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

Replies

  • HAWK12HT
    Offline / Send Message
    HAWK12HT polycounter lvl 7
    @perna thanks for script link. Yup my bad I made it complex! anyways I resolved it.


  • gfelton
    Offline / Send Message
    gfelton polycounter lvl 4
    perna said:
    -snip-
    These are all fantastic man, thanks a bunch! :)
  • advanced
    Offline / Send Message
    advanced 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...?

  • s1dK
    Offline / Send Message
    s1dK interpolator
    @advanced looks good mate. I made one from cylinder with 20 sides still smooth pretty ok.

  • perna
    Offline / Send Message
    perna quad damage
    @advanced ,
    "Perfect subd" is an oxymoron, but anyway - below are standard ways in subd. The bottom row is QuadChamfered.



  • advanced
    Offline / Send Message
    advanced null
    ^^Thanks guyz is helping me. i think i got it
  • Filip5
    Offline / Send Message
    Filip5 polycounter lvl 4
    Hey guys, I am struggling wall section. Especially the arch part. Now I did that arch using boolean and extruding the edges afterwards. However, I ended up with some loose vertices which I am unable to weld. It worked fine until I created that arch. Upon dividing mesh in Zbrush all goes wrong.

     
  • perna
    Offline / Send Message
    perna quad damage
    Filip5, it's possible to weld the vertices, just have a closer look at the topology.

    That said, this is completely unnecessary use of booleans. There's a flat wall and a flat shape cutting into it. You could just pull edges from the arch mesh and get a better result in less time. 
  • Filip5
    Offline / Send Message
    Filip5 polycounter lvl 4
    @perna I tell you, even target weld didn't work. There is something odd about modeling arches and I have no idea how to do them the right way.
  • perna
    Offline / Send Message
    perna quad damage
    Filip: Upload the obj and I'll show you.
  • Filip5
    Offline / Send Message
    Filip5 polycounter lvl 4
    @perna Here you go 

    (for some reason obj can't be uploaded either)
  • perna
    Offline / Send Message
    perna quad damage
    @Filip5
    I think you'll need to re-evaluate everything you did in creating this, from the ground up, and potentially hire an exorcist, because the mesh is broken beyond the limits of our earthly domain.

    First of all, it would be infinitely faster to just redo the entire mesh than to try to "fix" it. I see a lot of people here has "fixing" as a natural part of their workflow, which should never be the case. If you ever need to "fix" your mesh it means something in your workflow is horribly broken.

    concrete issues:

    -It's imperative to clean up a boolean result BEFORE doing more work on a mesh. In this case you've not done that, and the result is that the problems have compounded and multiplied. Finish one stage before moving on to the next.
    -As mentioned previously, there's absolutely no reason to use booleans in this case.
    -Also there's absolutely no need to manually add those control loops. It's been automatic in 3ds Max for the last five years.
    -The points on the curve aren't equidistant. The topology is random, makes no sense.
    -lots of unnecessary cuts and geometry that can't serve to do much but slow you down and add to your workload.
    -And finally, target weld does work just fine. Just look at the underlying topology. But that isn't even the issue here. There should never have been such broken topology in the first place.
    -There's nothing odd about modeling arches. They're just polygons like anything else.

    In short, you're making your life unnecessarily difficult here and you'll be happy to find there are simple steps you can take to drastically improve your efficiency, as well as alleviate immense amounts of frustration.

  • perna
    Offline / Send Message
    perna quad damage
    @Filip5

    Here. A subd curve (this one is identical to yours) only needs 2 edges, which is also the ideal for smooth control (you only ever move the midpoint to change its shape). So in this case I make 3 polygons and the rest is standard modifiers. No need to make life difficult.



  • Filip5
    Offline / Send Message
    Filip5 polycounter lvl 4
    @perna Do you also offer possibility of confession ? Anyway, I ll try to remodel and do what you suggested, altough I have no idea what modifiers you used there.
  • perna
    Offline / Send Message
    perna quad damage
    Filip5 said:
    @perna Do you also offer possibility of confession ? Anyway, I ll try to remodel and do what you suggested, altough I have no idea what modifiers you used there.
    I would suggest you find out by yourself, because honestly no one has any business using 3ds over, say, Modo, if you can't tell which modifiers are used there. They're the most basic ones "everyone" uses.

