[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]
Board
Settings Home
/3/ - 3DCG



Thread archived.
You cannot reply anymore.



File: 2.jpg (141 KB, 600x436)
141 KB
141 KB JPG
Is there really no clean solution to cutting square shapes into a curved surface for subd like in this picture? I can't find any solution with clean topology
>>
>>612436
The structure seems like it's made of wooden beams, so it's not really something you need to subdivide. Any particular reason you need it to be subd? Making the shape is easy enough, and you can always smooth the corners off with chamfers.
>>
very trivial problem if you know anything about modelling. but since you don't even tell us the software that you're using, how are we supposed to help.
>>
>>612443
this is a topology question, but I understand that if you were to help it would help to be able to name specific tools I guess. I've used Blender for 3 years but I'm using Maya now.

The main problem is any solution produces weird pinching, i'll try and post some examples of how other people try and solve this and point out the problems.
>>
>>612442
this is just one example of the shape I happened to have at the time, I've encountered this problem time and time again.
Let's just imagine in this picture, the main egg shape of the ship is connected to the "wooden beams". It's one mesh. No separate parts. (In this case, separating them would be the easiest solution but I'm merely using this picture as an example of the problem. Thank you)
>>
>>612442
>>612443
This is pretty much the closest thing I've come to a solution. All quads, everything seems to be perfect, but I'll show you what happens when adding support loops/edges in the next picture.
>>
File: 7.jpg (16 KB, 602x349)
16 KB
16 KB JPG
Note the pinching.
>>
File: 8.png (14 KB, 685x575)
14 KB
14 KB PNG
Clearer view of topology.
Again, there might be a better way of doing it this just happens to be the cleanest way I've found, even though it's still not perfect. Thank you so much for any and all help.
>>
>>612526
i think it would be much better if you model this lowpoly first, you can slowly optimize it to the point where you are capable of using subd
>>
>>612535
That picture you replied to is from a Youtube video attempting to solve this problem. Theirs works because it's a bathtub with a more natural curve into the square shape. The other pictures I just whipped up in Blender. (Just used Blender because I'm faster in it, software not important)
I don't quite understand what you mean. How much more low poly can this get? And why would making it lower in polys help. I feel like this is the bare minimum for the shape I'm trying to achieve.
>>
>>612541
>And why would making it lower in polys help.
because when you start with sub surface, you are trying to fit the shape to the modifier and this is confusing. even expert modelers don't start making shapes with subd unless they are modeling pure organics like bottles or vases.
start with simple primitives, slowly place the loops and slide them until you get desirable shape for subdivision.
>>
>>612542
adding to this (cont):

if you want to make it easy use mirror,insert,bridge,slide etc. anything that can you help generate the shape faster, try to place the loops to match the skeleton and extrude it outwards
>>
>>612544
I understand and would ordinarily do this. I can get that square shape alright but the only problem is beveling/adding support loops when it comes time to subdivide. I apologise if i'm missing something you're saying. Thank you for your time.
>>
>>612436
Learn to model with NURBS surfaces you cunt
>>
>>612546
let's say this sphere with square shape is the final shape i want and I want to just crisp those edges without pinching or horrible topology
>>
>>612547
I can model with NURBS, at a basic level. I don't see how NURBS would help soften those edges however.
>>
>>612533
>>612531
>>
>>612546
if the shape is really killing you then finish it with booleans, then do retopology to fix it, then subdivide.
its basically alot of modeling, but an experienced modeler should do this without booleans
>>
>>612549
why not use edge crease or mark sharps? you don't have to add any loops. this works in subd
>>
>>612551
this would work perfectly but the problem arises when you try to do this on a curved surface since extra edge loops distort the curved section. I'll save this image tho, thank you.
>>
>>612552
That's why I'd like to learn to do this without booleans. I know exactly what you mean though.
>>
>>612554
I could definitely do this. I was just wondering if there was really no perfect way of doing this with the support loop/bevel method.
Creases and mark sharps can cause problems with shading and normals unless you subdivide a level or two extra, which kinda sucks if everything else looks fine without the extra couple of subdivisions.
>>
>>612559
double inset on the extrude also helps in pinching, but only to a certain extent.
just find whats best for you
>>
>>612560
I've found that helps a bit, thanks.
>>
I just tried mark sharp and creasing and still have pinching. I refuse to believe there is no way to achieve this, it's seemingly such a simple exercise.
>>
File: 9.png (44 KB, 769x636)
44 KB
44 KB PNG
>>612573
The end goal is something like this, but with subdivision. The insetting in this picture could be replaced with a more grid-like topology to extrude from, but the control edges would affect the shape of the curved surface. I just don't see how this is possible.
>>
I'm beginning to think this is impossible to achieve. I've done a lot of research and I've asked in a lot of places and everyone says how they think it should be easily done but I have never seen a solution. People just stop replying. I refuse to just let this go, we can do so much in 3D why is this seemingly simple shape impossible? I understand why there is pinching etc, I just don't understand why there isn't a way to avoid it. You'd think this would be a super common problem with topology reference images everywhere like >>612551
>>
>>612574
i think you should construct the ship from multiple parts. making everything from a single shape will only hurt you.
you can even retopo all these parts later into a single shape
>>
>>612436

