. The big win with image-based VR environments is the ability to very
. quickly create complex environments. No you can't easily manipulate
. everything but you can represent environments in photo-realistic detail.
. Hot spots can be defined on top of these environment which let you
. take actions such as linking via URLs to other sites. Ultimatly I
. would expect environments with combinations of polygonal and image-based
. sites to exploit the advantages of both.
We've put some thought into this over here in Gopher Land...
Definitely there should be a place for Image based 3d images. They
can provide a far greater level of realism at the cost of network
bandwidth and memory usage.
Case in point, assume that I want to render a static image of an
elephant in 3D. I could either:
a) Create a polygon mesh to represent the elephant, perhaps make a
scale model of the elephant out of clay and stick it into one of
those 3D-digitizers. The result: a whole lot of polygons and
texture maps.
b) Use a framegrabber and an elephant and capture images from a
reasonable number of angles. If we assume a 6 degree spread
between image captures we would have a total of 1800 images that
would need to be cached to allow viewing the elephant at any
angle. (I'm assuming 90 great circles about the object, with 90
samples along each circle) If each image is a 20k jpeg file,
then you've got a four megabyte elephant.
Basic conclusion, case b is only useful if you limit the range of
motion afforded the viewer. Technique b is used by lots of games for
that very reason, it's a cheap and easy way to do 3D when you can
control the viewpoint.
-- | Paul Lindner | [email protected] | Slipping into madness | | Distributed Computing Services | is good for the sake | Gophermaster | University of Minnesota | of comparison. ///// / / / /////// / / / / / / / / //// / / / / / / / /