Really? Raymundo burgueno finishes cars in only one week.
Hes usually given some car modeled with flat panels and from there he begins doing fillets and transitions for client.
A full car body with all details for manufacturing (the so-called ābody-in-whiteā) takes months, sometimes years to complete. Nobody is capable of finishing that task in just several days, especially regular mass production cars that have a huge amount of body panels and small details attached to them. Even one headlight could take more than a month to design until itās ready for production. This is why actual cars are developed by a huge team of design engineers. A single person would need decades to do that task alone.
If you mean just the basic exterior surfacing for 3d rendering that lacks any engineering details usually not seen from outside, then yes, a simple car body could be made in days.
3 Likes
It was made solely in Rhino 5 several years ago. Itās a plug model for the inner structural panel of a car door. Some areas could be considered almost class-A as they are finished with G2 continuity, though a few rounded edges are G1 fillets due to the requirement to wrap around a round rubber sealing.
The 3-, 5- and 6-sided surfaces are 4-sided surfaces that were split in certain areas.
1 Like
do you really think that class A is about having G2 continuity?
G1 = tangent at end points.
G2 = G1 + same curvature at end points. inverse radius of osculating circle at end point.
G3 = G2 + same torsional curvature at end points. product of tangent and normal vector of three dimensional curve in space. multiplying vectors give plane, torsion measures the twisting of that plane as a function or arc length of curve.
by definition its mathematically impossible to have G2 and G3 continuity on flat surfaces because its curvature is cero. G2 can only be applied when curvature results in one dir, while G3 only when compound curvature.
rhinoās adjustable curve blend command is crap. do you really think that by aligning 2 or 3 controlPoints you are guaranteed to have curvature or torsional curvature? it does not work that way. besides if you āadjust controlPointsā you are breaking G2 continuity because in order to satisfy that kind of continuity, controlPoints need to be positioned at very precise locations.
alias is very smart because when matching for curvature on flat surface it automatically displays an error message suggesting that curvature failed.
just saying because that file of yours looks very mechanical, not sculptural. very boxy shape with no curvature, that G2 continuity is fake. besides from engineering perspective, cnc milling G2 is a waste of time and money, not ideal for your BiW file.
Thank you for your expertise opinion on the matter. The 3d model I created worked beautifully, both for rendering and CNC-milling, and so the real door panel is just as good. Thatās what I think is the very purpose of this kind of design work. 
i could not agree more. class a means nothing
"by definition its mathematically impossible to have G2 and G3 continuity on flat surfaces because its curvature is cero. "
The definition of class-A surface quality is about TRANSITIONAL quality and overall smoothness of a free-form shape thatās critical for delivering nice, beautiful reflections. You can easily have G2 continuity between a flat surface and a free-form one.
the G2 curvature matching will fail in flat side.
Donāt worry, Rhinoās āMatch surfaceā tool does it right.
alias matching command does not. 
Nonsense. A flat surface has curvature with a continuous value of zero. Derivatives of the curvature of a flat surface have a continuous value of zero.
Original source document on Geomtric Continuity: https://apps.dtic.mil/dtic/tr/fulltext/u2/a611758.pdf
For an indepth discussion of how NURBS curves and surfaces work, what is possible and what is not possible using NURBS curves and surfaces see āThe NURBS Bookā The NURBS Book | SpringerLink It discusses requirements for various levels of geometric continuity between surfaces.
Ultimately how the surface worked for itās intended purpose determines if it was suitable, not the details of the math used.
1 Like
cero curvature means
1/r = 0
r = no solution, undefined (mathematically impossible) 
well, at least Alias works that way, thus producing best results possible, because its math is consistent. maybe thats why high-end design studios adopt this kind of softwares.
Actually their math is not āconsistentā especially in areas around trimmed surfaces.
they have a long recurring problem with trimmed corners that makes a little mocroscopic āflagā shaped mess, and causes a singularity to be unjoined. (the dreaded pink dot of frustration)
I worked with several very good class A modelers over the years and had to use Rhino to fix this problem in every alias model they ever sent me. Rhino does not have this problem, and is the single reason I started using it 20 years ago. (To fix my own alias models that wouldnāt print or cnc due to tiny holes in the data)
I was level 3 alias trained in toronto and worked sqeaky clean back then, and nothing I made in alias was watertight because of this issue. They still have not fixed this as of about 4 years agoā¦(which was the lat alias model I worked on)
The fact that people think Alias math is somehow ābetterā is just nonsense. The basics of it were written in the 70ās and it has been bodged together over the last few decades without a real modern rewrite ever. (rhino was completely re written from scratch in v3)
Fun factā¦there are (far) less than 5000 alias auto studio users worldwide.
