STL file import for sculpting edit?

[Anonymous]

i have built a model in another design program and exported as stl, now i import stl into fusion,can i edit it using the sculpting workspace?

[Anonymous]

beautiful quads, by the way all of this model show up on my F360 so i dont even have to be going to the website to see the progress though if your wouldn't have mentioned it i wouldn't have noticed )

[PhilProcarioJr]

I post it in case anyone else is interested.....I know it's a long shot, but......

[Anonymous]

I found this pretty interesting. 

Posting here as the keyword search will probably drag more of us meshies in...

good tutorial on meshmixer and conversion to solids...

https://www.youtube.com/watch?v=yMaih-vPNr4

[PhilProcarioJr]

@Anonymous

Meshmixer definitely has it's uses, I rarely ever use it because I need precision.

That is after all why I use CAD, because Maya, lightwave would be 100x faster for mesh modeling things.

During this whole process I am using a workflow that allows for precision work.

Sometimes it can be a little tricky but the end results are worth it.

[nkloski]

I love Meshmixer! (I love it enough I created a 17+ hour eLearning course on it! 🙂

I'm not sure if I posted this before in this thread but here is a great workflow for Meshes in Fusion:

• start with mesh in tri's (normal .STL)
• reduce mesh using meshmixer as much as possible (10K tri's or less)
• use Memento or instant meshes (https://github.com/wjakob/instant-meshes) to convert to quads
  start fusion
• create base feature
• insert mesh
• finish base feature
• turn off history-based modeling
• convert mesh to tspline
• turn on history based modeling
• have t-spline fun!

If you have a truly complex file, you can do this with multiplr 10K tri- files and then combine them later on once they are t-splines.....but Fusion will bog down, obviously.

[PhilProcarioJr]

@nkloski

See using the methods I use I have never had performance issues in Fusion and I have made some really complex models.

I have a file now with almost 100 inserted meshes all meshes are at least 3,000 polygons each and still have no slow downs.

Meshmixer and Momento are ok but for really complex models they fall short IMO.

But if they work for you that's great.

[nkloski]

That's great! I've never seen a slowdown with meshes that are inserted into Fusion...."converted" meshes (t-splines), then that is where I can see slowdowns, yes. I've inserted and NOT converted, 200K (tri's) meshes into Fusion and since Fusion treats it as a "lump" it does not really have to manipulate it very much.

In Meshmixer directly, it is fine handling (on my machine) meshes up to about 2M triangles, then some "manipulate the entire mesh all at one time" commands start to bog down. Reliably, I have worked on 5M tri models if I work on specific areas at a time, no problem.

[PhilProcarioJr]

Plus the key point here is Momento is going to be another paid subscription app soon.

The free version will be limited. So no point in backing myself into that trap.

I already pay for enough software as it is.

[nkloski]

yep about Memento...I try to use the lower-cost or free software as much as I can, if it works.  Try that "instant meshes" program above...it is great at turning tri's into quad's and is free (windows only).

[TrippyLighting]

---

@nkloski wrote:

yep about Memento...I try to use the lower-cost or free software as much as I can, if it works.  Try that "instant meshes" program above...it is great at turning tri's into quad's and is free (windows only).

---

Instant Meshes works quite fine on the mac as well.

If you scroll down on the GitHub page it lists executables for the three main operating systems, Mac, Linux, Windows. 

[TrippyLighting]

---

@nkloski wrote:

I love Meshmixer! (I love it enough I created a 17+ hour eLearning course on it! 🙂

Please post a link, or it did not happen 😉

Thanks in advance!

[nkloski]

Happy to provide links...if you want a discount, please message me! 🙂

http://education.honeypoint3d.com/courses

[PhilProcarioJr]

This discussion seems off to me. For one thing what we are doing in this thread is not just about turning a triangulated mesh into a quad mesh. Both Momento and Instant meshes don't handle isolated definition. So unless your looking for a unified quad mesh neither work. This thread is not just about how many polygons can be cramed into a scene before Fusion slows down.

This thread is about efficiant work flows between mediums.

Setting up polygon models in the most efficiant way to get accurate CAD representations that can be used to build off of and be rebuilt. 

Using scan data in ways that critical details are not lost in the transition.

Building a topology base that has points in critical areas that will be used for snapping reference later.

This seems to be a case where this thread is not understood.

[PhilProcarioJr]

@Anonymous

So this skeleton was bugging me....I need a more accurate model... what do you think about this Femur?

