Page 6 of 8

Posted: Sun Oct 01, 2006 8:38 pm
by LetterRip
an ngon subsurf smoothes differently than an fgon will (an fgon use the original geometry of the cut) so more than just auto fgon creation is needed.

LetterRip

Posted: Mon Oct 02, 2006 8:34 pm
by kitsu
I have a rhetorical question:
Arn't fgons the same thing as ngons, only without the automatic tessellation algorithm?

I still think that if Blender could automatically create and smooth fgons it would solve most of the problems their absence causes. Making the connect and cut tools work with them as expected would solve the rest.

Can a winged/half edge data structure be built just for the edges surrounding a fgon? IIRC Anthony D'Agostino (Scorpius) was building winged edge data structures out of Blender meshes in some of his (Python) export scripts. So at least there is enough data there, it just needs to be organized. It can't take that much time/memory to create an HE struct for just a few dozen edges? There may need to be a rule that no fgon can share an edge with another though...

Posted: Mon Nov 27, 2006 10:55 am
by aws357
I really don't think implementing ngons (or a to z gons) will help people do things better.
As a character modeler, I do find that a nice topology gives a nice object. And even if Stahlberg do use ngons, he do so scarcely and always when he is sure the planarity won't mess with how the object looks.

I prefer the work of Cortina, because it looks better and you have to look very carefully to find any strange polys...

http://cortinadigital.com/

I don't think it is wise or wisheable to add ngons to subdivision modelling. After all, even in package like Maya or Max, once you are finished with the modelling and freeze the object, it has to tessellate the ngons into tri or quads...

A skilled modeller only need quads. And tris are not that bad on Blender because the catmul rom subdiv convert them in quads afterwards.

Posted: Tue Nov 28, 2006 7:56 am
by LetterRip
ngons are useful in the intermediate stages of modeling in that you can do face cuts etc on them without having to reorganize manually a bunch of geometry. You don't need or want them for final geometry.

LetterRip

Posted: Tue Nov 28, 2006 9:52 am
by aws357
Well... In the beginning, I thought ngons where great. It would allow much freedom etc...

But it comes at a price when you have to tweak and tweak :)
I prefer to do the things okay right at the beginning.
Make the geometry flow continuously on my topology... (it sound a bit like a surfer speaking of surf...)

But that's from someone who mainly does organic modeling.

Ngon might be great to make boolean object for twisted mechanic shapes.
:? That's all I can think of so far.

NGons

Posted: Tue Dec 05, 2006 3:36 am
by Azrael
How does the presence of NGons stop you from doing things right the first time? ;-)

Posted: Tue Dec 05, 2006 10:20 am
by aws357
Well, when I was a newbie in 3D, I played with maya for some times :)

I was very excited by the possibility to use polygons with more than 3 or 4 side.

I remember doing a human face (don't tell me you never made a "newbie face don't flame" post ) :wink:

It looked ok, but when I rendered the head, there was strange things on display. After posting the image, some local guru said :
"aaaack! tri all those gons and quad as much as you can, and it should be better"

And I spent an afternoon cleaning my mesh ;)
Plus there was some hidden ngons (badly done 5 side poly that looked like quad)

Okay... maybe I am too concerned for the 3D newcomers... Experimented users should have no problem dealing with ngons and doing nice things.

And probably, newbies have to do mistakes to understand why quads are better for organic modelling... :roll:

Ngons

Posted: Tue Dec 05, 2006 3:57 pm
by Azrael
Sure, and a mesh polluted by NGons can be a problem.

However, having NGons for retopologizing the mesh is quite nice.

Posted: Wed Dec 06, 2006 4:13 pm
by Bellorum
This debate is and always will be stupid, because people just can't get past the fact that you don't have to use n-gons if you don't want to.
aws357 wrote:I really don't think implementing ngons (or a to z gons) will help people do things better.
Not necessarily better, no - but for some it's a more efficient way of working.
aws357 wrote: I don't think it is wise or wisheable to add ngons to subdivision modelling. After all, even in package like Maya or Max, once you are finished with the modelling and freeze the object, it has to tessellate the ngons into tri or quads...
A flawed argument, since an n-gon isn't subdivided in the same way as a quad and a tri, for instance. They're treated very differently.
aws357 wrote: A skilled modeller only need quads. And tris are not that bad on Blender because the catmul rom subdiv convert them in quads afterwards.
A downright derogatory statement, especially considering many great modelers utilize n-gons in their workflow because it's the best way for them to get the job done.
But it comes at a price when you have to tweak and tweak
Really? Not having n-gons comes with the price of constantly having to clean up extra geometry created by the cut-tools.
I prefer to do the things okay right at the beginning.
Good for you. I'd say it's a fair bet most of us try to do that. But there's more than one right way.
I remember doing a human face (don't tell me you never made a "newbie face don't flame" post ) Wink

It looked ok, but when I rendered the head, there was strange things on display. After posting the image, some local guru said :
"aaaack! tri all those gons and quad as much as you can, and it should be better"

And I spent an afternoon cleaning my mesh Wink
Well, that's because you suck - not n-gons :wink:

In the end, all that matters is that someone has to be interested enough to code a feauture, and then you don't even have to use that feature if you want it. It's a win-win situation. Leave it to the open snores community to see everything in black and white. It's not the end of the world as we know it!! You commy bastards!!! <- That's a joke :roll:

Posted: Thu Dec 07, 2006 3:13 pm
by aws357
Well... I believe in France, there is a proverb that says only fools never change their mind...

http://forums.cgsociety.org/showthread. ... ge=3&pp=15

Look at these wire... There are n-gons inside. Carefully placed. And not much.

However, notice that the guy does use them most exclusively for mechanical part (armor and all) while the organic parts are quads exclusively. And he takes a great care about planarity of these ngons.

Moderator note: another proverb says huge images increase page load time and break formatting. Can we have a smaller pic, please?

:wink: Okay... okay... n-gons must be useful (else they wouldn't have been invented). 8) Must have had a "sectarian" surge all of a sudden...

If it permits me to do things like this guy, I won't be the one to complain later :D

Ngons

Posted: Fri Dec 08, 2006 3:59 am
by Azrael
aws357: One thing, in this portfolio: http://cortinadigital.com/ the model from the final flight of the Osiris has quite a few Ngons, in the hands and the legs, when Subdivided they create a desired pinch.

Having NGons, as Bellorum said could be used at your discretion.

And about the cosmonaut, how do you know the original base mesh did not contain NGons? :wink:

The point of the matter is, having the option of NGons is better than not having it at all.

Posted: Fri Dec 08, 2006 9:54 am
by aws357
"touché"

Posted: Fri Dec 08, 2006 5:01 pm
by Bellorum
@aws357

No hard feelings I hope, mate! :) I don't know anything about your modeling skills - for all I know you're probably better than me :oops: I just can't resist the occassional cheap shot. :roll:

Posted: Wed Dec 13, 2006 6:02 am
by mjordan
So, where this project is going?

project

Posted: Sun Dec 24, 2006 12:09 am
by Azrael
I think it is going to eventually get NGONS in blender, :-)