# How The F*#% Do I Model This? - Reply for help with specific shapes - (Post attempt before asking)

## Replies

• Offline / Send Message
polycounter lvl 8
8th Scepter yes you ll thank us later, I have made a full reference folder filled with cool tricks posted here from page 1 to 146 took me couple of hours to do so. Would suggest you do so same Also someone shared all the pics too from this thread somewhere here.
• Offline / Send Message
polycounter lvl 4
A suggestion ( i could be wrong) http://image.prntscr.com/image/159eac11504146c1852b905a35f056a1.png In ref image red dot surface and green line surface are not of same height. Give supporting loops like blue line rather than making triangle.

Edit: In most cases beginners like me see things wrong or incorrect because of less ref images study. Just few days ago i had problem with with gun belt, perna and one another guy helped me to put on path, it took me hours to model what i wanted. Practice,practice and practice makes a man perfect
• Online / Send Message
@8th_Scepter , the idea is to be able to break down an issue. To identify what it actually is on a conceptual level.

Issue #1: Curve/hard intersection. This issue comes up every 2-3 days.

Issue #2: Simply ignoring the basic principle that smoothly curved patches need edges of equal length. Yours are not as the segments next to the cutout get thinner towards the top, and all the quads are stretched.

• Offline / Send Message
vertex
Being both brand-new to sub-d's and thoroughly intimidated by the responses in the thread, I'll ask my regrettably basic question as a simple yes or no:

It's all quads, but it's leading to pinching in the reflection.  Is the only way to avoid this simply having a denser base mesh so the distance between vertexes is shorter?

(However, even increasing the base mesh resolution substantially doesn't fully address the problem:)

I keep seeing the principle that on curved surfaces, use existing base geometry as control loops, but this would appear to need an insanely high level of detail to accomplish that.  Is this nonetheless the solution?

• Online / Send Message
@mverta, you should include a link as not everyone is aware of what Peter Stammbach's cavities are about or how they relate to your question.
• Offline / Send Message
vertex
Thank you - in retrospect it was an entirely irrelevant detail so I've edited it out.  Simply looking at the quads I've modeled suffices. I will endeavor to be more efficient in the future.
• Offline / Send Message
null
Hey guys, any thoughts on how to improve this twisted cord edge topology? Because my current attempt looks terrible:

Ref:

Thanks.
• Offline / Send Message
interpolator
Algecir said:
Hey guys, any thoughts on how to improve this twisted cord edge topology? Because my current attempt looks terrible:

Thanks.
pretty straight forward. i have to go to work so here is a 3 min try.
https://www.dropbox.com/s/56y0wymjl3832sv/hereyougo.max?dl=0
and the file there, to see how i broke it down.

• Offline / Send Message
polycounter lvl 8
@mverta  hope this help you understand and solve your problem. Also removing this kind of detail wont help you learn mate! put it back on. Also note that in order to get mine L corner on curve super smooth I manually adjusted vertices on edge constraints.

• Offline / Send Message
vertex
Thank you - but I'm not sure how it directly applies.  My dilemma is that I have dozens of cylinders to make, each with notches cut out of the ends as seen above.  The notches are of all different sizes and depths, so if I was to attempt to use base geometry as control loops, I'd need to start with a cylinder with about 1000 sides just to make sure I was covered, which is retarded.  But the cylinders will ultimately be reflective, so it needs to have as little distortion in it as possible. So far increasing the base mesh's subdivisions doesn't really solve it.  I have also tried using a high-res target mesh to snap points to after the notches are cut out, but this too fails.
• Online / Send Message
@mverta , yours is a question which comes up and is answered up to several times a week here. See this thread for a discussion on the issue.
• Online / Send Message
greentooth
@8th_Scepter, part of why you're having a hard time is that you're modeling two separate pieces as one, so you're either suffering because you didn't study your references well enough or simply didn't find good ones. In future, try this: http://deicidenbf.thanez.net/
Credits to anything found there go to DeicideNBF. I just host the pix.
• Offline / Send Message
polycounter lvl 4
Algecir
Sliced cylinder http://prntscr.com/f5th54
Pull vertices and shell,http://prntscr.com/f5thmw
Turbosmooth http://prntscr.com/f5thze

Not sure if it looks same as this.

