ZBrushCentral

Dynamesh resolution problems with very elongated form

ZBrush Usage Questions
#1

Hi Friends!
So as the title says I’ve been stuck with a problem with Dynamesh, I’m doing a very elongated beast and I need to extend it more (like 3 time or so). Each time I refresh it, the polycount decrease drastically (from 800.000 to 100.000) no matter the resolution or the size of the mesh.
As I said I’ve tried to increase both the size of the mesh and the dynamesh resolution. I even tried to resize it in a third party software, and freezing the transformations.

So do you, gals and guys, have a good solution? or a workaround?
The only thing I’ve thought so far is separating the mesh in 2 subtools but it’s not a very practical nor elegant solution :confused:

For the forum admin, sorry if I’ve begin a thread in the wrong section but I couldn’t find an appropriate section to ask this question. And I’ve had no luck searching for an concluant answer in the forum either.

Thanks!

#2

Did you try to scale your object down?

#3

yup I’ve tried it too…well in fact, when I hit the “deformation>unify” button the mesh is scaled down. Then I have refreshed it without any success…

#4

You have to understand that dynamesh is voxel based and thereby limited by the bounds of the voxel grid which implicitly encloses your geometry. When you elongate your object, zbrush adapts its dynamesh(voxel) grid to this new bounding box size, but it does not increase the resolution of the grid. When you then remesh your object, the amount of voxels which score hits on the grid decrease, because in relation to the geometry, the grid’s voxel size (not resolution) has increased. The result, depending on how much you’ve gone beyond your initial bounds, has less polys than the original. You can mitigate this somewhat by increasing resolution of the dynamesh before remeshing, but this only goes so far and is still limited by the maximum dynamesh resolution. At some point, you have to turn dynamesh off and then go forward with your sculpting with the standard geometry tools so you can maintain your resolution. That’s why pixologic calls Dynamesh a base meshing tool and not a final sculpting or refinement tool. What would be nice would be some sort of visualization tool, so you could see your mesh’s interactions with the voxel grid more directly and be able to predict when this sort of problem might occur.

#5

Thank you both for your answers !

Yes, I think I see a bit clearer in the dynamesh logic now, thanks to you, nimajneb.
Still a question remain.
You said :
-“When you then remesh your object, the amount of voxels which score hits on the grid decrease, because in relation to the geometry,the grid’s voxel size (not resolution) has increased.”
So if the size of the voxel grid is dependent of the initial mesh’s shape, if I import the mesh in Zbrush, after being elongated, in a new workspace and hit dynamesh, won’t Zbrush create a new grid according to the new bounding box?

Hope I’m clear…I’m always unclear even in my mother tongue :wink:

#6

dynamesh_triage.jpg
I think I spoke imprecisely when I said “bounding box.” Just changing the size is not enough to cause this recalculation of the voxel grid. It has to be beyond some specific point. I did some tests (see image above) using a unit cube. I elongated the cube using the Size X deformation repeatedly, then dynameshed to see if the resolution of the resulting mesh shifted. It held steady until the mesh had been elongated 6 times, then on the seventh, there was a dramatic shift in resolution. So it’s not simply that the bounding box of the sub-tool is changing. It’s that it exceeds some size specified within zbrush itself, this causes the recalculation of the voxel grid. Unfortunately, there’s no way to say when this will happen. As this test shows, mesh density was pretty much stable until the 7th Size X deformation.

So, if you brought your mesh in from an external editor after elongating it there, what the resulting resolution from your dynamesh would be totally dependent how your mesh got associated with the voxel grid on import. I can’t really predict or say exactly what that relationship will be. You will just have to test and see if you lose excessive resolution.

#7

Thanks a lot for your time !
Your explanations were very clear ! You seem to have a knack for it, I mean it.

I’ve tried to export and elongate the mesh in another soft wihtout any better success. It does not really matter though, as I said in my first post, I’ll just have to cut my form into two subtools, detail it as much as it is interresting to do so in dynamesh then merge it classicaly and lastly do a nice and homogeneous retopology to fine tune the model.

Again thanks for this good post.