EVERYTHING You Need to Know About Topology

hey i'm jay and this is everything  you'll ever need to know about topology all right so we're gonna get super nerdy in  this video all right so if you're into 3d   modeling uh getting better at 3d modeling making  characters that sort of thing stay tuned hopefully   you find some useful info in this video if  you're not really into those sort of things   go into like a deep dive then this video is  probably not going to be for you my goal is   to make this video packed with all the useful  information on topology address kind of the   most common questions misconceptions and mistakes  that i've seen throughout my years modeling and   teaching and trying to make a video that just  encapsulates the whole everything you need to know   about topology that i could send to people that  you can send to people so i just want to be on my   channel for all time an overview of what topology  is how you can make it better and how it can help   you be a better modeler i've separated the video  into chapters down below so you can skip around   if you want to get info on a specific  topic or refer to something in the future   okay so first up what is topology seems like we  should start there so let's define topology also   referred to as edge flow topology is the term used  for the arrangement of the verts edges and faces   of a 3d model all of the verts of a 3d model  are connected by edges and those edges are   like the scaffolding that make up the faces  which is the actual shell of the model itself   the faces of the model is what you see and  the topology is like the web that holds   everything together 3d artists starting out uh  usually don't think enough about topology or   on the other hand some artists think about  it way too much kind of overthink it so   we're going to try to describe what makes good  topology what makes bad topology and try to talk   about the things you don't really need to worry  about and the things that are best to focus on   okay so why is topology important anyway like  why do you need to care about it if you're a   3d modeler i think of three main reasons number  one is the poly count the topology is the mesh so   it is what determines the poly count now i'm a  character artist in the game industry that's like   my day job so poly count is more important to me  than other types of modeling let's say because it   has to run at real time and with all the other  assets in a game so it's something that i think   about maybe a little bit more so than if you're  modeling for a different reason but polycount's   important no matter what kind of thing your model  is for and we'll talk a little bit more about that   but i think number one is polycount you don't  want to have a bunch of faces that aren't doing   anything you know you don't want to be wasteful  and you need to have enough to make your model   look good so it's all about this balance which  we'll talk more about too of of getting the   model right to do what it needs to do and also  fit under your budget for whatever you're doing   okay so number one poly count okay second reason  why topology is really important is deformation so   especially if you're making characters characters  are meant to move right they're meant to perform   so the topology is the actual mesh that's  animating your final product that you make   as a modeler so not only do you have like simple  joints which are pretty straightforward but you   have things like a face you know that can make  all these like squishy squashy poses and so it's   really important that you build your mesh in a way  that does everything that it needs to do which is   like renders correctly it's nice and smooth and  then can be manipulated and deformed by technical   animators and animators to give a performance  you know as a modeler that's kind of the   ultimate is that you can make a mesh that when  collaborating with others can come to life and   give a performance so it's important that your  topology can deform well in a believable way in a   naturalistic way to do what your model is meant to  do i don't want to make it sound harder than it is   we're going to talk about just a few of  the key principles you need to keep in mind   when you are building a mesh that's meant  to animate and deform so we'll look at some   examples a little bit further in the video  again don't want to make it sound like it's   you know super hard it's just a couple things you  need to be keeping in mind and pay attention to   when you're building those meshes okay third  reason maybe the most important reason it's   how it looks the topology that you build is your  mesh so how smooth it is uh how detailed it is   whether or not the you know maybe a parts of a  costume are detailed and so the modeling there   is what will make it look realistic if you're  making a game model it's whether or not the   silhouette looks jagged or smooth you know how  it's shading how it's rendering what is actually   3d what is texture so this is all going to be  part of like a balance that you're considering   as you build your model for whatever it's doing  so having a good understanding of topology is   going to help you to make better meshes that  are the correct poly count and under budget   can deform well and animate and look good when  they're rendered nice and smooth okay now let's   talk about a few of the most common ways that  you will be making meshes thus making topology   generating topology and kind of the stages of how  involved you have to be and how specific topology   would need to be and for what reason okay so let's  talk about making topology all right so there's   two main ways to make topology that's procedurally  and manually so let's first