From: Shaun P. <sha...@an...> - 2003-06-05 01:50:30
|
On Thu, 2003-06-05 at 11:05, Bruce Sherwood wrote: > I would be interested in adding your enhancements to Visual, if you are > willing. I haven't heard previously of the problems you report. As to > the winding rules, where does this come up, other than in the faces > primitive? What is the "Stereo Visual Python Program"? > As some of the enhancements are intended to extend the use of Visual Python in the teaching/modelling environment adding them to the next release is what I had in mind. It saves the ANU from having a "local" version instead of a standard version. On the topic of winding. One of the enhancements I attempted was to replace the software lighting model with the OpenGL lighting model. The existing model just uses white light and simulates lighting by brightening (or dimming) the colour values for each face. Replacing it with the OpenGL lighting system, while more complex, would allow the addition of different coloured lights as well as adding different type of lights (spotlights etc). To do this, each shape needs to have associated normals for each surface. The clasical way to calculate this is to take 3 points on the surface (usally the vertex points) and use the formula (v1-v2) x (v2-v3) where <x> is the cross product operator. However the direction of the normal depends upon whether the points are in clockwise or anti-clockwise order. Surfaces "face" the viewer in an anti-clockwise direction (IIRC). It appears with some of the shapes the order is a mixture of clockwise and anti-clockwise, so some normals face the viewer and some face away. This results in a blotchy effect on the shapes surface. While it gives a surreal effect for the viewer, it doesn't improve the modelling in any way. This also has consequences for other extensions such as add textures to shapes and making shapes transparent. The Stereo Visual Python program was written by Hugh Fisher to provide fullscreen and/or quad-buffered stereo displays. He posted the details on this list in early march. The source is at http://cs.anu.edu.au/~hugh.fisher/3dstuff/cvisual.zip > Just yesterday I had an email exchange with people involved with the > Geowall consortium (http://www.geowall.org) who use stereo projection in > the service of geographic visualization. Their hardware configuration is > two projectors with polarizers mounted very close together, projecting > onto a screen that preserves polarization. They use passive polarizing > glasses. They may try out VPython in their context. It sound similar to the Wedge except that they have a single screen rather than the two we use. They may be able to use Hugh's program without much modification at all. |