Excuse the scale. It's made according to the original specifications.
absolute boss. If you dont mind me asking but I assume you twisted the entire cylinder? I skimmed through a page prior and some people had messy topology. Wouldnt twist do the job?
Can we please stop posting scripts that automate the proccess. AFAIK this thread is about topology and how to handle subdivision effeciently. If people keep posting script solutions, noone really benefits because one doesnt know (especially newbies) how topology affects the surface etc. These tools are awesome and time saving but lets keep this to the roots
Meh. Even if scripts/shortcuts are used, what is important for anyone learning is to see the final wireframe of the thing, if they don't have scripts they sure know how to reproduce things manually. + 1 with what @Joao Sapiro said, the whole thread already tells about the fundamental of SUBD execution ..
absolute boss. If you dont mind me asking but I assume you twisted the
entire cylinder? I skimmed through a page prior and some people had
messy topology. Wouldnt twist do the job?
"messy" meehh. OK @Bumblebee here is how he did it though, @perna correct me if I'm wrong :
Can we please stop posting scripts that automate the proccess. AFAIK this thread is about topology and how to handle subdivision effeciently. If people keep posting script solutions, noone really benefits because one doesnt know (especially newbies) how topology affects the surface etc. These tools are awesome and time saving but lets keep this to the roots
im not sure i follow your logic. if you are worried with newbies i think that the info in this whole thread is more than enough in terms of topology/subd modeling. using scripts that make things faster/non destructive should be encouraged unless they over complicate stuff...
think about it this way : if we share what scripts we use etc maybe a technicall artist in a studio sees something intrresting and tries to replicate on their tools, wich in turn makes artists life easy.
I guess I wasn't very clear. I am not criticizing these tools in any way. I run a youtube channel and sometimes I get emails from beginners. They ve watched people use quad chamfer (chamfer zone) over and over again and when they want to tackle different shapes, they get messy results etc because you simply cannot run things like quad chamfer on dense or overly complicated meshes. Tools have a limit and this is where the knowledge of subdivision should come in. However, theres rarely anyone out there these days that explains how subdivision actually works and how to control it. You see topology so dense, that you sometimes wonder why they use subdivision in first place.
@perna Are you sure you could have dove it cleanier? @Yamanote Nothing is impossible to poly model (unless it's something that can't exist in a 3d space), you just must find the correct approach.
@Dreyzie I would Have do that with Zbrush though (for the welds) My workflow is always : Blockout - Make it SUBD Worthy Here is a super Quick attempt, not exactly the shape but :
as you can see I'm not afraid to make Aggressive Bools as long as I have my blockout for SUBD
Tips : There are welds at intersections = you can take that to your advantage = don't need hyper clean geo around intersections Tips2 : Proboolean ? Edit : under the circle cut there is some pinching issue , I've just overlooked this area but you get it
Quick ugly edit : You need more support loops , then you can connect the corners and points... But It's also for the Smooth Mesh Quality , because even tho you made it with bevels, the surface will bent with the creased edges still hard in the smooth mesh preview... @perna made an image on that can't remember where, but if you want to resolve your problems quickly make a little research on pinterest.com
@WaYWO Interesting. It looks like your setup would allow for iteration control (less/more edges). I'll have to try that method out. This was the older attempt (here). Thanks for the breakdown!
In some more modelling trouble.. How would I re-top these tank treads for UE4? Im currently hand-modelling each piece but it's gonna take me days. Any suggestions?
@perna if you dont or anyone else minds, are there tips on how to clean up geo such as for details like this?
The problem I have is always when cutting out the shape, I have triangles and other ngons that ruin my shading. When I try to channel the loops to avoid the tris, the object gets pinching because the segments are not evenly distributed anymore. Here are some attempts:
as you can see i sometimes end up with tris and during clean up, theres a lot of moving things manually involved. Im just wondering if this is the "correct way" to do. But mostly its the tris and other ngons i have problems getting rid of.
I think this is another variant of a problem that often gets posted on here. I'm getting better at solving it, but I'm not sure I got it entirely, so I would appreciate some feedback.
I'm trying to merge some shelves into a curved surface. Here is the low poly (with sloppy topology and smoothing groups) to give you the idea.
I gave a quick shot at doing an all quads version. I wasn't really going to use a subdivide on this, so I don't think all quads is necessarily super important. I'm just going to bevel the edges for the high poly later. The problem with this is obviously that it breaks the curvature of the outer surface.
Since I'm not worried about subdiv, I just did away with the extra edges that wrap around the outer curve of the shelf. This obviously leaves a bunch of ngons on the two sides, but is this actually a problem? Is there a better way to go about this? Most importantly, is this preserving curvature on the inside curved walls of those shelves?
Thanks Ausonian, that really helps. I think I will put one more edge loop in there so that those edges don't pinch out as close to the corner, but it works either way really.
The other thing I could use some advice for is ramp-shapes either popping out of or inset into a surface. Here are examples (and my attempts) at each.
It's harder to see on this one since it's so narrow, but it's the same topo as the triangle above, just with that diamond far more pinched.
