Topology research


Ya you right. I guess my memory is poor.

Anyway I’m no modeler, just someone who enjoys reading stuff. It’s interesting to see how theories have evolved. In my opinion this thread is a little advanced for beginners and for anything to click for them they really have to experiment and do things on their own.

There seems to be about 52 pages for this thread. I haven’t seen anything revolutionary that come out of it or maybe I’m just too lazy to read bit by bit.


I’ve read nearly all of this and almost all of the identical thread in the maya forum (which is longer) - the general form of topology for a face and other bend-y organics has been sharpened, polished to such a fine edge that I think these two threads tend to be better used to ask for CC on a models topology if you’re new and having trouble getting the flow right. As such, I’d say these two threads are extremely important for beginners.

Laa-Yosh seems like someone who’s getting kinda tired of beating the dead horse of whether the basic rules of topology - the basic loop structure that you’ll find in any quality model is the one to use: Check out the Hardcore Modeling forum: none of the models there are for animation. Yet they all follow the same basic topology.

As the Gollum model - it’s worth studding in all it’s intricate detail - you can even see the use of poles to control edge flow as described in some hugely long post linked somewhere here. There’s an awesome eye loop structure that you generally don’t see. But it’s super cool.

Until such time as a new surfacing system or a whole wacky new system for animation - basically a paradigm shift in the whole approach, like subdivision surfaces were some years back, I don’t think we’re going to suddenly suddenly realize that really it we’ve been doing this topology thing all crabbed for the last 5 or 6 years.

The biggest change, and it’s an incredible and awesomely shiny change has been the introduction of zbrush / mudbox (and topogun) being able to take as inhuman a mesh as this: and then sculpt out an awesome person, then use topogun to get the second or third level subdivision as the control cage for rigging and animating that has the topology required to support the detail level of the high res mesh is … well it’s freakin’ amazingly awesome.


I personally don’t recommend ZBrush for beginners…especially if they want to animate with the model. ZBrush will seriously screw up your entire model if you don’t know what you’re doing. It’s brilliant if you KNOW what you’re doing with it…but for people who have never even heard of topology I always tell them to leave it alone (except for if they want to texture)…

…I personally don’t even use it all that much. I’m still trying to learn from it, but at the same time I want to teach myself how to model a dense, highly detailed mesh without having to rely on ZBrush…I dabble in and out of it, but until I figure out this monster called topology, and getting it under my control the way I want it I rarely use ZBrush…because if you go in and sculpt all this detail, or completely make a model from scratch you gotta re-topo the thing anyway.

As for beginners…its kind of a tricky thing. I think this thread is vital, but as far as comprehension is concerned I don’t think it’ll be there for a first time user. It took me a while to grasp the terminology (vertices, poles, Ngons, etc), and learn how to move the polygons around to get what you want. Took me almost 2 years to get to where I’m at today, and even though I consider myself at an intermediate level, some of this stuff is still just an inch off the comprehension meter for me.


I’ve posted one of my models here before, but I personally feel that I’ve come a very long way since then. I’m able to get a lot more control over my mesh than before, but I don’t want to rely on my own personal crit to make improvements, so I would like for you guys to rip this apart. Please be as harsh, mean, blunt, etc. as possible. What do you think is working? What don’t you think is working? What should my focus be?

Note: I know there are Ngons and other no-no’s and boo-boo’s in the topo. I’ve yet to correct them. They’re merely “place holders” if you will. :slight_smile: And the model isn’t this dense. I just upped the polys so the topo could be read a little more clearer and I didn’t want to post something in Sub-Ds…


The flow looks fine, shuggs. Ignoring the obvious place holders, the inner brow region strikes me is overly complicated - you may want that edge flow, but it strikes me as unnecessary in general.

I suspect it’s being caused by something in the lower resolution.

I wouldn’t suggest topogun + zbrush / mudbox for a beginner either - it important to understand the whole edge flow thing. You’ve got to climb the mountain before getting to take express elevator. Though, that should be tempered by the whole process. Running every loop and moving every vert, to borrow a phrase, seems a lot more like surface management then modeling.


