The Illusion of Reality (Re: Why 3D)

[P3D Brian Reynolds <reynolds@xxxxxxxxx>]

Greg W. wrote:
> 
> Julius Martin wrote:
> 
> >I've often thought about setting up a wall with two eye holes
> >and a well lit stereoscopic view on the other side of the wall to demon-
> >strate the idea. If there is nothing in the view that would move,
> >I don't think anyone would be able to discern whether what was seen
> >was real or a photograph.
> 
[good reasons why it is hard to fool people with pictures snipped]
>
> Your idea *might* work with full-color holograms (which are rare
> because of the difficulty of producing them).  I saw one of a doll,
> mounted in a frame, where I would have sworn it was a real, solid
> object in a box, except that it was obvious the "box" wasn't deep
> enough to hold the object!  Stereoscopy is fun and can be realistic,
> but it isn't THAT realistic, unfortunately.
> 

Back in the 1980's at Cornell University the computer graphics lab did
some experiments in order to determine how good their new method of
rendering (radiosity) was.  The previous champion of photorealistic
rendering (raytracing) was good at specular reflections and
transparency, but bad at diffuse reflections and inter-reflections.
For the work they were doing (architectural lighting renderings of
office spaces and interiors) diffuse reflections were more important
and specular ones.

The team working on radiosity set up a scene that mimicked a modern
artist's work.  This artist used white shapes and colored lights to
form the art work.  The inter-reflection of the colored lights between
the various white surfaces produced colors and shadings that were not
present in the objects themselves.  The computer graphics researchers
made computer models of these shapes and used their radiosity system
(sort of a thermodynamics equalibrium system applied to light energy
distributions) to render views of the art work.  They then set up a
pair of video monitors and showed the origial artwork (sitting in
another room) and their rendering.  They were able to get a
significant number of people to pick their rendering as the real
scene.  Looking in the proceedings of SIGGRAPH (the ACM's Special
Interest Group on computer GRAPHics) from that time should turn up
their papers on this.

By the way, looking at SIGGRAPH proceedings from the 1980's and later
should also turn up papers on converting 2D photographs to 3D computer
models and renderings (I believe they even did stero pairs or
interleaved scan line work for use with shutter glasses and head
mounted screens).

-- 
Brian Reynolds                  | "Humans explore the Universe with five
reynolds@xxxxxxxxx              |   senses and call the adventure science."
http://www.panix.com/~reynolds/ | - Edwin P. Hubble


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