@Ausonian Thanks for the suggestion. I may have to try that. The only problem I'm having with it is that I can't bevel those edges hardly at all without the geometry flying out of control at the end of the ramp. It looks great with a subdiv, but I don't want to use a subdiv to get the high poly because it will result in that wavy normals thing on the curved parts. EDIT: I haven't yet figured out exactly why, but that suggested change actually fixed the only remaining problem I was having, Aus! Bevels work much better. so thanks!!
@Yamanote I think it depends. I tend to start working with flat faces, but sometimes I have to go back and really redo the geometry when I need to start worrying about shading later on. Something that I've found really helpful is to work with an edge split modifier on. You can tweak the angle cut off to suit the project, and it gives you a good approximation of what the end shading will look like without the need for supporting edges.
Replies
the only thing i did was to utilize a cylinder primitive to make sure its round
"messy" meehh. OK @Bumblebee here is how he did it though, @perna correct me if I'm wrong :
Are you sure you could have dove it cleanier?
@Yamanote
Nothing is impossible to poly model (unless it's something that can't exist in a 3d space), you just must find the correct approach.
[noparse]https://youtu.be/k_S1INdEmdI
I've been trying to model this propane tank, but the handle has been giving me some trouble. I can't figure out a good way to approach it.
This is what I have so far.
This is the handle I'm trying to create.
Any help would be appreciated. Thanks.
My workflow is always : Blockout - Make it SUBD Worthy
Here is a super Quick attempt, not exactly the shape but :
as you can see I'm not afraid to make Aggressive Bools as long as I have my blockout for SUBD
Tips : There are welds at intersections = you can take that to your advantage = don't need hyper clean geo around intersections
Tips2 : Proboolean ?
Edit : under the circle cut there is some pinching issue , I've just overlooked this area but you get it
If not post a picture with entire model so we can see what it looks like.
Quick ugly edit : You need more support loops , then you can connect the corners and points... But It's also for the Smooth Mesh Quality , because even tho you made it with bevels, the surface will bent with the creased edges still hard in the smooth mesh preview... @perna made an image on that can't remember where, but if you want to resolve your problems quickly make a little research on pinterest.com
@JoeWilson maybe we should rename this threas to :
How do i model this ? ( post attempt before asking )
by having that on the thread topic there wont be an excuse not to do it as i feel the current name can be offputting
I've attempted to model this lights a few times, but never been satisfied of the results...
How would you do it ?
Ah, that clears things up for me. I'm still trying to get the hang of a good workflow.
Thanks.
This is the same geo, shown 3 times. Is there a better way to do this so it doesn't seem to pull from the corner that much?
Im trying to model this detail for a VR scene for a client but I can't for the life of me work out a simple way to model it.
I've tried using Boolean on a Helix spline but the cleanup required will take hours and it also produces poor results.
Any Suggestions?
Thank you advance!
Hey man, the indents are a spiral so I need it to be one continue edge flow
Amazing script - Thank you!
Current attempt is colored red
The problem I have is always when cutting out the shape, I have triangles and other ngons that ruin my shading. When I try to channel the loops to avoid the tris, the object gets pinching because the segments are not evenly distributed anymore. Here are some attempts:
as you can see i sometimes end up with tris and during clean up, theres a lot of moving things manually involved. Im just wondering if this is the "correct way" to do. But mostly its the tris and other ngons i have problems getting rid of.
Heres my further cleaned up try:
For example here, i want to have an rectangle after aplying a subdivider.
Either spread out the edge loops, or add control (edge) loops horizontally .
This is a very common question, and there are 2 other images that have the same solution on this page btw.
Try this
I'm trying to merge some shelves into a curved surface. Here is the low poly (with sloppy topology and smoothing groups) to give you the idea.
I gave a quick shot at doing an all quads version. I wasn't really going to use a subdivide on this, so I don't think all quads is necessarily super important. I'm just going to bevel the edges for the high poly later. The problem with this is obviously that it breaks the curvature of the outer surface.
Since I'm not worried about subdiv, I just did away with the extra edges that wrap around the outer curve of the shelf. This obviously leaves a bunch of ngons on the two sides, but is this actually a problem? Is there a better way to go about this? Most importantly, is this preserving curvature on the inside curved walls of those shelves?
Erase the red edges and make the green ones
The other thing I could use some advice for is ramp-shapes either popping out of or inset into a surface. Here are examples (and my attempts) at each.
It's harder to see on this one since it's so narrow, but it's the same topo as the triangle above, just with that diamond far more pinched.
You could move the red vertices like this, but yours is already fine, I think.
EDIT: I haven't yet figured out exactly why, but that suggested change actually fixed the only remaining problem I was having, Aus! Bevels work much better. so thanks!!
@Yamanote I think it depends. I tend to start working with flat faces, but sometimes I have to go back and really redo the geometry when I need to start worrying about shading later on. Something that I've found really helpful is to work with an edge split modifier on. You can tweak the angle cut off to suit the project, and it gives you a good approximation of what the end shading will look like without the need for supporting edges.