Srf continuity

00.3dm (225.2 KB)

how do i match it better?
basically i made the srf reference to the video, and use the same setting of matchsrf.
the srf is similar but turn out its continuity is bad

What level of continuity do you want?
G0 - Position
G1 - Tangency
G2 - Curvature

What tutorial have you been following? Dividing the surface strategy is incorrect. I’m not sure how that person managed to accomplish the continuity (or maybe not). In other words, you should stop learning from the video unless you want to improve your muscle memory, modeling speed, etc., and you now know that the tutorial shows a bad way of modeling (especially for cars).

I added the curve example. To get the quadsrf, just sweep 1 or 2 and then project the trim curve onto the object and trim it.

One advantage of your model is that the zebra flow is consistent, which saved me a bit of time.


Hello- I’d be inclined to make a single trimmed surface there -

Incidentally, the parts in your file are minuscule - you are modeling in mm with what looks like possibly meters in mind.

Here is another possible trimmed surface configuration, which might keep the flow a little bit better:

And the other way to do this with an untrimmed surface is to put a singularity at the far end:


Rhinoceros cannot achieve G2 on its own, and G3 is impossible in this particular case. Zebra is not intended to verify continuity. Considering the adjacent surfaces and relationships, your approach will fail regardless.

Hello- I am not really sure what you mean - G2 is the continuity for those surfaces - the Zebra display does indicate that, as well as EdgeContinuty.



I have no idea what happened to your model, as we built it differently based on the Zebra flow alone. Again, Zebra is not intended to verify continuity, or more precisely, it is intended to briefly verify continuity such as G0-01, which is very obvious to everyone, and it is an informal approach to check G2 +.

The following is a comparison of the Rhinoceros and the Kajto.

Rhinoceros showed G2 continuity, whereas Kajto did not. My prior model demonstrated the same issue, but I did not retain each significant step. Nevertheless, this model indicated the same issue.
In this instance, G3 can be achieved by dividing either surfaces or by increasing the existing surface to 7 degrees with 8cps, but 5 degrees with 6cps is not achievable as depicted in the image.

Are we clear?

That’s what I was thinking, bad tutorial especially with that corner surface yikes!

also it might be the continuity settings because when I checked its not close to tangent or curvature for me in rhino.

In Alias:


  1. I believe you measured it incorrectly. It should be cross section check curvature, which Rhinoceros 3D does not provide.
  2. I can’t say the modeling approach is incorrect, especially given the little half-circular form is designed to simulate the same flow with the top edge on the surface. However, that precise circular gemotry is a completely incorrect method. It should resemble the top edge curve, making a quad surface. With the precise Zebra flow, you can develop your model. I can’t investigate further because I don’t have any reference images.

IMG2 :

Thank you for the confirmation in Alias. :crazy_face:

A bird’s-eye view is breathtaking, yet it may be isolating at times.

G1,G2 is ok

Is there any difference between creating a surface with 3 curves and 4 curves? Since in the video, he is suggesting do not creat srf with srfedge by 3 crv.

P.s. maybe isn’t this video, but I can sure he had talked about that in his videos

Any clue for about that?

Anyway, I forget to upload the video I follow. This is the attachment

The video is by @sgreenawalt who contributes to this forum. Perhaps he could provide input if he sees this thread.

The approach in the video to modeling a geometrically simple, smooth surface is interesting and more complicated than needed for the surface shown in the video. But it may be a logical solution to following certain rules.

Very interesting video, and yeah SkyG has a lot of great content on here so apologies in advance for calling it a “bad tutorial” just seemed odd when I saw the post as to why a corner srf is needed as opposed to trimming away the outside? hopefully he jumps on the thread.

I think he was demonstrating one approach to modeling a surface with the exterior edges are not trimmed and thus can exactly match input curves or the edges of other surfaces. Here is a thread on various approaches to modeling this type of geometry in the context of the side of a boat. Boat hulls with smoothly curve stems - how to model

Added: I added a post in the linked thread with an approach using multiple untrimmed surfaces without singularities.

Wait… the OP intended to make a car model, but the video showed a hull model with identical shapes. Then JACK was attempting to incorporate the hull model concept into the car. That’s dope. :face_with_thermometer:

The purpose of that video is to introduce the concept of the trimmed corner, which is a cornerstone of NURBS modeling. It is not to say that the trimmed corner is the ONLY solution - but it’s A solution that is often the best solution. There are certainly times when collapsing one edge of the surface down to a point is advantageous, but neither solution is right 100% of the time. What I’ve seen over and over here and elsewhere is folks forcing NURBS patches to be 3 sided, and then (naturally) having issues with continuity at the collapsed edge. A trimmed corner is merely another tool in a modeler’s toolbox, it’s up to the modeler to decide if it’s the right tool for the job.

1 Like