talk about   procedurally so procedurally obviously automatic  that's its strength it's doing a lot for you   a lot of people may not be thinking of  it this way but dynamesh and z remesher   are procedural and they're fast and that's the  strength is you're giving up control for speed   and this makes it quick to iterate that's why this  is awesome and used all at the beginning so you   can use it to make base meshes even more refined  meshes with z-remesher something that's really   good is with z-remesher as opposed to dynamesh  is you can give it some control so dynamesh is   just throwing topology at your meshes which again  is fine for like building quick shapes but then   typically i'll be transitioning to z-remesher  it gives better topology for the shape whereas   dynamesh is totally agnostic of the shape so zero  mesh will help support the shape you'll get better   topology that way and you can control it a little  bit you know you can have poly groups and say   keep these groups and kind of control it in that  way you can also kind of manipulate the mesh   because it's a little bit predictable in that it's  looking for strong edges and things like that so   still doing a lot for you but you have a little  bit of control in some of the use cases of z   remesher but it's not optimal that's the downside  if there is one but it's used all the time   also i'm using it for things like decimating  you know with decimation master if i want   to export high polys out for references  or even for baking and also procedural   topology is used for lods in games that's  something that we used to have to do by hand but   now it's super common to just use some kind of  a third party tool like simply gone or something   like that and do your reductions of your game  characters procedurally it's way faster since   it's far away from camera you don't need as  much control as you do on the lod0 the main   model okay so that's procedural now let's  talk a little bit about manually doing it   obviously manually meaning doing it by hand so  typically that means drawing or creating a mesh   on top of your source mesh this is also referred  to as re-topology probably seen this like in maya   you can use quadra so maya comes with a tool made  for this purpose and then there are third-party   software like topogun that are it's an entire  software made for this practice so you get like   extra bells and whistles uh using topogun but  you have to pay you know separate costs it's a   separate piece of software so if you're serious  about retopology you're doing a lot of work   uh it makes sense but for the most part you  know you don't need that i also know blender can   do re topology but this is generally the practice  you're drawing your polygons on top of some sort   of live surface of your source mesh that you  made and you probably made that source mesh   with dynamesh and or z-remesher okay so why would  you do this it gives you the most control you're   making polygon by polygon in some cases you know  you're adding and removing and editing every edge   so this means for production models it's essential  still all series productions essentially do this   unless you know for some reason they can  be wasteful and and in some cases maybe you   could do that but for games every polygon  counts really that's becoming a little bit   less the case but still you know it's not  very common to procedurally generate meshes   that are going to be used in some kind of game  production so the downside to this obviously   time consuming the most time consuming way to  make topology but gives you the most control so   it's the most optimal mesh usually right it's the  best for deformation it's just the right amount   of polygons that supports your shape without  being too wasteful so the silhouette is clean   and it can animate and you're left with an optimal  amount of polygons which you wouldn't be able to   as easily get with procedural modeling so  those are the common ways of making meshes   from auto generating it and just not caring about  it that much because it's all about the shape   to the far end where it is about the topology so  if you're making a real-time mesh that's meant to   deform and animate the best way is to do it by  hand because it gives you the most control so   you're going to make an optimal mesh it animates  as good as it can and it's like as close to the   budget as possible and that's how you get the  most efficient mesh that does as much as it can   do and we'll do a little quick example of of what  that would look like in practice okay so how do   you determine a polycount that's a tricky one uh  probably the most common question that i get and   that's you know unfortunately impossible to answer  the real answer is it depends which uh you know is   is kind of unsatisfactory uh and understandably so  so let's talk a little bit about what it depends   on and how you can better determine what a optimum  polycount is for your particular model okay so the   polycount how can you determine a polycount for  your model it's very situational uh so there's   not going to be a one size fits all but this is  the overall strategy now in an efficient model you   know one that's built optimally and isn't super  wasteful but one specific like game characters or   efficient rendered models the topology is really  serving three major important things and that's   the deformation the silhouette and the detail of  this model the deformation you know can it animate   uh believably does it look good when it's moving  around does it have enough loops to do that   okay check silhouette is it jagged or is it  smooth enough to be believable the textures   are going to take care of the inside details but  does the model