Advanced Member Skaven252 Posted February 10, 2009 Advanced Member Report Share Posted February 10, 2009 This is just a thought / question that occurred me when fiddling around with voxels. How are the voxel volumes handled in memory? I would reckon that if there are large areas of solid fill or empty space, maybe they could be packed somehow to save memory. Also, if there are very small details in one part of the model, and only smooth undetailed surface in another part, could the voxel resolution change dynamically, so that the densely detailed area would have more resolution than the smooth area (kind of like "smart voxel subdivision" if you catch my drift)? Would the constant packing/unpacking of the data use a lot more CPU than just processing an unpacked n x n x n voxel matrix? I found a somewhat related paper on the subject, but I'm not sure how helpful it is (I think it has more to do with noise reduction than dynamic resolution): http://cds.ismrm.org/ismrm-2004/Files/002214.pdf Quote Link to comment Share on other sites More sharing options...
Member Crusoe Posted February 15, 2009 Member Report Share Posted February 15, 2009 They are already 'packed' using a sparse voxel tree. Otherwise, you couldn't hold that many in memory. Quote Link to comment Share on other sites More sharing options...
Advanced Member Skaven252 Posted February 16, 2009 Author Advanced Member Report Share Posted February 16, 2009 Oh, okay. I'll have to google what that means. Hmm, yeah, this explains it pretty well: http://en.wikipedia.org/wiki/Octree Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.