|
|
|
|
|
|
|
|
The
OpenWorld technology is a custom 3D rendering technology based on a
totally novel 3D concept I had had back in 1996. This concept was initially
inspired by the popular Cyan game Myst. I was moved by the photo-realistic
scenes and really wanted to freely navigate these beautiful environments
without any movement restrictions and without loosing any graphic
realism in the detail. I knew nothing about 3D graphics at the time
and ended up learning from books and experimentation. After having excellent
success with the fundamental idea, I started a company named Aria Software,
Inc. and secured limited seed funding while further working on developing
the technology into a preliminary demo. I had also applied for a patent
with a firm in Newport Beach. I managed to get a demo
|
together
and later personally presented prototypes to a number of people such
as 3DFX's CTO Gary Tarolli, former Apple CEO Gil
Amelio, and Shiny Entertainments CEO David Perry among others. Unfortunately
a combination of inefficient self management, technical hurdles, and
industry un-acceptance lead to the eventual failure of the start-up
(yet an excellent learning experience). As a result, this technology
was back-burnerd for a while as I pursued other projects. Some time
has passed and I now would like to revive OpenWorld because I believe
there is something here of great benefit. I am looking for a sponsor
with technical and financial resources to help further the technology
into a full-blown 3D rendering engine under which entertainment products
would be created.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Features o Software only real-time photo-realistic non-polygon renderer o Somewhat similar to voxel rendering but OpenWorld's atom has been dubbed the trixel - a 3D pixel - and has a number of significantly different characteristics that enable real-time rendering o Extremely fine detail permits things as slender as a human hair to be rendered even if the hair is less than 1 pixel in width o Applicable to objects as well as environments o Antialiased edges so objects appear to be part of a scene and not just place on top of it o Automatic variable tesselation without the need for LOD's o Pliable deformable surfaces like skin o Easy particle effects because the atom is a 1x1 'unit' trixel o Movement hierarchy to handle characters limbs with the torso being the root o Separate tools for objects/character creation and runtime environment o Was patent pending o Unfinished technology with a lot of promise if nurtured in the right environment |
Limitations o Is not polygonal so deviates from the standard which means few artistic tools, minimal 3D hardware support, and little industry acceptance o Slower frame rates than typical polygonal renderers so not optimal for fast shooters. Geared toward real-time photo-realistic rendering o Concept proven in preliminary demos but not in a final form like a functional 3D engine so a lot of work still needs to be done o Needs better lighting model. Some thinking has gone into this but few tests have been performed as of yet. No tests performed on shading model |
|
|
|
|
|
|
|
|
But
they're heading in this direction anyway... |
|
|
One
criticism I received from people was that OpenWorld is not based on
industry standards such as OpenGL or the polygon. While this is true
I maintain that sometimes it is OK to deviate from the norm if there
is just reason: The 3D industry is also trying to achieve real-time
photo-realistic imagery, but I believe they're going about it inefficiently
mainly due to standards momentum and being locked into one way of thinking
and doing things. They're slowly heading towards a pixel per polygon
where polygons are becoming smaller and smaller and eventually they
will achieve pixel and sub-pixel per polygon resolutions but at this
point (and actually even before this point) it becomes inefficient to
render pixel sized polygons using current polygon texturing
|
techniques.
Many unnecessary calculations go into rendering a single pixel sized
polygon -- calculations that are required if redering larger textured
polygons but not required for smaller pixel sized polygons. There exists
a cross over point where it makes sense to break up objects into smaller
pieces and render them using new techniques. If you subdivide an object
into trixels for example and render them efficiently (where non-viewable
trixels are not computed) then none of these unnecessary computations
need occur and that processing power can be utilized in other more important
tasks such as lighting and shadows. This is what OpenWorld does now
entirely in software albeit in prototype form.
|
 
|
|
|
||
|
|
 |
 |