ZBrushCentral

Advice on baking normal maps

ZBrush Usage Questions
#1

I’ve been using zmapper for a while and in general am very happy with it for single things - It’s brilliant for doing a body, for example.

What I’d like some advice on is how people go about baking multiple objects (such as a woman, her hair, and, say, her clothing all of which are different sub groups) into a normal map.

Looking through epic games’ work flow I see they tend to import the resulting high poly meshes into 3dsmax (or another 3d package) and use a projection bake onto a low poly mesh. In practice I’ve found this tends to create less than ideal results the more sub objects I’m pulling in. Firstly it often creates a normal map that needs a LOT of fixing up (especially with any part of the model with hard edges such as armor, belt buckles and so forth). Secondly I find any areas where two objects push through each other (hair on the head for example) often can lead to a lot of weird artifacts as can tight areas, turns and so on (which aren’t nearly as much of a problem in zmapper renders).

I can apply a projection map in zmapper from a high poly source to a low poly equivalent but the results are much the same as I describe above.

I’m wondering how people approach this problem and what folks regard as a decent workflow. If, say, you’re making the aforementioned woman do you render out her skin subtool, her hair subtool and her cloths subtool as seperate passes? One big one? In zbrush or in your 3d package?

Any advice would be most welcome! :slight_smile:

#2

I get the best results with zmapper.

If your highres model’s polycount can stand as one tool I recommend making your normal map as one tool.

If you got a lot of highres subtools instead ,
I recommend making a normal map for each subtools since they’re gonna retain their respective places on the Uv sheet it’s easy to merge them in Photoshop afterward.
And having each of your parts as a Photoshop layer also got the advantage of allowing to tweak your maps on a subtool-based level ,
if you are the kind of person that modify your normal maps using Ps or Crazybump ect…

At least this is the way I work.I think a lot of studios
work using different pipelines.You have to adapt to wherever you’re workin.

#3

Aye - I could see that working. However if you look at the typical low poly mesh it might go across large subtool subsections. So you might have a few polys which defines a backpack, rope and bottles but those would be a number of subtools. So would people use a zmapper projection to bake first the backpack, then moving outwards the bottles and so on?

#4

Yeah exactly.
But the order doesn’t matter.
It can seem tedious but it’s something you can do while preparing the children lunchboxes.
That would be cool if we could merge subtools while retaining their subdivisions levels to be used in zmapper.

Still,you can save you some work by just planning it a little.

EX:You have a belt with some accessories hanging on it.
Here’s what I mean by planning:
when you’re at level1 and didn’t even start sculpting
decide which parts will be able to stand together at their highest level.
So let’s say if you plan that the belt and all it’s accessories won’t go above 10 millions polys altogether and your machine is able to handle that count then DON’T EVEN MAKE SUBTOOLS WITH THEM AND USE POLYGROUPS INSTEAD then, when you’ll go in Zmapper you can generate a map for the whole bunch.So that would mean using only subtools when you’re forced to do so.

But really, doing it on a per subtool basis is not that long.