[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: 1500465709547.png (408 KB, 666x666)
408 KB
408 KB PNG
Noobass question.

How the fuck do people get such a clean topology when doing hard surface modelling? Character stuff is easy since I have a lot of control and I can smooth and snap the topology back to the sculpt.
But if you start off with an object and start booleaning the shit out of it, that seems to ruin the topology, or when I attempt to fix mistakes the edges are no longer clean and it results in jagged lines inside a flat surface at times.

Now I assume this is partly software dependent as some programs could probably generate the clean topology as you work and allows you to fix mistakes without irreversible damage.

Now ok, here I'm going to say that I'm going to work with Blender. Can anyone show me how to do proper hardsurface modelling (links to tutorials would be nice but I've seen some bad ones where they just don't give a fuck about topology). Or I guess the only option could be to use another program like solidworks or something to generate the model first.

I'm in need of directions.
>>
Blender 101 Hard Surface Modeling by Darrin Lile. A lot of the stuff on that motor is covered by the tutorial, plus it's really fun and you get a result good enough to feel like you're on your way to become a pro.

If you can't buy you can get it on CGPeers.
>>
A little offtopic. Can anyone explain why would anybody need a clean topology on a hard surface model? If it's being deformed, fine, but otherwise I don't get it.
>>
>>611212
yeah if it's on a flat surface I don't really care but usually there are often curves where I would like a very clean and consistent topology. Like tubes for example, although those are pretty easy to make, the edges usually need to be evenly spaced out unless you're trying to make them have an odd shape on purpose.
In addition, for texturing it's nicer to work with clean topology so I can organize the UV map better.
>>
>>611213
I guess strictly speaking they don't necessarily have to have clean topology at first but when making changes it could be a source of frustration later.
>>
>>611214
oh and one more thing. Having good quad flow means I can make appropriate loop cuts and bevels or whatever else.
>>
>>611176
>How the fuck do people get such a clean topology when doing hard surface modelling?
you just git gud at hard surface, that's all
there is no special button or a program or anything else
you just
git
gud
>>
I wouldn’t spend so much time doing something you’ll probably never get paid for
>>
>>611221
Now I can see you are a faggot
>>
>>611176
Supporting edge loops and subdivision surfaces my man. I don't understand why anyone would want to use booleans, they are deceiving because although they give you the shape that you want they give you n-gons, which will make your subdivided mesh look like ass.
>>
>>611470
Unfortunately some people care about measurements and your deflating subdivision surfaces are unacceptable.
>>
>>611470
You wouldn't use subdiv with booleans, dummy.
>>
>>611487
>>611470
Of course i would. WTF are you talking about...too stupid to cut some support loops and keep them all quads after the boolean operation?
>>
>>611491
Ofc you can use it, but there are also different workflows in the existence, ya know. You wouldn't have any probs with bools if you used a different method like baking bevel shader e.g.
>>
>>611485
that's why engineering models and "pretty 3d models" are completely different things

tldr use nurbs
>>
>>611470
if you make highpoly model you would eventually do retopology as well.
so it doesn't matter what you use, at first
>>
>>611470

