Polygon efficiency vs. Closed Meshes?

Alright, this has bugged me for some time, which of the two meshes beneath is the “better” one?

The one to the left is simply two boxes (12 polys), and the one to the right is a closed mesh (22 polys), if I were to make a model for say the Source engine, which type should I use, and what are the downsides (if there are any) of using the one to the left? Wouldn’t the result be the same? (except maybe the UV mapping)

They should work the same. I prefer the second one though.

Not seen much affect from the poly count until you reach around 40000 tris (then it starts to lag a bit, but not always).

You can always use the lower poly one as your collision model for the higher poly one ^^

But wouldn’t it be a lot simpler to model with the one to the left in some cases, if you’re going to attach something that you’ve modeled separately to a larger mesh, for example?

The one on the left has more to render, but the one on the right has more polygons to go through. I don’t know much more than that though.

I don’t think there’s a right or wrong way. Some meshes have to be done closed and some left open, it depends largely on the model.

This.

It really depends on what you are using with the model. If that connection between the two boxes is important to the model (normal mapping, skeleton deformations), then you should leave it in.

The right one is much easier to uv map and texture, because you know exactly where the seams are.

You’d know easier with the left one if you baked ambient occlusions.

you would get horrible vertex lighting on the one on the left tbh.

i would use the first model for static purposes, like physboxing or inanimate objects with repetitive textures. going advanced with that one, however, is taboo. if you want to live, and make peace with those who hate you, go fullmodel.

I used to close them, but I’m forcing myself to use the method on the left.

Use the one on the right, rendering intersecting polygons is annoying for the renderer anyway as far as I know, I really doubt it’s much better than having a few extra poly’s. ALSO as I said earlier, source uses vertex lighting, and if you are using any sort of vertex shadowing the first option looks like ASS. for instance, if the upper half is in light and the lower half in shadow then with the first option you get a shitty looking glow appearing underneath the intersecting box because it only has 2 vertex points to shade between. However in the second example it can light the model more accurately.