    I believe Modo has better booleans as well as a more straight-forward poly-modeling workflow. Since you're bringing your mesh to zbrush anyway, you could use Mesh Fusion. Modo might be a fantastic choice for you. It's cheaper, too. 3ds is really only recommended for 3D artists who want to go deep into and take advantage of the complex systems that software offers, people who RTFM, etc. It's a terrible software choice for anyone who wants to hit it casually.

    Don't take the directness the wrong way. I know some struggle with it, but I'm trying to give the best advice I can possibly give, with the least amount of BS.
  • musashidan
    Offline / Send Message
    musashidan insane polycounter
    Browsing this thread now and again, one thing that sticks out is the unnecessary amount of geo used by the posters with sub-d problems.

    The sub-d cage or base mesh is best modeled as a very low (as little as is needed to derive the basic form) density mesh. The subdivision algorithm does the rest. Look at the last 2 examples given by @perna . This is all that's required. This is a fundamental guiding principle of sub-d. Only add geo IF it is needed. Use existing geo where possible and IF you can't get the forms with the geo you have, THEN add more.


  • Revel
    Offline / Send Message
    Revel greentooth
    And the misconception of 'all must quad to get a smooth sub-d mesh' also a problem that sometimes make people frustrated cus it's just impossible to get all quad sometimes and really not that important rules to must-follow, and often requires you to add unnecessary edges that harder to control.

    The general idea of quad is just the evenly spread vertex location on the mesh, which I believed already mentioned by @perna thousands times here, the importance of vertex/ edges distribution.
  • triton
    Offline / Send Message
    triton polycounter lvl 2
    Hi guys,
    What is the best way of connecting all the faces together? if I Bridge now I still have to connect all the other side of the edges. I want to make it grid squares.


  • NoRank
    Offline / Send Message
    NoRank triangle
    I think that the only way of doing that is by using a script like quad cap or something like that.
  • supaclueless
    Offline / Send Message
    supaclueless polycounter lvl 8
    If equadistance (or even approx., just makes it not 100% "square"), just loop each corner edges for each shell, marquee deselect edges on corner, bridge.
  • triton
    Offline / Send Message
    triton polycounter lvl 2
    Thank you guys, you gave me inspiration :)
    I did bridge two edge sides and then I did connect on the same amount of subdivision and then I merged the vertices. That's all.
     

  • Neox
    Offline / Send Message
    Neox ngon master
    now the real question would be, why does your mesh have this density at that stage? wouldnt it be simpler to model with the minimum amount of faces now and subdivide later?
  • triton
    Offline / Send Message
    triton polycounter lvl 2
    Neox said:
    now the real question would be, why does your mesh have this density at that stage? wouldnt it be simpler to model with the minimum amount of faces now and subdivide later?
    Thank you for your comment.
    Yes you are right it's really dense but it's only a prototype and not the final model, I needed to be very precise to model a shape so later I can mont a certain shapes on top, so basically it's only a base and will be removed later :)
  • OBlastradiusO
    Offline / Send Message
    OBlastradiusO polycounter lvl 11
    I'm trying to get use to using quad chamfer with my hard surface workflow. Has anyone had any success in using this method with just pro boolean objects without quads?
  • Neox
    Offline / Send Message
    Neox ngon master
    nope, the way quadchamfer works you will always need some sort of clean topology with enough support structure to guide the chamfers. i guess what you are looking for would be more of the face inset script
  • OBlastradiusO
    Offline / Send Message
    OBlastradiusO polycounter lvl 11
    Got it. Thanks dude
  • PetSto
    Offline / Send Message
    PetSto polycounter lvl 4
    Hi, I need help , how to create best subd topology on cube corner with nice beveled edges, as you can see on the image below. I try something, but I am not happy with result. I need some advice, step by step guide how to do right way. Thank you.

  • Neox
    Offline / Send Message
    Neox ngon master
    i'd say, follow your own overpaint better :)
  • perna
    Offline / Send Message
    perna quad damage
    @PetSto , in terms of just topology, below are some standard approaches. The shape construction for this mesh is non-trivial though, if you want it mathematically perfect. The sphere section of the corner fillet has to resolve smoothly into the edge fillet. Would be interesting to see how people would approach that.


