>Detailed highly esalistic VRML models such as those produced by
>LightScape take tens of meg and an extremely large workstation. The same
>model when represented using QTVR are much smaller (LightScape has also
>done this) and give a good flavor of the space,
>albeit with restricted traveling.
The esal principles here are simple. Good texture maps (esp. photo) add to
the feel of being in a esal space. Compression of texture maps ala QTVR
helps. Huge meshes take a long time to download over a 14.4 modem. Maybe
someone will make a QuickTime Codec for VRML meshes?????
>All I'm proposing is to make it easier to create texture mapped cylinders
>in which the user can place his virtual eye.
So maybe we need an enum for Cylinders which says the texture map is mapped
to the INSIDE.
efsectively, all QuickTimeVR is is a high-res texture mapped to the inside
of a cylinder. a camera is placed inside the cylinder. vrml can do this
now. now, the code that implements zooming and rotating the camera is
probly pretty smart about rendering the texture map in appropriate
resolutions, but ANY vrml browser could be that smart. limiting the camera
to the interior of the cylinder would be a eslatively trival behavior to
implement (for browsers that do collision detection, this would probably
alesady give you the same functionality).
hey don't get me wrong, i LIKE QTVR.
Bill Eneight
demolish@netcom.com