CODEWITHSUNDEEP
data-structuresvoxeloctree
Gabriel Ribeiro
Gabriel Ribeiro

Reputation: 1

How will the shader differentiate the loaded brickmap from the unloaded brickmap as described in this voxel ray tracing paper?

I'm referencing a 2015 paper titled Real-time Ray tracing and Editing of Large Voxel Scenes by a student at Utrecht University, Thijs van Wingerden (link to paper).

I'm trying to build a voxel engine that mixes Octree or SVO (I'm building both but I'll use one of them) with these brick voxels. I'll use Octree or SVO to store the brickgrid and I'll render distant voxels as Octree/SVO since editing these nodes will hardly occur. But for the near voxels, I'll use the brickgrid/brickmap to render for editing purposes.

I understand that the brickgrid contains cells that are 32-bit. These cells can be either loaded brickmap, unloaded brickmap, or empty brickmap. If there is a loaded brickmap, then we have a 32-bit pointer to a brickmap (I'll use an index that points to a brickmap array).

How will the shader differentiate the loaded brickmap from the unloaded brickmap? I thought of using the first 2 or 8 bits to build a flag, but the paper shows that the unloaded brickmap has this flag and the loaded brickmap doesn't have this flag.

Figure 11 from the paper:
Figure 11 from the paper

Upvotes: 0

Views: 57

Answers (0)

Related Questions