Ngons can be fixed dummy. I always use booleans and I don't have any issues when subdividing.
>>
>>611485
If you have issues with subdivision modeling deflating then you aren't modeling properly.
When I was new there were times I was convinced certain components couldn't be built without throwing immense abouts of polygons at the issue.
Years later however, I struggle to imagine any object that couldn't be sub-d modelled.
Learn how to isolate edges with kiteforms, don't just smash edgeloops all over your work.
At the end of the day that solved almost all my problems, learning that.
High quality hard surface modeling isn't hard, it's just tedious and takes experience and comfort in whatever program you're using.
I use Maya personally.
>>
File: eeee.png (37 KB, 595x386)
37 KB
37 KB PNG
>>611485
>>611487
>>611491
>>611496
>>611498
Sure, you can use booleans and fix topology later, but isn't it more fun without them?
>>
File: topology.png (78 KB, 736x834)
78 KB
78 KB PNG
>>611833
Also how else would you do model a hole? I'm curious
>>
>>611834
Not that poster but by hand. Multi cut more than likely being a Maya user.
I never use booleans because it takes just as long to clean as it does to build it by hand.
>>
>>611834
B-b-b-b-b-BEEEVEEEL SHAAAAADAAAHHH
>>
>>611834
In your contrived You forgot to mention that your "better topology" deflated by an unknown amont and the hole expanded by an unknown amount and is therefore useless for any kind of production purpose.
>>
>>611837
Can't replace real/existing topology, promotes laziness and bad practice
>>
>>611843
Bevel is an essential tool in certain situations.
>>
File: woah.png (95 KB, 366x623)
95 KB
95 KB PNG
>>611841
What do you mean by production purpose? Do you mean modelling for industrial visualisation? Because I assume that OP meant hard-surface modelling for artistic purposes. Otherwise my point is moot. If the purpose is arch viz or something why use subdivision surfaces at all? Why not use CAD? I have no experience in that side of cg so I don't have a clue.
>>
>>611491
I don't know but to me that seems like more work
>>
>>611836
Can you provide an example? I think people on here think I'm being condescending but I'm genuinely interested in how others approach this.
>>
File: heyy.png (24 KB, 599x561)
24 KB
24 KB PNG
>>611834
More efficient that the one in that poster
>>
>>611848
I mean modeling for industrial production. To eventually obtain some kind of mold out of it. I just wanted to point out that approximating surface subdivisions (the only ones available in Blender) do no preserve measurements, including volumes and are therefore useless for anything more than looking at the model.
>>
>>611833
>but isn't it more fun without them?
you can't model every shape in existence, this is where booleans come in
>>
>>611853
>To eventually obtain some kind of mold out of it
You use CAD for that, not artistic modeling suites.
>>
What I see most often when people use booleans is that the model and the object used to cut a shape in it have vastly different polycount and/or topology. If you're careful to keep everything roughly similar in polycount then booleans work great.

I always cringe when I see someone cutting a circular hole in a 200,000 poly mesh with a 16 poly cylinder.
>>
>>611850
When I was new I would import a primitive cylinder, enable transparency and use that as a stencil to multi cut the shape into whatever faces needed it, with an orthographic camera.
It's also good practice when sub-d modelling something like this to use the minimum amount of edges possible in your cylinder, obviously this makes incorporating it into the original geo much easier.

I dont use stencils anymore, but I use custom pivot orientation/snapping and scaling to keep my edges clean now when im multicutting.
Not sure if you can do that in other programs, I do all my retopo and hard surface in maya exclusively.
>>
>>611850

Also I would provide a visual example but it would look exactly like what this guy posted, the hole would anyway.
- >>611852
>>
>>611856
Model motorcycle tire tread manually and tell me that again.
>>
>>611930
Why would you need booleans for that?
>>
File: 8.png (372 KB, 1028x540)
372 KB
372 KB PNG
>>611856
>can't model every shape in existence

...can't you?
>>
>>611930
You wouldn't, the point was that if you can model (good) tread, without just half assing it without deformers, you'd realize you can model anything. Didn't think it needed to be explained.
>>
>>611947
Replied to my own post rip my career.
>>
>>611929
Except I'd have another border around the outside of the circle to connect the kites to.
>>
>>611947
>>611949
Gotcha. I agree.
>>
>>611956
Modo seems to make booleans less of a pain in the ass though, but I know nothing about it. Might be worth checking out if you like using them. Foundry makes the best compositing program in the world imo, and Tor Frik uses it a lot.
>>
>>611176
i mean you just can just say fuck it to topology and make the shape and then retopo if you are new to hard surface
>>
>>611824
>Learn how to isolate edges with kiteforms, don't just smash edgeloops all over your work.
Can you please elaborate more what these kiteforms are? I tried to google but found nothing
>>
ok OP here I'm back after a busy week. I forgot to ask this question and people have already commented stuff mentioning it:

So how common is it that people generate a hard surface model disregarding topology only to retopologize after? And is this a "recommended" approach or should I avoid doing this unless I have to?
>>
>>611978
the way he made the case insert is even more impressive than the engine itself imo
>>
>>611212
normals and lighting
>>
>>611978
But how do you retopo hard surface?
You can't just use quad draw or something like that, because the model will just look lumpy.
>>
>>611212
normals, lighting, uv mapping, texturing, looks like it was done by someone who knows their shit.

If I was given the assignment of UV mapping and texturing an asset and the topo was a fucking mess, I would march down to the modeler and beat the shit out of them with a wacom tablet
>>
>>611176
Don't use booleans.




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.