What software should be used for painting 3D objects created by Rhino?

unhandled

(Alex) #1

Hi guys,
In many cases I need to add texture or paint some Alpha maps over the 3D objects created in Rhino. Zbrush and Mudbox are capable of doing so but they need to subdivide model enough prior to do painting and Rhino models are messed up when they are subdivided so I cannot use them as texturing softwares in the case of Rhino exported meshes. I know that I can use photoshop for adding texture or Alpha map over the UV map but in many cases I have to customize painting over the 3d object and painting directly is needed. What is the remedy?


(Brian James) #2

One workflow would be to layout your UVs in Rhino and then save out an Obj with the texture coordinates. Then open that in a program like Blender (max, maya, c4d etc.) and paint to the UV’s directly in a new image texture. Save out the image and load into the material back in Rhino. This method is particularly nice since it doesn’t require subdivision and the result can be applied back onto the polysrf with the same UVs.


(Alex) #3

Thanks, but I meant 3D painting. None of those softwares you mentioned can do 3D painting.


(Brian James) #4

I can paint to a UV texture directly on the model in 3D here in Blender using Texture Paint mode. Perhaps you mean something other than painting a 2D image over a 3D model in the viewport?


(Alex) #5

No, Like what you can do in Mudbox or Zbrush. You can paint or use stencils directly over 3D object and it’s projected into the UV automatically as you are painting. Is there any solution?


(Brian James) #6

Ah, you mean mesh sculpting with image stencils. ZB is the one I’d use for that although it’s possible in Blender and I’m sure other apps too.

It sounds like you’re running into issues when sub-dividing. In ZB, make sure not to sub-divide the tool (imported Rhino model) directly. Optimize the mesh first, weld points and create a dynamesh from it before sculpting.

Or you can use the texture painting workflow to create 2D bump maps for use back in Rhino if the actual geometry isn’t needed.


#7

It is actually possible to perform projection based 3D painting inside Blender. It’s not Vertex Painting as in Zbrush, hence there’s no need for Catmull Clark subdivision. 3DCoat is another feature-rich option, which again lets you 3D-paint meshes derived from Rhino-models.


#8

Is MeshMixer a candidate? (if so, it’s free at that)

http://www.meshmixer.com/

// Rolf


#9

Hi Brian, I’ve been losing my mind trying to import a very simple geometrice form out of Rhino and into 3dcoat. I’m glazing over at this point as no matter what I try, there’s never a workable mesh. Trying to increase the resolution in 3dc just makes matters worse. All the edges are destroyed. There’s a retopo tool, but I’m not getting anywhere with it and what’s imported doesn’t even show up in the sculpting room. That’s beside th epoint but what I’d like to do is created a more uniform and dense mesh that retains edges. Is this possible to do inside Rhino before exporting to .obj?

GWL_Scaled.3dm (5.9 MB)

I tried to ul the .obj exported but its not allowed.


#10

3DCoat is excellent for painting models from Rhino.

However a bit of pre-processing in Rhino will make your life a lot easier.

Firstly you want to join everything that is one object into a polysurface.
Then mesh it. tweak your mesh settings to get a density that is suitable for your shapes but don’t go crazy with density as that just slows things down. Make sure “jagged seams” is UNchecked so that seams are nice with coincident verts.

Go here for a bit of discussion on mesh settings: https://wiki.mcneel.com/rhino/meshfaqdetails

Then select just the mesh and export that.

If you have various bits that will be different materials, then you can assign material names in Rhino before exporting and create multiple UV sets in 3Dcoat

UVMap in 3DCoat and off you go.

Note that 3DCoat has two very different types of painting; Per Pixel and PTEX.
Unless you have a very good reason to the contrary, use PerPixel painting as the resolution of the image is controlled by the texture map and is independent of your polygon density.

Hope that helps, Steve


#11

3DCoat may be a bit overwhelming in terms of import options.
The attached settings should work with typical geometry from Rhino. Set the texture resolution as needed.

Remeshing should not be necessary, unless you want to output for Realtime applications. You will get best results with relatively dense and evenly meshed geometry. That’s easy to do within the advanced controls of the Mesher.


#12

Hi Steve,
I think this good reason would be any sort of treatment which deforms the surface considerably. If one only wants to paint some decals and other 2D detailing one can indeed use low res meshes. But any 3D painting which later gets rendered with a displacement map will profit from a good start-resolution.


(Brian James) #13

Thanks @Steve_Howden & @hifred I’ve never tried 3dc myself. @JKayten It sounds like the mesh settings wiki page that Steve provided will help get you a better mesh density. Since you also mentioned edges ripping apart in 3dc, I’d also suggest doing some tests with the Weld command in Rhino after you create the mesh from the NURBS model in Rhino with the Mesh command. This can probably also be done in 3dc but I’m not certain. More info on Weld here if you need…
http://docs.mcneel.com/rhino/5/help/en-us/commands/weld.htm


#14

Indeed! After reading what you and Brian pointed me too - I realized I’m more of a doofus than I imagined. I expected the meshing tool to output something better than what the density slider produces - which is to say no difference whatsoever in this case unless other aspects are tweeked.

Jeez!

I used the weld option in the export menu. Is that suitable or does that also need more attention?

I want to suggest numerical input boxes for mesh parameters be changed to a slider - it would make experimenting much easier - and preset the relevent ranges in each - maybe even indicate the range in parenthesis to the side. Maybe auto preview would be nice as well.