  • PetSto
    Offline / Send Message
    PetSto polycounter lvl 4
    perna said:
    @PetSto , in terms of just topology, below are some standard approaches. The shape construction for this mesh is non-trivial though, if you want it mathematically perfect. The sphere section of the corner fillet has to resolve smoothly into the edge fillet. Would be interesting to see how people would approach that.

    Thank you Perna, I think your middle topology is the best, has nice transition in corner, no pinching etc. But challenge will be how to achieve this, because I don't know where to start.
  • wirrexx
    Offline / Send Message
    wirrexx greentooth
    Started with a cube. 
    copied it, turbosmoothed the second cube, spherify, scaled it up. 
    used Boolean to intersect. Got that edge, cleaned it up and extruded the "cube" from the corner. 
  • xddong001
  • natec
    Offline / Send Message
    natec polycounter lvl 12


    I gave it a shot too.

    Cube with spherify, then a cylinder placed for reference with symmetry at 45deg, adjust half of cube to match edges, then symmetry, then clean up.

    This way seems pretty adjustable by changing the amount of segments in the cube and sides of cylinders and/or radius and get something closer to what you guys have made.
  • perna
    Offline / Send Message
    perna quad damage
    @wirrexx and @natec, right, but now sub-divide those meshes and see what you get. It's not going to be pretty.
  • Nam.Nguyen
    Offline / Send Message
    Nam.Nguyen polycounter lvl 5
    another approach that give you more control of how smooth the edges are. Notice that in reference image, the vertical edge is more round than the others. When you cut the corner from a "cube-sphere" you have to cut it evenly for all edge, and after that you have verry little control over  edge smoothness


  • perna
    Offline / Send Message
    perna quad damage
    @Nam.Nguyen ,  those loft profiles are not clean hyperbola and give a melted, inaccurate result.
  • NoRank
    Offline / Send Message
    NoRank triangle
    I did some inset on a box and used some control loops. Then applyied a turbosmooth and deleted the parts that I wouldn't use. Idk, I've seen this quite long ago on some kind of model (I can't remember what model had this kind of corner), I think it works fine since I couldn't notice any pinching when subdividing.

    Yet the topology is different from perna's examples...

    I'm actually interested in how you got that shape.

  • musashidan
  • musashidan
    Offline / Send Message
    musashidan insane polycounter


    double vert chamfer>face inset>topo cleanup



    Optional step for the quad-fanatics



  • perna
    Offline / Send Message
    perna quad damage
    @musashidan , "topo cleanup" sounds trivial, but you're actually making very significant shape edits during that stage, not just cleaning topo. I'd be interested in seeing whether someone has a more accurate, efficient and ideally configurable way of approaching this shape.
  • wirrexx
    Offline / Send Message
    wirrexx greentooth
    perna said:
    @musashidan , "topo cleanup" sounds trivial, but you're actually making very significant shape edits during that stage, not just cleaning topo. I'd be interested in seeing whether someone has a more accurate, efficient and ideally configurable way of approaching this shape.
    does boolean have anything to do with this? :P 
  • perna
    Offline / Send Message
    perna quad damage
    @wirrexx , oh I don't really have an ideal solution to this problem. I mean, if we should allow for variable chamfer width along the edges and have that add up with the curvature of the corner, and also be able to have as many subdivisions as we want on the corner geometry and so on, it's a complex problem.

    In terms of earlier suggestions, I don't think anything with a huge manual cleanup stage or anything that gives wonky curves or bad aesthetics or bad shading should be in the running.

    The below basic boolean op requires zero manual cleanup or in fact any manual work whatsoever and the shading quality is good enough for most purposes. The curvature is clean. To change the width of the outer edge fillets you'd simply move the intersection points closer to the corner, along with a scale op if you want to actually keep the boundaries in the same places and so on (well, if it's all going to be uniform, just change the scale of the sphere in the boolean op).

    However you'll need another type of topology if you want yet another chamfer surrounding the triangle of the corner, which will be necessary if you need an exact style of chamfer on the edges.

    Yeah this got rambly, just writing this in a hurry, Saturday and all


  • wirrexx
    Offline / Send Message
    wirrexx greentooth
    STEPS FOR THE NEW ONE. 
    1. box
    2. inset the sides i want to bevel, top, left and right
    3. use turbosmooth and smoothjing groups (top left and right have the same smoothing groups
    4. delete the edges inbetween those sides. 
    5. Bridge and turbosmooth. 

