So, I was wondering how you get an image to tile across a certain part of a model.
For example, if you have a model of a house, you may want an image of carpet to tile across the floor.
Would you make the floor a separate object from the rest of the house, tile the image across the floor, uv map the house, and then join the objects together?
I just don’t really know the best way and was hoping someone could help out.
The easiest was to do tiling that I know of is to overlap all of the polygons into the same UV space.
This also works for making symmetrical things. With a symmetrical object, you can mirror the UVs, and again, overlap the UVs. Except one half will now be oriented “backwards”.
I get what you are saying. I usually overlap the polys for things that are symmetrical. But, for example, the floor of the house model was one big polygon. Maybe a rectangle. Then there would be nothing to overlap.
In that case, I’d set the corner of the quad to be at the four corners of the UV map. Then I’d pin the UV at (0,0), and scale until you have the amount of tiling you want. This will go far outside the 0-1 UV space. I’d probably just leave the floor as a separate mesh, to avoid wasting one of the mesh’s UV maps. Otherwise, as far as I know, you can’t use a lightmap.
Hmmm. So far it seems the best solution is just to leave it as a separate mesh then.
Is there a problem with that? If you’re on a full-fledged computer, that’s only one or a handful more draw calls. No big deal. If you’re on the iPhone, you can use static or dynamic batching. In the 3D software, you can parent the rest of the house to it, or vice versa.
Right. And no, there’s no problem. Didn’t mean for my post to sound like there was. 
I figured there was some sort of trick I was missing, but I suppose not!