Can't select a body

[robduarte]

Hi, I'm having trouble selecting a body for operations like "split body", "combine", etc. It's a mesh body that I imported from an STL. Tried with an OBJ and the same thing happened. Here's me trying it... http://youtu.be/Hv0-BNZ6dHE

Thanks for any help,

Rob

[laughingcreek]

Rob-there is pretty much nothing you can do with a mesh model.  you can snap tsplines to it, which is nice.  you can't snap the tsplines directly to the mesh vertices though, so not so nice.  In theory you can convert a mesh to a Brep surface.  All the meshes that I have tried that with have failed to convert (a size issue I think).  Again, not real usefull.  You can't directly edit the mesh itself.  nothing. nada. no moving vertices, slicing and dicing, removing bits or adding. nothing.

To effectivly work with a mesh in fusion, your best bet is to use another program to convert it to quads and reduce the number of meshes.  Then in fusion you (might) be able to convert to either a tspline or a brep body.

[TrippyLighting]

True. Actually given the similarity in modelling workfow between mesh (subdivision ) modelers and Fusion 360 the ability to retopo a mesh in Fusion 360 with T-Splines would be very helpful!

[daniel_lyall]

I have been trying for days to find a way to get mesh bodys to do something have not found a good way to make them usefull yet been close but can get ride of the last fault in the mesh

[deyop · Accepted solution]

I'd like to give you some background on our mesh strategy.  We are currentlly working on several fronts to provide a broader range of capabilities with mesh data.

First, where we started from.  Our original introduction of mesh supported only a short list of capabilities.

• Mesh as a visual asset.  Use the mesh to visually compare and to a limited extent measure to a developing design.  Nothing spectacular here.
• Mesh as a target for T-Spline snapping.  We used meshes to take advantage of the Object Snapping and Pull capabilities of T-Splines.  This is one approach for utilizing scan data.  This approach is still dependent on a lot of user interaction and is quite "dumb".  By that I mean the distribution of T-Spline vertices is dependent on the user and the snapping is based totally on finding the closest point on the mesh from the vertex.  We didn't consider this an ideal reverse engineering approach but worth our effort to handle some cases.
• Converting mesh faces to BRep faces (Mesh to BRep).  Again this is not an ideal solution because the number of faces in the mesh must be relatively small (10,000) faces in order to convert and utilize in subsequent operations.  A limited solution but one we had gotten requests for.

Where are we headed?  Based on forum postings, interviews with users and experience with other Autodesk applications we are developing a mesh workspace for Fusion.  The basic requirements started with providing an interaction that would be familiar to Fusion users while supporting great performance with the large datasets that are typical of reality capture (scanning).  We now have working prototypes that are capable with hundreds of thousand of facets.  Our focus is based on this list of workflows in this order of priority.

• Analyze and prepare meshes for downstream operations.  Whether meshes are generated from scans or from converted CAD data there are certain capabilities needed to leverage their use.  We are utilizing many Meshmixer capabilities within the mesh workspace to support alignment, subsetting and repairing meshes, optimizing the layout of facets and combining meshes in order to utilize the mesh for downstream operations.  These operations provide control over the structure of the mesh which will ensure it being more capably used for 3D printing, converting to solids or extracting lines and surfaces.  There are a number of applications for utilizing the mesh in early rapid prototyping that will reduce the need to for investing in modeling operations.  Here is a posting that discusses the topic.
• Utilize the mesh for reverse engineering.  Once the mesh is prepared and positioned it can be used as a template for creating an "ideal" form.   To this end we are taking multiple approaches.  First we will allow the user to define mesh sections by defining planar intersections.  These mesh sections can be used as a template for line and spline development.  We are also developing a line fitting capability that will allow the user to define a region of the mesh section to fit lines, arcs and spline within a specified tolerance.  These extractions can then be used as input to modeling operations.  Simarly we will allow the user to specify a set of facets which can then be input to a surface fitting command that will fit planes, cylinders and surfaces within a specified tolerance.  T-Spline snapping will continue to be used but with the ability to subset and optimize the mesh the results and performance should be much better.
• Convert to BRep or T-Spline.  When reverse engineering from mesh data the user is often required to rebuild the entire mesh from scratch in order to put the data into a format that is useful of manufacturing requirements.  We have a project underway that will extract BRep or T-Splines directly from the mesh.  This is far more sophisticated than the current Mesh to BRep.  As always the result will be dependent on the quality of the input but we have seen promising results.
• Sculpting directly with the mesh.  There certain form definition capabilities that can be best handled with a mesh editing capability.  You may have seen result from Mudbox or Meshmixer that would be nearly impossible to achieve with conventional CAD modeling techniques.  Utilizing images to deform a mesh is also a method developing form that expands the designers toolbox.  As mentioned earlier these techniques will only be valuable when integrated into a modeling process that can utilize the results.  We are considering that as we move forward.

