Of course there’s plenty of room for optimisation, the first tetrahedra computation is a bit weird for instance and could be done better I guess. The reason why I’m sharing this is that I was pretty pleased with the processing speed.
I used the incremental version, the slowest and the easiest to implement.
#PSYKOPAINT ANDROID FULL#
It’s very heavy to render as the full convex hull is computed onEnterframe, it’s just to give you the idea. Here’s a demo: click to respawn, when the hulling time is 0, you can click again to respawn the thing. it was fun, it gave me a chance to study the Half edge data structure, a nifty little tool used for computational geometry.Īs I was not doing much yesterday I thought I’d rewrite the hulling algo from scratch, 3D-engine free (for the pitch I used minko, so my implementation was strongly coupled to it ), only with the native API. I wanted to break apart 3D objects on the fly (a dynamic fracture algorithm ), so I had to do a 3D convex hull. Last year, my pitch was about 3D and stuff. I’d like to stress the fact that this year’s elevator pitch is going to break some jaws I’ve heard that simo would be there along with devboy and the psykopaint dude. Time to meet up with some good friends, hang around with talented bastards and get shit faced for 4 days :) Very soon I’ll attend the Flash On The Beach for the second time, last year I was lucky enough to show something during the elevator pitch and now I’m going as an entusiastic attendee. it can be useful in some cases, like when you want to perform coarse collision detection on a complex volume. For those who don’t know, a convex hull is the tightest volume enclosing a given set of vertices so that none of them lies outside the volume.
0 kommentar(er)
