640x192

triangle-quad-ngo....png

๐Ÿงต N-gons

Anonymous No. 887127

Do you use n-gons in your modeling? Why or why not?

Anonymous No. 887131

No
Because autism.

Anonymous No. 887132

1. Yes.
2. Topological requirements. Doesn't happen too often.

Anonymous No. 887135

Yes for concept and blockout, no for retop or subdivision.

Sometimes if I know I will be smoothing the mesh to add more detail later I will leave it in if I find an area shades best with it soon as it will be converted to all quad by smoothing

Anonymous No. 887136

>>887127
No because I'm not a retard.

Anonymous No. 887145

>>887127
Yes, of course. The sneedustry standard modeling workflow requires meshes to consist solely of ngons. Quads should be avoided unless they're mostly out of sight. Triangles are acceptable, but you should always try to form an ngon out of them if possible.

Anonymous No. 887151

>>887127
I model exclusively in ngons. Even for a simple cube shape I add additional vertices along edges just to get nice ngons everywhere.

Anonymous No. 887152

>>887127
No
Subdivision doesn't like them
Makes it easier to select edge loops

1103x778

yes.jpg

Anonymous No. 887168

yes.

Anonymous No. 887177

>>887151
king

Anonymous No. 887195

>>887127
They're all triangles.
Quad is 2 triangles.
Ngon is 3+ triangles.

Anonymous No. 887200

>>887195
Yeah idgi, it's all fucking triangles anyway. No easier way to get planar surfaces when it inevitably needs to be crunched by the software.

307x475

file.png

Anonymous No. 887201

>subdivides once
what ngons?

Anonymous No. 887202

>>887195
3edges pole = a smoothed triangle
5edges pole = a smoothed pentagon

Anonymous No. 887295

>>887127
fine for hard surface since it's non deforming bad for organic

Anonymous No. 887310

>>887152
>Subdivision
just add more ngons til it's smooth

Anonymous No. 887327

>>887295
Uuuhhm aktchusually, lots of organics are also not deforming, like any kind of prop in nature

Anonymous No. 887364

>>887127
always. as long as an ngon stays completely planar and manifold, they behave perfectly fine. if the geo really really needs to be quad, it's easier to cut up an ngon into tidy poles and loops than it is to exclusively model in quads and clean up accidentally weird topology. triangles are also absolutely fine to use, especially in subdiv modelling. modelling in only quads is a sure trait of someone who isn't confident with topology fundamentals and swallowed the quad pill because it was an easy way out.

Anonymous No. 887368

>>887127
Yes, realtime mechanical shapes that won't deform like guns, vehicles or various environment static meshes etc can benefit a lot
from employing n-gons on large flat surfaces that have to connect to surrounding geometry with any number of vertices.

As long as you understand how shading works and you use them in the right place not using N-gons means you're a noob that have yet to learn.

Anonymous No. 887371

>>887195
>>887200
The issue is if you don't specify where an edge should be put (in other words specify which things are triangles), the engine (game, render, viewport, whatever) will try to use an automated process to do it for you.
The problem there is that:
1) it may end up recalculating it every time a change occurs (change of camera position, model moving, etc) and this may cause a flickering effect as it decides to put the edges somewhere else.
2) You my have (no pun intended) what's called an "edge case" where, mathematically, both values or placements are correct; this tends to cause crazy artifacts.

Most engines have some kind of triangulate pass when you import, but again this doesn't solve the problem; an automated process is going to pick where to place the edges and that may or may not cause problems.

>haha lmao we were joking of course we already knew! imagine writing a real response or worse, an effort post!
My post is for any noobs that blundered into this thread

Anonymous No. 887376

>>887368
>As long as you understand how shading works and you use them in the right place not using N-gons means you're a noob that have yet to learn.

If you're using ngons in the first place you need to learn basic topology. Don't project your lack of knowledge on others faggot

1350x2295

MultipleNgonNoSup....jpg

Anonymous No. 887388

>>887376
If you're not using N-gons you need to learn advanced topology. Don't expect everybody to be as novice as you are faggatron

Anonymous No. 887390

>>887388

The topology is horrendous holy shit. You can't be serious. Ngons are extremely easy to fix but you choose to cope and remain bad.

