Meshing Question Why is my ship's exterior striped?

[thepenguin]

I have recently been meshing a new ship (called the Excelsior). However, the outside of the ship looks like this:

As you can see, the middle of the ship looks uniform, but the ends are strangely colored. The entire ship is made of the same material, but those parts are different mesh groups. It should look something like this:

Here's the mesh file itself: https://www.dropbox.com/s/vvl2ssr4s7jg9hk/Excelsior.msh

So, what's wrong with it? I'm pretty new to the whole meshing business, so any help would be appreciated.

 

[orb]

From a quick look at the .msh file in a text editor, I think it's because there are no normals included and maybe some of the auto-calculated normals get a wrong direction or something.

Try (re-)calculating normals in the meshing program and then reexporting the mesh again.

 

[thepenguin]

Try (re-)calculating normals in the meshing program and then reexporting the mesh again.

Orb, when I tried to calculate the normals in the OMW mesh viewer yesterday, I got the same result. Do you have a better way to calculate normals? Is there a specialized program that most people would use to do this kind of mesh work? Right now I'm looking at 10,000 find and replace operations to deal with the normals, but I wouldn't want to have to use that script if there were another option.

 

[Loru]

Why not use Gmax or Blender to create meshes?

I tried to import your mesh into 3d studio but it turned out error

---------- Post added at 05:41 PM ---------- Previous post was at 05:26 PM ----------

BTW - your issue looks like multiple vertices in the nose tip.

In this tutorial around from 4:30 there is same issue and way I deal with it.

 

[RisingFury]

Use Blender, not GMax. The learning curve is much easier and as far as functionality goes, Blender has everything that GMax has when it comes to Orbiter.

There are a few specific tools that GMax does better, but overall Blender is probably easier to work with.

 

[Hielor]

Of course, if you're wanting "easy to use" and "fully compatible with Orbiter," there's always Wings3D. Wings also has the advantage of its face model being basically unable to come up with the kind of malformed geometry you have there...

 

[n0mad23]

I see no advocates of anim8or remain these days. Must be a paleolithic thing. (Going to have a look at Wings3D as it's been on my "to do" list for a while now.)

 

[Pipcard]

The members of Forum Orbiter Italia tend to use Anim8or for their add-ons.

 

[PhantomCruiser]

I use anim8tor, but I'm suffering from developer block at the moment. Mostly a time constraint though. Once the kid is again away at school I'll have more time available for my own stuff.

 

[BruceJohnJennerLawso]

I see no advocates of anim8or remain these days. Must be a paleolithic thing. (Going to have a look at Wings3D as it's been on my "to do" list for a while now.)

I try, but then they tell me to go back to my cave :lol:

Admittedly, its not really a great modelling workflow, and the process of getting it into Orbiter is awfully tricky. Im not really doing any modelling right now anyways though, so it doesnt really matter that much.

 

[gattispilot]

I use Anim8or.

 

[liber]

How much I can see your model have z-fighting at front and back section. In individual group you have merged front and back face so this might cause you a problem.
Try to make front and back part as you made a middle section from only one face never mind if is in group but just single face in one group,facing out.

 

[Loru]

Yeah - Z fight is another possibility. In what program you've created this mesh? Info would be bit helpful to diagnose problem.

 

[thepenguin]

How much I can see your model have z-fighting at front and back section. In individual group you have merged front and back face so this might cause you a problem.
Try to make front and back part as you made a middle section from only one face never mind if is in group but just single face in one group,facing out.

I don't exactly understand what you are saying, liber. What is z-fighting? I've never heard of it.

Okay, google answered that for me. I don't see why any two things would be defined in the same plane to cause a z-fight, though.

There is a reason that there are all of these different groups, by the way. I was planning to animate them.

Yeah - Z fight is another possibility. In what program you've created this mesh? Info would be bit helpful to diagnose problem.

Loru, if you wanted to know the program, it would be Microsoft Excel. I could implement the same generation algorithm in just about any programming language, though.

 

[liber]

Of course for animate group but you have modeling glitch,in one group is merge front and back face,make back and front face but each face back or front put in it's own individual group.

For smooth front and end you need to remodel it. But check this mesh.

http://speedy.sh/wVam6/Exc.msh

 

[Cizurator]

I'd say it's about parameters you put in to create normals. You have to raise the parameter that sets how sharp angle is required to make edges "hard". I experience this with Wings3D often, but i don't know how to figure this out in your program.

 

[Hielor]

Out of curiosity, why would you go to the trouble of making the mesh in Excel (or anything else, really) when you could do the same thing in a meshing program for significantly less time and with fewer errors?

 

[thepenguin]

Out of curiosity, why would you go to the trouble of making the mesh in Excel (or anything else, really) when you could do the same thing in a meshing program for significantly less time and with fewer errors?

I would have to say that it's just because of how I think, really.

In my mind, the ship's hull is defined as being a 16-gon with radius:
[math]R=51-1.02^{|Z|-1.45}[/math]
So the vertices are like this, where i is the point's index:
[math]Z=Floor(i/16)-200[/math]
[math]n=i mod 16[/math]
[math]\theta=\frac{\pi}{8}*(n+Z/200)[/math]
[math]X=R*sin(\theta)[/math]
[math]Y=R*cos(\theta)[/math]
so the vertex is:
[math]<51-1.02^{|Z|-1.45}*sin(\frac{\pi}{8}*((i mod 16)+Z/200)),51-1.02^{|Z|-1.45}*cos(\frac{\pi}{8}*((i mod 16)+Z/200)),Floor(i/16)-200>[/math]
Which may seem complicated, but it makes a lot more sense to me than trying to get something right.

However, If there were a modeling program that would let me put in a 3D equation, I would certainly use it instead of excel.

TL;DR: If I could specify Equations in Blender or something, I would use it. Is there something that I should have looked for?