Help diagnosing mesh issues

I am getting closer to accepting a loss on this model i’ve been hard at work on in the hopes of getting it to 3d print via Cura 5.1. This project was one of the first things I’ve made during my 1.5 month self taught zbrush journey so far so I understand that it may be full of mesh issues I am not experienced enough yet to comprehend. I understand the concept of making models manifold, and 3d printers needing things to be designed in certain ways so i made sure to closely comb over the parts to check and reinforce anything that wasn’t attached to something else. I’ve used zbrush’s mesh integrity option, fill holes, and even tried some 3rd party options like ms 3d which has a auto-fix mesh option, and tried mesh mixer which was making things unnecessarily complicated, and still Cura is warning me that there are mesh errors and printing may be a risk.

I will provide the images below of the model I am speaking about for reference.

When I was making this, I wasn’t using dynamesh or subdivisions much because I didn’t want to alter the geometry, and specifically the sharp lines caused by manipulating IMM brushes. I wonder if there is not a better way to be considering my issue here? I have decimated the final merged subtool, and can’t really use any zremesher or dynamesh as it destroys the intended final design results. Can anyone help me with my headache here? I have been trying for the last week to figure out if this should just be scrapped and I should move on, or if I can somehow maintain this, get the mesh corrected but keep things preserved and intact?

Screenshot (7) Screenshot (8)

You haven’t mentioned physical size but in any event the model looks incredibly complex if you want to print with an FDM printer (Cura is an FDM slicer). You might need to consider printing it in pieces. Even printing with a resin printer would be difficult although they’re better suited for high detail.

Your best option for making a single water tight mesh is to use boolean, either Live Boolean if you have many subtools, or Gizmo:Remesh by Union if there are many individual pieces in a single subtool. Boolean will not change any topology except where pieces overlap so you won’t lose any detail (which can be a problem with Dynamesh).

Make sure each piece is a valid solid for best boolean results. Weld Points, Close Holes and Fix Mesh used in combination should ensure that. Then run the boolean of your choice.

When you have the booleaned mesh (Umesh_) then run Decimation Master to triangulate the mesh and reduce the polycount. It does a great job of retaining detail where it exists. You will not lose any of your sharp edges and you’ll have a nice triangle mesh which can be exported as STL for printing via the 3D Print Hub plugin. Make sure to set the proper scale, either before with Scale Master or during STL export with 3D Print Hub.

Good Luck.

thanks for the reply, so I have done nearly everything you suggested, before merging I doubled or filled the subtools and corrected any spaces between parts, merged down, ran fix mesh, close holes, used a dummy append to create a new boolean mesh, deleted dummy, decimated down, and exported to Cura for printing. Cura is showing me some errors in areas that would be impossible for me to go back and re-do without starting from scratch, I get the same error message after using MS 3d’s built in mesh-auto correct as well. So I was wondering if there is any alternate way to remesh everything and keep the original intact? I just struggling how to word this correctly to even look up a solution online.

I should note does decimating before mesh correction make a difference?

OK, just stating the obvious, but have you checked that none of the subtools/parts are inverted i.e. normals pointing inwards? Disable Display:Double and check. Flip if required. It may also be that a few polygons in a single subtool got flipped during the modeling process so look out for that.

You could try Autodesk Meshmixer for mesh repair. I’ve found that pretty robust.

Bear in mind that any app saying there is a problem could also be a false positive. I’ve had meshes that reported OK in Zbrush, invalid in Meshmixer and valid in my slicer.

Don’t know if decimation before mesh correction would make a difference, although all polygons would be triangles, and therefore somewhat easier to treat mathematically. Worth a try. It’s just CPU cycles.

Another option would be to decimate subtools before boolean. A reduced polycount and simpler polygon shape (triangles) might help with the boolean process. If all your subtools are merged into a single subtool, decimation will keep the individual parts separate. Then just Gizmo:Remesh by Union.

Good Luck

Hi @ndsugi ,

There are two separate issues here for the purpose of print. Issues with the geometry (non-manifold, not watertight, etc), and issues with the “physical” form of your mesh.

I concur with @tobor8man–the mesh in the form you have on display here will be problematic on a fundamental level for an FDM print process for reasons completely aside from errors in the geometry. In particular the open sections of mesh that are made up of long, thin panels of geometry seem impossible to print as a single piece in the way they are now.

To be clear–even if you successfully fuse all this geometry together in the form it is now in ZBrush or otherwise, the mesh will still be problematic for print. The basic form of this mesh, not just the geometry, needs to be consolidated far more than it is now for the purposes of print. OR the mesh needs to be broken up and printed as a series of small individual pieces that will be fitted together after printing. Prepping a piece like this for that sort of process would be a daunting task.

