Home Technical Talk

Head Topology question?

polycounter lvl 10
Offline / Send Message
melviso polycounter lvl 10
So looking at the polycount wiki u have this:

and then this:

There are some differences here especially around the mouth and eyebrow knot areas.
Which one is better or is the first one okay to replicate?I also wish there were more views of the first one like a side (to show the ears), and back as well but they are not available in the wiki.

Replies

  • RN
    Offline / Send Message
    RN sublime tool
    Hi. They do have slight differences, but they are both fine.
    You can safely pick either.

    The bottom one follows the Pixar face topology design, so personally I would go with that one for the extra reassurance.
  • musashidan
    Offline / Send Message
    musashidan high dynamic range
    Here's the Pixar topo laid out flat - note the locations of the poles. They are key to how the topology deforms. As for the ears and back of the head, these areas don't directly deform so you don't have to worry about the topology as much.

  • Alex_J
    Offline / Send Message
    Alex_J grand marshal polycounter
    does that pixar one translate well for more realistic characters? anybody do anything like that? I'd think so, but the eyes are so big maybe that effeccts where the poles should go.

    great GIF there btw
  • RN
    Offline / Send Message
    RN sublime tool
    There isn't a 'one face topology to rule them all'.
    You can start from a base topology, but the moment you change the size of facial features to suit different character designs (the size of the eyes, the size of the nose, mouth, forehead etc.), some of the edge loops will need to be reconnected to others, some will need to be removed, otherwise the geometry density and flow will be messed up (some loops going on crazy curvy paths, instead of clean, short paths).

    For example, a huge dwarf-nose that crosses the space of the eyes, that reference topology would need to be modified to fit this kind of design.
    httpsvignettewikianocookienetwowwikiimages776Alliance_posterjpgrevisionlatestcb20110714034742

    PS: That GIF is from Brian Tindall's (http://hippydrome.com/), it also has references for full body topology and other useful stuff.
  • melviso
    Offline / Send Message
    melviso polycounter lvl 10
    Tbh, I felt the first one works more for photorealistic heads to me but wasn't sure since the pixar one has loops at the mouth edges where the smile lines form. But the first image is lower poly, easier to read and follow. I need the other views of the head to study how the loops connect to the face from the other areas especially the ears.
    The pixar one has a lot of loops around the eyes due to the big size of the eyes, therefore more consideration needs to be given to tidy these loops up to places like the ears, cheeks, e.t.c.
    I feel the first one is great for understanding topology since its lower poly and reads well when it comes to the forms and facial features.

  • melviso
    Offline / Send Message
    melviso polycounter lvl 10
    Compiled some shots  from Sergi Caballer's vid on topology:

    Shows a low poly version of the pixar/disney face loops. Thought someone might find it useful. All credit goes to Sergi Caballer.
  • RN
    Offline / Send Message
    RN sublime tool
    Oh I remember that, it's so cool, it's the first time I saw someone using sculpting tools on low poly / already retopo-ed meshes, as a quick soft-selection way of changing shapes.

    This page of his portfolio has some examples of blendshapes under that topology (or similar enough), scroll down a bit:
    https://www.sergicaballer.com/portfolio/mario-marian-animum/

    There's also Josh Sobel's classic facial rigging course: https://www.joshsobelrigs.com/expressive-facial-rigging
    He uses Maya, but most is doable in Blender too. Great material.
  • musashidan
    Offline / Send Message
    musashidan high dynamic range
    A thread like this is never complete without the classic Gollum WETA Workshop wires. :D Not great resolution, but these are by probably THE most legendary character topology modeling pioneer, Bay Raitt. This is full film/VFX topology and built for much more complex facial rigging and render-time sub-div, but definitely worth studying. From Avatar to this day WETA went much more dense on the base facial topology rather than relying on animated displacement maps and Bay Raitt was pretty much the pioneer on the now popular technique of projecting organised topology to scanned maquettes and sculpts.


    And here's the reworked, more dense topology that WETA later adopted.

  • melviso
    Offline / Send Message
    melviso polycounter lvl 10
    Thanks guys :- )


  • Octo
    Offline / Send Message
    Octo polycounter lvl 17
    I think the first image is a better example of good topology compared to Pixars one. The main reason being that the Pixar version has the loops from the cheekbones running in under the nose, while the other one leads down to the mouse which more closely mimics facial muscle anatomy.
  • melviso
    Offline / Send Message
    melviso polycounter lvl 10
    @Octo I think the pixar one looks for ways to get good deformation when muscles stretch during deformations with the loops. 
Sign In or Register to comment.