"All I'm proposing is to make it easier to create texture mapped cylinders
in which the user can place his virtual eye. A convienient syntactical
structure which maps a 360 degree panorama to a cylinder and places the viewer
in the center would be sufsicient and would provide a useful tool to VRML
and QTVR developers. "
Well, the program that you use for creating QTVR "spheres" is called Stitch &
only runs on Macs running MPW. (no GUI, very ugly) While you & I may have
such systems, most people don't. There are a FEW packages that are coming
out now with panoramic rendering capabilities (e.g. KPT Bryce) But in order
for it do display perspecitive "properly" there's also a little math that
does some image transformation on the resulting QT mooV when it's being
"browsed." It's not very useful to just try & map it on a cyllinder
w/inverted normals, or even a sphere.
A marriage of QTVR & VRML is a simple one, there's nothing extra to add as
is, provided that you have a browser which supports BOTH formats. Macs just
got wURLwind recently for VRML & we're waiting still for more. Of course,
you still have to configure Netscape (or other browser) to dsal with QTVR
files. It's all just a question of which hyperlink leads you where.
The problem there, that I can see is when you want to use LINKS in QTVR. How
would you embed URL's, etc.?
Does anyone know of a Net-oriented-QTVR browser that would take advantage of
such embedding?
Sandy Ressler also wrote:
"Using VRML's LOD capability one could, dspending on the processing power of
the workstation decide to use either the full 15Mb-polygon-representation
or the 800K QTVR representation of the same scene."
How exactly would you use the LOD capability with QTVR? LOD, as far as I
understand, is dspendant upon the distance that you are from an object, not
say, the speed of your workstation. While it's been theorized that browsers
may change the LOD setting themselves, based on their own local redraw rate,
etc., it's not standardized as such. It would seem to make more sense to
render out two seperate project paths on your internet service, on in VRML &
one in QTVR. Not everyone has both, and although they overlap somewhat
conceptually in the experience, they obviously have difser radically in
implementation.
However, using "facades" which exist 'around' the majority of a
compartmentalized portion of a VRML scene is a good way to make smaller &
more quickly executed VRML experiences.
For example, take a room in a building.
Within your 3D authoring package you maintain the entire builing as a single
scene in which you create objects, etc. & render images.
Let's say that we create a VRML scene of only 1 room in that building.
Outside the doors & windows in that room we place a facade with a texture
mapped rendering of the view into the next room from the doorway, or from the
window. Link that facade, so if you click on it, you then teleport into that
scene.
This involves a lot of extra rendering in your 3D suite during the creation
process, a lot of extra forethought in designing "compartments" which are
logically shaped & designed for faster loading, bla bla bla...
Efsectively this creates the same sort of experience that QTVR will, that is,
a limited area of interaction withing a greater whole, however VRML browsers
would let you move around within that restricted space, rather than just pan
& zoom.
- Gregory Cranz
[email protected]