[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vr / w / wg] [i / ic] [r9k / s4s / vip / qa] [cm / hm / lgbt / y] [3 / aco / adv / an / asp / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / qst / sci / soc / sp / tg / toy / trv / tv / vp / wsg / wsr / x] [Settings] [Search] [Home]
Settings Home
/3/ - 3DCG

Thread archived.
You cannot reply anymore.

Alright, so forgive me for asking what is a stupid question. Why do I see abominations such as quadrilateral polys in 3D work when I've never seen any such unholy entity in a model made by professional developers?
File: triangulation.jpg (525 KB, 1736x1024)
525 KB
525 KB JPG
Possibly because you are ripping models from games and wonder why they are made entirely of triangles and not quads.
The reason is that all 3D models are made up of triangles but artist creates them using an 'isoline display' that hides certain edges
in order to enchance the human authors understanding of the flow of the surface.

Marking which lines are to be visable and hidden is of no use to the machine rendering the mesh so this data is typically thrown away when you export to formats game engines use.
Pic related depicts what's going on there.
Wouldn't the ideal thing be to use as few polys as possible? Working with quads means you'll end up doubling it, right? I don't know myself since I'm a total noob and only really do low poly stuff anyway where I find triangles suit my needs much better.
It's the same regardless, say you need to make a simple plane if you do it with quads it takes one, if you worked in triangles there is no way of reducing it less then 2. In the same way when making a character with 100% quads you will always double the amount when going to triangles and if you where to work in triangles from the start you wouldn't be able to reduce the amount
>abominations such as quadrilateral polys
Man, I hope this is satire.

>only really do low poly stuff anyway
I guess this is where you fall apart. This isn't a valid excuse, but I'll turn a blind eye since you're starting out.

So there are three primary reasons why you want to work in quads.

>Human Readability
It's easier to work in quads. It's as simple as that.
The four-edged nature of a quadrilateral means it tiles cleanly (when your topology is done right).
Take a look at the image posted in >>654786. Check out those clean flowing lines in the comparison on the left. Contrast that with those on the right. The right is much more difficult to pick out edge flow, as it's called, compared to the left.

Remember those little arts-and-crafts doodads where they had those "fold along the line" markings? That more or less carries over to 3D. Having everything in quads means you have consistent, predictable lines to where your model can "fold" when it's deformed. Compare that to a bunch of haphazard triangles, where if you try to fold it you'd just get a bunch of odd spikes and everything like that.

This one's unrelated for you, but if you ever move up to working on bigger more detailed models that aren't just for games, this is important. Quads split predictably and easily. You cut a rectangle into four smaller rectangles with ease. But can you do that with a triangle? Sure, you can, but it's much more difficult and won't look as pretty. And imagine having to do that on a bunch more triangles. It gets ugly quick. (for an example, create a cylinder, give it some sub-d, look at how the caps become all weird and spiky because the caps are made out of triangles)
> Wouldn't the ideal thing be to use as few polys as possible?
contemporary graphics cards can handle several million of triangles while maintaining smooth frame-rates.
The era when you really worried about polybudget is long gone, today you just use as many as you need for the thing you're making (within reason).
if you go over your designated polybudget it typically isn't a big deal unless we talk magnitudes and all the time.

Way it works polygons that are all part of the same model will go thru the graphics-pipe like greased lightning, the problems starts to happen when you have lot's and lots of
separate objects that all needs to be sorted. Like a single object of 8 million polys will be smooth and interactive while 8000 objects of 1000 polys each will crawl.

Only reason you'd really worry about polycount to a high degree is if you deliberately go for a retro aesthetic.
> Working with quads means you'll end up doubling it, right?
No. You use the same amount of triangles regardless, confusion comes from how polycounters sometime show 'polygons' instead of 'triangles'.
What matters is the tri-count, and when people generally talk about 'polygon count' they really mean 'triangle count', not the number of actual mathematical polygons.

A quad is just 2 triangles with one line hidden for your viewing pleasure.

The reason you work mainly with quads is that when you start to make higher resolution surfaces that looks more realistic it is useful to flow and space your edges in loops
of stripples that can curve, bend and flow as needed, as quads tile predictably in two directions it making them a useful base building block to work with.
Whenever you need a single triangle on the surface, you will just go ahead and use a single triangle instead of a quad.
>I think quads suck, my PROFESSIONAL© loli low poly models are all tris
>I'm a noob but I totally know better than all those 20-year veteran AAAAA+++ devs that's why I made this thread to tell ya'll.

A thread died for this.
While it may not seem so at first glance, rendering a triangle is simple and efficient, while rendering a polygon with more vertices than that is in fact rife with problems and ambiguities, so it turns out a bunch of triangles is more efficient than a single quad. This is because a triangle is known as a simplex, which essentially means it has a bunch of useful properties (for example all its vertices are always coplanar even if there's math errors and imprecisions, which is a huge thing in rendering).

So, when you see polygons, for rendering purposes they're always triangulated before being sent to GPUs (which only accept points, lines and triangles as primitives, and even lines and points are actually treated as triangles internally).
Why do I see abominations such as you breathing air when the greatest people in history are dead?
Just hear me out
wouldn't working with the lowest amount of tris possible be ideal even just as a base for a higher poly model?
did you not read the rest of the replies?
Goddamn, dude, no. That's completely wrong. Unless you're building in quads, going from low to high is going to be an absolute mess. Like said above, triangles don't subdivide well. Especially when triangles are next to each other. Even then, if your low-poly is built in quads, you're going to get a mostly shapeless blob unless you're skilled in sub-d modelling. You lose a lot of detail as you make things smoother and smoother, unless you really know how to set things up right for sub-d.

And here: want me to blow your mind?
Ideally, you want to start with the highest amount of tris possible.
Why? Look into sculpting. Like working with digital clay, it gives you unparalleled ability to add as much detail as you like. And it works in tris too, since triangles are ultimately the most malleable shape.
>wait, anon, aren't you contradicting yourself?
That's where you learn retopology. Sculpts are going to be in the thousands to millions of polys large. Retopology brings them down to a sane count as well as enforcing good quad flow.
>don't i lose all of my sculpt detail?
Baking, my guy. You bake the sculpt to your retopo, the high-poly to the low-poly. Baking essentially transfers all that sculpt detail into a texture.
Ya'll already wasted too many words on a thread that was never supposed to exist in the first place.
OP is dumb af, stop posting and let the thread die already.

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.