XGen tube based grooming... nothing happens when I click "test"

[Anonymous]

I select my scalp geo, create a new description, load all my tube geo into the widget and click test, but nothing happens.  No x's, no error, nothing.  Still reads "??" next to guides.  What am I missing?

[Michael_Todd]

If you look at the XGen Log, it will it reports any issues found with the tubes. They need to be closed around the tube but with open ends with valid UV's. The UV direction determines the Guide root and tip.

I'll have to check, but it maybe that the tubes cannot be in a namespace, though.

Cheers

[Anonymous]

I tried it with just one piece of geo from the bangs and the scalp mesh; the bangs geo has UVW.  Same issue...

The errors I see in the log are "XGen(2):  No boundary edges found." and "XGen:  Errors when creating guides from poly."

The geo of the bangs is open at both ends.

[Anonymous]

Ugh, this is driving me crazy...

I've narrowed it down to an issue with the geometry, I think.  I created a cylinder, gave it auto uvws and was able to add it to the groom tool widget and generate guides, no problem.

The problem is, I created the original hair geo and uvws using Max and Zbrush, importing it into Maya as an obj -- these are the programs I'm comfortable with -- I'm completely new to Maya and was only planning to use it for hair/fur and look dev, since the only tuts I can find for the look I'm after are all Maya.  So, I'd rather not have to completely redo all the hair geo in Maya, since it'd take me forever.

I've tried freezing the transforms and deleting the history on the imported geo (to the best of my knowledge), but I still get the error "Errors when creating guides from poly."  when I click test on the widget.  I've tried ungrouping the geo too, in case that had something to do with it somehow, but I get "Can't ungroup leaf-level transforms", which, for all I can tell searching around, means it's grouped or parented somehow -- but deleting the history makes no difference.

Does anyone have any ideas?

[Michael_Todd]

If all the tubes are separate  but are all consolidated to one mesh, use the mesh/seperate tool to make all the tubes back into individual meshes. The tubes can be grouped but they have to be separated from each other,

[Michael_Todd]

The tubes need to be closed around the circumference. It's not a tube otherwise, but a folded plane.

cheers

[Anonymous]

The tubes are all separate.  They should all be closed around circumference, since they were modeled from boxes...  at this point, I may just give up and try to recreate with spline guides.

[Michael_Todd]

What errors does the XGen Log report when you hit test? There must be a problem with the geo or the UV's The UV's drive the root to tip orientation so that the guide bases are at the surface and the tips at the ends of the tubes, so the UV directionality is important. If there are unmerged verts at border edges around the tube, that will also causeissues.

Also, have you tried removing the namespace from the geometry for the tubes?

[Anonymous]

Well, these are the errors I get:

I selected all the verts on the groom tube and merged them (I think), but I don't see how they could not have been single verts to begin with.

Finally, I checked the UVWs and I think they're correct to the best of my knowledge (I've tried them this way and the other way... narrower side is the tip):

Thank you... hopefully this sheds more light and helps you help me.

[Michael_Todd]

Setting the log to 2 or higher will give you detailed info on what happens as XGen evaluates the tubes as it tests them.

The tubes can't have caps either, they have to be open. 

 Could you share the scene or the geo?

Cheers

[Anonymous]

Sure, here's the scene.  Thank you.

[Michael_Todd · Accepted solution]

OK, I found the problem with the geo

The tubes have to be open ended and only have quads. There are tri's at both ends of the tubes that are causing the evaluation of the tubes to fail:

Another thing you might want to consider is that the tube goes into the scalp, which will lead to guides that dive into the head. It might be worth cutting off the base of the tube so that the first span is on or above the surface. The bases of the generated guides will be snapped to the surface.

removing the first Span and snapping the cv's to the scalp (in live mode) gives a good clean set of guides

I should have pointed out the Quad requirement earlier.

Cheers

[Anonymous]

Right you are.  Funny thing is, I saw those tris when I was cutting off the ends and in the back of my mind thought, "You know... those might be an issue..."  Obviously, that idea did not stick.  Time to fix the rest of the tubes.  Now comes the challenge of getting it to look right.  Thanks a bunch for your help!