ZBrushCentral

All ZBrush Animation Pipeline

Community
#1

Hi folks, after a huge brak, I am back in the 3D world. And still a noob, of course. :lol:

I want to know if anybody uses ZBrush for animation.

What I mean is: does anybody rely on ZBrush only for the modelling of fully animatable characters (ones that will be rigged and animated in another program, of course)?

If anybody is using ZBrush exclusively for he preparation of animatable characters, I would like to know their pipeline, tips, tricks, etc.

I am learing Cinema 4D now (because it LOOKS easy) and I want to make a small animation. Don’t want to do any modelling in C4D, though. I want to do all the modelling in ZBrush.
Possible? Stupid idea? Else?

Thank you!

#2

Ye, people do it all the time. It’s a very common practise. You can use zspheres to block out ur character, even use them to rig in zbrush so you can pose it differently.

#3

Thanks, it is good to know that modelling for animation could be done entirely in ZBrush.

Any particular tips or tricks on how to deal with those edge loops in ZBush and also how to tweak te polygons so not to have problems when you use the ZBrush mesh in the animating program?

Any video tutorials on this subject matter?
Any short animations that use All-done-in-ZBrush models?

Thanks!

#4

The problem is, if you use zspheres or another primative as a base it will have grid topology, which looks fine at 2million polys but it wont work if u drop it down to a low count. So what you n eed to do is retopologize the model so that it can be used in animation. You can do this in your 3d package or use Zbrush

http://www.youtube.com/watch?v=z6eZ-QWv5f4

Theres a tutorial on how to do it in Zbrush. Hope this helps :smiley:

#5

^ This retopology thing sounds cool.

So, all I have to do is create a detailed model in ZBrush without worrying about the poly count and then re-topologize it in ZBrush again?

Do I then use the re-topologized mesh as my low poly in the animation program (C4D in my case) ?
In other words, do I add the bones to the re-topologized object and only use my detailed mesh as a bump map?

Or did I get this wrong? :rolleyes:

#6

No thats pretty much it, you can also use it as a displacement map. This will actually render your high poly model onto your low poly mesh that you have rigged and animated in c4d.

Don’t ask me how to do that tho, I’ve only ever applied displacement in XSI. But you can also apply a normal map using the zmapper tool in Zbrush. Normal maps use colour information so it’s more of a 3d bump map effect. Displacement will add a lot more render time compared to bump and normal which are more or less instant.

#7

Cool.
So I will just create a low poly via retopology and then animate it.
Sounds easy.

How low should the retopologized low poly be?

I mean, is it OK if the low poly is all squared, faceted and with visible polygons and hard edges? Will the normal / bump maps hide those rigid lines when I render the model in the animation program?

One more thing:

I have noticed that when people export bump maps from ZBrush, they lower the subdivision levels down to the lowest one.
I find this strange.
Isn’t it the model’s highest subdivision level that should be exported as a bump map? :rolleyes:

_

#8

The more polygons you have from the original model the better the map will look. So it’s about finding a good balance or optimisation and quality. Your geometry sillhouette is what shows through at the end of the day. Displacement maps will give you the high detailed models geometry. I’ve never really tried it with a really really low model. I always took at least level 2 or 3 out for high resolution image stuff.

Displacement works on alpha so it uses the white as the max distance out and the black and the max distance inwards and everything inbetween. Normal maps and bump maps are just surface textures. So if you have hard edges or a cube for the guys head thats what you’ll see.

As for the subdivision level, I’m not sure why they are doing this to be honest. Say I have a 2 million polymesh in Zbrush, Ill increase the model to level 4 around 100,000 polys, something that wont kill my computer to render but will also give a good result. I delete the lower levels and create my maps. As far as I’m aware this gives a better result but to be honest I’m not 100% certain. I’d have to test it on level 1 to see how it looks.

#9

I hear ya.

When using retopology, is there a risk of losing detail? I mean, yeah, a low poly mesh always has less detail. And it looks like the retopology process is creating an edgeloop-optimized low poly (so it must be less detailed). But does retopology make it even less detailed as compared to the model's lowest subdivision level? Does it "smear" the basic shapes in the surface geometry? What I am trying to ask here is: will I end up with a different model after using retopology? Or does it kind of "stick" tightly to the original polygons?
#10

In my experience with working with retopology tools in Zbrush, as long as you retopologize properly, you’ll maintain the larger details. For example, if I have a character that has a mole underneath his eye, and I retopologize, there’s a chance that that mole might not be available depending on how dense I’m doing the new topology. Alpha stamps for skin pores also are more likely to be lost, but the main points - musculature, length, width, the sense of weight - will still be there. In fact, it’s easier to get better detail in areas after retopologizing, especially noticeable in the head and facial features. With a bad topology, or a grid topology, it’s hard to get proper looking lips without a high poly count, but with the right edge loops on the face, you can have a model that barely hits the hundred thousand mark in polys and still have discernable lips. With those edge loops in place, you can pack more detail in those areas while maintaining a lower poly count. My advise is to get the main forms where you want them, then retopologize, then put in the small details like alpha stamps, textures, and the like.

#11

Normal maps are good for distanced characters, ie. characters that you aren’t going to see from very close up. The rule of thumb is, the closer the character will be, the higher the poly count to make that character look okay. The bump maps also help in those regards, so the character could be a bit closer with the same low poly count than if it had only the normal map applied.

Normal maps and bump maps can’t be taken from the highest subdivision level of the model - you have to go at least one level down to get them. It’s better to take them from the lowest level you can find that still retains the silhouette of the end result model. For a bump map, however, you can simply take your normal map, go up a subdivision level or two, and then take another normal map and convert it to greyscale for your bump map. If you want to work with high levels of detail but don’t want it to bog down your computer as you work, you might want to use a displacement map, which keeps the geometry down until render time. At render time, it will create more geometry to displace the details across the model, which increases the amount of time it takes to render, but allows you to work with the low poly model when positioning, lighting, and so forth.