For the purpose of print output you need to work towards closed, watertight solids with enough thickness for your printer’s capabilities. You also have * a lot* of deeply inset undercuts and overhangs that would frankly be a nightmare to create support for. You want to be working towards chunky, single piece solids with a level of surface detail that does not create overly complicated, deep cavity recesses.

Setting all that aside however, if you produce a mesh of sufficient thickness in ZBrush that is a closed, wateright volume, and that mesh reports no errors with Tool> Geometry> Mesh integrity> Check Mesh, chances are good that the geometry itself will print alright even if your slicer reports minor geometry errors. Those programs can be over sensitive.
Watertightness can be checked with the “Analyze Selected Subtool” report in the Transform palette.

That is purely in regard to errors in the geometry (non-manifold, not watertight, etc). Your slicer could be warning you about issues with the form of your mesh–the issues I mentioned above. Those warnings seem valid to me in this case.

Re: Decimation. Decimation is typically the last step in the process before export. Its primary purpose is to reduce polycount to make the file perform better in programs that may be limited in their ability to handle high polycount meshes. Decimated geometry is difficult to work with for many purposes in ZBrush though, which is generally why it is a final, pre export step. Any errors in your mesh geometry should be corrected before Decimating, as those errors may prevent the piece from Decimating correctly.

This piece has too many long thin sections of geometry to be fused with Dynamesh. That feature is likely to produce a very broken mesh if you attempt to use it on this model in its entirely.

Live Boolean can be used to fuse everything together, but the potential for Mesh Integrity errors is quite high. Geometry> Mesh Integrity > Fix Mesh may correct them, but I consider it likely that any attempt to fuse this geometry together in the form it is now will result in problem sections of mesh that are fundamentally unsound. In that case the mesh may continue to break or generate errors until those sections of mesh are identified and either redrawn in a more stable fashion, or cut away entirely.

What really needs to happen here for the best results is that you need to completely re-evaluate the form of your mesh and consider how to create it as more of a unibody form with surface detail that does not intrude too far into the surface of the mesh. Sort of squint at your mesh and try to see it as a vastly simplified, solid form. How would you carve that solid, one piece form out of a block of wood? For best results, you’ll want as much detail as possible to be simple surface detail on that basic form, and not a series of tiny individual mesh pieces. The cleaner the form of your mesh, the better it will print, the easier it will be to prep for print, and the easier it will be to avoid geometry errors.

In order to do this you may need to retopologize completely new, vastly simplified geometry over the top of your model as a solid piece, and recreate the detail on the surface of that new mesh piece.

In other words…

Good luck!

Thanks for your detailed reply, I assumed this would technically work because even though the thin pieces appear to be floating, they are all connected to something as one unified piece, meaning no non-manifold areas. I recently was reading one artist who has been printing complex fine detailed mini’s in one piece on his fdm printer with geometry and overhangs that seem to be far beyond the simple solid models I’ve seen, so I assumed fdm must be able to work with more than i expected.

I have two questions then regarding how I could move forward with this without scrapping it for some other future use.

If I mounted this mask onto a human head sculpt, would that make any difference or is it the same problem?

Would a resin printer be able to handle something like this?

Printers are bound by the laws of physics and the limitations of their mechanical parts and printing process. FDM printers work from a printer bed > up process. Every layer that is printed must be built on top of another layer. If nothing is underneath it, it will need to be supported by a physical structure that you must create and print. You can expand outward as long as there is is a gradual upward expansion at not too severe of an angle. You can sometimes get away with expanding straight outward as long as you are not extending very far from the preceding layer. But any sort of substantial overhang will require additional support.

The problem with all the long thin pieces with minimal connections elsewhere on the model is the level of additional support they will require. An FDM process also requires a minimal thickness. Depending on how large or small you are printing, your printer may simply not be able to create those pieces reliably with all the support connections they will require. Unless you are printing extremely large or as a series of small pieces that will be fitted together post-print, this is going to be a mess.

Long slender pieces are also going to be very sensitive to vibrations or mechanical disruptions in the process. A minor targeting error during the print can destabilize everything that comes after it. In other words, they will be extremely unforgiving and for FDM print you generally want a form that can absorb some minor mechanical errors.

Overlapped and redundant interior geometry will also complicate your model. You really want to be sending a single watertight volume to your slicer/printer.

A resin printer has some similar concerns and some different concerns as well. This piece as it is now is not likely to print well on either as a single unbroken piece.