void pcsampleleaf(int handle, float sample)
This function can only be used with the pcopenlod() function, and then only within a pciterate() loop. It replaces the current iteration point with an importance sampled leaf descendant of that point. The weighting used to select the leaf point is the “area” channel provided to the pcopenlod() function’s “measure” parameter, or a uniform weight if no area channel was specified when opening the point cloud. The sample parameter is expected to be a uniform random value between 0 and 1.
If the current iteration point is already a leaf point or the point cloud was not opened with pcopenlod(), pcsampleleaf() has no effect.
This function is useful when aggregate point information cannot be used in a meaningful way, and provides a mechanism to access the information contained in child nodes in the point tree. For example, it wouldn’t make sense to trace shadow rays from an averaged point position, but it is useful to choose one of the child points and then send the shadow ray to that point.
Example: Shadow Rays ¶
// Open a point cloud and retrieve a single aggregate point representing the // entire cloud string texturename = "points.pc"; int handle = pcopenlod(texturename, "P", P, 8, "measure", "solidangle", "area", "A", "samples", 1, "aggregate:A", "sum", "aggregate:P", "mean"); Cf = 0; // This loop will iterate only once while (pciterate(handle)) { // Query A from the averaged point float ptarea; pcimport(handle, "A", ptarea); pcsampleleaf(handle, nrandom()); // Query P from a sampled leaf point vector pos; pcimport(handle, "P", pos); if (trace(pos, P-pos, Time)) Cf += ptarea / length2(P-pos); }
| See also | |
| file | |
| ptcloud |
|