I’m quite new to doing modeling, and I’m following the tutorial up on Blender Cloud about making game assets. I’m trying to do the normal map baking but things aren’t quite turning out right.
This is what I’m getting:
It’s all coming up quite pixelated, and I suspect it’s because the texture is 512x512, and every part of this model has it’s UVs squeezed onto the same map.
Assuming I’m correct to believe it’s a matter of resolution, what should I be doing instead? If this were a model designed to be released to TF2 (the goal of the tutorial), or a model to be used in the Unreal Engine, would it be practical to have multiple UV maps for the different parts? Or perhaps just split the parts to two maps…
(practical as in, won’t blow the budget allowed for a small asset in the game like a hat)
You can split the UVs by material instead if needed. I don’t think that’s needed for a model like this though, still, 512x512 is pretty small, you can make it at least 1k or 2k if it doesn’t look sharp enough.
Also, for the normal map, do you have a high poly model with more detail? If you’re baking a normal map from your 39 poly model you won’t get good results. The normal workflow is to bake the normals from your high poly model to the low poly model so it looks almost as good but has lower poly counts.
Yeah that’s what the tutorial has me doing. The high poly is the the same model but with the edges beveled smooth. I’ll try these solutions and see what works then, thanks!
Checking “Normal Map” under Image Sampling smoothing things right out.
I also snapped all my UVs to pixels, which helped reduce banding near the edges of the faces. However it didn’t eliminate the iffy normals right where the edge of the faces are. You can see what I mean here http://i.imgur.com/8dejKz9.jpg