Hello,
I wanna throw a big n - gon on the ends of a helmet.
Its a hard surface with no deformation at all.
Just a hard plastic cup basically.
I tried this and, triangulated inside Blender and brought into UE.
Seems o.k.
I haven’t tried to UV map it yet.
Can anyone tell me if this is a very bad idea and why?
Thank you!!
From what I understand n-gons are o.k. on hard surfaces and the model is going to be triangulated anyways before export
from Blender. So I dont see why not.
I would guess the only issue would be how it affects the texture you plan to use. Because it will create some funky triangles. Any detail will be distorted.
Yeah I am going to UV map and texture it now and see what it does. I will post results. Thank you.
Here is the result. Looks fine to me. No distortion or pulling.
The image is a little soft that is fine.
ngons are fine on flat surfaces. unless u plan for the peace to bend/morph from its original shape.
O.k. thanks that is what all the forums / blogs say too. That is what I thought.
Looks like ngons vs triangles vs quads only has to do with deformations anyways.
I find a combination of triangles and quads to work for me.
Sticking to quads only to me is a little inefficient and you tend to create more polys than needed.
Since you triangulate everything anyways for engine’s calculations I dont see why not use triangles.
As long as their placement is thought out so you dont get weird tearing or edges that deforms in a weird
way that breaks the topography of your model.
If I HAD to break that flat surface up in quads just to stick to the “rule” then I would end up with a ton
of polys adding “weight” the file size.
what rule are you talking about?
I just get the idea from looking at examples and reading forums that with 3D modeling its a “rule” that you stick to
quads. Maybe I am wrong 
yeah they tend to say its a rule when learning to model as its a good modelling practice to stick to all quads. But once u get to a certain skill point u can get away with knowing when and when not to stick to the quad rule.
ah, that’s what I figured you meant, its not really a rule but more of a guideline, don’t get me wrong in most cases using quads is the best/easiest thing to do(it also helps with shading/smoothing (for example making a cars body is easiest to make using 99.99% quads but some parts like door creases require the use of tris)) but in almost every decent mesh I have ever seen whether it was made by a hobbyist/freelancer/indie or AAA studio the modeller has had to use tris for certain parts (its almost unavoidable:)) so basically just use what you need to, as you know they all end up as tris in UE4 anyway;).
Yes thank you. There are many articles online if you Google around where they post the most common mistakes beginner modelers make.
I guess they never Google “common mistakes of beginner 3D modelers”
If you’re going to bake you should triangulate first. Normal maps are generated based on the normals of the low poly mesh, which will vary depending on how the mesh is triangulated. Since triangulation methods differ between applications you can sometimes get inconsistent results if you try to bake an untriangulated mesh then reexport it for another program (like UE4)
Also, the rule should be “keep your topology clean and manageable” not “generally use quads”. It just so happens that if you keep your topology clean and manageable then yes, you will usually always end up with a mostly quad mesh. The problem is that the opposite is not true, a mostly/all quad mesh does not necessarily have good topology. In fact I would usually say it doesn’t, in order to keep a complex mesh all quads people typically have to do really bizarre things with their edge flow in order to make it all work.
If anything the most common beginner mistake I see is people religiously adhering to all quad topology with no regard for the quality of how they’ve laid out their geometry because they think triangles are “bad topology”. Clean topology doesn’t have anything to do with the shape of the polygons.
The main argument against it is control - you can’t be sure how the engine will tesselate it. If you want to make sure it’s done in a way that won’t lead to any artifacts just turn it into some proper quads before export.
O.k. noted. Thank you.
The rule mostly comes from old hardware back in the PS2 days. Sony hardware has always been a work horse poly pusher from the psp on up but shared edges per vert were areas that use to hit the systems hard (I believe the GameCube took a similar hit). I always clean up those faces manually out of habit these days. When you are counting milliseconds to get a level into frame, every tiny bit helps.
I see. But here is a question though … if a ngon or triangle saves extra vertices from being created isn’t that a good thing though?
I find with quads only I have to create a lot more faces.
In terms of game engine calculations for those nano seconds.
