General advice or tips for low-poly modeling?

coolguyslims

Hello,

I've finished a couple projects with good results, but I see people mention sometimes to maintain good edge flow on low-poly models. Does edge flow matter if the model is not being subdivided? I haven't had any problems yet with what I've been doing.

Are there any general tips to keep in mind when create low-poly props for normal maps? Does edge flow matter for low-poly? Any modeling principles to keep in mind besides the obvious? Just want to make sure I'm not missing anything.

(I'll include some wireframe images of a model I made when I get home)

Thanks :)

Eric Chadwick

Depends. Often it's best to refine the topology to match the flow of the silhouettes.

If the model deforms at all, for example a floppy hat with ragdoll constraints, then topology flow really matters, it's helpful to have good edge flow wherever it bends.

More on the wiki

http://wiki.polycount.com/wiki/Topology#Principles_of_Topology

Also some wireframe examples & animated gifs:

http://wiki.polycount.com/wiki/Limb_Topology

MilkMaster

As Eric put it, the importance of edge flow depends. However, it's probably best practice to have it as good as you can, whenever you can.

As for just general low-poly modeling tips, here are a few I can think of right now:

• - One of the most, if not the most important factor, is to keep the silhouette of the low-poly as close as you can to the high-poly. This of course doesn't mean making a low-poly have a ton of edges, but trying to bake a perfectly smooth cylinder onto an 8-sided cylinder will cause noticeable issues like scalloping. It might take some experimenting to see how low you can go before there's a serious quality drop.
• - If your low-poly includes several crashed objects, try to attach them wherever it makes sense. I know this sounds vague but that's because it depends on many factors (object shape, orientation, size, etc.). For example, if your low-poly consists of one cube crashing through the face of a larger cube, try to "stitch" them together. This will help save UV resolution space and also usually gives better-looking bakes.
• - While this might be more of a preference and there might be times when you don't want to do this, I usually conform my low-poly to my high-poly. This means that when you have your low-poly and high-poly overlapping, the high-poly completely covers the low-poly. You can achieve this using a conform or skin wrap tool that most programs should have. This again I usually do to try and achieve better bakes, but as I said, different situations might call for other solutions.

Eric Chadwick

Great tips.

Re: bullet point 1, check out Understanding averaged normals and ray projection/Who put waviness in my normal map?

Re: bullet 2, here's a great example: http://wiki.polycount.com/wiki/Normal_Map_Modeling#Low-Poly_Mesh