• Advertisement
Sign in to follow this  

Octree question

This topic is 4975 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

Hi, i have read much papers and books about different variations of octrees... but all those articles demonstrate cubic octrees (each side of a node´s bounding box has the same length). however, in my own implementation, my bounding box sides can have variable lengths, depending on the scene input. would that be a disadvantage or advantage or does it just don´t matter? thanks Gammastrahler

Share this post


Link to post
Share on other sites
Advertisement
Octrees is all about dividing each node to 8 smaller nodes with the same size.Why do you want to do it differently?I mean,how are you going to determine the size of each node?

Share this post


Link to post
Share on other sites
i use non-cubic size nodes, runs fine for me.

advantage: more accurate node visibility testing
why? because sometimes a node may contain verticies that dont even remotly resemble a cube, eg. a flat plane. if any part of the cube is in the view fustrum, its rendered, but if the node boundingbox wer more closly matched vis testing would be more accurate

disadvantages: slightly more memory usage.
because you have to store center, min, max where a cube you only have to store center, length.

btw, an oct-tree with non-cubic nodes i belive is similar to a data structure known as an N-Tree (try google?)

Share this post


Link to post
Share on other sites
Sign in to follow this  

  • Advertisement