What is class A?
Its a term originated in automotive industry and it is an alias for surface modelling under highest quality aspects. In class A you optimise a model so much, that light reflections are totally controlled and flowing the right way. This is needed for the highly reflective skin of car exteriors and interiors. Class A modelling is very time consuming, that’s why its only used on visible surfaces. Constructive parts are class B, which do have a technical relevance and are not visible, and class C surfaces are just to fill gaps or as temporary surfaces used for production.
There are also many other aspects…
One of this aspect is continuity between surfaces. G2 (Curvature continuity) means light reflections are at least tangent aligned. But having continuous reflections everywhere does not mean to control them -> this is the criticism of T-Splines regarding class A.
Potentially every CAD software using Nurbs and thus Bezier can achieve that, but practically you need many specialised functionality.
1 Like
Gracias ! / thanks
Saludos / greetings
…
Hello,
I understand that McNeel has to follow the money. There are likely many more Rhino users who couldn’t care less about Class A than do care. I’m not sure there is a business reason for McNeel to bother with these sort of tools in Rhino. While I admit I have a tendency to grossly oversimplify the mathematical and development aspects of this problem I do think that what the industry is lacking is the proper UX for these tools. I think there are ways to relieve the decision making burden from the application and rely on the human that make sense, but I don’t think it’s in McNeels DNA to do this. I’m not taking a swipe at McNeel, just thinking it’s not in their wheelhouse given their foundation so I’m not expecting it from them.
Rhino 6 actually does have a setback fillet command. It’s a test command, but you might want to give it a spin: Testsetbackblendedge. You need to type it entirely, test commands don’t autocomplete.
5 Likes
I just did this in the V7 WIP:
2 Likes
You beat me to it. I just did it with R6 ( I had it on the clip board when I saw your post):
So this was a “one shot” command. i modified nothing, I just ran “Testsetbackblendedge” and selected three edges. That’s all.
// Rolf
Right. The corner will be smooth but if the surfaces are flat (degree=1), and they use Fillets (degree=2), the edges will be G1 Tangent as the fillet can not support G2 curvature continuity.
The user will need a Blend instead of a Fillet to get higher degree continuity.
1 Like
Question from the unwashed masses here: Why is the corner made with Testsetbackblendedge considered superior to the one made with FilletEdge? (See attached). The Testsetbackblendedge corner seems to introduce distortion into the flat surfaces.
Thanks,
Dennis
Who said that?
Perhaps there’s a reason to why the command is named Testsetbackblendedge ?
I think this came up as a claim that something exist, or is in the making (by implication).
// Rolf
How would the flat surfaces be distorted when they are only trimmed and not otherwise modified? The differences you see are probably the differences in the shapes of the fillets and G1 vs G2 continuity between the fillets and the flat surfaces.
Edit: The trimmed edges of the flat surfaces are different shapes between the two shapes. The corners are rounded in one and square in the other.
That ain’t no setback corner.
It is not a question whether a setback corner is “better” than a linear corner. They are simply two different design solutions and it depends on your intended formal aesthetics which one you prefer.
The problem with Rhino’s automatic solution is that the surface quality is not so good, once you look at the highlights and analyse the curvature change. Also, a proper setback corner should have some lead towards the six surfaces of the actual corner to avoid that mechanical engineering software type look; it is actually 9 surfaces that are involved for a top-notch solution.
2 Likes
Yes of course it is!
And the answer is related to why it is better in a given situation with given requirements. And those requirements are usually related to aesthetics and how reflections flow along the surface, or how the object feels to the touch. A normal filletedge corner will give a sudden change of surface direction and can give the impression of imperfection, just like the “barreltop” effect, where a flat plane can appear concave because of it’s surroundings. (So designers often make those surfaces slightly convex to fight that optical illusion)
I think a setback fillet isn’t a nice solution either. But both versions are the best types of corner fillets which work on nearly every 3-edges-situation. So they are not the best looking, but best suited ones for generic algorithms.
I think setback fillets are looking very big, which lets the eye always focusing on the corner of a shape. If you want that, then its good. So yeah, there is no black and white.
The Boxfillet.gh example certainly is no setback corner. What Mikey showed above is.
And to claim that one formal aesthetic is better than another is futile, because perception remains forever in the eye of the beholder. It would amount to saying a zucchini is superior to a cucumber.
2 Likes
Well, zucchini is better than cucumber on the grill, and cucumber better than zucchini in cold soup…
I could be wrong, however…
1 Like
Futile? No. Never. What’s the fun in that?
Any aesthetic claim is obviously in the eye of the beholder since it is based on emotions acquired through knowledge and experience through a life time, so there are no universal constants defining beauty. But that has never stopped mankind in trying to define it through harmony and equations. And based on those “schools” we can easily argue why one solution is better than another.
In the same way you CAN easily define and argue that a zucchini is superior to a cucumber for any given dish that “requires” that given vegetable for a certain texture, taste or the ability to be fried with out ending up as a watery substance (all driven by emotions by the cook or the consumer), but the moment you mix universal laws into the equation of a words true meaning in the context of the grammar it sits in, then we are all lost in an endless shade of gray and need more time and more than one beer to be satisfied with our own and each others point of view 
Thank good it’s friday! I’m off to the beach so Cheers!
See? Thanks for proving my point : )
Off to the pub now, almost 17:00!
1 Like