Low Poly - Stylised Modelling Workflow

Alex1

Hi guys,

Can someone please help me out a bit here? I can't seem to get my head around the workflow for creating low poly game assets such as these:



I've searched these forums and watched a number of tutorials from 3DMotive to try and understand the whole process, but each time, some part of the process seems to be left out or skipped.

So, could someone please help me out by explaining or simply listing the usual workflow for creating assets such as those shown above?

I'm not expecting anyone to reveal all of the secrets involved in this workflow, just perhaps help me fill in the gaps in knowledge. This is what I know so far:

1. Model high-resolution mesh - Ensure even edge distribution across mesh (consistent poly density for good sculpting results).

2. Sculpt fine detail into the high-res mesh (e.g. cracks, noise).

3. Retopologise high-res mesh to produce a low poly version of it.

4. Unwrap low-res mesh.

5. Bake out texture maps (normal (optional) and ambient occlusion) from high-res mesh onto the low-res. ???

6. Produce diffuse (handpaint) and specular maps with the help of the AO map.

7. Apply texture maps to low-res mesh.

I know there's no right or wrong way of doing this, but is this the typical workflow? The main area I struggle to understand is the baking part, so if anyone could share some knowledge in this area I'd really appreciate it.

Many thanks.

AtticusMars

Option A
- Model low poly mesh
- UV it
- Paint textures onto it, adjusting geometry/UVs as necessary

Option B
- Model/sculpt high poly mesh
- Retopo
- UV
- Bake some textures (I'd bake a normal/ao/concavity/convexity plus whatever else you want and some matcaps) to the low poly
- Paint over baked textures using them as a guide (and as masks)

Generally you wouldn't bother making a specular map unless you have a normal map also, you'll just paint your specular highlights into the diffuse. Some exceptions to this do apply (WoW uses specular maps for its terrain)

Option A is (probably) the more "typical" workflow, or at least has been until recently. The main advantage to this is that if you're good at painting, it's fast since you don't have to create two models.

Option B is more modeling work but less texturing work since has the advantage of being able to bake down consistent lighting to the model without the need to paint it. A good choice for people who don't have a strong sense of shape and lighting, and even for people who do have a good sense, if they tend to paint slowly this also might be the faster option.

Option B also has more project specific benefits such as making your work more reusable since your sculpted pieces can be repurposed in other models at different orientations and baked again with =consistent lighting and no quality loss.

Really just depends on how you prefer to work and the particular needs of your project.

pixelb

Yep, you’ve got it. Although the old way of painting directly over your low-poly is still alive and well- it looks like all of your examples were painted that way. There are some artists at big studios (Riot, Blizzard) that will go to the effort of making sculpts just for the AO bake, but it’s still a luxury, not the standard.

From what I’ve observed, it does seem like what maps are baked out and how they’re used is the least standardized part of the process. It’s not just limited to normal and AO, you can also get useful information out of bent normals, cavities, material ids, baked specular highlights, even matcaps…although most people won’t use all of these. Ultimately, what baked maps you use will be dictated by the style you’re aiming for.

Alex1

Thanks for the clarification guys!