I'm drunk off my ass right now because I hate myself and wish every waking moment that the sweet release of death would come for me, yet even in my inebriated state your challenge doesn't seem that much like a challenge to me. This took me only a couple of minutes and even the mistakes in it I realized during rendering but couldn't be arsed enough to go back and fix.
>>
>>612587
lmao thank you for the effort but you haven't sharpened those corners inside the extrusions which is the problem here.
>>
File: Fixed it.jpg (36 KB, 896x649)
36 KB
36 KB JPG
>>612588

There... fixed it.
>>
>>612590
Okay I already showed this earlier in the thread, could you subdivide that and show me the perfect corners with no pinching?
>>
>>612581
I could model different parts but the main point of this, is being able to smoothly subd this extrusion with no pinching or artifacts. And if I wanted a displacement map for panelling to run across the rounded part into the area I extruded, the "beams", there would be seams breaking up the pattern. Or say I wanted to model an extruded strip all the way around from beam part to rounded part, I would have to model that on 2 separate objects and try match them up where they meet.
>>
File: dont disrupt my drinking.jpg (79 KB, 1000x1250)
79 KB
79 KB JPG
>>612591

That wasn't me earlier in the thread but it looks okay when I subdivide it.
>>
>>612594
Interesting!

How is this different from what I did here? I don't see what I'm doing wrong.
>>612533
>>
>>612593
you can separate this part and work on it from a different layer.
this will not receive subd, you can just bevel it
>>
>>612598

I don't know, like I said I'm really drunk right now and not really thinking that clearly. Except about how I'm a fraud and a failure and the world would be better off without me.
It might be your subdivision settings.
>>
>>612436
>add edge loops
>select quads
>press delete

???
>>
>>612623
what quads are you deleting. the hard part of this is getting no pinching from square holes cut into curved surfaces when subdividing as seen in
>>612533
>>612531
>>
>>612603
if I didn't want it in one mesh that would be the best workaround, thank you
>>
>>612550
you can fillet nurbs surfaces
>>
>>612605
lol
>>
File: Curved T Connection.jpg (61 KB, 1000x1250)
61 KB
61 KB JPG
Okay took a crack at it again today sober and it was much easier to come up with a more elegant topology that looks better subdivided.
>>
File: curved t connection wires.jpg (221 KB, 1920x1080)
221 KB
221 KB JPG
And really once I sobered up and realized how to do it I was all "well DUH, of course that would work much better".

Remember /3/ drinking and 3D don't mix unless you're sculpting abominations with big dicks.
>>
>>612598
It looks like you are using Inset Faces. This makes the 45' angles. If you get the quad the correct size then Extrude then add loop cuts. I usually select all the Inner quads and tap E > right click ( in Blender). That creates very close loop cuts.
>>
>>612783
Amazing. Thank you so much. It's just beautiful. I'll have to check if this allows for super sharp edges but it looks great thanks again.
>>
>>613935

You're welcome, now if you'll excuse me I've got a bottle of gin and some self loathing to do. Big dicked abominations aren't going to sculpt themselves.
>>
>>613943
What do you prefer for drunksculpting? ZBrush or Blender?
>>
>>613989

3D Coat. It's easier to use when drunk.
>>
>>612436
chris foss my nibba
>>
i wish nintendo would just hire chris foss as a concept artist for metroid.




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.