Jump to content

Archived

This topic is now archived and is closed to further replies.

DanielCFilho

Subdivision Surface Sphericals

Recommended Posts

  • Topic Author
  • Hello, all.

    I've searched the Cafe and the internet in general about this, but coudn't find a perfect solution. I'd like to know any method you guys use in order to model spherical objects when apllying them to SDS. I know that a standard sphere is not good for subdivision because of the poles of triangles that produce pinching and artifacts after SDS. I'm currently using the hexahedron to model any kind of spherical objects for SDS, cos' it's all made of quads and that way doesn't produce artifacts. But the problem is that an hexahedron, after subdivided, doesn't end up as a perfect sphere. It results in very subtle bulges that deform the shape of an exact sphere. So, if you guys have any suggestions, tips, on how to get a perfect sphere using subdivision It would be greatly appreciated. Thanks.

    Share this post


    Link to post
    Share on other sites

    The only way to achieve perfectly even subdivision on a sphere is to use all triangles. It cannot be done with quads just because of the nature of the shape.

    You are right to use hexa rather than standard sphere, but that is the closest you're gonna get if you want to maintain quads. But you can get that closer to a perfect sphere using the spherify deformer. 

     

    CBR

    Share this post


    Link to post
    Share on other sites
  • Topic Author
  • Thanks, Cerbera.

    Share this post


    Link to post
    Share on other sites
    On 8/24/2017 at 0:55 PM, Cerbera said:

    The only way to achieve perfectly even subdivision on a sphere is to use all triangles.

    Is this one of those instances where you have to accept the evil to get the desired results? The ends justify the means? 

     

    Share this post


    Link to post
    Share on other sites
    19 minutes ago, Monstrphil said:

    Is this one of those instances where you have to accept the evil to get the desired results? The ends justify the means? 

     

    In this case the ends would justify the means if you need your tris for softbody squishy deformation for example. If not, then no, because hexasphere (all quads) with spherify deformer  = perfect sphere, so there'd still be no need for any triangles ;) With all that said, SDS does a better result with all tris than it does with tris and quads, so that's in there to complicate things further :)

    Share this post


    Link to post
    Share on other sites

    DEATH TO TRIANGLES

    Share this post


    Link to post
    Share on other sites
  • Topic Author
  • The problem with the spherify deformer is that it can be idealy used only for a generic sphere. That will be almost never the case when modelling real objects. The general case will be objects WITH spherical parts. For example, the generic light bulb will end sphericaly, but the whole shape is something else. I keep finding online examples of techniques to achieve a spherical part under SDS, but never a good tutorial. For example, I've noticed that one thing that people do is to substitute the triangulated pole in the end of a standard spherical with a kind of quadrangular filling. But when I see that and try to emulate by myself, I never get the right result. I'll attach an example I found of a light bulb for instance. (I found those images on turbosquid. I don't know the protocol, but if I can't post it here, just let me know) - The thing being: if anyone knows any tutorials or anything about sphericals under SDS, I would appreciate it.

    Light bulb SDS 1.jpg

    Light bulb SDS 2.jpg

    Share this post


    Link to post
    Share on other sites

    I too would be interested in knowing how your light bulb was modeled. I've been trying to duplicate the top-but have not found an easy way to get your results. There must be a simple solution.

    Share this post


    Link to post
    Share on other sites
    On 28/08/2017 at 9:38 PM, DanielCFilho said:

     

    Light bulb SDS 1.jpg

    Light bulb SDS 2.jpg

    This is just a standard sphere but terminating in a quad patch rather than a pole. 

     

    If you have the HB modelling bundle with its quad-caps option it's literally 3 or 4 clicks to do this, but can also be done without it easily enough using a small plane as your centerpiece, the points-to-circle script, some quick brush smoothing and then using soft selection in dome mode to gently pull out the curve. You can optionally apply a spherify deformer to the sphere before you model the lower half if you want the top absolutely spherical.

     

    CBR

    Share this post


    Link to post
    Share on other sites
  • Topic Author
  • Thanks for the recipe, Cerbera. I'll try that. As I don't have the point-to-circle script, I've been using the FFD deformer to transform that small plane you say into a circle. I think I can get close. But I keep forgetting about the soft selection, I almost never use it for anything. Anyway, I've been searching the web for more clues about this, and what I've understood is that whatever method people use, the end result seems to be always just an aproximation, never the perfect spherical. Maybe one lucky day I'll find the perfect method by chance. Until then, thanks for the tips.

    Share this post


    Link to post
    Share on other sites
    1 hour ago, DanielCFilho said:

    Thanks for the recipe, Cerbera. I'll try that.

    You're welcome. If you run into problems with yours, let me know, and upload the scene file so i can check what you did...

     

    1 hour ago, DanielCFilho said:

    what I've understood is that whatever method people use, the end result seems to be always just an aproximation, never the perfect spherical.

    No, that's not true. Provided you use the spherify on the subdivided result of your model before you delete the lower half and model the bottom of the glass you  get the top as perfectly spherical as spherical goes ! That isn't necessarily the case if you apply it directly to the base (unsubdivided) model, which may not have enough segmentation to conform to the curve the spherify is demanding, whereas the SDS result mesh almost certainly will.

     

    CBR

    Share this post


    Link to post
    Share on other sites

    For the script point to circle you can use maybe this one ,

     

    Share this post


    Link to post
    Share on other sites

    • Recently Browsing   0 members

      No registered users viewing this page.

    ×
    ×
    • Create New...