Laa Yosh,

I consider you to be an elitist in a negative way. It is very difficult to engage in a dialogue with you without you being cynical and non-emphatic toward others. Granted not everyone can whip up a world class model like you or the other Z-brush guy, but nobody is born a master modeler. Through some kind a process one reaches this excellent goal. But you have nothing than destructive critique to offer. It’s guys like you I’m crusading against.

There is a lot to be said on the topic of topology. Lots of techniques that master modelers are aware and unaware of, but nonetheless they keep those for themselves. That’s why I try to analyze every aspect of what makes a good topology and how to get there. I think that this thread is getting derailed by guys like you who shun away from the technicalities but indulge in the vague.

I’m sure you’re a swell guy, I’m just commenting on your attitude.
My 2 cents.


I am just pissed off a bit, it’s that simple.
I’m swamped with work but take some time to post the image that I’ve learnt the most from and it gets immediately dismissed. Well, some would probably leave in silence, some would say “frak you all”, others might apologize for trying to mess with the topic. I get a bit rude, some sleep will probably help it.

Then again I’ll probably shut up for now.

  First of this is no way of discussing, You say certain people derail this topic.But Laa-Yosh posted a vailuable classic image of good topology with some input on it, nothing vague about that. 
  Everyone can give his or her opinion on topology and new findings, but dont come in degrading someone else his input because you find it to vague, when you yourself dont ask him or her a question to clarify.  I think everything in here is pretty clear.
  And i clarifyed some things in my earlier post. 
  Now im just going to discus further on the one thing that was on topic in you're post 
  [i][b]""Lots of techniques that master modelers  are aware and unaware of, but nonetheless they keep those for themselves.""  [/b][/i]
  Aldo this could be seen as true for a beginner, this is not the case. For all people reading this thread and strugling with translating there design to 3d, just look around , making off's wip work and so on.
  All the masters are really know for sharing there topology, you see it allmost in every making of on even regular dvd's how the greatest cg characters in 3d history are layed out.  You must not expect all those people coming in here and posting all there stuff.
  Just look for it and you will find it all.

you’re flow/ polygon direction is descent, loop around the eye area, mouth, but try to have the mouth loop star above the point of the nose, around the mouth, over the chin. Make sure you’re polys are even distributed, now you have very dense areas. Remove the sharp loops around the nose. In he neck and forhead you have some loops construction. Try not to model in detail right away, start with the flow, where are the big muscles and bulges going. Model in the fine detail later, dont make it hard for yourself by going to a dens polycount on specific areas right away.

  For you now, the best way to get more experience fast, is try to model body parts, as good as you can, study them, and eventually connect that, go step by step.
  Create a good mouth , nose, eyesockets etc.  You can do everything in one go, but its important to know what edgeflow works best for every part.  Note,  model these things with not to much polys, less is more, just get it right, if less still atteans the same shape, than dont ad edges for nothing. Maybe in the edn yuo can connect it all, and have a face.

One of the wires that inspired me years ago, here you can see in a good manner how a main flow influances all forms, even the more detailed parts. So it is easy starting with a less dens layout, and filling it in. Aldo modelling highdetail can be done as well… Do it myself most of the time, but to learn and understand what you are doing best, it is wise to start low, with the most important shapes.

Btw this example is not flawless, there are some poles i would not let in there, but it show what everything does very well.

  Good luck.


First of this is no way of discussing, You say certain people derail this topic.But Laa-Yosh posted a vailuable classic image of good topology with some input on it, nothing vague about that.

Ja, na heel veel gezeik eromheen en lijntjes trekken op bestaande lijntjes kan mijn tante ook!

As the Gollum model - it’s worth studding in all it’s intricate detail - you can even see the use of poles to control edge flow as described in some hugely long post linked somewhere here. There’s an awesome eye loop structure that you generally don’t see. But it’s super cool.

I’ve seen this before and I’m seeing it again. The pentagon shape around that pole propagates a nice diagonal topology all the way up.

I think that one of those pentagons is all you need to slant the topology. Gollum’s mastoid is almost poleless. I will look further into that because the location of it’s complentary N-pole plays an important role too I think.

This solution will permit veins running across the mastoid which can be seen too in Gollum’s neck.