• Offline / Send Message
vertex
@perna  Thank you, and yes I've seen it asked and answered in various ways, none of which seem to make sense enough to me to directly apply.  As I said, the first suggestion (more base resolution) is impractical, and when it comes to things like you've illustrated in that link, they seem to point to early termination and n-gons/triangles/whatever to localize the problem.  I've tried that, too, and seem to come up with artifacts, so perhaps it's just that I'm not terminating things the right way.  I suppose I'll just have to keep playing.  I have to learn to do this and can't give up.
• Online / Send Message
@mverta, good call on realizing that simply increasing base res is not the way to go.

Subdivision modeling is all about approximation. That means you won't achieve the precision found in CAD, but it also means you have the highest levels of freedom and efficiency. In approximating a shape you choose the approach most suitable for each specific case. You're always at the mercy of surrounding geometry, which means there are a lot of these cases and it takes a while to build a good repertoire of solutions.

Below are two approaches, out of many, for your case.

• Online / Send Message
Here's a couple which use QuadChamfer. In other words no manual tweaking of control loops. The cylinder uses 24 segments.
stack:
• meshsmooth
• smooth
• shell

edit: The last version allows for a very soft edge without breaking the smoothing.

• Offline / Send Message
polycounter lvl 2
Hey everyone. Second time posting a problem here. I'm practicing some hard-surface shapes and I encounter this type of shape and I have no idea how to block out the shape. I tried different methods but they don't work perfectly. Thanks in advance.

• Online / Send Message
ngon master

take a box, angle it, subtract bool it from your shape, clean up, profit
• Online / Send Message
@admiralpixel15 , just boolean subtract a box from the base shape.

edit: Ninja'd by Neox. To expand on the advice, it seems you need to practice booleans and get used to thinking in terms of which operands make up a shape. Do a ton of quick hard-surface block-outs, or simply look at machined shapes and try to figure them out in your head, etc
• Offline / Send Message
polycounter lvl 2
@perna @Neox Hey there Sir, sorry for the late reply. Thanks for solving my problem with the shape. Here's my attempt of making it. I tried to increase the loops to compensate with the shading problem for the corners. Is that bad? and please critique my work in case there's flaws with it.

@perna Yes sir I do really need more practice with booleans as I rarely use them. I will definitely do some hard-surface block-outs as well. Thanks for the advice Sir PER.

• Offline / Send Message
polycounter lvl 2
@perna Hey there sir and sorry for late reply. I just want to say your methodologies/techniques are the most easiest and most efficient way of modeling shapes. Before I posted my problem I did try your method with the three edges but I was having difficulties getting  the hard edge in the middle and I was desperate to perfect it but fails every time. I tried to use booleans as you told me solves it, but not the most efficient but it's effective. I keep forgetting the fundamental part of subD and that increasing the resolution of the model is not the right path to go to. Anyways thanks for making an effort solving my problem and showing your way of approaching the shape.

Seeing your model makes me wanna kill myself. I wonder how did you get the hard edge in the middle part. Was it one of your modifiers? I used Maya instead of 3DS Max because it's the university's main tool for 3D stuffs. Sorry to bother you.
• Online / Send Message

OK, you use Maya. Well, the modifiers just do stuff automatically which you would do by hand in Maya. Most importantly Max has a quad-chamfer modifier ( either the one in recent versions of max called "chamfer", or this one by Marius Silaghi: https://mariussilaghi.com/products/quad-chamfer-modifier) which automatically creates control edges based on smoothing groups (equivalent of hard/soft edges in Maya). I believe there may be something similar for your software but I'm not sure.
• Offline / Send Message
polycounter lvl 2
@perna The edge with the boolean cut-out, the one circled in the image from the post above. I was really hoping that some features in Max is implemented in Maya. Like the Quad-chamfer for edges instead of Bevel and Shell for thickness instead of Extrude because it gives you more control in my opinion. I need practice in terms of edge loop efficiency like how Vitaly Bulgarov do to his 3D models and some block-outs. Blocking out the shape is my enemy especially if there is a complex/complicated cut/inset. Any tips/advice from that one?

