Using xatlas instead of the uvs from marching_tet

[xk-huang]

Thanks for this wonderful work! I have a question about the design choice here, which is using xatlas to create uv coordinates instead of taking the uvs coordinates from marching_tet.

[jmunkberg]

Hello,

Thanks for your feedback! To avoid texture seams (or use texture filtering/mip maps), we ideally want large, continuous regions in UV-space, which is the reason for computing a parametrization. Simply allocating a small fraction of the UV domain for each triangle in the dense tet-grid works as well, but this (naïve) parametrization texture seams nearly everywhere, and produces sparser, lower-quality textures.

That said, xatlas is quite slow for very dense meshes, so for some applications, you may want to revert to a "per-triangle" parametrization, or rely on volumetric texturing (as we do in the first pass of optimization).

[xk-huang]

Thank you so much for the helpful reply! 😄