AI imported beziers extruded to solid - problem with Variable Radious Fillets

Hi all,
I’m 2D designer with polygonal modeling experience 3dsMax & Cinema 4D.
I was hoping to get a head start in Rhino by using my designs from Adobe Illustrator for a while, rather than tracing them. It seems I am able to make decent extrusions; however, Variable Radius Fillets are finicky. Even if I’m not importing compound curves, it’s a crapshoot - sometimes they work and other times they go wild with crazy extrusions. That usually happens when the diameter setting is bumped up “just a little too much”. Sometimes fillets appear to work; but they are actually separate from the solid - they sit there just under the surface of the corner edge as a separate entity. I attached a v5 file. The middle and left examples show this behavior while the right example was created within Rhino and worked just fine.

And sometimes one section of the curve loop reacts to the fillet command in some way (good or bad) while another section of the loop seems to completely ignore it. That is exhibited in my left & middle examples as well (pointed end of pick shaped objects). I thought that the cause might be from me accidentally selecting my original curve(s), but I can’t imaging the fillet for solids even working on curves. Just in case, I made sure I selected only the solid; and with similar results. I tried with AI v8 files as well (left example) that seems to work for Cinema 4D but it made no dif here… And I tried exploding, rebuilding edges on top surface, and extruding from that. That fared no better in the “fillet buffet”.
My questions:

  1. So are my imported AI bezier curves remaining beziers, and the problem stemming from Rhino not having robustness in this fillet command when it comes to dealing with beziers?.
  2. Or are my beziers getting converted to NURBS during import; but the converted curve geometry still has some structural issues?
  3. Is there a way to clean up / validate the curve geometry so that it behaves like it’s a native Rhino citizen?
  4. Is there any way to look into the geometry to tell what it is going on technically / structurally giving insight for predicting potential fillet problems?

Thanks for reading this. I hate being a noob

AI Extrude & Fillet.3dm (456.4 KB)


There is one major problem with your curves that can be easily spotted with the Analyze, curvature graph command. At the bottom of heart shape, the curves break tangency. This is causing the fillet to fail. You can fix this with Match curve command with the Average box ticked. The second minor issue is the slight break in curvature of the the big sweeping curves.

Just to be picky, the area you’ve called out is curvature continuous- The outer line of the graph represents curvature and in your image this line is not broken - it has a hard corner, meaning rate of change of curvature is not continuous.

You’d see a step in the graph at a curvature break. Like so-


As Stratosfear pointed out, CurvatureGraph is your friend here (see Help for more details) - any curves that you are not sure of, and I’d say in general that includes any curves from AI, as well as any font curves, whether generated by Rhino or any other software, should be anlayzed with curvatureGraph if you want to make clean surfaces from them.
Match is a good tool for cleaning up, see also ChangeDegree, Rebuild, InsertKnot, InsertControlPoint, InsertKink, Split(Point option) and CurveBoolean. Among others.


Thanks @Stratosfear.
You aren’t discussing the object I was referring to in my questions; but it does show that my Rhino curves technique
needs some best practices applied as well…

Thanks @ Pascal,
This is the direction I needed. I’ll look into what you suggested.