This is in contrast to the solution below where the edges of a N-pole is made into a triangle to slant the topology, but it seems that you need to ‘draw’ a succession of those to achieve that. Furthermore if you want to add detail across the mastoid it will make the base of the neck needlessly denser too.

By the way, I have seen a lot of similarities between Gollum’s and Gerry’s topology.


This is so helpful. Thank you, man! I really appreciate it! I completely looked over the brow region for some reason…

@LY: Just know that this learner hasn’t dismissed a thing. I’ve saved that image and have been studying it ever since…


No problem - helps me to see other people’s stuff.

I also save piratically every image that pops up in this thread - it’s all gold.


What you pointed out in gollem and you’re example image are ways to break the current flow, to end up with lesser polys, to have a clear view of what is there, to later on ad detail, and most of the time, than that poly construction will dissapear.

Aldo the lower image is a bad example, because of it’s layout. The gollem image gets the pentagon shape because of the intersection of two loops. It is a litrle bit visible, these shapes cant really be helped, you get them as soon as teo opposit flows of youre polys/ edges, (how ever you wist to see it) connect to each other, a loop in rigid area, see chin, or two loops connecting, the part in the neck you pointed out.

This mainly are the areas that you want to pinch, or folds occure when making a blendshape, it incurages the mesh to move a certain way when moved. Helping ,making blendshapes a little bit easyer, when movement is predefined by its mesh, just as it is by skull, muscle and fat. A chin can there for be a more square modeled part/ less restricted to the main flow, because the chin is rigid and controlled more by the jaw bone than muscle and skin movement. - This differs per character, mainly influances by it’s thickness/ fat tissue. and skin properties.


i went through all the diss , every modeling artist should read this thread


Well, it’s hard to pin-point one exact universal topology that works for every head/body. And topology is a subjective sport when learning it. I usually put some of my works away for awhile and come back to them later so I can look more objectively at them. But no two of my models share the same topology.


This is brilliant. I can grasp most of it, but some of it zips right on pass my radar. The 5 region drop off? Anyone have a clue what that’s about? The closest thing I could think of was giving that specific region (arm, forearm, shoulders, etc) a reasonable drop off region to deform and move through space properly so they don’t run into the other regions of the body and creat something off key.

Am I in the ballpark with that or off?


The denser the mesh the better it will hold volume. So finding a balance between density and convenience is important.



The pages have changed to:

The fade and drop are just some b.s. terms that I use to try to help people when they are just
starting to learn model for movement. Basiclly a fade is when you need to have spans in an area like
the elbow and the knee. The drop is when you have a direction change and is used in the arm/shldr
area and the hips/thigh area.

To see some of this in motion with this type of modeling:





I just saw under your name that you work for Pixar?! Is this true?! And thank you for the updates! I’m learning so, so, so, so much from this. It’s inspiring. I was up until 4 A.M. last night reading!

I’m preparing to send my demo reel package to Pixar for their ‘Recent Graduates’ internship program. I’ve been so stoked about it ever since this past summer. If I get into it it’ll be a tremendous dream come true. Ever since Toy Story I’ve told myself I’m going to be working for them.

And thanks for explaining the drop off terms. That really helps. I take it that the black loops are the major stress areas of deformation while the grayer-to-white areas help support them?


Yes, the dark areas of the mesh are the areas that start the fall offs. They are there to help me see
my mesh movement, but these fall offs are not set in stone. Is it the right way?, not sure but is the way that i work. every one has a diff approach. I use a modeled bone structure under the mesh to help me
keep my volumes consistent while I articulate the movements.

The site is just a leaping pt. for people to see the way that I approach pts. in movement.




Hippydrome’s stuff is pretty cool, although I recall you originally used clusters a few years ago, right? Is this simple blendshapes now, or the pose based deformer from Michael Comet?

Anyway, here’s some more food for thought.

An interesting new approach where you just don’t terminate detail at all. Sort of combining the best of polygons and NURBS, where the position of your 5-sided vertices and number of spans/loops aren’t locked together. And of course you don’t have to worry about stitching either.

Keep in mind though, topology for a young Brad Pitt would certainly be quite different…