Anonymous No. 887392

>>887390
>The topology is horrendous

What the topology looks like on an object that doesn't deform doesn't matter one tiny bit. As you can see it subdivides and shades perfectly fine despite them numerous N-gons.

If I only showed you the shaded version you'd have no idea it had messy topology with numerous poles and N-gons. Everywhere that model needs to be quaded it is quaded. Where it doesn't need to be quaded - it isn't quaded.

If you know how and where to use them - use them.

1386x834

bakedNgons.jpg

Anonymous No. 887393

>>887392
Once you bake something the topology means even less. The lowpoly that goes to the engine has even more N-gons as this is the most effective way of building it. The 'never N-gon' is a total meme, it is only true when you're new to modelling and have to learn to understand edgeflow and topology.

Anonymous No. 887405

>>887390
t. dunning kruger
the only people ive seen who constantly spout "wtf your stuff isnt all quads!!!" are people who have never worked on anything of substance
no one gives a shit about how you do things as long as it renders out fine

Anonymous No. 887406

>>887371
this is why you triangulate before export in your modifier stack

Anonymous No. 887409

>>887393
>The lowpoly that goes to the engine has even more N-gons as this is the most effective way of building it
There are no n-gons going to the engine because at that point there are only triangles.
Seethe, cope and dial eight.

851x1021

3-angulated.jpg

Anonymous No. 887422

>>887409
>There are no n-gons going to the engine because at that point there are only triangles.

Your inexperience shows again. Ofcourse there are only triangles in the engine, N-gons are always "only triangles".
There is nothing that happens when you export your model that triangulate your model, it just throws away your 'isoline data'.
Extra data that is there for you the human to more easily manipulate your surface, that is just waste data for how the machine draws the surface.
By your logic N-gons and quads (paired triangles) don't even exist. An 'N-gon'is just a name for the isoline display of a polygon with 'N' number of edges.
Maybe there's some exotic CAD/nurb format that have true Ngons but in the usual suspect modelling softwares all polygons are always made of triangels.

If you set your software to display all edges instead of your isoline you'll see the triangulated mesh.
>So you spread your cheeks apart and dilate that third eye of yours anon, you don't have to seethe you don't have to cope.
>Just spread your mind wide open at full goatse FOV and take it all in while enjoying the learning experience laid bare before you
>like the N-gon n0_0b you where, but no longer forced to remain.

Anonymous No. 887423

>>887422
Thank you for proving my point.

Anonymous No. 887424

>>887376
look at this clueless quad fetishist and laugh
topology can benefit from Ngons and triangles but you don't know that yet

Anonymous No. 887431

>>887168
looks good, are you modeling with subdiv or retopologizing an existing model?

Anonymous No. 887456

>>887431
modeling

Anonymous No. 887466

>>887424

There's nothing wrong with triangles on a hard surface model, ngons are fine if you triangulate the model before export. But you don't do that shit with subd modeling.

Anonymous No. 887472

>>887127
no, not in the finished mesh anyways.

i'll occasionally put one in while im still working out some complicated edgeflow on a retopo project, but i try to resolve it as quickly as possible. they can create problems and my render guy has complained about them in the past, so its best practice to just not use them. i know everything eventually tesselates into tri's, its just quads are so much more predictable, you dont get any pinching or artifacts when you smooth the mesh, and at our studio we are constantly moving meshes between maya and max, so its best to have those meshes as clean as possible

Anonymous No. 887473

>>887393
if its a non deforming object topology practically doesnt matter. it really only matters on things that need to bend or deform

Anonymous No. 887491

>>887127
I use ngons exclusively. Artifacts are much faster than calculating ambient occlusion.

Anonymous No. 887495

My renderer doesn't suck(unlike blenders who spazzes out if you subdivide ngons) so IDGAF

Anonymous No. 888075

>>887371
>>887406
This

Anonymous No. 888077

Not usually, because I'm already remeshing and sculpting by the time it would come up.
I should probably spend more time in the block-out phase in the future, though. My modeling skills need more polish.
I don't have any problems for ngons before retopo.