# Learning topology

SO I understand the face topology. but I have issues with the checks muscles because front and side ortho just help with the shape. doesn’t blender now have a bridge that you can add divisions to help curve for thing like this? I am learning about topology on detail, fill and premiter loops.
(I would like to find a image reference for this) which I need to know which loop does what .

now also, I know quads are good topology but, for animation is that when you want poles and ngons etc? can you still do animation with a all quad mesh? I wouldn’t think so like face topology correct because of the mouth, eyes etc moving?

for an example if you have edge loops even though there quads when there is a pole how come the edge loop wont continue. i know it’s because of the pole but wont I dont understand is, it still looks like a quad unless I learn how to see the topology major loops like the detail, fill and premitier loop.

I can use the skin modifier and i know how to edge model but, this doesn’t help me with detail so, this is why I want to learn these type of loops.

For organic shapes you want to animate you should strive for all quads, as even a density as possible, and as square in shape as possible. That said quads on the inner lips will be smaller than the cheeks, they just will. In creating an all quad human face poles are unavoidable and neccesary. Whenever possible try for for a 5 pole or a tri pole (3) A pole will always terminate a loop. When traversing an edge loop, and you come to an intersection containing exactly 3 different edge choices, the loop will continue down the middle edge. If you have a 5 pole, there are 4 choices, and thus the loop ends. Same with tri pole. For a (realistic) human face you would never use an n-gon or triangle. n-gons can only be used in specific cases in hard surface modeling where the n-gon is completely flat, and has containing edges around it.

Figuring out where the points on the cheeks are is diffucult. It it not clear from front and side reference photos where to place those verts on the y axis. It takes some trial and error, make sure you get a 3/4 reference shot so you have a better idea of how the cheeks forms.

There are two major ways to model human faces. The first is tradition modeling, extrusion, using front/side (and 3/4) shots.

The second way is to sculpt the face using dyntopo, usually I get more reference images from different angles so I can pan around see how my face is looking from many different angles. For me it takes about 3-4 hours of dyntopo scultping to match a likeness. After that you retopo the sculpt, you may be able to shrinkwwrap a low poly face around it and save some time. Then it’s another multires sculpt to get medium details, crease behind the nose, brow ridges /creases… The general rule is, if you can’t sketch or draw and match a likeness pretty well you’ll be hard pressed to sculpt it convincingly. Sketching and drawing while away from the computer is how a lot of artists increase their skills.

Keep at it. The human face is one of the most difficult things to model and render convincingly. The major phases go like this:
Grotesque, Creepy, Uncanny, Realistic.

Topology is what makes the model function and supports next stages in the pipeline. It’s about clarity, predictability and control of forms. Not just for modeling but also for unwrapping, texturing, and rigging. All models are purpose built and it’s the intended use of the model that drops model requirements back in the pipeline, and for modeling that means taking those into account in the model structure.

Quads are preferred because only with those the model has a readable structure (clarity). Edge and face loops have a direction and with those it’s possible to have flows for static forms or deformation. It’s the vertices that move but limiting that movement and using the flow that edge loops provide allows to make new forms with a surface that make as much sense as the original (predictability, control).

Animated models also often use subdivision surface on top of the control cage to have more refined curves and surfaces, and quads subdivide very predictably: A quad gets subdivided to four quads on each subdivision level, or two edges get a new edge in between when looking at it only in one direction (Catmull-Clark subdivision) (predictability, control).

Poles are important to direct how the edge and face loops flow, and with that also the density on an area. The pole itself can cause problems because it’s a point where multiple spans end, a turning point for multiple surface flows, and trying to use that vertex for completely new predictable form by moving it is either difficult or impossible. That’s why those are strategically placed on areas with less or no movement so they won’t interrupt surfaces that result from deformation. Triangles and n-gons introduce poles and n-gons are worse with subdivision because the resulting pattern makes a pole with many more turning points than what the quads around would need.

And practice, practice and again practice, don’t just read or watch things.

From your past threads it looks you’re still stuck on the words and their meaning instead of actual modelling.
Reading about all the theory is a thing, but getting first hand manual experience is much more important in order to learn.

So go and sculpt/retopo/model things, you’ll learn topology in no time this way.

I am practicing and also trying new things like the skin modifier and dyno-topology. I asked this question before but forgot what someone said. I can see the face loops but what I dont understand is how to mesh replicates someones face, its there structure correct? so basically once I know how to do the basics if I apply this to any face there facial features will replicate the mesh like this Angela Joli picture? because some face topology pictures I have seen dont look like there face until they apply a image picture of there face over it. so basically the topology doesn’t change only the face unless its is Disney where the yes are bigger and mouth is smaller?