Serious trouble understanding the workflow "Low to high to low Poly"
Hello,
So I've made this human like character in 3ds Max, retopologized it, so it has a clean topology. I then Unwrapped it's UV's to two different maps. One for most of the armor parts, helmet and stuff, the other one for the links between the limbs so to say. In total I have quite an amount of different small objects.
My Understanding (for sculpting details) would be to bring in all objects per UV Map (Meaning two different sets of objects) into zbrush, sculpt them, bake the normal map , and aplly the two different normal maps to the two sets of low poly objects later on.
So I creased all neccessary edges of each set, and imported them into zbrush. If I divide the objects (of the first set e.g.) now to about maybe 3-5 million polys, then not every piece has enough poly to smoothely sculpt on.
If I split by polygroup and divide individually, I can sculpt exactly as I like, but If I later on merge the subtools I get this massive 12-20 million object, which I cannot export as obj for baking.
If I export every subtool on its own, I dont know how merge the different normal maps, I would get.
I'm kinda lost on what to do now..
I know its much text
but it would be really cool, if someone could help me out here.
Cheers!
So I've made this human like character in 3ds Max, retopologized it, so it has a clean topology. I then Unwrapped it's UV's to two different maps. One for most of the armor parts, helmet and stuff, the other one for the links between the limbs so to say. In total I have quite an amount of different small objects.
My Understanding (for sculpting details) would be to bring in all objects per UV Map (Meaning two different sets of objects) into zbrush, sculpt them, bake the normal map , and aplly the two different normal maps to the two sets of low poly objects later on.
So I creased all neccessary edges of each set, and imported them into zbrush. If I divide the objects (of the first set e.g.) now to about maybe 3-5 million polys, then not every piece has enough poly to smoothely sculpt on.
If I split by polygroup and divide individually, I can sculpt exactly as I like, but If I later on merge the subtools I get this massive 12-20 million object, which I cannot export as obj for baking.
If I export every subtool on its own, I dont know how merge the different normal maps, I would get.
I'm kinda lost on what to do now..
I know its much text
but it would be really cool, if someone could help me out here.Cheers!
Replies
If I have to merge a few normal maps, I do it like this. First I bake all of the pieces.
1. I make everything white from the normal map except the island that I want
This way when I copy the whole image the island will be in the right place
on the UV Map and I can just delete the white.
2. CTRL-A, CTRL-C
Now I have the whole layer copied
3. Paste it on top of my my normal map, then delete the white out of that layer
4. Rinse and repeat until I have all of the islands
5. Merge it into one normal map layer (optional)
Use Decimation Master on it first to cut the size down.
If you're baking in a program like xnormal, you can have several individual models loaded up at the same time when creating a single map.
If you're baking in zbrush (not a recommendation), you could use the MultiMap Exporter plugin to do the same.
You don't need to merge all those separate highpoly parts in zbrush, just export them into a single obj for baking.
xNormal supports a cage but you have to import it separately from what I remember, max makes it automatically and you can tweak it right in max, with xNormal you have to go back to your 3d app tweak the cage, export, import THEN bake. With max it's just tweak and bake.
In max I also bake custom lighting setups into the models, I can bake ]head to toe gradients into the models.
See how the pyro has an overall gradient applied to it? Darker on the bottom than on the top? These are really easy to add in max. Which is great because I'm not trying to paint that into a jumbled up UV map with pieces going different directions.
Wouldn't that destroy my initial sculpts then? I don't understand the point.
Awesome! So I just export every High Poly Subtool and fill in the hp and lw in xnormal in the right order and it automatically bakes it to one map?
EDIT: Don't mind the "right order".
How? I just found the way with merging them first..
The point is a decimated mesh at 200k could look as good as a 3 millions tris mesh and will be easier for max or xnormal to handle.
I don't wanna sound rude, but I find that hard to believe.
I think the normal map of the sculpted detail of the x mio. mesh would look different, then the normal map of the same sculpt decimated to few hundred k. Wouldn't it?
Enlighten me please
EDIT: Okay, sorry if I DID sounded rude. I just tested a decimation to 10% of a ultra fine sculpted detail and I literally see no difference at all switching back between 4 mio. and 400k polys. Not a single bit. So, thanks for the tip on this front
If I use the decimation master on one of the subtools, then the subtool's position (I think scale too) is not the same anymore. It does not fit to the rest of the model. How can this be avoided?
EDIT:
Seems to be the set pivot option, not the decimation master at all..
You can decide how much you want to decimate the model. You're not going to notice a big difference (depending on the model) if you go from 2 million polys down to 1.5 million right? But at some point, yes, you'll begin to see a difference.
xNormals does a pretty good job of handling high poly obj's, you shouldn't need to worry about losing too much detail.
If your computer can't handle a few hundred thousand polys, thats a bummer. It means you're going to have to sacrifice some detail.
Normally what gets hacked out is nothing but fluff that isn't actually contributing to the tiny details at all but was there because the mesh was subdivided like crazy to get detail in another area.
Obviously if you go too low it starts biting into the detail but anyone who's used has been impressed with how well it works.