4 Likes
Well, hello from one of the 5000* (although I just got an e-mail from Autodesk that our license is EOL so weāll see what happens). 
Itās so funny, because some Alias users say exactly the same about Rhino. āNothing is watertight and the math sucks.ā 
In the end, itās a tool, and you use the tool youāre the most comfortable with. Iāve used Alias since 2010, and it took me years to get comfortable with it, and I still feel creatively constrained (but not technically). In Rhino, itās the opposite. It took me just weeks to get comfortable with it, and I feel completely creatively free, but very technically constrained (just look at my post history here).
Weāve both printed (with auto thickness that was added a few years ago) and CNC milled with no issue from Alias, but also done the same with no issue from Rhino.
*Ups, misreadā¦ we donāt use āauto studioāā¦ we used one of the lesser tiers.
1 Like
Agreed. Alias is NOT the best software out there for sure. widely used by industrial designers who care only for styling and rendering artistic concepts. Rhino is my favorite tool for modeling but i have to admit that there are aspects to consider. CATIAās matchingSrf command is perfect.
curvature crv blend match to line deg 1 does not produce professional results.
if you mirror a curve, u have 2 identical curves. both curves are G2 by definition (tangent and identical curvature). controlPoint count is insignificant. if you modify any curve u are producing a difference in curvature radius, thus breaking G2 continuity.
in command history section, rhino calculates the Curvature radius difference of both curves. with this criteria, how do you explain G2 continuity match to line deg 1? in order to calculate this, rhino calculates curvature at intersection point between curves and then substracts maximum val to minimum.
max - min = difference.
if difference is 0.0 units, it means that both curvature radiuses are equal, thus G2.
you cannot have true mathematical G2 continuity in a lineā¦
max - 0 curvature = 0.0 units (no real solution)
the same when establishing G3 continuity between 2 dir curve and 1 dir curve. its impossible to have equal torsional curvature between curves.
i really doubt rhinoās continuityā¦
not quite sure what it does, but somehow does not seem correct.
1 Like
Nobody cares which CAD program someone uses and there is not formal definition of class A. But if you are embedded in an industrial development process, you will need to deliver quality and functional surface models, no matter if you disagree about certain rules or not. Something what this discussion lacks is the fact that modelling something like a car is teamwork, involving many companies and across multiple countries. This is why surface models will be validated, by software and people just doing this sort of work . If you canāt pass this, nobody cares if you can model a car in 1 week. Besides this, what you are probably referring to is a concept model. But such a model is not equal to a production model because it doesnāt satisfies any technical requirement, whereas a production model is far more technical, then artisticā¦ Technical models have very strict constrains. E.g. Take a PDC sensor and place it in the orientation without loosing your nice highlights in your bumpers. Can be very tough, much harder then a complex corner blendā¦ Does this answers your question?
3 Likes
adjusting controlPoints to make highlights and reflections appear sleek cannot be technical work, it may be hard and everything, but not technical.
like it or not, car bodies are not that tecnical nor aerodynamic. not convinced? search for famous car designers: giles Taylor (rolls royce) + Bespoke, Sasha Selipanov (bugatti & koenigsegg), etc. they are not technical people, only artists and artisans. they starts by sketching random psicodelic lines and then hire a team of NURBS experts to surface it shortly after that.
formula 1 cars do have intensive CFD analysis (aerodynamics) and consider the properties of materials used and security to far greater extent. formulated carsā¦ very expensive, designed in specialized labs. notice that f1 cars are not that hard to surface compared to common cars.
Its not the only thing you do. You create a shape to fulfill technical needs, not only aesthetic ones. This is what the stage after the conceptional design is all about. That is why it takes so long to create a functional model. If you create a grille you need to make sure enough air comes in, you create a window, you need to make sure you can pull it up and down, you create a door, it should open. Your air vent should work and your sound system as well. There are regulations on how sharp a shape can be, if you create a headlight there are tons of regulations. And so onā¦ There are so many functional parts in a car, most concept designer donāt even care about. Which is okay in that stage. But this is why a concept car looks different to what it becomes later.
Furthermore whats so technical in running a CFD analysis and why do you think a modern regular car doesnāt need optimization on that one?
Of course not everyone doing this kind of job is an engineer, some havenāt even seen an university. But in any case, those people have strong understanding on how to combine highest aesthetic requirements with its actual functionality, and they also have enough expertise to identify problems.
Anyway. Iām not going to convince you. Iām not doing this kind of work anymore anyways, but in case you are an automotive designer (or you want to become one), I would highly recommend to visit those people who make a design idea become reality. You might be surprised how much effort it actually is. And how much āfameā those people should become as well. All those people you mentioned (and their work) are only very tiny parts in the process of creating a car.
6 Likes