THE N-GONS PROJECT

[Azrael]

NGONS usually look like polygons. But if you could add some kind of option to "Highlight " Ngons in Orange or some other color. Then you wouldn't need to hunt for them in the final cleanup.

[Bellorum]

N-GONS==Control

N-GONS let me place geometry where it is needed at that time. The end-end result is usually a mesh with 4 and 3 sided faces. But I love having N-GONS there when I am initially sculpting. It's a great thing for control. It helps to keep the base mesh nice and clean.

If blender added N-GONS I would (and others) feel more comfortable modeling in it. Especially when cutting into a single face.

100% agreed. Unfortunately some people bashing n-gons have hardly ever even tried working with them. You can argue all gets converted to tris at rendertime, but it's just not that simple. When subsurfed, a 5-sided n-gon doesn't render the same as a quad and a tri. But it's workflow that makes me intrigued. I can sketch up a model in Wings 3D faster than in Blender, and still I've used Blender for a longer time. On the other hand, there are things about Blender workflow I like better, so of course I want the best of both worlds in the same package.

[cessen]

N-Gons are always fake anyway. In the end all your graphic card understands is triangles and vertices.

Wrong. By that argument we should eliminate quads from Blender, too, since triangles are obviously the only "real" polygon.

Yes, video cards and many 3D renderers do tesselate the geometry before rendering it, but that doesn't mean the geometry is fake. Are quads fake? No. Are NURBs fake? No. Just because they are tesselated at rendertime doesn't mean they're fake.

In the case of N-Gons there aren't standard rules on how to interpret them if they're non-planar (that I know of), but that still doesn't mean they're fake. They're definitely not fake as a data structure.

And as has been reitterated many times here, they are useful as a modeling tool. Their main use is for box modeling, where maintaining clear, clean edge loops during the modeling process is often more important than keeping the model all-quads. It also lets the artist easily dice up the N-gons into quads themselves, instead of having to undo the tesselation done by the program first (which is a nasty work-flow barrier that slows things down). It also lets the artist more easily localize detail in the mesh while modeling, and propigate the edge loops etc. later.

Also, it been mentioned that N-gons have a predictable effect on subdivision surfaces that can sometimes be exploited to good effect.

[Azrael]

Any word on when NGONS will be implemented? I am looking forward to creating a model 100% in blender.

[jesterKing]

Any word on when NGONS will be implemented? I am looking forward to creating a model 100% in blender.

But you already can!

/Nathan

[Azrael]

Blender's lack of NGONS leads to some serious control issues when it comes to cutting my mesh. These unfortunate issues lead me away from modeling in it. I, and many others tend to model things in a quick and dirty way, that just lets me let loose on my work.



That image began in a process that looked much like this:



This illustrates how NGONS can be used in the construction of a mesh, and how we can use them in our favour to work on something relatively quickly.

this a tutorial shows another example of box modeling using NGONS in construction: http://ambient-whisper.cgcommunity.com/ ... age-1.html

[Sonic TH]

Yes that was my first reaction when I discovered Blender didn't have N-Gons, "well I can't use this". But you gotta think about this, Blender is open source, if people shun it because it doesn't have a feature they want then it most likely never will. I hope that sooner rather then later, we will have N-Gons.

[Azrael]

It leads me away from modeling in it, but not from UVW, and rigging. But if it had NGONS, modeling in it would be superb, on par with all the other great features.

[artisanicview]

Just a short remark. Actually this thread have visited 15246 times. I think is one of the most vieved threads of this forum. And I think this is normally beacouse of the importance of NGONS implementation into the Blender.

[Youssef|3d]

some ideas for blender

http://www.nevercenter.com/videos/index ... review.mov

realy like that tab where you can edit your primitives and the split tool

[Eternl_Knight]

NGONs are very important for modelling (and I agree - the number of hits this thread seems to get implies how many people would like to have them implemented).

I currently model completely in Wings due to its incredibly smooth workflow for just that task. It's support for NGONs makes re-topologizing a mesh for animation very simple. While the Blender modelling workflow is not as smooth as I would like, the lack of NGON support is actually the deal-breaker for me.

Has there been any progress on this or has it been abandoned?

--EK

[greboide]

http://mediawiki.blender.org/index.php/ ... Dev/HeMesh

i think that it can interest someone willing to help in this project.

[James Morris]

having ngons would definately attract people, but the truth is that almost all of them would abandon them if they really learned how to model (calm down). blender's fgons make ngons almost unnecessary, but I think what we really need is for blender to create the f-gons automatically. for example, take one side of a cube and subdivide it (noobies expect four faces but are horrified at the diagnol cuts across the other sides of the cube), but have the fgons created already (should be a toggle button). The result to the user would be that they would think blender has ngons.

[psyborgue]

having ngons would definately attract people, but the truth is that almost all of them would abandon them if they really learned how to model

eeh. I think it's okay to use ngons as long as they are not still there when you are finished with your mesh. I used to model primarily in wings and was very used to doing this. It's nice to be able to delete an edge (creating an n-gon temporarily), and then reconnect the vertices in another manner resulting in better topology (without having to re-create faces). Before i finish, i make sure the mesh is 100% quads.

That being said, there are better ways to model which result in better planned meshes, as well as other ways around the n-gon limitation in blender (such as creating a polyline cage). It took me a while to get used to modeling in blender withoug n-gons but now that i am used to it, i have to admit that i do plan the topology of my meshes better.

However, for those who come from other programs and really don't want to re-learn to model, it would probably be very useful to have ngons (or f-gons if they work well enough). It's hard to argue not having a feature makes a program better. After all, almost every single other 3d app i can think of supports n-gons. Are they all wrong?

[BackwardsDown]

Blender does "need" to have NGONS in my oppinion. They are very handy and are used in a lot of programming-tutorials. Also, people would abandon blender less, becouse in my opinion, NGONS is the 1 big thing I miss in blender.