Given that SVO would be a great structure for real-time deformation of models, can we expect to be able to blow dynamically generated chunks out of the flesh of our next non-human adversaries?
I can see SVO, with scalable voxelized normal maps, as being the biggest breakthrough in global illumination since the lightmap, but have you found a way around the issue of matrix blending? Voxels can't be transformed and the tracing ray must stay linear to be efficient, so it would be relatively difficult to implement an animated model without requiring a laggy cluster of inter-bone transform regions at every joint...
Slashdot Mirror
Given that SVO would be a great structure for real-time deformation of models, can we expect to be able to blow dynamically generated chunks out of the flesh of our next non-human adversaries?
I can see SVO, with scalable voxelized normal maps, as being the biggest breakthrough in global illumination since the lightmap, but have you found a way around the issue of matrix blending? Voxels can't be transformed and the tracing ray must stay linear to be efficient, so it would be relatively difficult to implement an animated model without requiring a laggy cluster of inter-bone transform regions at every joint...