Mudbox, MeshMixer or Z-Brush?

[3Dpcb]

I want to make some organic details on the surfaces of my 3D model.
As a reference, I would like to use an underlying artwork. It should also be possible to import the updated mesh back to Fusion, so I can edit with the tools in Fusion.

I am considering purchasing Autodesk Mudbox (€695.75) or maybe just use MeshMixer (free). There is also Z-Brush which is a lot cheaper.
In Mesh Mixer It does not look like it is possible to import a reference image Mesh Mixer.

I gues that some of MeshMixer tools at some point be part of Fusion.

Does anyone have experience with this software?

[bob_holland]

3Dpcb,

Just so that we are all on the same page, you want to take a Fusion 360 model and export it for use into a different product (Mudbox, MeshMixer or Z-Brush) make organic changes and then re-import the resulting 3D form back into Fusion.

Thank you.

[3Dpcb]

Yes, that's exactly what I want.

[deyop · Accepted solution]

Those organic changes you want to incorporate in your model may lead to some challenges in how the data is formatted.  In any of those mesh or polygon based workflows the results will always be polygons.  Integrating these back in to the original model is difficult because the original representations will be T-Splines or solid models.  There is a Mesh to BRep command that will take every facet of the mesh and convert to a BRep face but that is quite limited in terms of the number of faces that can be practically used.

We are working on projects to incorporate the mesh more directly into Fusion workflows.  I would suggest experimenting with Meshmixer as it has quite of bit of mesh editing functionality at the right price point.  We are also working on getting those same features into a mesh workspace in Fusion.

What will you do with the final model?  Is it just for visualization or will you want to 3D print or manufacture parts?

[cekuhnen]

What do you want to create with MudBox?

Mudbox is nice because it uses quads ZBrush will like MeshMixer create a dense triangular mesh.

You can I htink use Meshmixer to make some basic sculpted shapes, and in Fusion use that imported mesh to sketch over either

via TS or Brept tools.

[3Dpcb]

I can see the problem with exchanging Mudbox, Mesh Mixer or Z-Brush geometry with Fusion. It sounds perfect to incorporate the mesh Tools directly into Fusion.
I will start by experimenting with MeshMixer and wait and see what will be possible in a future Fusion / Mesh Mixer updates.

What will you do with the final model?  Is it just for visualization or will you want to 3D print or manufacture parts? :

1. First I would like to render the model.
2. Then print it on my MDX40A, Makerbot or Form1+.
3. Finally, I will hopefully get the product manufactured.

[3Dpcb]

I would like to be able to easily be able to add a more detailed surfaces and with many small details to the models I create in Fusion.

Sometimes inspired by nature, other times I need add a surface some patterns or ornaments.

You can I htink use Meshmixer to make some basic sculpted shapes, and in Fusion use that imported mesh to sketch over either

via TS or Brept tools.

I suppose you also could use the Pull Command ("Snap to Mesh data") to redraw geometry from Mudbox, MeshMixer or Z-Brush in Fusion?

Like in this video:

http://fusion360.autodesk.com/resources

[cekuhnen]

yes and now

It sounds as you want to create a product that is more polygon based and less NURBS. So in this case your workflow should be Fusion 2 Poly app and then send the shell to print.

Specifically for displacement sculpting (textures) you cannot do this in Fusion. The snap to mesh data is only good for coarse details like blobs but not fine details.

If money is tight you could also look into Blender which offers a sculpting workflow similar to Mudbox(quads) or Z-Brush(tries) but it costs nothing. ZBrush however can deal with more complex models.(millions and millions of polygons)

[deyop]

I have experimented with Meshmixer and basically you can take any CAD data convert it to mesh and integrate it with textured or organic mesh data.  You could try the zipper brush, Erase and Fill or Make Solid as strategies for integrating the two pieces.  As a hybrid mesh it will be 3D printable without much issue.  It may even be possible to machine the mesh to create tooling.  The limitations will be that parametric operations will be difficult to support as the mesh is still a "dumb" bunch of facets.

We are definitely interested in hearing any specifics anyone would like to share about mesh modeling requirements as it is an area of active development currently.  Based on feedback from other users we have broken it down into these areas.

1. Prepare model for 3D printing.  Analyze mesh problems that would compromise the ability to print the mesh body.
2. Open,analyze and repair meshes for downstream operations.  This is especially the case with scan data where the quality of the mesh may not be very good with holes and missing areas.
3. Utilize mesh for traditional modeling and analysis tasks.  Use the mesh as a body as criteria for design development.  Measuring between a developing design and the mesh.  Extracting curve data.
4. Convert mesh for model and form development.  This would include creating analytic and surface fitting with the intention of creating an ideal manufacturable solid model.
5. Provide a mesh sculpting capability through brush interface. 

[cekuhnen]

I am not sure if mesh mixer can actually do texture based sculpting, I did not see that option today.

I was involved in few NURBS Sculpt projects and what I would loved to have is this:

a modeler that can deal with NURBS but also with polygons fine.

Specifically for polys the following tools were helpful (Blender Rhino):
Boolean operations, trim mesh, etc besides fill holes ...

What we did was build the main design in Rhino convert it to polygons and then we sculpted the textured surface where needed and send that to the casting place.

Now I must say the poly tools in Rhino where sad, pretty sad, which is such we did in Blender, and later after the concept phace build the design in Rhino with

keys etc so I could texture the model later back in Blender.

For Fusion I am not sure if this might make much sense. dealing with high density meshes seems always to be a problem with Rhino. I am not sure about Fusion

as the UI there is slower by a lot already. But I think Fusion Brep to poly sculpting app is a good idea.

Maybe meshmixer could get out of the root of what it was first designed for and become a light version of MudBox.

[deyop]

I have been experimenting with some of the Meshmixer capabilities in anticipation of introducing them into Fusion.  That work is underway now.  The team has been spending considerable effort to ensure good performance with large mesh files.  

Where can the mesh come from?  Mesh can come from converted CAD data, reality capture or polygonal modelers like Mudbox.  Currently the output will always be a mesh which can be visualized, 3d printed or machined.  Part of the long term goals of the Fusion Mesh project include extracting lines and surfaces for a more standard reverse engineering approach.  We are also working on methods of automating the process of mesh to CAD conversion.  

Here is an example of a "mashup" between CAD generated data and scan data.

The "Make Solid"  command is the basis of creating these geometric "mashups".  Here I have a scan from a Next Generation scanner and a new base created as a solid in Fusion and converted to mesh.

I repaired areas of the scanned mesh with the Erase and Fill command which will repair holes while evaluting the surrounding mesh to flow the patch.  I also cut the scanned base off and extruded the remaining latch so that it extended into the CAD base. 

I combined the two meshes and used the Make Solid command.  I didn't need to attach the two meshes.  They only had to intersect each other as the command envelops the mesh with voxels which define a closed region.  Now there is one integrated mesh which is perfectly watertight and ready to print.  Finally I used the Hollow command to hollow the interior.

The purpose of this experiment was to describe a workflow that can incorporate both mesh and CAD data.  In some cases the mesh may not satisfy manufacturing criteria but there is a lot of power in using this strategy for prototyping and custom designs.  

Let us know what you think.

[cekuhnen]

that's nice to see.

it would be fantastic in case I have a high res laser scan to like in ICEM be able to rebuild NURBS with good fitting onto the mesh data.