XNurbs releases a ground-breaking NURBS software

unhandled

#61

Is there something we can try without having solidworks?..


#62

Okay. I’m impressed.
The multi-sided patches are of very good quality.
There are some less continuous spots in your example surfaces, but this seems to be confined to cases with strange surface topology.
As long as one sticks to a reasonable patch layout the surface quality is very good!

So bring on your plug-in as soon as possible.
Regarding price: of course spending less is always appreciated, but I think 595 $ seems reasonable for what this software seems to offer.

Looking foreward to the release date :grinning:

Cheers, Norbert


#63

Hi Toshiaki_Takano

You will need SolidWorks. Perhaps, you could get a trial version of SolidWorks.


#64

Hi Norbert and all,

Thanks.

XNurbs manual discusses the problem - “Conflict Constraints”: Many existing models have some tolerances. For example, in the image below, the original curves/edges in the triangle loop have a gap/overlap, and they actually do not connect with each other. The same problem can occur to Tangent/Curvature continuity. patchhole demos this kind of problems. This is why we don’t want users to set an arbitrary tolerance, or inexperienced users may run into problems. With our current UI, XNurbs can guarantee to generate surfaces with good quality.

Positional Constraints


#65

Sorry if it’s been answered in the thread… it’s become a bit long to search…
Will the plugin be for both V5 or V6?


#66

XNurbs, you should have a trial version as a plugin for Rhino, otherwise we can not do any test. Word and images seem all right, but at times it is not like this: they would be concrete examples.

If this technology were optimal, I believe that Autodesk would start eating, as it did with Tsplines.


(Tom) #67

[quote=“XNurbs, post:60, topic:49049”]

Hello,

first of all from a standpoint of someone who has also written surface plugins I have to say its technically a masterpiece. So congratulations on this one. I believe there are dozens of practical applications for it. So well done.

From a standpoint of somebody who does class A surface modelling, the resulting surface quality is simply not good enough to argue its outstanding, but at least it is good enough to satisfy areas where surface quality isn’t that important, such as engineering and architecture.

[…]
“… no human-being can generate a better result by modifying them. The surface generated by XNurbs is the one among all possible solutions that has the minimum energy and satisfies all input constraints, i.e., the smoothest surface.”

human can generate better results! Not by trying to fit in a single patch surface by some boundary conditions, but instead by choosing a logical patch layout in order to reduce open edges down to 3 or 4. There are several strategies in solving these kinds of blends, but they all have in common to iso-align as much sub surfaces as possible, so that the overall curvature is flowing right. This is more important as getting a perfect g2 connection.
But in general Rhino and Solidworks are not the tools to do deal with such blends very well. You don’t even find the right tools to evaluate matching quality, at least without plugin usage. Can you even modify cvs in solidworks?

But still I think there are enough potential users and real world applications, and I believe that the CAD platform you are targeting are suited very well.

I hope to see this plugin getting started on Rhino, and I hope it’s getting better and better. It might be too expensive for the group you are targeting, although I understand the efforts involved and the need to refund. I would argue to lower the price, at least for the start. Companies who want high quality results, can also buy ALIAS or ICEM SURF in order get better results. So I would rather go for people not being able to purchase these cad-softwares.

In addition, it would be interesting to see how the plugin performs when all boundary surface are double curved. In all of your examples you have chosen simplified shapes, being non-curved or only curved in one direction.


#68

The cost should not be very exaggerated, let’s talk about a ‘multiblend’ in less words, to make substantially fittings n-edge (VSR, for example, is less powerful but more complete in instruments and cost could be justified).


#69

XNurbs, let’s start by testing your tool with this simple example (this is an example that has been circulating for a long time on Internet, where Alias is compared with Rhino. Needless to say, but as zebras show, Alias does his job well).
test patch.3dm (122.0 KB)


#70

Absolutely.

This is what in the end it’s always about. Badly planned patches, then badly drawn curves, resulting in oddball surfacing situations.


#71

I did some further evaluation. The patch could be smoother.


(Marc Gibeault) #72

Untrim the surface and it will be nice and smooth


#73

And the continuity of the five surfaces?


#74

Hello,

Some odd bunching going on with the XNurbs surface relative to the VSR one. The VSR surface is using multi blend which isn’t a great solution but the vanilla VSR result looks better to the naked eye.


#75

I did that comparison as well. The xnurbs result is superior to the VSR multi blend.


#76

Hello,

@Stratosfear, mathematically, or?


#77

VSR continuity analysis is roughly the same for both.


#78

One thing is certain: xnurbs or multiblend of VSR, McNeel should have made such a tool, useful for generating complex fittings. After so many years … and there are not even SubD.
The patch tool is very ridiculous at the comparison!

The points of force of a surface modeler, at the state-of-the-art, are these, not the rendering or the thousand views, or the least options added here and there.


#79

They do have a tool, the patch command. It just doesn’t work well when matching continuity to input surfaces.


#80

It does little to patch: it often creates many folds and continuity is reduced to G1.
The patch of Rhino 6 is the same as Rhino 5. It has been slightly improved compared to that in Rhino 4. At the end of little evolution!
Developers should understand where to concentrate resources, that is, in the basic aspects of surface modeling!

Rhino is an exceptional modeler, but you can not think of using a fillet plug-in, one for rendering, one for multiblend, etc
Eventually you first buy CATIA … :wink: ehehehehe