I don’t know how to handle the red area, and I can’t completely use quadrilaterals for topology, which is a bit difficult for me
If you know the solution, please help me. Because it’s really important to me.
Thank you.
I have placed the file below, you can change it in your favorite way. SCOPE.blend (2.5 MB)
Even without triangles those topology would not being suitable for subd anyway, since its not a subd model + triangulation. It looks like a bit wierd lowpoly model from CAD without fillets.
Before anything, I would like to ask: what is this model going to be used for? Maybe it doesn’t need subdivision? I would subdivide this only if it was for super high-quality closeup renders.
If you do want to retopologize the part in the image, here is how I would do it.
First, set the object to shade auto-smooth so you can easily see the sharp edges. The shading is a bit messed up right now.
For the retopology itself, I would do it by surrounding the sharp features with two edge loops, one on each side, like this.
This is all quads and will automatically give you support edges for later when you subdivide the model. The tightness of those starting edge loops will decide how sharp the edges will be when subdividing the final mesh.
Once you have surrounded every important feature with 2 edge loops, you then fill the spaces in between them with a grid of quads. Those spaces in between don’t need to be especially precise in their topology, as they are mostly flat. As long as the important edge loops from the previous step are well made, it should look good.
An other thing, you have lots of repeating features on this object, so do duplicate identical parts you already retopologized.
Personally, I wouldn’t use subdivision for something like this. I would just model it at geometry density that I think is suitable for my purpose to begin with. Subdivision would not be suitable for the rail for example. It would be much easier and more efficient to just model it with bevels. There is no point in using Subdivision modifier here in my opinion. You could probably use subdivision for the scope more efficiently, but it would create unneeded topology. You don’t need to use subdivision absolutely everywhere.
No. It’s just a a tool that you can use. It smooths things and adds subdivisions to faces in a particular way - that’s it. Use it or don’t. You can use it for anything including hard surface modelling, but it’s more popular and useful when dealing with smooth organic forms. It’s not required for anything. It’s very often misused when simple bevels are needed. Not in a way that there are any rules against it, but it’s often just easier to use bevel functionality for that and you don’t create extra geometry that you don’t actually need. There are no rules. You can still do it if you want… Just think about what’s easier for you and have a reason for doing stuff. Like right now - why are you retopologising it at all? What’s the reason? You want it to be better? Better for what purpose? You should know why you are doing something. Otherwise it’s really pointless.
Maybe you’re right。
To be honest, this is also something that I remembered on a whim, I don’t plan to do anything with this model, I just want to make it look cleaner and neater
You can form an edge around the corner to maintain its shape even with the use of Subdivision modifiers.
It’s a simple method, but it also requires a lot of thought.
The way to hold the corners is simple, but it should not use too many edges, and should be formed as evenly as possible.
For this reason, it is a simple but demanding task.
Please refer to the channel in the video link below.
From what I can see from the screenshots, the mesh looks ok (for an hard surface).
To “clean up”, you can just select the edges that divide each quad into triangles and dissolve them (X + Dissolve Edges).
If, while in the process, you find some poly that can’t really be turned into a quad, just leave it as a triangle. In this case, a triangle can be considered a good topology.
(Note that this won’t remove the triangulation of the surface… it will just redefine face-loops, from 3 loops to 4)
For the above i used tris to quads… and also a bevel and a weighted normals modifier… (only defaults). The quadrification does almost change nothing at the look.
In an interesting way, you can also try using Bevel modifiers.
Rather than forming a topology, this method creates edges to release subdivision directions or clumps.
You can try it when you need a quick and easy method for a form similar to that modeling.
It’s one way to do it… I don’t know, maybe it’s just me, but what I see there is just loads and loads of unneeded topology.
It’s all nice and good, but once you have a large scene and plenty of assets and add something made like this… It’s just a waste of resources. …in my opinion.
All those loops are doing nothing to hold the forms:
Sure some are needed as support loops for subdivision, but still. And the worst thing: all of them will turn into a whole lot more loops after subdivision that will do a whole lot more of nothing to hold the forms.
One is 2 556 tris, the other 26 624 tris and is soft and round… So… 10 times more geometry for a worse result for something that is supposed to be precisely machined… I don’t know… Personally, I don’t like assets like that in my scenes.
How you advice to remove those “unnecessary” loops if they help to build a circular hole?
And putting another set of extra support loops are pretty common practive for subd modeling, since you cant just make a “corean bevel” style support loops and call it a day. It will affect the shading and it needed to be properly managed.
Yeah, and subd modelling is wasteful. Not in all cases. For most organic smooth forms all the created geometry is useful. For hard surface, it’s often hugely wasteful. I think it’s overused.
