Lecture 9: Ray Tracing & Acceleration Structures (86)
noah-ku
Compared to using a grid, this method is much more efficient because it uses computing power in more concentrated areas. Knowing that the objects are in clusters, we can put them into buckets (in this case, 32) to better make calculations in the right areas. This is a good balance between accuracy and efficiency, and is a popular method that people use today.
jefforee
How are the bucket sizes/widths determined? Is it something constant or will it differ for each set of triangles?
zepluc
Could this bucketing method be adapted for real-time rendering applications, or is it more suited for offline processing due to the computation involved?
JunoLee128
I'm confused why axis alignment is much more efficient than non-axis aligned. Is it a convenience thing, in terms of coding? Or is there a relation to the basis we're working with (since axes are arbitrary in the end)
Compared to using a grid, this method is much more efficient because it uses computing power in more concentrated areas. Knowing that the objects are in clusters, we can put them into buckets (in this case, 32) to better make calculations in the right areas. This is a good balance between accuracy and efficiency, and is a popular method that people use today.
How are the bucket sizes/widths determined? Is it something constant or will it differ for each set of triangles?
Could this bucketing method be adapted for real-time rendering applications, or is it more suited for offline processing due to the computation involved?
I'm confused why axis alignment is much more efficient than non-axis aligned. Is it a convenience thing, in terms of coding? Or is there a relation to the basis we're working with (since axes are arbitrary in the end)