network or other

Next
 From:  Apophis
1881.1 
Hello

I wanna find a way for a better surface result. I've choice the network function to create the surface. Unfortunately the result isn't really good. The mesh is horrible. Is there another way to create the surface with better results ( mesh )?

Regards














  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  PaQ
1881.2 
Hello, here's my method for this kind of surface.

For technical explanation -> Michael will maybe answer you.

Based on my little xp, I would say network like to work with 4 curves.
So I split (trim) the base curve with the middle one.
The curves selection order is important in this case (I thought Moi is resolving it automatically, but some selection combo don't work I don't know why).

This method will put the pole surface on the front of the object, so the result is quite ok.




(I think it's the worse .gif capture I ever made)

EDITED: 3 Feb 2010 by PAQ

  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  BurrMan
1881.3 In reply to 1881.1 
I did this one by using sweep. Then play with angle and divide larger than in export to adjust mesh as you want.

I think network may be more for shapes that are more wavy and cloth like. when you have a smooth surface, look for sweep and revolve.

I was originally trying to do it with revolve as the surface from this function is beautiful, but I couldnt figure out to get the top profile in it. (maybe non-symetrical cant be revolved, but rail revovle with "Scaling Rail" would have done it! Dont laugh, please! :) )

Talk later,
Burr

EDITED: 19 Jun 2012 by BURRMAN

  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  PaQ
1881.4 In reply to 1881.3 
Hi Burr,

Looks like there is an ugly pinch in the front pole surface in your sweep method :P
  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  BurrMan
1881.5 In reply to 1881.3 
I also noticed you were using quads and triangle with the export. you may preffer this so I tried with that.

The best result I could get with the model I posted was angle 7.5 divide larger than 3 and avoid smaller than 2. I still was left with some hatching on the back and a smalll one on the nose bottom.

Do these make a difference? does it render better than the version created by paq's proper network method? Maybe a render guy can chime in here.

Thanks for the time,
Burr
  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  BurrMan
1881.6 In reply to 1881.5 
Ah I see, so all that coming to a point there is bad. Thanks paq.

Burr
  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  PaQ
1881.7 In reply to 1881.6 
Well the meshing output between the 2 method are quite similar, it's just the shapes at the pole that are a little bit different


EDITED: 3 Feb 2010 by PAQ

  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  BurrMan
1881.8 In reply to 1881.7 
Looks like the network one is a much better set to try and material. I was focusing on only the patchwork of the mesh (narrow minded).

Thanks,
Burr
  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Message 1881.9 deleted 16 Aug 2008 by BURRMAN

Previous
Next
 From:  Michael Gibson
1881.10 
Places where things come down to a single point can tend to have awkward shaping.

Sweep and Network calculate the surface in fairly different ways - Network is kind of more like a loft through 2 directions at once, and sweep produces a shape by sliding the profiles along the rails, which can kind of give the rails more of a kind of "weight" to the final shape I guess.

Sweep would have probably made less of a problem there if the center line had more of the same kind of curvature as the bottom rails - in this case the center line is sort of more flat than the bottom part.

- Michael
  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  Michael Gibson
1881.11 In reply to 1881.2 
Hi PaQ,

> The curves selection order is important in this case (I thought
> Moi is resolving it automatically, but some selection combo
> don't work I don't know why).

This is a bug - it is getting confused when looking at curves in a certain order and running the coons patch method instead, which is a special kind of network that can be done with 3 or 4 curves, but not in this kind of arrangement.

If the curves are ordered in a certain way then it does see that it should not do a coons patch and does the regular network.

I should be able to fix this up shortly.

- Michael
  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  BurrMan
1881.12 In reply to 1881.10 
Would appreciate any comments here for my learning.

Since I like the revolve surfacing, I did this by rail revolving the bottom curve with the arc, then turned on points and stretched surface to match top curve.

Do I have the same result? Is it better or worse? Is it a 50/50 for the texture program?

Thanks Paq and Michael,
Burr

EDITED: 19 Jun 2012 by BURRMAN

  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  Michael Gibson
1881.13 In reply to 1881.12 
Hi Burr, the revolve method is certainly a fine surface, but not quite exactly the same shape as the other ones, if you look from the side:



You can see that the side-view profile of the surface does not quite match that middle rail.

But if you want a nicely formed rounded tip and don't mind about such exact control over that "down the middle" profile curve, then rail revolve is a great way to get a rounded shape and it has an advantage that it uses a small number of curves to define it.

- Michael
Attachments:

  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  BurrMan
1881.14 In reply to 1881.13 
I see said the blind man. Once again, no attn to detail on my part... I'm learning though!

Thanks again,
Burr
  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  Michael Gibson
1881.15 In reply to 1881.14 
Hi Burr, it is a good idea to experiment with the different surfacing tools to see what you get - and it is certainly not a bad idea for when you want a rounded tip to think of revolve or rail revolve.

You just give up some of that "down the middle of the surface" control. So it is a little different, but still for many purposes that could be fine.

The nice thing is that revolve and rail revolve tend to make a nicely formed surface throughout, including right at the tip points. They don't tend to really have the same kind of problems near the tips as other commands do, for other commands the curves tend to go through a sort of rapid changing shape near the tip there which can make for some more difficult shapes in that area.

- Michael
  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  Michael Gibson
1881.16 In reply to 1881.2 
Hi PaQ,

> The curves selection order is important in this case (I thought
> Moi is resolving it automatically, but some selection combo
> don't work I don't know why).

This bug is now fixed for the next beta - it was not properly testing every single combination to see if the 4 curves had a curve down the middle like that (and so was not good for Coons patch) but it would get it right in certain orderings.

In the next beta it will test all combinations properly so the ordering will not make any difference for this shape which is what was supposed to happen.

Thanks for reporting the problem!

- Michael
  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  Brian (BWTR)
1881.17 In reply to 1881.16 
I think it was with a bit of luck, but this wing profile worked out well for me in MoI eventually.
Render is in Carrara.
Brian

EDITED: 31 Dec 2008 by BWTR

  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
Next
 From:  Apophis
1881.18 
Morning

So many answers with different possibility. I will try some of your suggestion.

Many Thanx
  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged

Previous
 From:  BurrMan
1881.19 In reply to 1881.18 
Morning Apophis,
Mostly the first from PaQ was a legit answer. The rest was me trying things that didnt quite work for your exact curves. The only thing I found was if your curve shape is slightly flexable then the revolve would produce a nicer surface. Just revolve cant work on those exact curves.

Good luck,
Burr
  Reply Reply More Options
Post Options
Reply as PM Reply as PM
Print Print
Mark as unread Mark as unread
Relationship Relationship
IP Logged
 

Reply to All Reply to All