Anyways here's a sample of Bulgarovs 3D stuffs from his Ultraborg kitbash. Currently studying edge loops efficiency to make the shape but also terminating tris to make it quads. Anyways thanks for your time. I won't ask anymore and abuse you.
• Offline / Send Message
I really liked the simplicity of Perna's approach so I decided to have a go at it without the Quad Chamfer modifier, and using again using Open Sub-Div. I'm trying to get used to using Open SubDiv so this was a learning experiment!

Without Quad chamfer both required a fair about of cleanup to control edge hardness. Using Max's chamfer modifier I was really limited to how soft I could get the edges in tight spots without the chamfers overlapping each other. Using open subdiv there wasn't enough mesh density around the 'slice' to give a nice bevel. I ended up adding loops later on to clean this up.

The other problem with Open Subdiv I found was editing the poly added new hard edges that I had to clean up, or it messed up my crease settings. I also had to really fine tune my creasing on an edge by edge basis that was quite tedious.

@perna by the way, do you have anything set up to stop max from smoothing the entire mesh by default when you chamfer an edge? having to always uncheck that box is annoying as hell. Especially when if I miss it I might not notice for a while.

anyway this shape is fun, I'm going to detail it out!
• Offline / Send Message
@perna you absolute saint. Yes setting all those up manually is constant annoyance. I can't wait to give this a run!

Yeah in reality I could get more consistent edges at the smoothness I wanted if I just built the model with that in mind. But I wasn't really paying attention to proportions when I threw it together and stuck in some too-thin bevels. Mind you, having consistent edges is something I always struggle with!!
There are all sorts of tricks and approaches to avoid overlapping chamfers, so you should strive to learn those instead of turning to Creases, which is an absolute nightmarish disaster for hard-surface and I strongly recommend against using it for anything but unimportant background objects.
What a ringing endorsement! /s. Ha! Yes, from detailing out the rest of the model yesterday I eventually started getting sick of having to constantly reset crease values on edges, remove creasing from flat edges that where adjacent to a creased edge I had chamfered. I wouldn't go so far as to call it a disaster, but the Open SubDiv tool in max are not solid enough IMO, and it definitely lacks the control you get from support geo, to the point that I ended up putting in loops in place regardless.
• Online / Send Message
fearian said:
What a ringing endorsement! /s. Ha! Yes, from detailing out the rest of the model yesterday I eventually started getting sick of having to constantly reset crease values on edges, remove creasing from flat edges that where adjacent to a creased edge I had chamfered. I wouldn't go so far as to call it a disaster, but the Open SubDiv tool in max are not solid enough IMO, and it definitely lacks the control you get from support geo, to the point that I ended up putting in loops in place regardless.
@fearian, Interface is not actually the problem at all. I don't have time to get into detail, but for anyone to say Creases are not a disaster, for the kind of hard surface challenges we deal with here, they simply don't understand the math behind creases nor Pixar's intended purpose for Creases. In our context they're a joke, a silliness and an insult to my family four generations back.
• Offline / Send Message
polycounter lvl 2
Hello everyone, I'm trying to model 2 twisted pipes but I'm struggling with the ends and the pinch I'm getting, if any of you kind sirs could help me out that'd be awesome
Here is what I'm looking for :

And here's my attempt :

As you can see I'm not getting the shape i want and I faces overlapping. Thank you everyone
• Online / Send Message
@sanfuer , this is confusing because it seems you already have 90% of the shape. What's stopping you from just doing more of exactly the same to complete it?

As for the overlapping faces, just fix it manually using basic modeling tools.
• Offline / Send Message
interpolator
perna said:
@admiralpixel15 , I zoned out and went overboard with this

I just wanted to address you increasing the density of the rounded corner geometry. For a clean, controllable curve all you need is three edges, as shown above.

This way it's extremely easy to tweak the angle of each curve.  Simply move one single edge up or down:

It's also extremely easy to change the profile of the entire curve, make it sharper, etc.

Below is the cage model for the complete mesh, without modifiers. I manually remove the edge created by the symmetry modifier for the rightmost edge (you can see in the above wireframe images that there's no centerline running through the rounded corner.

