Loading...
> >

Common modeling mistakes?

JonnyRayJonnyRay Posts: 1,744
in Hexagon Discussion

Since the holidays are over, I've been picking up some of my modeling projects again and noticed some things I needed to clean up in my mesh. It got me thinking about doing a tutorial or something about common mistakes that beginning modelers make and how to resolve them. So far I have a couple:

  1. N-gons (our old nemesis!)
  2. Similar to above, but poles / t-verts
  3. Non-manifold mesh

Anyone else have suggestions from personal experience or helping others?

Comments

  • AscaniaAscania Posts: 1,849

    Too few/too many polygons for what you are trying to achieve.

  • andya_b341b7c5f5andya_b341b7c5f5 Posts: 694
    • Trying to make it (whatever it is) all out of one, single, seamless mesh: split it up.
    • Adding subdivisions too early, and/or not modeling with later subdivision in mind
    • Doing no design work at all beforehand - often results in poor topology, nasty edge loop flow.
  • ShawnDriscollShawnDriscoll Posts: 375

    Reversed normals confuses a lot of new people to the hobby.

  • JonnyRayJonnyRay Posts: 1,744
    Ascania said:

    Too few/too many polygons for what you are trying to achieve.

    Nice. Goes a bit with Andya's theme of poor planning. I see "too many" a lot. People thinking they want a "lot of detail" in their model and don't realize they're adding unnecessary polygons. Seems to happen especially when they start with the primative and create a lot of subdivisions right off the bat.

    andya_b341b7c5f5 said:
    • Trying to make it (whatever it is) all out of one, single, seamless mesh: split it up.
    • Adding subdivisions too early, and/or not modeling with later subdivision in mind
    • Doing no design work at all beforehand - often results in poor topology, nasty edge loop flow.

    Gathered some input from other sources and I think the theme I'm going for is more about planning the project rather than getting into "how to model". I've made a few of those mistakes myself. :) Experience is a good teacher though!

    ShawnDriscoll said:

    Reversed normals confuses a lot of new people to the hobby.

    Good one! :) This also goes along with one of the points I want to bring up in regards to identifying when to UV map the model. I remember working on my baby crib model, the first time I tried to map it too late in the process and it was a lot harder than if I'd worked on the map for the individual parts then cloned them. But going to UV map without first checking your normals can be a major issue too!

  • DiomedeDiomede Posts: 15,169

    If model is to have morphs, not anticipating edge flow to match

  • MaxHancockMaxHancock Posts: 226

    A common mistake is trying to make every part with one mesh. You can get over a lot of topology hurdles by separating the model into more pieces. I know it sounds obvious, but many think of an object as one whole and wrestle with the topology when adding details to it. 

  • RaytownmikeRaytownmike Posts: 61

    The boolean trap. Some folks think they can use boolean functions to sculpt forms, over and over. The geometry can get messy very quickly usin booleans to many times on one mesh.

     

     

  • RoygeeRoygee Posts: 2,247

    1.  Too much detail too early

    2.  Not thinking ahead to how it will be textured, therefore how it will be UV mapped - planning again

  • jjmainorjjmainor Posts: 485

    Repurposing your models instead of starting from scratch.

    Make a library of every piece in addition to the finished projects because all those door handles, drawer pulls, windows, stairs, railings, etc. can be re-used in later projects, thus saving you time.

  • Wee Dangerous JohnWee Dangerous John Posts: 1,605
    edited May 2020

    Good idea jjmainor, I have a few "starter" objects which I adapt and use in projects. I also have a set of router profiles (2D line drawings) which I use for picture frames and dado rails etc, saves a lot of time.

    One thing I have made which comes in very useful is a scaled yard rule, great for making furniture (anything which needs to be real-life size).

    Edit - Just to point you to a "step by step" tutorial I did in conjunction with Patience55. This shows how to save DS figures with a hand pose (sitting poses are very useful too).

    https://www.daz3d.com/forums/discussion/44296/hexagon-sword-tutorial/p1

    Post edited by Wee Dangerous John on
  • gregsgraphics_cafd0962a0gregsgraphics_cafd0962a0 Posts: 87
    andya_b341b7c5f5 said:
    • Trying to make it (whatever it is) all out of one, single, seamless mesh: split it up.
    • Adding subdivisions too early, and/or not modeling with later subdivision in mind
    • Doing no design work at all beforehand - often results in poor topology, nasty edge loop flow.

    I agree with all, but with some caveats.  Making things out of one single mesh is good practice if the model you're making exists in the real world as one single piece.  If it doesn't, then you should model it with the same parts it has in life.  If your model is fantasy, then it's up to you how to break it up, if at all.  Modeling things as one piece is usually done for game engines or animations where physics or other simulation properties can break up the polygons that aren't welded.  3d printing is also another area where you'd want the model welded together usually.

    Model objects as they would be fabricated in real life is a good way to look at it.  Most furniture comes as separate pieces, so it should be modeled as such.  However, things that are permanently welded together (metals) are much easier and more sensible to model as one single piece.  Also, parts that are put together in such a way that one surface flows seamlessly into the other should probably be modeled as one piece in CG.  So there are exceptions, but basically model things as you might fabricate it in the real world.

Sign In or Register to comment.