Loft not despite all curve seams aligned and sorted correctly

Hi everybody,

I always do my best to deep dive on all resources I can manage to find before asking anything, but I am stuck.

The problem is that I am not able to create a proper loft in Grasshopper out of seams aligned and sorted closed curves.

Please see GH file for futher details (data internalized) (112.2 KB)

A little history:
These curves extracted out of a noise implemented Brep.
But since bottom of the shape also has the same noise and I want it to be perfect circle,
I thought it would be a good idea;

  • to cull first N curves
  • inject the circle into the list
  • and finally loft them so it would be a flowless transition.

As you will see on both screenshot and the GH files, it doesn’t work in any case.

I am truly out of ideas.

Thank you in advance.


Looks like the only thing you need is the Loft option ‘Rebuild’ (white group below). The purple group moves the seam (start/end point) of the bottom curve (blue group) and all other curves above it using Crv CP (Curve Closest Point). (112.5 KB)

Hi Joseph,

Thank you for you quick response.

It’s my bed not to tell the whole story so let me give more details as to why “rebuild” is not a solution for this specific case,

That Brep is part of a bigger object and splitted using “Split Brep” to make those operations.

So the ultimate goal is to join this piece back to where it belongs to.

Since “rebuild” is slightly changing the curvature form, there appear small gaps when joining the pieces back together.

I don’t know if there is another way to handle this but all the resources I come across so far suggests that having “sorted” and “all seams aligned” set of curves shall form a loft without any problem.

Since I have that already, do you have any idea as to why it might not be working?

Thank you.

Loft is failing because your curves (except for the circle) have way too many control points.

So rebuild is required. Joining edges of lofted surfaces can be a challenge. Good luck.

You can use surface from points (113.8 KB)


Nice thinking “out of the box”!

You could also use NetSurf (Network Surface): (114.5 KB)

We can’t know if these methods work without the original geometry.



Thank you for the answer and it obviously constructs closest form.

However I guess it’s not constructing the same curve so it would fit/match the other piece perfectly.

Actually, the overall goal is to be able to have a gradually dying noise effect on the object towards the bottom and/or top.

There is a way I found on this page where the Graphmapper used a filter but that is using “fields” and it doesn’t work in my case using Noise on Mesh or Fast Noise (at least I couldn’t figure it out)
Planter pot with attractor points

Hi @Joseph_Oster

Thank for sharing. That solution also looked so promising in the first place as it is using the curves itself to construct a new mesh.

However, I tried comparing the resultant curves measuring their center points of off each other but the result is not “0” so I guess it won’t work as expexted.

Here is the file for screenshot: (123.6 KB)

Thank you

What, why and where this matters?
You know that a number like 3.89e-6 means 3.89*10^-6 or 0.00000389 units in your document.
Even if those were meters (architecture), it would mean an error of 0.00389 millimeters
or 3.89 μm.

For comparison, you are saying that an error of about a hair dust particles width is relevant while designing a ice hockey field. #baldeaglescreeching

What are you doing?

You can increase the rebuild count or subdivision count to thin the error, but it’s meaningless already, and it will never be “0”.
Re-consider what you are doing/asking.


Hi Riccardo,

Yes, you are %100 right. That tolerance does absolutely mean nothing for fabrication/production purposes. We all do agree.

I started with GH just 84 days ago (my license is going to expire in 6 days. LOL)

The reason why I am questioning this tolerances is because to understand if this is something possible to achieve or not, that’s all.

Otherwise, the case is solved with several alternatives thanks to @Joseph_Oster and @laurent_delrieu.

The result is this: (1.7 MB)


Am I happy? Yes and No.

YES, because the starting point of all this was to overcome the challenge of a “fading noise effect” and now it’s there.

NO, because I know this whole shape can be constructed from scratch with fading effect included.

So the next step is to find out how this could be achieved. I will start a new topic as @laurent_delrieu advised.

Thank you.