visual cues, the brain makes the wrong judgement call.

These external visual cues are very important and are not to be
under estimated. In the mouse example, if one now moves the mouse to the
left, its depth will depend on the visual cues other than due to the
obvious motion. Moving it left over the characters (or other objects)
it will be hard to see any depth change. But if the mouse pointer is
moved left, immediately below the characters, so as to not cover them,
the pointer will seem to be behind the monitor window. This is due
to the brain over riding the stereoscopic depth clues (indicating
the pointer is behind the characters) in favor of the interposition
clues. The brain from experience cannot reconile the two conflicting
clues.

>In a car chase across the screen where one vehicle is outpacing the
>other, either the faster vehicle looks unrealistically wide or the
>slower vehicle looks unrealistically narrow.

Again, this is a distortion but distortions are not unique to Pulfrich,
but rather to all forms of stereo.

>I would agree with the
>comments by George Themelis characterizing Pulfrich as "artificial"
>(i.e. "does not relate to variations in the 3d dimension in the real
>object"); but I would go further in the case with the Pulfrich
>Effect (i.e. does not relate to variations in the 3d dimension in the
>real object and may portray depth which is contradictory or misleading
>when compared to variations in the 3d dimension in the real object).

Again I disagree with the whole assumption. IMHO it is true stereo with
of course, limitations, distortions or whatever.

>I made the point regarding two dimensional cues to indicate that the
>brain will take the misrepresentational depth due to the Pulfrich Effect
>and use two dimensional cues to resolve it into a more realistic
>depiction of the scene. An extreme example of the brain's power to do
>this is the inability of most people to see human faces as hollow when
>viewed with prisms reversing disparity.

Well here, I will go full circle back to the pendulum or mouse pointer
where there are no visual clues to counter the two dimensional cues that
you refer to.

Having said all that, it might seem that I love Pulfrich! This is not
the case, it has severe limitations but still I think it doesn't deserve
the reputation it gets. Seems to be worse than the anaglyph reputation!
It has distortions but so does all stereo. I agree with you that
there are other strong visual cues that help us determine depth but
Pulfrich 3D, distorted or not, is still stereoscopic anyway you slice
it and is not due to other depth cues. Sorry, that's the way I see
(no pun intended) it. I gave my reasons why and you have also. Don't
know what else to add to this. 

>As to Pulfrich being the only game in town, I would refer you to the
>article in the April 1974 edition of American Cinematographer entitled
>"The Video West, Inc. Three Dimensional Photographic System". The still
>photographs made with their technique accompanying the article are
>clearly viewable in 2D without glasses and in 3D with glasses. The
>technique was designed for video and film.

That sounds interesting! Can you elaborate on it? I hope your not
refering to chromadepth or whatever.

>My understanding of experimental evidence is that given the limitations
>of stereoscopic acuity, most individuals make judgements relying
>upon two dimensional cues when viewing distant objects. The issue is
>whether the Pulfrich Effect provides realistic information or an
>accurate portrayal of the true depth of objects. A pendulum moving in a
>straight arc is seen to move elliptically when viewed with a darkened
>lens over on eye, in fact it moves in straight arc.. An anaglyphic film
>of the pendulum viewed with anaglyphic glasses shows the pendulum moving
>in a straight arc. Which is more accurate or realistic?

This is hard to answer. It depends. As I mentioned the pendulum is not
a good example, since it depends on motion for achieving stereopsis.
Since it is dependent on movement, any lack of other visual cues will
arise in distortions or error in judgement. An anaglyph is not accurate
or realistic in the color it portrays. So what is the final important
factor in determining if it's accurate or realist?

>Again, as I indicated in my response to Gabriel Jones, I made the point
>regarding two dimensional cues to indicate that the brain will take the
>misrepresentational depth due to the Pulfrich Effect and use two
>dimensional cues to resolve it into a realistic portrayal of the scene.

So are you saying Pulfrich is not due to stereopsis?

>The issue is whether you can tell an effective story or present an
>effective story with all the constraints that compensating for the
>misrepresentational effects of movement with the Pulfrich Effect. I
>doubt whether they are aiming for an accurate representation of depth, I
>think "a jazzy effect, with occasional pronounced depth effects" sounds
>more Hollywood probable to me.

Now that is another factor. Now one is saying that it is an accurate 
representation of depth at all. All anyone is saying, is that it works
and is true stereo and not due soley to 2D cues. Of course these 2D
cues are required to make heads or tails of the visual disparities.

>As indicated in my responses above anaglyphic methods viewable in 2D or
>3D, which would much more accurately depict the actual depth
>relationships of the objects photographed exist. Without totalling
>controlling all movement by the camera, actors, and scenic elements (to
>the exclusion of entertaining storytelling) misrepresentational depth
>relationships are endemic to the Pulfrich Effect. The television industry
>has chosen the cheap thrill approach, but is promoting it as an accurate
>portrayal of depth.

Well you seem to be concerned about achieving an accurate protrayal of
depth, but we're talking about entertainment here and not medical science
or whatever. Even then, in certain cases these distortions or exaggeration
of depth is sometimes used to stress certain features. Example is in
NASA's flybys of planets such as Venus where the mountain elevations are
exaggerated. You also seem to favor anaglyphs. Don't get me wrong, I love
anaglyphs also, but did you really prefer ABC's "accurate" protrayal of
stereoscopic depth over Pulfrich?

>I agree to the extent that I believe that the requisite attention to
>motion is a crippling artifact of utilizing the Pulfrich Effect in
>television and film presentations.

Ahhhh, at least we agree on something! ;-)

>> Why did the Pulfrich Chicken cross the road?
>
>To pump up the ratings of NBC as inexpensively as possible.

I am glad to see you have some humor! :-)
Seriously speaking though, I never thought I would be discussing
Pulfrich to this extent. Ouch!

Regards
Gabriel, wondering how he got to be a proponent of Pulfrich!



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

End of PHOTO-3D Digest 2120
***************************
***************************
 Trouble? Send e-mail to 
 wier@xxxxxxxxxxxxxxx 
 To unsubscribe select one of the following,
 place it in the BODY of a message and send it to:
 listserv@xxxxxxxxxxxxxxx 
   unsubscribe photo-3d
   unsubscribe sell-3d
   unsubscribe mc68hc11
   unsubscribe overland-trails
   unsubscribe icom
 ***************************