Visual discontinuity in seemingly simple surfaces

There is another way to make the G2 flow really smooth. However, it will not preserve the original middle blend portion of the surface. If you want to try it, simply select your original model, then press the “Make uniform” icon and see the magic. The surface’s control points remain unchanged, but the surface itself becomes much more smooth.

Yeah, that is true too - a uniform curve or surface will tend to be ‘better behaved’ for this type of thing.

-Pascal

Hi Pascal,

I understand that you weren’t saying that going the other way will get me exactly what I want. I’m trying to learn how to approach these things to get predictable, quality results in the future: why does choosing one direction maintain the curvature of my inputs, while choosing the other one doesn’t? I understand if this information is covered elsewhere online, and would be happy to be pointed in the correct direction to get this information.

I don’t really care about G2 vs G1; I’m aiming for high quality reflections for renderings, as most of my work is rebuilt by engineers in other CAD programs.

The surface you built is beautiful, would you be able to upload your file so that I can compare it to my attempts? Compatibility with V5 would be awesome.

The input on the long side is relatively complex - multiple spans - if you sweep that on the short simple curve, the far edge is exactly the same as the rail, just moved (more or less, but as opposed to offset in the sense of the Offset command) While the far edge from the rail does change, there is only one span in the curve - three points, so there is no danger of breaking any internal continuity.

If you sweep the arc-ish thing on the longer curve, your original plan, the far edge from the rail is an offset of the rail - this is where things are different because the offset loses internal continuity compared to the original.(as I mentioned, Offset > Loose will produce the same result)

In any case, for good reflections, in general, I’d shoot for transitions that are ‘progressive’ -dunno if that is the right word - - they accelerate once have a single well defined peak curvature, then come back - the graphs above show this.

SweepTheOtherWay.3dm (476.9 KB)

That said, the previous advice of making the two straight surfaces and a separate transition surface is really the way to go…

-Pascal

1 Like