Advanced Vis Question [message #-986432] |
Sun, 12 January 2003 14:20 |
|
This is how I have always thought vis worked: You place the manual vis point over the vis sector. LevelEdit writes on that sector which meshes should be visible while the camera is over it. Now, it is said that your vis sectors have to be an exact copy of the terrain, but I was thinking of an alternate method. Suppose I were to manually create the vis sectors for various parts of the level to prevent as many vis errors. After building the terrain, I make a separate mesh with a minimal amount of polygons that generally keeps the shape of the terrain. I make this mesh with, say, 20 polygons. I explode the mesh, and set it's collision properties to vis. Now, I only have to worry about 20 vis sectors and have a minimal amount of vis errors. It would be just as effective if this were to work from what I understand. Has anybody ever tried this, and if you have, what were your results? Thanks.
|
|
|
Advanced Vis Question [message #-986431] |
Sun, 12 January 2003 14:22 |
|
More VIS sectors mean more areas where things will "VIS out", or disappear... Resulting in better framerates. [ January 12, 2003, 21:24: Message edited by: aircraftkiller2001 ]
|
|
|
Advanced Vis Question [message #-986430] |
Sun, 12 January 2003 14:26 |
|
True, but you see, having 50 vis sectors in the middle of an open area wherin no new meshes are visible doesn't help at all, it simply creates more unmarked vis sectors to hunt down.
|
|
|
Advanced Vis Question [message #-986429] |
Sun, 12 January 2003 14:27 |
|
quote: Originally posted by SomeRhino: True, but you see, having 50 vis sectors in the middle of an open area wherin no new meshes are visible doesn't help at all, it simply creates more unmarked vis sectors to hunt down.
Sometimes VIS work requires that. I strike a balance... Usually stick with about 50-200 VIS sectors for a small map, 200-400 for a large one.
|
|
|
|