Thnaks very much for the screen shot - pictures are always the best! What I ultimately want to accomplish is a 3d printed structure that looks like a stone-built fireplace. After I manage to get something workable into 3dc, I’ll need to figure out how to sculpt it with I think some sort of template overlay. That’s another week of learning, but at least I’ll be feeling great when it delivers results. I won’t have to worry much about coloring layers as all painting will be done in old world of analog.

Thanks again to all!

eidt - actually I will have to make it all look nice with layers as I’ll need a rendering for out of state clients to review… eek.


#15

Thanks Hifred. Yes, I totally agree.


(Brian James) #16

On your GWL_Scaled model, I’d first use MergeAllFaces to combine the co-planar surfaces at your mirror axis. The I’d use some mesh settings like this… Only density and Max edge length are used so it’s in my opinion the easiest way to use the detailed options.

After that, you may or may not need to weld the mesh depending on what tools you are going to use and in what programs. So I can’t say for sure what angle of welding is needed for what you’ll be doing in 3dc but Weld at 45 would be my best guess at a good setting.

One last thought on your model since it has thin walls. This can require masking or grouping by topology in mesh sculpting programs to prevent moving the polys you don’t want to. It increases what you’ll need to learn but it should be possible in any mesh painting or sculpting app to prevent the sculpting of occluded polys given the current view or to create mask groups. Good luck!


#17

I’m finding that even though I select what I think is correct the model imports as a single item - that is there’s only one layer in 3dcoat.

I’m really quite confused with this matter, as what I think exists in the 3dc work space will disappear if I move into another - for instance what shows up in the paint room disappears if I move over to the sculpt room. I need a better conceptual understanding of what’s takine place in these different aspects of the process - that includes within Rhino as well. I came to Rhino and the 3d CAD world from the old analog paradigm - you know actually making things with actual tools…

       Anyway, if you've any links to tutorials or such that a old noob like me should learn I'd be grateful!

#18

These all are actually purely 3DCoat specific issues – and better asked in their forums.

On Layers: What 3DCoat inside its Paint-Workspace calls Layers is rather meant in the Photoshop sense. This Panel controls visibility of paint and surface structures and gives access to Blend modes – even across objects. If you have various physically separate object inside your Rhino file and import into the 3DC Paint-Workspace they do remain separate. The Paint Objects Panel found in Windows/Popups controls their visibility.

The whole paint workspace is there to colourize and surface-structure (=sculpt moderately) models which are mostly finished and which already have UV’s assigned.

The Sculpt workspace is intended for creating models from scratch and for introducing large form changes on raw models without UV’s. That workspace has its own object tree > Vox Tree. Bringing models from the Paint Workspace to the Sculpt workspace to twist and bend them would quickly render UVs and existing textures worthless. That’s the reason why 3DCoat doesn’t let you do this. It’s common though to start models in the Sculpt workspace and to detail them inside the Paint workspace, after remeshing.

You said you want to immitate some stone built structures. In case you only need the deformation you should import the model to the Sculpt workspace and switch to Surface mode. In case you want to 3D print with textures that’s also possible – but this would require some more explanation.

There’s of course the Help-File. And Youtube is full of 3DCoat Videos, just have a look.


#19

Thanks, I have been watching a great number of vids on 3dc - but I’m not quite getting the sort of info you just shared. I’m guessing meshes exported from Rhino have UVs - or do they? I think there’s also an issue with the relatively thin walls in the model which is only representing a textured cladding/facade.

Sculpting moderately is exactly what I’m after to generate a render in Keyshot, however I’d want to also have the option to print the texture for actual painting.

Then there’s also the retopo room … As far as I can understand it, that is where one can make adjustments to the wrapping after sculpting but there’s also vids that seem to use that room as a starting point.

Also, I’m creating a model in Rhino that has meshes on seperate layers but they don’t seem to show up in the export. I can’t upload the .obj file as it’s not allowed.

GWL_Scaled.3dm (11.8 MB)


#20

Yes, those clips are created by people with a mesh modeling background for mesh modelers. The learning curve is usually quite a bit higher for those, who come from CAD.

Yes, all Nurbs geometry get basic UV’s assigned at creation time and this assignment is inherited when converting the model to most mesh formats. Mesh models originally don’t have UV’s, one needs to unwrap them or assign a UV-Projector.

On your output goals:
One may understand the Paint workspace as the place where the render trickery takes place. Deformation applied here appears inside Normal Maps and Displacement maps, Colour information is output to texture maps. In everything you are doing you are locked to the input resolution, the initially loaded model geometry actually remains unchanged.

If you want deformation to appear in the mesh you will need the Sculpt room. Here you can also increase mesh resolution when drawing small details. When the model is finished one may decrease the resolution for 3D-printing – the algorithm will retain a dense mesh where needed and uses larger triangles in flat areas.

In order to paint that finished deformed model you need to recreate UV’s and mesh modelers would at that point usually perform quad re-meshing, either by hand or automatically – but these steps require some skill as well, especially for thin walled parts. For your workpiece it might be sufficient and a lot quicker to simply re-import your finished mesh to Rhino and to assign box-mapping. Then export out again as .obj and load into the 3DC Paint workspace for texture work.

If you can’t upload a particular file format simply create a zip file. That should work.