Loft command: perplexity

Is it normal that by performing a “loft” of the five curves the surface is generated with those isocurves (ugly)?
(furthermore, it is not possible to change the direction of the curves, nor the seam of closed curves).
loft.3dm (312.3 KB)

Is the complex curves and simple curve combination needed?

Yes, as they are

So I guess rebuild is not an option…

I guess a an option to add a slash for loft like sweep2 would help…

If you can’t rebuild the curves then you have to live with the result IF you need to use loft.
Or add an interpolated point cuve on each end and use networksrf with desired tolerance:

I have often wished for a similarly advanced tolerance option for loft, so let’s put surface tools back on the wish list for V7 :slight_smile:

You can also loft with rebuild, but that doesn’t tell you the tolerance it hits, so if that is important then you have to rebuild the curves and check the tolerance there.

Cause of the isocurve shapes is the different parameterization of the input curves. Solution is to rebuild the curves with the same parameterization. Use FitCrv with the desired accuracy tolerance. Or Rebuild with sufficient points to keep the deviation within the desired tolerance. The loft the rebuilt curves.

Thanks Holo, it is clear that some commands still have deficiencies. The loft is one of this.

I do not understand why the seam of closed curves can not be modified in the loft command (it would be convenient and functional to be able to drag these points to the position that the user chooses …).

NetworkSrf effectively rebuilds the curves before creating the surface.

I tried to re-parameterize the curves (with the same control points) the result is better!

Reparametize rescales and offsets the parameter. It does not change the distribution of the parameter along the curve. If it changed the distribution of the parameter while keeping the same control points then the shape of the curve would change.

Loft is a very old command and is simple and predictable if you use simple curves with few controlpoints, but it does not have any sophisticated rebuild option. So rebuilding the curves with the same amount of controlpoints would be your best bet to get the desired result if you HAVE to use loft.

I think we should review this command (if a curve has more control points then the result becomes not very good).

Regardless of this example, the commands to generate surfaces (loft, sweep, networks …) should work at the state of the art in Rhino. This is not acceptable for a surface modeler! Regardless of the cost of the software …

No need to Rebuild with the same number of control points. Similar distribution of the parameter along each curve is needed which Rebuild or FitCrv provides.

Yes, if you want to control the output before running loft then it is.
But you can use Loft’s own rebuild, but as stated that does not give you an indication of the tolerance you get.

Oh, and loft’s fit command can give you some odd results so I never use that:

result on the right is not acceptable (only one curve has more control points)
loft.3dm (388.2 KB)

Number of control points is not the cause. Differences in parameterization distribution is the cause, which can occur with a small number of control points.

Try to make an example …
The command must be reviewed! (I hope the developer will take care of it without having to wait for Rhino 7, Rhino 8, etc.).
At this point I would be curious to see how it behaves MOI with the same curves …

Here is my point illustrated:
Rebuild all curves together and choose the number of points that gives you the deviation you can live with:

Then Loft the curves and see how each isocurve matches the points.

You get the same result by using loft with rebuild, but you don’t get the deviation output.

Use Rebuild on each curve with a different number of control points on each curve. The waves in the isocurves go away. The cause was not the different number of control points, but the different parameter distribution.

Different parameter distribution can occur with the same number of control points on each curve.

1 Like