Quick modeling question for you guys.
I'm just wondering what's more efficient out of the two models in the below image. one on the right is intersecting and is 14 tris lower.
I've been looking at the World of Warcraft models in a model viewer, and some models the artist tried to merge as many verts as possible, whereas others didn't seem to care and used a lot of intersecting geo. I get that intersecting isn't entirely bad and does provide lower poly counts.
So yeah, just wondering what people's takes is on this: what you prefer and why.
Replies
tl;dr
Depends on the situation. Doesn't really matter.
Intersections can cause problems with z-fighting, shadow-casting, lightmapping, etc.
Manifold meshes can cause long thin triangles, can take longer to create, etc.
Depends! Test in your situation!
I see no problem with either technique, there are pros and cons to both. Personally I prefer to model my some stuff with method A, and other stuff with method b, but really it all comes down to the functionality of the art in game.
If it is just for a portfolio piece, I would say use what makes the most sense to what your working on.
EDIT: Really, i think the biggest con to option B, is that it would not bake down, perfectly. And I love me some really clean bakes, so I would choose option A, for a piece like that!
Yeah, totally. I'd go with A as well for normal maps; however, this stuff I'm working on now doesn't use normal maps.
The one on the right is 14 tris lower.
I'm only seeing a difference of about 6 triangles, if I'm seeing things right. But while the one on the right has fewer triangles, it quite possibly has more vertices in the end as it must have more UV splits.
But still, for such a simple object does it really matter? I mean, i don't know what the hipoly is suposed to look like, but even 6 sides seems low >D
These are for a Unity game. I wouldn't say our specs are super low.
I read through the thread Eric posted. Got some good stuff out of that. I'm probably just over-thinking things. I know the extra triangles won't make much difference. I was just curious is all.
Thanks guys.
Yep, got ya.
I think for this game its fine either way.
If you're going to and bake a normal map for it, you're going to have noticeable aliasing along that edge, that you would otherwise not have.
Yeah, I definitely would approach it differently if it were normal mapped. I'm not texturing these props. The texture artist is using 3Dcoat, and its not causing any inconvenience.