POV-Ray : Newsgroups : povray.binaries.images : WIP: Sterograms of a cube! : Re: Sterograms of a cube! Server Time
14 Aug 2024 07:10:51 EDT (-0400)
  Re: Sterograms of a cube!  
From: gimi
Date: 5 Jan 2003 15:59:24
Message: <3E189C89.4030304@psico.ch>
Jaap Frank wrote:
> I think the SIRDS are the same principle as the red-green
> double pictures that you have to view with a red and a green
> glass for your eyes. Only the different colors are replaced
> with a spaced random dot pattern and then superimposed.
> Am I correct?

nope. with rg-stereograms, you don't use the exact same
information for both eyes, but the results from the "filter
process", which makes red spots look black when seen through
the green glass and vice versa. - you do not look at a different
spot in the image with each eye!

[...]
>>so you approximate a parallel projection for either of the two
>>objects in order to avoid the "curvature" that andrew mentioned.
>>- did you try using orthographic projection instead?
> 
> It's one picture with two identical objects placed beside each
> other without overlap and then viewed with one camera with a
> small viewing angle. The camera is a normal perspective one.
> You can do it yourself easily.
> There's one drawback with this method. You can't make a
> 3D picture from inside an object, because the objects will
> overlap then.

i understood well how you do it! - but do you see how you
can make your "crossed-eye" into a "non-crossed-eye" stereo-
gram? - the angle from which an eye sees the object does
not have to be the result of it's position, but you can rotate
it instead, so that you don't have to cross your eyes..!

 > If you use an orthographic camera, indeed you have to rotate
 > the objects to get the right configurations. I've never done
 > that, but it's another possibility. As long as you get the correct
 > differences it's oke. Maybe the 'perspective method' gives
 > you the normal distortions and the 'orthographic method'
 > will not give the right distortions for depth, but you can try
 > both with the same object and compare then.
 > If the orthographic camera doesn't work, you can rotate
 > the objects to make "magic eye" pictures.

yes, the reason why i proposed the orthographic camera is
exactly that it will not distort the image if you choose
the "magic eye" variant.

but the orthographic camera will *only* work *if* you rotate
the object!

ok, maybe this shows what i mean (this is not intended to be
a stereogram of any kind - don't hurt your eyes! ;) :

       ----        ----
      /   /|      |\   \
      ---- |      | ----
      |  |/        \|  |
      ----          ----

this is how you do it - the same object looks different,
because one is to the left and thus seen from a different
angle as the one to the right. like when you look at the
object in real space. - you could cross your eyes and...
"hey! this looks like 3D!" (don't try it! ;)

note that the difference between both images is the result
of both their position and the kind of projection used
(like a normal camera does).


      ----           ----
     |\   \         /   /|
     | ----         ---- |
      \|  |         |  |/
       ----         ----

now if you rotate the left one counterclockwise and the right
one clockwise a little, you get the opposite effect. when
you use it as a stereogram, you won't have to cross your eyes,
but look "through" the image instead to get the effect.

this will look "wrong" when used with a "normal" camera,
however.

       ----       ----
       |  |       |  |
       ----       ----
       |  |       |  |
       ----       ----

and finally, this is how the objects would look with orthographic
projection but without any rotation; namely the exact same object
twice. - ok?


i guess we all agree now - so lets go and render some!! :D


g.

-- 
mailto:gim### [at] psicoch
http://www.psico.ch/ 
http://psico.servehttp.com/


Post a reply to this message

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.