game character workflow

ashleysparling

I have been looking at the Making of the giant tutorial. I was looking for different work flows on how to make characters and I was wounding if this work flow was typical in the industry.


here is the work flow.





the bit that I am confused about is the uvmapping after the zbrush stuff.


can anyone unconfuse me

Marine

where it says retopology it means build low poly

rooster

yeah, uvs don't matter whatsoever on your high poly mesh only the one you bake your normals to

EarthQuake

Generally everything we do is baked and texture before it gets rigged.

Emil Mujanovic

I think it really depends on your workflow. I know of some people who like to unwrap their basemesh and bring it into ZBrush and sculpt to the level they are happy with. They'll then export their lowest division and touch up the UVs to minimise stretching then re-import in the base back into ZBrush. From there they can bake out their normals, AO and whatever within ZBrush.
I personally only unwrap my lo-poly mesh (pretty much the worflow you mentioned in your first post).

-caseyjones

Mark Dygert

EarthQuake wrote: »

Generally everything we do is baked and texture before it gets rigged.

Agreed. The workflow as outlined in the first post is a good workflow for people just starting out and are uncertain how a model will deform when its rigged.

After about 4-5 models (especially bipedal humanoids) you have a pretty good idea how to properly topo joints. Its perfectly fine to UV and do materials then rig.

Also sometimes when you get into doing the materials you find out you need some extra definition or detail. Which means you'll probably need to re-rig. Keeping the rigging at the end, will give you more flexibility to add changes.

EarthQuake

One thing to mention is that its rather common for this workflow to pass through 3, or even 4 people. A modeler who will do the highres, lowres, uvs and bakes. Optionally a texture artist who will finish off the materials(this will be done by the modeler in a lot of cases). Then off to a rigger to create the rig and skin the mesh. And finally off to the animator. So its good to keep back-and-forth to a minimuim here.

ashleysparling

Thank you for your help every one this has made sorted me out. I am going to give this way ago. I was looking the d'artist Character 2 book as well. and it seams that epic game does a similer thing. I saw a talk with Epic game and they used this type of technique but I couldnt rember about the UV mapping.

thanks again guys.

Ash

benclark

hi

when I put rigging in there I meant it in the most basic sense. It was a very basic rig just to get the character out of the modeling pose. Its by no means a rig to use for animation.

The maps were still baked using the tposed mesh, not the posed one. The textures were still painted using the Tposed mesh too.

I usually just rig the character and save that file for use at the end but it definetely doesnt have to be in that order. If anything it was in that order just to get it over and done with because my rigging ability is arse. Autodesk just asked me to write out everystep I took in the making of the model, but its not really supposed to be a tutorial on how it should be done

ashleysparling

wow Hello Ben. Thank you for posting on this post. I have to say a very good making of. I was just looking at different work flows and I like the look of yours but was confused on a few bits but thanks to every one that is all sorted. thanks again

Ash

danpants

[quote=rooster
yeah, uvs don't matter whatsoever on your high poly mesh only the one you bake your normals to]vBulletin[/quote]

Sorry this has been bugging forever. I keep hearing people saying they create a low res version after they finished the high res version. How is that possible? Don't you need the UV's from the high res to match the low res model? If you create a new low res model doesn't that alter the normal map on the high res?

Can someone explain that please. Because I usually UV my model first then take it to Z Brush and then use Z Mapper to generate a normal map. But I don't want to do that, I want to work on the high res then make a low res version, but then how do I bake the normal map without messing up the high res? I know I have to create a low res model and then UV it. But then how do I generate the high res with no UV's to the low res with new UV's?

Kovac

Create the hipoly off of zspheres or a base mesh then find a good, Max (or other) friendly subdivided resolution that maintains the basic forms then export it out and build your lowpoly topology off of that. Unwrap that sucker then go back into zbrush to bake.

danpants

ok Let me get this straight.

I got the export a good high res out and start building my low res. But does that mean I start to delete edges and reduce my polycount from the exported high res or does that mean start a new mesh completely from scratch over the exported high res?

pior

Whatever works for the specific case you are confronted with.

Vailias

Dan: the reason it works is that the low poly UV layout is what is used to generate the normal map. The normal map is formed via projection through the geometry of the low poly onto the high poly and returning the surface normal at the location of each pixel in the map size you choose. You can map whatever you want into a normal map, the mesh doesn't have to be even close to the low poly, but it will just look odd. The thing being mapped just needs geometry, not a uv layout.

danpants

Thanks for you help.

Thanks Vailias much appreciated.

So to elaborate , all you really need is a low polygon mesh with UV's and then you bake the normal map from the high res (no UV's) to the low res model.

Thanks for clearing that up I always wondered what people meant by building a low res mesh.

I want to change my workflow that is why I am asking. I rather jump to the high res since starting with a low res takes me too long.