Difference in Games to VFX workflows?
polycounter lvl 3
I'm come from a video game background and I'm looking into VFX but I'm struggling to understand a few things.
In games we always convert our meshes to tris once they're ready for baking, and we use that same version going forward in any game engine to maintain consistency otherwise our normal maps can look off if the mesh is converted to tris differently between applications, and the baker.
Even animation work is done as per the normal way (quads), then we convert to tris once done as per the above.
With VFX I'm seeing that people work in all quads for the most part, and even when baking they wont convert to tris because the claim is their deformations will suffer. I'm trying to understand why this is? Is it because they use displacement maps instead of normal maps so converting to tris to bake really isn't required? Do VFX artists even use normal maps without converting to tris prior to baking as a normal workflow? I'm talked with a few that do...
Thanks for any help understanding this.
In games we always convert our meshes to tris once they're ready for baking, and we use that same version going forward in any game engine to maintain consistency otherwise our normal maps can look off if the mesh is converted to tris differently between applications, and the baker.
Even animation work is done as per the normal way (quads), then we convert to tris once done as per the above.
With VFX I'm seeing that people work in all quads for the most part, and even when baking they wont convert to tris because the claim is their deformations will suffer. I'm trying to understand why this is? Is it because they use displacement maps instead of normal maps so converting to tris to bake really isn't required? Do VFX artists even use normal maps without converting to tris prior to baking as a normal workflow? I'm talked with a few that do...
Thanks for any help understanding this.
Replies
And its no problem if you have "some" triangles in your lvl0 mesh. Animation meshes are really dens these days.
How come in VFX people don't have to worry about quad meshes looking different from application to application? One of the reasons in games we use a tri mesh between our applications as needed.
Do quads actually deform better than tris if the conversion to tris is done after all the animation work?
About the quads - They still get triangulated at render time, but on such high poly meshes, they get triangulated better and the "errors" are not visible. While they can show up visibly on low poly meshes. Normal maps and triangulation is also closely related in game assets.
In VFX how low do your low polys go for rigging? I assume you're not using high polys to animate with? Do you go as low as those in game development, or is it more of a medium poly mesh?
here's an example of a stand in model :