Hi, Im working on a simple 3d game and Im trying to figure out a good way to create a pre-calculated PVS map. The world map is a 2d array each element represents a "block" (with a fixed size 64x64x64). I have tried two approaches, 1) I go through the 2d map casting a ray from each corner of the current block to every other solid blocks courners. This method worked but very slow and I want to collect information on none solid blocks for ambient lightning on movie objects. 2) I go through the 2d map and for every element i create two cones (two half circles actualy) and go through the 2d map again checking if the block is inside the cone and then if the block is solid. If the block is solid a i need to split up the current cone into two too narrow the field of view. I haven't finished the second method because half way in I thought of a case where my teori didn't work. If a block is 64x64x64 then blocks visible from the center of the current block is different from another poin in the block (for isntance the edges), this can be fixed (well not fixed but a better approximation) by doing a "cone sweep test" for all 4 corners and one form the center but I'm afraid that will take too long. Both those two approaches seems to complicated for a simple 2d map to me, I think I'm making it harder for myself then it really is. Does anyone have a better method or an idea for a method ?

Sorry I wasn't clear enough, the game is a third person shooter (like Wolfenstein 3D, no raycaster though) so the camera view is changing (it is free moving camera in all direction not like wolfenstein where it fixed in one axis).

Also I want to use the result of the PVS calculation to make it easier to add ambient light not a block by just going through every visible block from a light source and just add the intensity.

gwihlidal: Do you have a suggestion on a method to use to calculate PVS on a 2d map ?

cast a ray along the left side of your view frustum, by 'walking' along this ray starting at the camera, see what gridlines it intersects till it finally hits a solid gridline(a wall).
this wall is visible
'slide' the ray to the right along that single wall to it's corner
now reapeat the ray walking check, if you hit something closer, then 'slide' along that wall for the next ray check
or it might hit some more distant wall, same deal tho...
and so on...

in this manner... rather than have rays everywhere (potentially wasteful as they might hit the same wall many times) your rays are chosen only to check the critical points of the map -the corners of walls where map features transition
and it only checks the nearest one; ignoring those that are hidden behind...

I will try the easiest method first and go from there, thanks guys.

gwihlidal: No code needed I rather read theoretical methods and figure the coding out myself (more fun that way)