I hope this gives you some idea of what direction we are heading with mesh data.  I expect to share more detail in the near future and enlist your help in evaluating the interaction and results.  I know you are likely anxious to try out these new capabilities but I ask for your patience to ensure good interaction and performance are delivered when released.  Here is an image that somewhat guides our approach illustrating the source of the mesh data and its use as a template for extracting ideal forms or to utilize as the design master.  As always we value your feedback.  Please feel free to respond.

Thanks

[big_red_frog]

Appreciating that there looks to be lots of mesh capability in the pipe line which is great, I am trying to use what I thought was an existing capability.

Specifically "We used meshes to take advantage of the Object Snapping and Pull capabilities of T-Splines"

I believe this is demonstrated in this video where snap to mesh vertices and points is exercised on the dinosaur model.

http://knowledge.autodesk.com/support/fusion-360/learn-explore/caas/CloudHelp/cloudhelp/ENU/Fusion-F...

However, when I import an obj mesh into fusion, I cannot get any snap to the mesh to work. It just seems to ignore the presence of the mesh as I try to construct new faces.

I had hoped I could laboriously generate new faces and eventually solids via selection of the underlying references from the mesh, but no luck, I simply get no snap.

Is there anything obvious I might be missing?

My mesh is predominantly triangles, rather than quads, could this have something to do with it?

Thanks.

[colin.smith]

Hi Big Red Frog, 

I assume you have the "object snap" option turned on in the T-Spline tool that you are using?

Also, have you tried Pull?

Check out this quick video and let me know how your experiece differs.

http://screencast.com/t/xYwgTbx9l7

Thanks

Colin

[big_red_frog]

Huge thanks for the video, its really appreciated.

Note you had some serious audio issues, sounds like fan noise on the mic!

I need to explore my experience a little further so I can qualify the issue more cleanly. Is there somewhere I can read up on the role of "create base feature"?

Meanwhile, that pull capability looks so powerful. What I want to do in my my case is create faces using the actual intersection points of the vertices of the mesh as reference. So buiding a realitively accurate solid, with the the mesh as the guide.

In your demo, it looks like you were tracking to the surface, but it was not snapping to the actual intersections, is this the case, or does it put the face corners on intersection. the underlying mesh you were working with was quite dense, so I couldnt be sure.

For my purposes, I have tried to work with a pure import of an STL. As well as taking the STL through 3DS MAX and export as a SAT file to try to make it a reference solid or mesh format that can be worked with.

Neither has worked for me.

I need to set aside an hour to get a good capture of what my workflow looks like and where I am struggling so I can give you a chance to help guide me.

Thanks again for the support.

[colin.smith]

Hi BRF, Gald the video helped.  Yes, I noticed that odd fan sound in the video also, sorry about that.

WRT your two questions:

1- Create base feature:  As you know, when Fusion 360 was first released it was history free. When we introduced the parametric version of Fusion not all of the tools that were in the history free version were brought over to the history side.  In order to access some of those tools you have click on "create base feature" to suspend the history and get access to some of the non-history tools - like Insert Mesh in the case of my video.  This is similar to how T-Splines operates in History.  We have been slowing bringing more features over to the history side.  If you think there are some tools that should have a high priority to be brought in to the history side let us know in the idea station and we can add it.  The other way to import a mesh in to your design is to upload it to your project file and then insert it from the data panel.

2- T-Spline object snap: It sounds like what you want is more of a conversion from mesh to solid.  We do have that feature but if your mesh file is very dense it will not engage the process.  The T-splines don't stick to verts on the mesh automatically you have to eyeball it.  

Hope that helps.

Colin

[big_red_frog]

Thanks again.

So much has happened in fusion since I last used in anger, I have a lot catching up to do.

The problem with convert mesh to solid, is I am generally working with video game assets that can be captured and exported as STL or OBJ, but a more a partial surface model, rather than a closed form, and will have many odd intersections and other anomolies, that are fine for a rendered video frame, but don't translate when trying to turn them into a solid.

The intent is to take the fractured mesh, use it as a reference, preferably selecting viable faces from it where they are good, and then stitching together the gaps, and clipping the anomolies, until you have a workable closed mesh, which can be converted to a solid.

At that point you can do many things such as hollowing it out, splitting it into seperate pieces suitable for printing and adding assembly / mounting features, or other such. As per the Halo Helmet effort that got me initially involved in Fusion 360.

Getting from that STL orignal to the workable solid with a sustainable workflow is the dream here.

I have cut up an STL to more managable chunks, imported to 3DS MAX, exported as SAT, imported to Fusion, this gives me more core mesh and some bodies, but it is still a mess, and I will have to join the seperate pieces back together again at the end, which may not be practical.

Just being able to use an original STL as the reference in Fusion to steal faces, and stitch them together would be the best of all worlds.

I don't think it exists.

I need to play some more with my 3DS MAX -> SAT -> Fusion workflow, and either prove it viable or abandon and start with direct meshes again.

Watch this space...