Tangency and Continuity

[San_Escobar]

Hey Folks

Happy new year to you all! Have been while since my last post 🙂

I recently been using this Sculp operation (T-Spline) and I came across some silly issue for me. Modelling something which in the eyes and zebra stripe looks good, until when I plot the curvature comb on the body It looks bad.



So I wonder if there is trick or technique to go around and get the tangency and continuity to respond well??? Does my Box Mode matter if it looks bad (like porcupine) lol but if on Smooth mode looks good. Is it the right principle? 

I personally don't like mess, as it tends to bite me later when I want have a minor adjust also I tend to think like when I model in Alias to keep everything nice and organised like have the cvs aligned and distributed evenly.

Looking forward to to hearing from you. 

Cheers!

[O.Tan]

@cekuhnen, I guess you'll be able to give an excellent explaination on this? 😄

I do get these as well but I'm not sure why as "visually" and in zebra it looks smooth 

[cekuhnen]

@San_Escobar

So here is the deal:

Zeba is a nice toolink tool to detect if edges are not aligned meaning detect G0 coniditions.

Otherwise Zebra is useless.

For that you will use the curvature comb. Alias has a specific tool that is like a plane slicing through an object and it shows you

the curvature along that cut section.

The kinck you see is because T-Splines can only offer G1 transitions and not G2.

[San_Escobar]

@cekuhnen thanks for your input. 

I do understand it, what I am after is a tip & trick with t-spline to get this smooth. also to enhance my principles 

If the Box Mode looks ugly and untidy does it affect in anyway? what technique do you normally use? Do you actually align your sections on Box Mode and get the shape required or you have professional standard that you could share?

Thanks again. 

[PhilProcarioJr]

The approximation control mesh can definately effect this in my experiences with T-Splines. I would start there. Care to share a screen cap of the control mesh?

[cekuhnen]

@San_Escobar I am lost now.

Do you want to know how to get the curvature comb consistently smooth?

[San_Escobar]

Hey @PhilProcarioJr and @cekuhnen

Phil I don’t get, what you are saying. “The approximation control mesh” what mesh? You mean the t-spline patches, in Box Mode??? Also I don’t get what is this control mesh, can you clarify so that I can better contribute w/ you.

Clas, No, I want to know, when I am on BOX MODE (Alt+1) the model looks all messed but on SMOOTH MODE (Alt +3) it looks good, and the curvature is close to continuity (Cuv. Comb) and nicer also on Zebra stripes. So does the alignment in Box Mode affect anything? Or by the book, I need to keep it tidy as possible too?

Also, what is your technique to have the T-spline Curvature continuous, avoiding max. to have any kink on surface joint.

Here is the piece of file for your review, If you check you will see that the Comb is resonable good, (PLEASE DONT JUDGE THE CURVATURE ACCELERATION IT IS JUST A SAMPLE) 

Thanks again

[cekuhnen]

Here is what T-Splines and a BREP body created from a TS body can show you

You will have kinks. The kinks mean that there the surface is only G1 not G0 and not G2.

You will not have perfectly G2 like combs when the technology (TS) can only generate G1 between patches.

The screen shot shows that all so called kinks are G1 indicators for edge flows.

How to do it better? Use Alias and do this in NURBS modeling if you want G2 or G3 transitions.

BUT what you have is G1 so it is kinda smooth at least it is not G0 so it should be visually fine in most cases.

[TrippyLighting]

I've never had a chance to study industrial design, so while I understand the concept behind G1,2 and 3 transitions, what do they provide in practical terms ?

@cekuhnen would you have some practical examples where these mathematical considerations make a difference ?

For example if I design a part for 3D print such as these, would it make any difference whether these are designed with T-Splines or with NURBS ?

[cekuhnen]

A good placed G1 can fake a G2 transition.

The best example is this:

make a cube in fusion and G1 round all edges

make a new cube and G2 round all edges

then compare you will see that with G2 the surface accelerates much slower generating a softer transition over a longer distance.

That results into more stable highlights.

That is why I find it pathetic that in 2015 ah 2016 you still see so many products on the market with disgusting G1 edge roundings.

And honesetly and embarassment for the engineer who did this or management that was ok with that.

In theory a G1 transition mold could be cut faster with the correct CNC milling bit while a G2 G3 transition will require more work.

However the question is if for organic shapes G2 really would be that much of a difference because with T-Splines we already deal

with non flat faces but curved once.

When I did print some products sculpted in Blender and then brough into Fusion the surface was fine. But I also made sure that

all faces are evenly spaced out helping with the smooth feeling. But you were not able to even see where patch edges were in the print.

I think G2 and specifically G3 is insanely important for car bodies to stabalize the shape of highlights.

[TrippyLighting]

Yes, but you know what to look at 😉

So for the most part as I had somewhat gathered, this is importeant for glossy and other highly reflective surfaces. As such for most currnent 3D printed stuff not so relevant.

The point you make with the even spacing is one that is often overloked when the general aim is to mdel something with the least amont of polyons and I believe that this mesh irregularity is also the problem with the OPs mesh.

[San_Escobar]

@cekuhnen thanks for your imput, Make sense that the curvature is broken as its only G1 ratio. (Tangent)

However, analysing the facts, I understand that it wont matter how the BOX Mode looks like as far as I get a nice SMOOTH Mode conform to. Right...

Have said this, I am trying to understand this TS principles and I came to terms that IMO it is not a feasible industry standard modelling practice, but may be used for concept purpose. As I image that if you can allow things to be un-tidy, for self-experience I know it can bite you later, when an update or precise move/adjust has to be made. Thanks again!

@TrippyLighting I normally tend to imply a great surface blending in all my models, even when comes to Solid modelling like in Inventor, SW, F360. or whatever, as you can see image attached. Again, I know for fact that if you leave a kink on surface depending of the grade of 3D printing to be produced it will show on the physical model. For instance I did work in a high end jewelry company and I use to print staff on SLA 0.003mm or 0.002mm layers. and some kink was just ugly on the eyes. Also I have seem some bad polygon model been printer and It looks ugly too. But I suppose for FDM it doesn't matter of course depending on the part to be manufactured and its purpose.

You made a good point, as I mentioned, I always tend to keep the models structure well organised and evenly spaced control points be it CV, Control Point or a BOX Mode.