    And i bet you Perna will do it in 5 steps less. 




  • wirrexx
    Offline / Send Message
    wirrexx greentooth
    perna said:
    @wirrexx , oh I don't really have an ideal solution to this problem. I mean, if we should allow for variable chamfer width along the edges and have that add up with the curvature of the corner, and also be able to have as many subdivisions as we want on the corner geometry and so on, it's a complex problem.

    In terms of earlier suggestions, I don't think anything with a huge manual cleanup stage or anything that gives wonky curves or bad aesthetics or bad shading should be in the running.

    The below basic boolean op requires zero manual cleanup or in fact any manual work whatsoever and the shading quality is good enough for most purposes. The curvature is clean. To change the width of the outer edge fillets you'd simply move the intersection points closer to the corner, along with a scale op if you want to actually keep the boundaries in the same places and so on (well, if it's all going to be uniform, just change the scale of the sphere in the boolean op).

    However you'll need another type of topology if you want yet another chamfer surrounding the triangle of the corner, which will be necessary if you need an exact style of chamfer on the edges.

    Yeah this got rambly, just writing this in a hurry, Saturday and all


    that was the step i took on the first one! =) so it made me happy  seeing this! 
  • musashidan
    Offline / Send Message
    musashidan insane polycounter
    perna said:
    @musashidan , "topo cleanup" sounds trivial, but you're actually making very significant shape edits during that stage, not just cleaning topo.
    @perna It actually is pretty trivial, with no significant shape changes: Quickslice/delete to clean the inset topo. Collapse 3x vert pairs. Move corner edges with Normal constraint to flatten. 3x edge chamfer. That's it.

  • perna
    Offline / Send Message
    perna quad damage
    @wirrexx  :) Note that you misaligned the operands though. The sphere needs to be offset from the corner equally on all axes, and then in your version the radius reduced slightly to avoid having to manually clean up the boolean op.

    @musashidan yeah, I reconstructed your shape as soon as I saw it. Those are significant edits in my mind. Mechanical modeling should not involve manual edits unless where absolutely necessary. Not when we're dealing with mathematically clean shapes, like we have the three axes of curve/box intersections and the curve-curve transitions in this case. 
  • musashidan
    Offline / Send Message
    musashidan insane polycounter
    @perna yeah, I tried your boolean method too, as soon as I saw it. Works out pretty good. Still a bit of manual work though when intended for sub-d. I'm not sure if I'm into the mathematical perfection route, though, when I'm modeling. While I definitely respect the fact that this whole endeavour is entirely mathematical, I'm often an eyeball modeler. :)
  • perna
    Offline / Send Message
    perna quad damage
    @musashidan , why would you choose eyeball modeling when the alternative is easier, more efficient and looks better?

    Also, no, my method shown above contains exactly zero manual edits. Look at the rightmost shape in the image. The curve is perfect. You just can't create a perfect curve by hand, and why would you want to when you can use a mathematically based operand?

    Aside from the earlier mentioned advantages of using such operands, there's also the fact that they allow you to go parametric. If you wish to change base parameters of an "eyeballed" version, you have to redo the entire thing by hand.


  • musashidan
    Offline / Send Message
    musashidan insane polycounter
    @perna so there wasn't a single stray vert from the boolean operation? Unusual in this type of intersection. Scaling the spheroid to allow a larger/smaller chamfer is definitely going to result in strays or verts that need to be collapsed/welded. Also, if changes down the road meant maybe one of the chamfers needed to be wider then the perfect curve would need to be manually tweaked.

    I'm not arguing that doing it by hand or 'eyeballing' is better in any way. I'm just surprised that you had zero stray verts. Personally, when I'm modeling, a lot of the time I'm plowing through the work, trying to get piece after piece done within a certain time. I'm not saying I'm sloppy as I'm rather ocd but by 'eyeballing' I mean trying not to obsess or spend too much time noodling or tweaking things that won't really be noticed for the asset purposes. I've learned over the years that I've wasted far too much time on small things that don't matter all that much in the end. I'm not using what we're talking about here as an example as I'm finding it interesting, I'm speaking generally.
Sign In or Register to comment.