Whilst clunky, it seems like most major 3d modelers use primarily polygons, and I've also heard that quads are the way to go. I feel like intuitively, implicit definitions have its good points as well; however, I don't think many modelling software uses them as its forte. Why is that? Is it because GPUs are specialized towards triangles? I thought they were also good at math?
Max
MeshEdit should help illuminate these tradeoffs - polygons are generally nicer to model with, and with some restrictions, support operations like subdivision & extrusion. Quads are particularly good for that. Some modeling programs support binary operations on shapes as well as signed distance fields, but it can be difficult and fragile to create complex/realistic objects with them. (But definitely not impossible! Check this and out and this for more.)
As for GPUs, they really only render triangles by default - everything else either has to be converted or use a more specialized rendering path like per-pixel ray marching/sphere tracing (for signed distance functions).
ryanmelon
This data structure is similar to how .obj files storing the data of a model. All the normal, texture coordinates information are also connected to vertexes with id.
Whilst clunky, it seems like most major 3d modelers use primarily polygons, and I've also heard that quads are the way to go. I feel like intuitively, implicit definitions have its good points as well; however, I don't think many modelling software uses them as its forte. Why is that? Is it because GPUs are specialized towards triangles? I thought they were also good at math?
MeshEdit should help illuminate these tradeoffs - polygons are generally nicer to model with, and with some restrictions, support operations like subdivision & extrusion. Quads are particularly good for that. Some modeling programs support binary operations on shapes as well as signed distance fields, but it can be difficult and fragile to create complex/realistic objects with them. (But definitely not impossible! Check this and out and this for more.)
As for GPUs, they really only render triangles by default - everything else either has to be converted or use a more specialized rendering path like per-pixel ray marching/sphere tracing (for signed distance functions).
This data structure is similar to how .obj files storing the data of a model. All the normal, texture coordinates information are also connected to vertexes with id.