Sculpting a photorealistic human face for animation: use the multiresolution modifier?

Question

I'm locking down a production workflow for sculpting photorealistic human faces for animation but having a conceptual problem about whether or not to involve the multiresolution modifier

the sculpt workflow is: character sculpted in actual real-world clay > form captured via photoscan to an ultra-dense triangulated mesh > clean-up, sharpening, additional details and skin textures done in blender > mesh retopologized to lightweight animation-friendly all-quads > wrinkles and details recovered via normal maps and displacement maps plus a subdivision modifier

there seems to be no reason to use the multiresolution modifier in this process - is there a strong reason to use it that I'm missing? - it's a very powerful modifier, it obviously must have some important uses

a related issue: I have sometimes had difficulties with data loss when relying on the multirez modifier to preserve high rez sculpt work, especially when trying to save/restore multirez data using the .btx format - I suspect I'm not understanding the place the multiresolution modifier occupies in a professional sculpt > animation pipeline - can someone clarify?

Answer 1

From what I understand about the development of blender and the workflows that are out there, the multi resolution modifier is no longer frequently used, and for good reason. I think it still has uses, but overall it's no longer an essential modifier for workflows such as the one you describe.

The multiresolution modifier was, for a long time, the only feasible way to sculpt in Blender. Manually creating a dense mesh or using the subsurf modifier (applied) was an option, but these are destructive methods that lead to objects that are so dense they are difficult to edit.

With the addition of dynamic topology, there is now a way for an artist to create highly detailed objects without this modifier. In your case, a scanned object replaces this, but the results are the same: You and up with a dense triangulated mesh, and the workflow you outlined is a great way of turning that into a final product.

Since dynamic topology is much more free from an artistic point of view than the multires modifier (you don't have to worry about topology at all) it's no longer used much for initial sculpts as there are better options out there.

However, the multiresolution modifier still has a place, and that place is in adding details to modeled meshes. Not everyone is a sculptor, and sometimes hand modeling an object is the best way to go about it. If you then want to add details, this is where the mutiresolution modifier is great. It allows the artist to add details non destructively, and gives the option of baking this detail to maps. Using an applied subsurf modifier or a dense mesh won't allow you to do this, as any detail you add is practically permanent.

Finally, the mutiresolution modifier can still be a good way to sculpt an object from scratch, provided the topology is OK, or you start with a simple model. All in all, this modifier has it's uses, mainly in workflows where a mesh with good topology is the starting point.

Answer 2

I'm learning sculpting myself right now, and I struggled with the same question! However, the multires modifier is a powerful tool and shouldn't be overlooked. To clarify, Dynamic Topology should be used for concepting, and the Multires Modifier for finished meshes.

Dynamic Topology is flexible and easy to use; for conceptual sculpting, it's the way to go. It acts the way zbrush might, and kinda like real life (adding clay to places you need more). However, it's gonna mess up your mesh!

The multires modifier, on the other hand, has the advantage of being able to toggle on and off, and it plays really nicely with other modifiers and vertex data. This becomes more helpful when you have a finished model with nice topology that you want to add detail too, because it's completely non-destructive.

So for your workflow, you might use: real-world clay sculpt > photoscanned to a dense mesh > mesh retopologized to animation-friendly > wrinkles, sharpening, additional details, and skin textures done via multires-modifier

And then just turn off the multi-res modifier during animation. This would let your details be 'real' geometry, and you wouldn't need the normal/displacement maps. The multires modifier is certainly more powerful/flexible than subsurf+displacement maps, but it's up to you if you need that extra flexibility.