Any suggestions on how to 3D sculpt model an easier way?
I've seen your other thread and now that you've shared a model I can see that your T-Spline has a number of problems.
I might have time to create a screencast later showing how to fix those and have the form adhere better to the original form of the mask.
In general, however, the problems you describe are somewhat independent of the T-spline problems. Whenever you change the topology of the mesh chances are that you have to readjust steps in the timeline afterward.
One way to reduce that is to create the features as standalone features/bodies.
I personally would not use T-Splines to design this mask. Using 2-3 curves and some lofting can get you a long way and is more gentle on the following timeline features.
Your very first sketch has a spline with this curvature. spikes in curvature at he ends are problematic.
the curvature is wonky because you pushed this control point up near the end with out adjusting the tangent handle
and also because you used nine control points instead of 3
I reworked that spline to have just 3 control points, now it looks like this
after doing that the shell worked. I went ahead and reworked the spline in the second sketch as well. the spline in the 3rd sketch needs some love also. I'll leave that to you. reworked file attached
@mgianzero wrote:
So I can understand why making my first profile with only 3 control points is a much better t-spline.
The curves you are working with are splines, NOT t-splines.
For t-splines you have to start a Form.
Having said that, T-Splines and spline (curve) based surfaces (NURBS) have strong mathematical similarities.
When you click on the "Finish form" button after designing your T-Spline it is converted into a NURBS surface.
@mgianzero wrote:
Regarding the third profile, you again showed me the curvature combs display, which I had read about, but don't really understand very well. What exactly didn't look right at those two control points that you highlighted with arrows? Are they not supposed to cross over the main curve itself? I reworked it slightly which may be a bit better (see attachment), but I still don't understand what my problem or my goal is here regarding looking the combs display. Sorry, but I'm learning here.
You want to use the "simplest" geometry you can to describe the shape of an object, so use the fewest amount of fit points you can get away with. The curvature comb will look less "wobbly" in a good spline curve. It will also create a "lighter" surface. When your loft includes splines as the basis you create a NURBS surface. The shape of a NURBS surface is created by a quadrilateral mesh of CVs (control vertices or control points).
Fewer spline control points will create a surface with fewer CVs and as such uses fewer data and fewer computational resources.
Most importantly, however, a number of tools in FUoisn 360 are very sensitive o curvature problems, namely thickening and shelling. If you keep your curves smooth they these tools will perform much better.
@mgianzero wrote:
Lastly, I noticed you used the stitch commands several times once your completed your loft. I think you combined the two halves in order to patch the center piece. Otherwise F360 won't see that edge as one continuous front edge, right? But then why'd you stitch the 4 bodies together before shelling it out? Does that mean that when you stitch bodies together in the patch or surface workspace that completely encloses an inside space, that the model will convert to a solid body from a shell? In a previous attempt at making this model (all with the surface environment) I made the entire shell and then used the thicken command instead of enclosing and shelling it out. Is this a better technique? It just seems counterintuitive to convert a surface model to a solid to then shell it out, when really all you want to do is "thicken" it, no?
Shelling a solid is indeed a much better workflow than individually thicken surfaces then hoping they will all combine happily into a thin-walled solid. Individually thickening surfaces can easily create near-tangent and near-coincidence situations where those solids border touch. Those are problems are difficult to workaround.
That is particularly true for beginners, who often create less than ideal curves 😉
Okay, it's starting to make more sense. But you say " The curvature comb will look less "wobbly" in a good spline curve. It will also create a "lighter" surface. "
You referred to this surface:
I added my own rail in a better place (more perpendicular to the profiles) and it improved the surface (curvature map is somewhat smoother). However, shelling still did not work.
Lofting often requires a little experimentation to get the surface to look like you want to.
In such cases, I work with offsetting and trimming surfaces.
Then I stitch everything and end up with a solid.
BTW, you did not follow all of the techniques I shared in that model. You forgot the helper surfaces, which is the reason why your mirrored surfaces are not even tangent across the mirrored boundary.
Model is attached for study.
@mgianzero wrote:
. So it appears that the shell command, like the thickness command, have limited capabilities using Fusion 360 with more complex models.
Most solid modeling techniques are somewhat limited 😉
@mgianzero wrote:
I wish I had a deeper understanding of the offsetting and trimming surfaces that you are doing here so as to know what to fix. For me, it's more like trial and error. Wish I had a better grasp on what I need to do to get these surfaces to do what I want.
If you keep on with the trial and error and do some reading you'll slowly nut surely will be able to connect the dots.
Here is a link to the Autodesk Alias Theory builders.
