The white lines are a rebuild single span degree 3 curve that blends the middle of the filet with the flat surface. They naturally intersect because they are the same curve that was copied with Orient3pt.
There are many ways how to draw this, it’s an exercise to see what works best for you.
About the math and the logical intersection of guidelines it wasn’t entirely clear to me at that time. Maybe it would now, not sure. I came across other situations where it does make sense and the point is for instance located at 1/3 between two points. Anyway to really understand I guess I should pick up my math but can’t find the time for it (or patience).
I’m really interested in the construction shown in image13.jpg above, and am trying to recreate it.
I can’t follow the process from this image though. @asdfsjal - Could you please describe the steps?
Or anyone else who has managed to reconstruct this?
I was able to do something similar to @Niels’ construction and gradually moving the point along the intersection of 2 planes, but it seems the earlier post implied an entirely constructive technique?
I’ve recently been trying some new stuff using relaxation with Kangaroo to maintain G2 between surfaces: https://vimeo.com/239818980
but so far only managed this when the edges are also kept G2, so it can’t do star points like this.
If I had a geometric construction based on intersections/reflections and so on, then I think it would be possible to include into the relaxation as a constraint to be enforced iteratively and bidirectionally.
More magic and sourcery from The Wizard DP. Sometimes I really do wonder if you’re using the same Rhino that I’m using.
Most impressive Daniel, even if it is a restricted case!
Is there a way to construct the three blend curves that make the central ‘star’ such that they intersect each other and the intersections are coincident? I’ve done it by picking an arbitrary point for the intersection and constructing the blend curves from there, but it is messy and often looks forced.
Yeah I’d also be curious to know more about how the CVs were laid out in image 9. If continuity can be solved by perturbing CV positions via K2 goals then I might have a pretty snappy way of extracting surface patches from arbitrary quad meshes.
If you extrude BD along a vector normal to plane ABD
and extrude CF normal to ACF
then intersect them,
I think point E has to lie on this intersection line. At least in the examples I’ve seen so far this is always the case.
…but what determines where along that line?
