Current mainstream technology does not really model the density functions very well. Yeah, we can use volumetric grids, FEM, voxel representations and particle clouds that define densities. But these files dont scale, worse theres no practical way to model the distributions effectively.
But we are, slowly but surely, entering a new age of manufacturing where it is not enough to know just what the surface looks like, but know what is inside the structure, not just that its one material. But rather some amalgamation of properties that can change smoothly or abruptly depending on your need.
But how would we handle this in a way that is still editable? Is anybody aware of any methods that might eventually lead us to next level of model definition? The best i have so far seen is some general implicit density representations like extended meta balls but its still not accurate enough, or usable for me.
