lynear

07-07-2002, 03:18 PM

Hello everyone most of you have probably mastered the geomipmapping technique but i am still lost on one part generating the error arrays. The concept is quite simple its just the implementation is not. I have no reference to look at because i know of one geomipmap implementation that is open source and it is really hard to understand. Ill show you how i am trying to generate the error array, I am probably totally off so here goes:

now a 5x5 cell (4x4 triangles) looks like this

|_|_|_|_|

|_|_|_|_|

|_|_|_|_|

|_|_|_|_|

each row is a triangle strip. so there are w*h number of vertices(x,y,z). and 2*w*h of indices (uint). the indices are the same for every tile the only thing i change is the start of the vertex buffer. I don't know if its fast or slow so don't kill me 8). So now to generate the error arrays i took the midpoint between the higher level and subtracted it from the lower giving me the height change then multiplied it by "C" (in Boer paper) but i can't seem to right the procedural code to it. I just need a code sample i guess to see how other people have done it. I need a reference and i need help.

Dan

now a 5x5 cell (4x4 triangles) looks like this

|_|_|_|_|

|_|_|_|_|

|_|_|_|_|

|_|_|_|_|

each row is a triangle strip. so there are w*h number of vertices(x,y,z). and 2*w*h of indices (uint). the indices are the same for every tile the only thing i change is the start of the vertex buffer. I don't know if its fast or slow so don't kill me 8). So now to generate the error arrays i took the midpoint between the higher level and subtracted it from the lower giving me the height change then multiplied it by "C" (in Boer paper) but i can't seem to right the procedural code to it. I just need a code sample i guess to see how other people have done it. I need a reference and i need help.

Dan