P3D Re: SI anaglyph

[Gabriel Jacob <jacob@xxxxxxxxxxxxxx>]

Sorry for the delay in replying. I usually take a day off on Sunday's
from the P3D discussions, to recharge my batteries! ;-)

I wrote:
>>This is confusing. Your saying "staircasing" is not the same thing, as
>>cardboarding? Can you elaborate?

Tony Alderson writes:
>By "staircasing" I am drawing an analogy between the 2D aliasing of
>pixel based images, and "stereo" aliasing. When the pixels get larger
>than the stereo acuity, then the image will appear to be made of a
>series of discrete planes, rather than a continuous curve. The "pixels"
>in the SI image are pretty small.

Thanks! That's exactly what I thought you meant, but before I put my
foot in my mouth, I said, "let's me see what exactly your referring to!"
:-) This staircasing, I still consider the same, as cardboarding (I'll
elaborate later).

>A carboarding cone will appear to be a plane, regardless of the
>resolution of the image or the size of the pixels:

You said in the previous paragraph, "when the pixels get larger than
the stereo acuity, the image will appear to be made of a series of
discrete planes, rather than a continuous curve." I agree with this
statement, but now your saying the pixel size is immaterial? Now, I'm
lost.

>Didn't say it was ortho. It's not ortho from close-up either. I still
>think the perspective match is better,

At some point it has to be ortho. Assuming that they used a normal
stereo base.

>but simply matching angles of view does not make an image orthoscopic,
>in my understanding of the term.

Actually the image is orthoscopic (as opposed to pseudoscopic). As to
simply matching angles of view (and stereo base), this would make it
ortho (orthostereoscopic).

>You can match viewing angles of a 500mm lens, but no matter what
>you use for a stereo base, you won't get an "ortho" image. (Except in
>the sense that we don't see stereo beyond a few hundred feet, so even a
>flattie image, far away, is arguably "ortho".)

Wouldn't there be only one correct "ortho" stereo base? Namely, around
65mm (in this particular case, anaglyphs)?

Gabriel continues:
>>Furthermore, there is no roundness to her arm (if that is what your
> referring to).

Tony replies:
>I disagree. I see roundness in her arm, but maybe I'm just being fooled
>by the shading.

I seriously think it's just an illusion (shading) or more of a factor
than any stereo information. 

>If "stretch" is happening on her arm, it's happening in her arm too.

It would, IF there was any stereo information. As just mentioned, I
think it's an illusion. I don't have to tell you, that often times
people will claim a 2-D movie or 2-D image "looks" 3-D if it is sharp
enough to give a lot more realistic 2-D cues of depth.

>IMHO, there is enough edges and shading in the red channels. But why
>have I seen cardboarding in stereo slides, where this issue of red
>channel detail does not apply?

I think it is for similar reasons. Lack of detail, resolution, parallax,
stereo acuity, pixelation, or whatever else you want to call it! :-)
In other words, crappy cameras. ;-)

>I disagree. You are comparing images with unknown variables. I believe a
>comparison with controlled variables will show otherwise.
 
This was in reference to the mostly B&W Universal Studio ad. True, there
are unknown variables, but still, isn't it interesting that you don't
have to stand back to see the depth in that image. If it was color, you
might have to stand back a mile (and stretch it) to see any depth. 

It looks to me, that the Universal image is most likely the correct angle
of view at arms length. It's very hard to tell with the Heidi picture,
since it most likely was cropped and the rocks are hard to judge (size
and distance wise). Although, her bottom bikini does seem to "float" a
few inches away from her body!

>I'll try to find the time and energy to demonstrate this, but I would
>encourage you to try your own independent experiments.

Good idea, although I'm pretty sure ;-) that for every example you give,
I'll find a counter example. Isn't that what science is all about! ;-)

As I had mentioned in my last post, I've noticed this cardboarding a lot
with cheap cameras. Granted I didn't do controlled tests, but I did take
enough pictures with them to observe this.

In closing, I should add that with respect to cardboarding, I think there
is also a problem of definition. I agree with you that the mismatch of
viewing angles will create a cardboarding effect. This is a well known
effect and can be seen when using binoculars. In that particular case,
yeah sure, viewing angle is of prime importance, but when I (wrongly
or rightly) think of cardboarding, I think of the lack of detail in the
image (be it due to color, or detail, resolution or what have you).

This got me interested to see what actually "cardboarding" is defined as.
I checked up on the net and found this.

 Cardboarding: A defect occuring in stereoscopic images, when because of
 inadequate depth resolution, a scene divides into a small number of 
 separated planes in depth (like cardboard cut-outs), instead of the planes
 being continously linked as in normal vision.

This can be found at,
http://www.itc.co.uk/mirage/Standard/glos.htm

Now of course, this is not the last word in definitions, and they missed
out on viewing angles, but I would like to find some more scholarly info
on this twisted matter, that the sinful SI issue has brought upon us. ;-)
Bruce or Abram might have something to share? Or the gang at ISU? Or some
other "uneducated" soul!?! ;-)

>Gabriel finished:
>> (...) I do greatly respect and enjoy,
>> Tony Alderson's In Depth 3-D work and knowledge!
>
>Thanks, nice of you to say so. But I notice you don't worship the ground
>I walk on!   :-(  or is it  ;-)  ?
>
>Tony

Ahhh, maybe one day, IF I make a pilgrimage to LA! ;-)

Gabriel