ZBrushCentral

How to Combine Multiple Body Parts into Single Seamless Mesh

ZBrush Usage Questions
#1

Hi there,

I was reading someone’s issue of 3D World Magazine at a freelance job, and I came across a very cool article . It had a nice solution for if you have modelled each body part individually, and you now want to Combine/Sew the meshes into one Body Mesh.

I can’t remember exactly how it worked, because I only had time to briefly look it over . I remember they suggested basically modelling each part up to high detail, then create discplacement maps for each part, then you drop to a lower SubD level and export the low poly body parts into Maya or MAx or whatever, then you sew the low poly model together . You then take the low poly model back into Zbrush and apply the displacement maps to get your high detail back onto your model.

This makes sense to me, but I am definintely at a stage in learning Zbrush where I need a more detailed explanation to get it right .

Can anyone point me to a tutorial of this kind ??

Any help greatly appreciated,

Best Wishes ,
essclock

#2

The problem that you’re most likely to run into is that when you get back into ZBrush you’re going to have a lot of overlapping UV’s. So the first thing to do is when you put the model together in your other app you need to also make sure that each part is a different group. Then in ZBrush, here’s what you’d do:

  1. Import the model.
  2. Turn on Polyframe. You should now see different colors for each group. Those groups each correspond to a different displacement map.
  3. Subdivide the model enough to support the displaced details.
  4. Ctrl+Shift+Click on one of your groups. This will hide everything else.
  5. Load your displacement map for that group into the Alpha palette.
  6. Apply any blank texture to your model.
  7. Turn off Transform>Quick3DEdit
  8. Set Tool>Display Properties>DSmooth to 1
  9. Turn on Tool>Displacement Map>Mode
  10. Set the displacement Intensity slider to the value found under Alpha>Alpha Depth Factor
  11. Press Tool>Displacement>ApplyDispMap

Your displacement map will now be converted back into geometry in the model. Now Ctrl+Shift+Click on the canvas to show the entire model again. Repeat steps 4, 5 and 9-11 for each remaining map.

Note that this will only work well if you didn’t subdivide the original parts to the maximum possible amount before you detailed them. Let’s say that your model has five pieces and each was at 2 million polygons when you created the displacement maps. That’s a total of 10 million polygons. You’re probably not going to be able to get the combined model to that high of a subdivision level, and may have some loss of detail as a result.

#3

Hi aurick,

Thanks for the detailed explanation . It makes a lot of sense to me :smiley:

Let’s say that your model has five pieces and each was at 2 million polygons when you created the displacement maps. That’s a total of 10 million polygons. You’re probably not going to be able to get the combined model to that high of a subdivision level, and may have some loss of detail as a result.

This is exactly what I was thinking about the past few days.

I am going to end up taking these models into Maya for Animating and Rendering. So, do you think this workflow would make sense ?

Model each part up to a high SubD in Zbrush and export Disp Maps for each part.

Then drop to a lower SubD and export to Maya.

Sew the parts together in Maya and export back to Zbrush ( Using a UV Layout where each body part has its own UV group like you suggested ) .

Apply each Disp Map to each body part ( After maybe going up 1 or 2 SubD level just to smooth out the shape ) .

Export Disp Map again as one big map .

Go back to Maya and apply the big Disp Map to the lower res model and Render.

I am not sure if that all makes sense. It is something I was just thinking of after reading your post. Do you think this could work ?

Thanks again for the quick and helpful reply.

Best Wishes,
Essclock

#4

Why go to all that trouble?

Instead, use Maya to lay your UV’s out in multiple regions. Use ZBrush to create 32-bit floating point maps (which is done using MultiDisplacement2 – part of Displacement Exporter). Then apply those multiple maps to your model in your rendering engine.

This allows you to use multiple maps, which are more memory efficient for your computer to handle and also allow greater overall detail. After all, you can get much more detail onto 5 (or however many) 2048x2048 maps than a single 4096x4096 map. And even 4 such maps are easier for your computer to work with than a single 4096x4096 map.

#5

aurick,

Thanks for the great info ! Makes perfect sense . I can see I was thinking too complicated. Gonna go try it now :smiley: u rock

Best Wishes,
essclock

#6

Dear Aurick,
I just have read the process of uniting meshes into one and thought if it would be possible to
compose a script (plugin) that would automate all steps into a coupla mouse clicks on the script interface buttons.
That would be great, wouldn’t it?

Anatom:D

#7

I’ve actually tried this multpiple ways, and every way ends up with seams. There is obviously a way to do it cause I’ve seen many models using the technique, but there is definitely a problem or step to get it seamless. Im rendering in mental ray which could have steps I dont know about, but I’ve rendered normals and displacments fine. The only problem is applying multiple maps to a single model, whether it is the original mesh or the multiple meshes stitched back together as one mesh. Any help is greatly appretiated.

#8

hi,

I am a newbie and am wondering that will the model with the displacement map has a bit different from the high detailed model in Zbrush? (I am not familiar with the displacemap.) :rolleyes:

thanks

#9

The way I thought of doing this (trying to recall from a quick glance from a tutorial by a user to connect an upper human body to an animal body.). Not sure if it might take time or not be perfect. But to start, merge all the parts into one subtool (still separate meshes) while at lowest but similar subdivision levels where they near match. Also make another tool with them all merged at their highest detail.

Take the merged low detail as a rigging for a zsphere, then go to edit topology, then hold down that key (forget which) to brush the topology lines of the rigged low poly model into the new topology and stopping at where each body part connects. Then manually do the rest of the topology to connect them togethor. Make a new polymesh from that topology with no subdivision.

Then use the merged high poly tool as rigging for a zsphere this time. Then under topology, choose append or import to grab the new tool you created with the parts you just manually connected with topology. Then basically try to project the detail of the highpoly into this new one piece mesh. Will require some settings to prevent polys from jumping out.

I suppose there could be a bit of a pinch where the parts connected. But that shouldn’t be too hard to fix.