Zbrush High-Poly to Cinematic Animation Workflow Question

DallanP

Heyo!

I know how to go from a high-poly down to a low for real time assets, but I am wondering how it works for cinematics? Do you still bake down to a lower poly mesh for animation/rigging? Not using normal maps but instead displacement/height maps? I've watched a few vfx breakdowns where they show the actual mesh, and its still pretty high, so I understand that part. The most confusing part for me is how to maintain the super highpoly detail down to the lower mesh for animation. I know this a pretty big question, if you all have any great tutorials that are worth watching/buying that would be amazing. There are so many tuts out there its hard to know which ones are good and worth my time.

Even a simple bullet-ed list of each step you take to get from high poly down to a lower mesh would be helpful.

Thanks!

kanga

It would depend on what application you are doing your cinematics in. More and more higher frequency game models are being used for some pretty spectacular realtime engine renders. Unreal has now 'sequence' for this purpose. If you plan on making a cinematic in standard 3d software I would recommend using a higher poly model than you would in games and use displacement mapping (bulk detail) and normal maps (high frequency detail) so you  can switch them off when animating and on for renders. Just any displacement map tutes should get you up and running.

I hope someone here who is actually at a studio that is producing cinematics right now can give you more detailed advice.
Cheerio

3levan

Bump as Im also trying to figure this out.
I've been using ugly low poly proxy meshes, no normal map or displacement.. I can only assume a lot of VFX breakdowns for movies are still just choosing to show us higher poly renders. Looking at animator portfolios for film they often just have these really basic proxy models to work with.

More information would be nice!

Chimp

Open subdiv?

musashidan

Basically, it's the density of the topology. A good example is Drake from Uncharted 4 . Game res mesh 30k tris. Cinematic res mesh 90k tris. Also, because the vfx/cinematic rig is more complex and allows for a much broader range of motion, the topology has to support that. Slapping on subdivision isn't going to do it as the underlying topology won't support the deformation.

But the real-time/offline gap continues to shrink.