Why does MatchSrf change the surface?

Hello, there is something incomprehensible, a normal curve, normal surfaces, which, for some unknown reason, change when they connect. The surface is also deteriorating (Zebra, new edges). Why does the bottom surface change when MatchSrf?
1.3dm (2.1 MB)


have you tried uncheking the refine match option? the modification of the surfaces after match is needed to join them better but it depends of the original surfaces anyway.

Refinement cleaned - the same. I just do not understand the meaning of surface modification. :thinking:

Well why are you matching these? The starting shape is fine. And if this is a simple round shape, the place to do the matching is at the curve level, you could do a revolve with history on and match the curves…

MatchSrf breaks up the surfaces because with a revolved shape like that the surfaces are Degree 2 with “kinks” at the quadrants, so the surface gets split at the kinks into a polysurface. It’s probably not totally necessary to do that, but MatchSrf is meant for fully freeform surfaces–meaning Degree 3 minimum, and “kinked” surfaces are considered a no-no–so that’s what it does.

I understand, but I still do not understand what is happening.

What in particular? The surface gets split into a polysurface because it has “kinks” in it. The surfaces don’t match up with each other after fact because it simply doesn’t “care” about that. This isn’t a sort of surface MatchSrf is set up to handle specifically, as I said you should just edit the curves before revolving. MatchSrf is meant for “freeform” shapes, where the surface being modified is at least of Degree 3 in both directions.

Well, I’m surprised at the change in the bottom surface after MatchSrf. I do not see the point and logic of change. And in general, periodically, this change is simply annoying (interferes).

Well I don’t actually know exactly which part is disturbing you. It could probably be changed but I’m just trying to explain what’s going on.

Any sort of revolved shape has “kinks” in it, kinked surfaces are not considered clean freeform geometry, and most Rhino tools will split up kinked surfaces when they are found and they certainly aren’t going to try to keep adjacent kinked Degree 2 shapes smooth.This is not a situation where you should be using MatchSrf. What are you trying to do? There are several better ways of going about this.

@JimCarruthers What is your definition of a kink? How can a surface with continuous curvature contain a “kink”? If the curve which is revolved has continuous curvature then Revolve will produce a revolved surface with continuous curvature.

Revolved surfaces are rational with control point weights which are not all equal to 1. The rational form is used because it enables circles to be represented exactly with four control points. The non-rational form with all control point weights equal to 1 cannot exactly represent circles.

(Incorrect statement deleted.)

I only care about one question, why does the bottom surface change after MatchSrf? This is somehow meaningless :confused: The surface is already matched (with a curve).

Well… you also ask for Rhino to REFINE the result, so it has to rebuild thus it is bound to be different.
And you ask for CURVATURE match, that is different from Tangency. How much does it change with Tangency as Continuity?

1 Like

For reasons I don’t understand. Matching degree 2 surfaces is always problematic. In this case the revolved forms also have stacked knots Degree 2/4 Spans/knots. Rhino splits these surfaces into a polysurface of 4 single spans after matching.

No changes at Tangency. However, the original curves already have continuity. I understand the rest.

I think the key is not degree 2 but rational- matching with rational surfaces is not straightforward - I suspect that when you’ve encountered degree 2 surface matching problems these have been revolved surfaces or similar where the control point are unequal.

The result is rational in this case - a middle control point, per 90 degree arc gets thrown out of line with the quadrant points and a kink develops there. It looks like the position is the one that would be correct for a non-rational surface, which is I guess what you were saying, @davidcockey.


I somehow still do not understand the arising kinks, because the surfaces are equivalent and continuous (from the curves). I also do not understand the change in the bottom surface from the top (after MatchSrf), that the tool can change when there is already continuity, what is the point of editing? I think the tool should not change the surface in such conditions.

Hello- the kink that is develops by moving the points is in the vertical direction - there are points that move such that the horizontal rings of points are not lined up across the quadrants of the circle. The thing to keep in mind that Nurbs circles have fully multiple knots on the quadrants and kinks are formed if the points are out of line:

That is what is happening on your revolve so the object is split up at the quadrants.


Good. Actually, I really care about it …
Curves are already continuous
Why does the final surface change when getting curvature? The surface already has continuity with the bottom surface (from the curves). What does MatchSrf want?


Hello - please try this on a surface that is not a revolved surface, or, make the revolve with the ‘Deformable’ option. Is the result the same?


Simple extrude :confused:

I decided to check such a surface, the changes are almost not fixed :confused: