@XNurbs I don’t get that weird zebra on the surface I sent either, and yes, I am well aware about singularities. I know when it is a problem and when to avoid them. In this case however, it is not the point of attention. The surface can easily be over-modeled and trimmed. This discussion is to get the best possible surfacing with the tools we have. So far the loft gave the overal best result in the shortest time, which is also an important factor. This is what I like about XNurbs the most: that in many cases you can have a very good result in almost no time where manual surfacing might be able to take it a step further but at the cost of a multitude of the time needed.
The method you propose gives an non symmetrical surface and with neither setting I can get a good flow:
I did try other ways though to get a better result with XNurbs. This one is quite interesting where the lower surface is a blend surface and then trimmed, then filled with a Xnurbs with an additional guide point:
Overall it would be nice if XNurbs allowed for more aggressive surface flow to create more G3/G4 like transitions. In other words, we need more ticks on the lowest slider (at the right side of the spectrum) so that we can create similar surface tension as we can with BlendSrf.
The attached files all have history enabled for the XNurbs surfaces to check them out.
Btw: the blend surface was a G4 blend
This GIF shows the weird zebra of singularities. Because of singularities, the quality and shape of the surface can be unpredictable and unreliable. I don’t think you could find such a design in the final car body for production.
To get Class-A model, I suggest that you split the surface based on curvature. However, this is not a typical surface like a car body. The slitting is useless for common surfacing.
The model above is QuadSided.3dm and the video is generated from the model. The original model is created by Alias or Catia, and it is a part of an automotive body for production, i.e., the final design.
Yes that’s a very tiny piece of a Catia model I sent to you
anyhow, the surface with singularity is not a surface I would call a final surface, but again, this was not the main concern. I’d rather have one point off than a complete wobbly surface, the point is easy to fix by slightly overmodeling the surface. We are looking for a way to create the tip of that surface with Rhino and/or XNurbs tools. Based on your screenshot I made another one which so far seems to give best flow (trimming the loft and filling the remainder with an XNurbs surface.
Yes, this is definitely better than the original one.
I am sure you won’t present such a design to your customers. You are a professional designer
@encephalon , I would ask the money back if this were a paid job. It only shows Rhino + xNURBS creates a better model.
We posted a message to another thread and asked Rhino to publicly reveal the guys’ names who vote xNURBS on Food4Rhino. The following is a part of the message:
Currently, xNURBS rating on Food4Rhino is 2.8 - genuine xNURBS users give 5-star rating on Food4Rhino and trolls use fake IDs or anonymously give 1-star rating. In other words, nearly half of rating are 1-stars and another half are 5-stars on Food4Rhino. For anyone with basic statistics knowledge, it means the rating on Food4Rhino is forged. In plain language, trolls are manipulating xNURBS rating on Food4Rhino… We have 18 votes in total on Food4Rhino. Those guy(s) didn’t realize that, with their “work”, we only have 1-star or 5-star rating on Food4Rhino, which makes it “awkward”.
The xNURBS development benefits the whole Rhino community and Rhino users. From the simple fact that nearly half of rating are 1-stars and another half are 5-stars on Food4Rhino, it is obvious that trolls are attacking us. Apparently, someone try to discourage us from further developing xNURBS for Rhino users. Without xNURBS, professional designers would use other software (or Rhino competitors) to design a production automotive model, e.g., the model below. Some person(s) are upset with the fact that, with xNURBS, talented industrial designers can use Rhino to deliver super impressive automotive models in a very short period of time. In my opinion, the trolls may actually work for Rhino’s competitors and it is Rhino’s job (not ours) to deal with them. Bob, could you ask Carlos to publicly reveal the guys’ names who rate xNURBS on Food4Rhino? I don’t see any reason why someone needs to use fake IDs or anonymously make 1-star rating.
I’m also curious to reveal the users who vote with 1 for the xNURBS patch tool, because then the @XNurbsXNurbs user will publicly apologize for calling me a “troll”, and will give me a free license for his plug-in as a compensation for his manner. I will also continue to help him make his plug-in better, by pointing out the errors that it produces.
Since the same user focuses on writing about quality, professional designers and asking money back for poorly executed job. The 3d car model posted above fails to obey the majority of industry-standard rules, despite the shiny renderings that may fool some people. If you want to understand what makes a car model to be considered good quality and manufacturable, here are the major rules that all truly professional designers follow:
They make sure the flow of the surfaces across the car’s body is consistent and free of any waviness (my post above including the screen-shots taken from the video posted by XNurbs clearly show the true nature of its surface flow);
They maintain constant radius for the door windows (be it single-curved or double-curved), in order to make it possible for the latter to slide along the window rails;
They make sure that the wing mirrors will let the driver see behind the car, following the official law requirements. This is especially important on cars with very wide rear fenders. Tip:
They would never design extremely low doors that would collide with the nearby sidewalk (unless suicide doors are being used).
They make sure that the tyres will not collide with the wheel well arches while moving across the suspension stroke.
They set at least 120-150 mm ground clearance on most passenger cars, but never less than 80 mm even on the most exotic sports cars, because of 3 main reasons: a) the bump stops in most countries are up to 70 mm tall; b) the ground clearance decreases by at least 10-20 mm while 2-5 people sit in the car; c) the front end dive a lot upon high-speed cornering or sudden braking.
Fix: “Without xNURBS, professional designers would use PROPER MODELING TECHNIQUES to design a production automotive model.”
@XNurbs I do like XNurbs, I have used it in many projects already. I’m not a car designer though and have never needed to produce surface models for exterior car parts, only interior panels.
I’m just generally interested in creating high quality surface models that are as clean as possible.
You and I have had many email exchanges already and I helped you improve XNurbs with sending models and feedback. I hope you can stay this open towards feedback, as that’s the best way to make things better.
I don’t quite understand why you think @Rhino_Bulgaria is a troll. He’s a very high respected member of this community and has shared a lot of his knowledge with us and continues to do so.
In fact his proposal is the best offer you can take: let him use XNurbs and give feedback on how to improve. He’s one of the very few people in the field that has proven production models made for exterior car parts in Rhino.
The car model you show in the video is Not a professional production car model, it’s a concept study and as already pointed out has many surfacing flaws. Totally fine for a concept but not as production surface data.
