An now for something totally different:

I spent the afternoon thinking about Worley noise. The question is, if you want to be sure to find the nearest N points, how many cells do you have to search? You don't know anything about the distribution of the points in the cells, so you have to assume they are in the most inconvenient positions possible. As it turns out:

- 0 cells (obviously)
- 12 cells
- 15 cells
- 17 cells
- 17 cells (again)
- 24 cells
- 26 cells
- 27 cells
- 27 cells (again)
- 28 cells
- 31 cells

My list goes on to 25 points (63 cells), but I don't fell like typing it all in. Facebook posts aren't good for this sort of thing.

All this assumes there is one point per cell. If there are

`k`

points, you can find `k`

times as many by searching the same number of cells. This is a step function, so there are no intermediate values for finding e.g. 3 points when `k = 2`

.
Back to essays page

Back to home page