Surface Matching - Dense Isocurves

Before surface match:

Even just for a tangent result, the isocurves are very dense in one area (after RefineMatch=On). See below the result as an example.

Is this simply because I’m matching an Untrimmed Surface to this trimmed down edge? Is it something poorly parameterised, since the surfaces have the same point count in the direction I’m matching at (x7).

I’m really just asking out of curiosity, for the purposes of learning to avoid similar results in future cases, and in the pursuit of something cleaner! Thanks!

surface-match-problem.3dm (236.0 KB)

Hello - the ‘Refine match’ setting is what allows the extra knots to be added. I’d not use this until you think it’s really needed (to be able to Join, say,) and even then, I prefer to add a few knots myself to getting a huge pile of them.


Thanks Pascal. The Join was okay, but zebra wasn’t without the refine match. Would there be any particular logic over the placement of extra knots added by hand?

Hi Johnathan - what I’d do here, I think, is matchSrf with History on. Then use InsertKnot, with MidPoints=Yes (allows you to hit mid-spans with the knot insertion, which helps keep things uniform) and add one or two knots in the area where zebra looks like it needs the most help. Repeat as needed, the Match will catch up with History and update the zebra. I usually find I can add just a relative few and evenly spaced knots and not the big pile the automatic way does.

BTW, unless there is a reason not to, I’d tidy up the mis-match on the edges before using MatchSrf

ExtendSrf is a good way in this case:

All that said, it might be possible to create that surface you’re matching with the exact same structure across that the other surface has - that would probably be a plus for matching.
surface-match-problem_Maybe.3dm (100.9 KB)


In many cases, if your typical model units’ tolerance is something like 0.001 mm, using |Match surface" with 0.002 mm in the “Refine match” option will be enough to successfully join the surface edges. Sometimes even 0.003 mm or 0.005 mm will do the job. In your case, 0.0025 mm is enough.

1 Like

Thanks Pascal - not something I would have thought of with regards to MatchSrf with history on and inserting knots on the fly. regards to ExtendSrf yes I did try this too before running MatchSrf - as even MatchEdgesByClosestPoints didn’t give a good result.

When you say exact same structure, do you mean same point counts in U and V directions? As mentioned I’d hoped that them having the same structure at the adjacent edge, degree 3 and 7 points, was what was most important for the match.

Hi Jonathan - the best way to match the structure is to use the surface edge (if untrimmed) as input to the new surface.

surface-match-problem_Maybe2.3dm (102.2 KB)