Lecture 9: Intro to Ray-Tracing & Accelerating Ray-Scene Intersection (44)
ld184
what does mailboxing here mean specifically?
adham-elarabawy
Another potential solution here would be to always check for objects in a consistent manner (from left to right for instance) in the direction of the cast ray such that intersections are implicitly found in the order that they occur.
waleedlatif1
I don't understand how caching the intersection of the objects to avoid re-testing would be an optimization here. Unless we implement the suggestion in the second comment here to always check for objects in a consistent manner, I don't see still how this would work. Even if we cache the intersection, how do we still know which object the ray hits first?
joeyzhao123
I'm a bit confused with the solution part of the slide. What cell should the intersection be inside?
Zc0in
I think the cell is the blue grids in the picture @joeyzhao123 The res point(intersection) is not in the cell, so it is not the nearest
justin-shao
@waleedlatif1
The idea is to implement the "intersection point within the cell" check first. This enables us to make sure that the detected intersections are within the order of our cell traversal. The optimization part is done to cache intersection points per polygon, which in the example in the slide, would save time calculating when we reach the cell with the red dot (this saves time when we encounter geometries that span multiple cells).
what does mailboxing here mean specifically?
Another potential solution here would be to always check for objects in a consistent manner (from left to right for instance) in the direction of the cast ray such that intersections are implicitly found in the order that they occur.
I don't understand how caching the intersection of the objects to avoid re-testing would be an optimization here. Unless we implement the suggestion in the second comment here to always check for objects in a consistent manner, I don't see still how this would work. Even if we cache the intersection, how do we still know which object the ray hits first?
I'm a bit confused with the solution part of the slide. What cell should the intersection be inside?
I think the cell is the blue grids in the picture @joeyzhao123 The res point(intersection) is not in the cell, so it is not the nearest
@waleedlatif1 The idea is to implement the "intersection point within the cell" check first. This enables us to make sure that the detected intersections are within the order of our cell traversal. The optimization part is done to cache intersection points per polygon, which in the example in the slide, would save time calculating when we reach the cell with the red dot (this saves time when we encounter geometries that span multiple cells).