[Anonymous]

found this video from a year ago in the tips section.

https://www.youtube.com/watch?v=nQr20fTQS5M&feature=youtu.be

I do wonder what progress has been made since and when this sort of capability will be made available?

agree that what we are talking about here is a discussion of repeatable workflow to move between coordinate systems,

in a reliable way with predictable results, 

whilie preserving the detail of the original input.

This is the only discussion I have seen so far that really identifies these topics in a deliberate way.

I think your constraints are slightly different from mine, in that you expect the results to be measurably consistent in a way that I may not entirely need,
but that just hightlights that the workflow needs to satisfy the needs of a wide variety of use cases,

some of which require measurable and meaningful consistency of resulting artifacts across transition events,

and others which require that the essential detail (loose term I know) of the original to be preserved.

Not really clear to me where we go from here as :

 - you have said that the free software path does not support the workflows required well or at all (I do not understand why not? are there IP / SW patent constraints or some other reason?)

 - A//Desk has not yet released the mesh workspace update ,so we do not know what will be in it beyond hints, and we do not know when it will become available

 - we do not know what the impact of the A/Desk Mesh workspace will be on the other tools like MeshMixer etc that A/Desk offer

is it useful to detail the free (meshmixer -> memento -> Fusion ) and non-free ( Maya SubD -> Maya Nurbs -> Fusion) paths and their steps, so at least we the alternatives are clearly defined? 

could we set up an automated service that provided access to the non-free path to people who cannot afford the SW?

sort of not in my character to do nothing in a siutation like this, but not sure it makes sense to do anything to generate an interim solution...

what do you think? 

[Anonymous]

i dont see nothing but nice quads

[nkloski]

Hi Phil,

I guess I did understand your stated point of this thread, but I did not connect my thinking properly.

I use this workflow quite often in my own business where I need to move from scan data to CAD/Surfaces while keeping as much detail as possible. Here is what I do:

• Generate scan data (STL in 3-sides polys/tri's).  the image below fits what you said, as it is only a part of a larger scan.
• In Meshmixer isolate the part I want to turn into CAD and delete the rest.  Sometimes it is the entire model, sometimes it is only a small part of the model. Export as OBJ
• Use instant meshes with the settings turned up as high as you need to go.  The model does not even have to be manifold. Export as OBJ/quads.
• Import into Fusion and use the technique I stated earlier.
• The nice thing about this workflow is that you do not have to use the TSpline "Pull" command.  When that command works, it works great...but more often than not, it does not work well for more complex shapes.  See the pictures below for a complex shape I scanned, and converted.  The individual features on this shape are just over 100 microns (.1mm) so the detail seems to be pretty well preserved when converting into Fusion/CAD/Surfaces/T-splines.  If I had tried to use the Pull command it would have been a mess.
• This process can be repeated if needed to get other areas of interest into the same Fusion design.

Original mesh:

Converted into T-splines in Fusion (around 8K quads created by Instant Meshes):

Then with Design History turned back on:

[Anonymous]

In the last image you have there, the image shows a shape with a series of faces which must be individually selected if you wish to manipulate them.

A basic question I have whether it is possible to unify those faces so you end up with a true solid whole, and if so how you woud go about doing that?

It may just be me, as I am quite new to Fusion: I can get this far in what I am doing, but this state seems problematic to me from an editing or CAM perspective.

[nkloski]

Yeah, that is weird...when you switch design history back on you no longer see individual T-Spline faces. You can try this with a normal T-spline object...in Sculpting mode you see all the T-splines, and when you exit the Sculpting mode, you get much larger quads...but once you go back into edit more, all of the individual faces are still there. I am not sure how Fusion determines how to delineate those larger areas...but it is just a visual thing, when you go back into edit more, all of the faces are there.

I guess, also, you can assign Appearances to each of those faces, that is one thing you can do to them, but again, I don't know the criteria for how Fusion determines what to lump together for that top-level visualization.

The object is indeed a solid, though. In MY example, I started with a non-manifold object and ended with a non-manifold body....but if you start with a manifold body you will end with a nice "Solid".

[Anonymous]

 I just tried it and realized that when I try that it fails because the meshes I am working with are always borked somehow when I import them.

so I have never seen it work, and a big part of the failure of the open source workflow to be repeatable is that I have not found a tool that can reliably fix the meshes so the subsequent steps will complete properly.