Modifier stack:
• TurboSmooth
• Symmetry (vertical)
• Symmetry (the special case one, this is hidden until I need to mirror some manual work)

Cage without Quad Chamfer and Turbosmooth:

This is probably the best and easiest expl. anybody ever got. It's so incredible powerful! Thank you once again Wise one!
• Offline / Send Message
polycounter lvl 8

Perna advised to check out this thread. I had a little pinch issue on a mesh in another thread. I  did a paint over on my image since im not home. So I think I need to use more edges to begin with so those extra edges can also be used as support loops. Also make sure to check the edge spacing. I originally created this flat, cut in the edges, cloned and bent the segments to get the shape.

• Online / Send Message
@ajr2764 , I mentioned for you to check this thread because your question gets answered here up to several times a week. The real issue isn't that you don't have enough geo, but that your topology is not of a consistent density. To make a smooth curve, the points on the curve should be equidistant. Try to never stray from that rule.

The first thing you should try is, well you have this weird geo making your results bad, so simply remove that geo.
• Offline / Send Message
polycounter lvl 8
Hi all, Ok so I may get a lot of flak for this but one must seek help from the polycount council and @perna .
Right I am using spline > outline > extrude > chamfer workflow and came across an issue of maintaining width on an angled surface due to outline parameter in edit spline. I know I can do old school poly modelling (grey plane) also max is not cad tool but would like to know if there is a way I can approach this in faster way while keeping equal width on an angled surface.

• Offline / Send Message
polycounter lvl 8
I fixed up the topology and looks like I'm getting a good result now.

• Offline / Send Message
node
Guys, how would you model large, continuous, smooth shapes like this kayak in 3ds max? I considered polygonal modeling with turbosmooth, but after adding loops it quickly becomes a mess to adjust size here and there, because it's one shape which cannot be broken in parts. I started modeling it with curves and then surface, but it's really clunky in 3ds max. Is there a better approach?

• Online / Send Message
@PeterBroof , it becomes a mess when trying to do everything at once. Build all meshes  (not just this one) in stages of large to small shapes. Otherwise you're going to have issues no matter what you do, as a dense mesh is a nightmare for controlling smooth curves, and a simple mesh isn't able to support all the detail.

Also, of course, block out.

• Offline / Send Message
vertex
Hi

Does anybody have a solution about that weirdness on the top of my sphere.

Thankx a lot
• Offline / Send Message
interpolator
Hikabuzy said:
Hi

Does anybody have a solution about that weirdness on the top of my sphere.

Thankx a lot
check back on this forum my frriend, there's alot about it!
• Offline / Send Message
interpolator
@Hikabuzy
I understand that this stuff is hard to learn. If it helps, I have a pinterest board on hard surface topology.
https://www.pinterest.com/Makkon06/hard-surface-topology/
If you get some time, I'd suggest going through it and recreating some of the shapes in there. Thank you for making the effort! It can take a lot of tries before things start to click.

Otherwise, I would suggest trying to have more uniform topology, in this case more subdivisions for the cylinder, making room for intersections to smooth better. Good luck.

• Offline / Send Message
polycounter lvl 8
Bump
uh little help fellas on my post above
• Offline / Send Message
vertex
Thanx a lot guys, next time im gonna put more effort on it and  check it before

ps: sorry for my bad english. Its not my mother tongue

• Online / Send Message
HAWK12HT said:
Bump
uh little help fellas on my post above
Honestly your question is a bit of a jumbled mess and I don't understand what half of it means. Could you rephrase it into something simple, like "how do I set the distance between two parallel splines?"

In general, spline tools in Max are very basic and out of the box it would be infinitely easier to do what you want with epoly (you can even convert that to splines afterwards if you want).

If I were to work more with splines I'd not hesitate to purchase Polyline Pro by @miauu.

• Offline / Send Message
polycounter lvl 5
perna said:

If I were to work more with splines I'd not hesitate to purchase Polyline Pro by @miauu.

You always seem to know the best plugins/scripts for max, which ones would you recommend that all max users should have?
• Online / Send Message
edit: no, sorry, I just can't this post editor. I'll try again. Guys...
• Offline / Send Message
polycounter lvl 8

• Offline / Send Message
polycounter lvl 5
perna said:
-snip-
These are all fantastic man, thanks a bunch!
• Offline / Send Message
null
Hi guyz,
Im bothering with this simple topology... and cant find out how to make this object with perfect topology... some help?
Im thinking at the end chamfer edges and apply turbosmooth

try from blender

made something but still think its not good...?