overall have enough to support   the three-dimensional look of the character and  in a way that isn't noticeably uh faceted okay   and then the detail does the model or character  have some costume elements or props or weapons or   little details like glasses or jewelry or  something that require in a disproportionate   amount of polygons so that it looks believable  okay rather than talk about numbers we're going   to think about this as proportional all right  since every model is going to be different but   if you think of the proportions of what the  overall polycount is it might look something like   this deformation is uh priority maybe priority  number one but doesn't take up a majority of the   count because it's just are there enough loops on  the joints that doesn't take a large amount it's   just placed properly now silhouette is going  to take a little bit more depending on your   character too maybe it's got a lot of hair maybe  it's got round elements and armor or something   and then detail could take a disproportionate  amount so let's say your polygons are spread   out like this well we have a budget so our  overall poly count might come in high if you   don't have a budget you have to set a budget per  game per character everything is so different um   my general suggestions to people is if you're  making a character for a portfolio then i would do   it anywhere between 50 000 and 200 000 triangles i  know that's a big range because um why not i mean   every year it gets better and faster if you make  a character between one and two hundred thousand   triangles it should look pretty good and so that's  generally what i would recommend to people if   you're making a certain type of character that's  for other games like you know an overwatch type   of thing a fortnight type of style then you don't  need you know 200 000 would be wasteful really a   lot of the poly count will come from the hair or  how smooth everything looks and that's really for   certain you know particular types of characters  so you don't really always need that many   so that's why i give such a broad range so it's  kind of depending on the style of character   that you're doing and you try to get it within  that range and then it would be reasonable uh   in a portfolio in my opinion all right so that's  my most general answer to what a poly count should   be so you can kind of think of it as your  overall poly count is like a point system   that you're allocating to the most important  parts of the model and that's what's going   to determine where you start placing these points  then you're going to have to start taking away to   get our overall budget down now you don't want to  take them away from deformation first like i say   that's a priority you want to be able to animate  that's the whole point and the silhouette needs   to still look pretty good so what are you going  to take the most from it's going to be detail   and you're going to take away topology in places  that aren't focal points you know like below the   belt and shoes and things like that it maybe they  have some props on their back so it depends on   the game and the camera but you're gonna start  where the camera doesn't see it most often and   you're gonna start taking away tiny details that  would be nice to have so when that happens then   we start to get below the budget and now maybe our  ratio looks like this but this is how you're going   to finalize your poly count you're not just going  to build your low poly and then it's done you're   going to build your low poly in a reasonable  way try to make as best predictions as you can   but ultimately you're there's going to be like  a cutting phase at the end when you're really   trying to get specific so with experience  comes kind of a general understanding of   how high a mesh will be as you're making it  but you're almost always going to come in   over and that's really what you want because  it's easier to remove than to add so it's   even good to come in a little over and bake it  because the textures will line up to high poly   a lot better and then you cut it to get it down  below budget whether you're making an object or a   character that's meant to deform or you're making  a character for a game or a cinematic or something   you're going to make different decisions on  your topology so let's talk about the difference   between making a character for offline rendering  cinematics movies things like that and then games   so we'll continue on with games for a little bit  since we're already talking about it in games it's   typically going to be lower obviously and when  it's actually rendered in the engine with the gpu   the entire model is thought of vertices and  triangles so that's why in games we measure   the poly polycount in triangles that's also  why it's totally normal to use triangles   when you make your game characters now you don't  want to triangulate your game character that would   make it really difficult to uv map to select edge  loops and edit the model so we still are typically   modeling in mostly quads but because it's going to  be triangles anyway it's really common that you'll   use triangles to end edges to make things cheaper  because obviously that's the name of the game when   it comes to making real-time art but also you'll  sometimes add a triangle yourself so that you know   exactly how a polygon will be split because  if you leave it up to the renderer it might   rotate the edge in a direction that you don't  necessarily want so sometimes you are adding   triangles to make sure you're getting the right  shape that you want now with models for offline   rendering cinematics and all that stuff it's  typical that the mesh will subdivide on render   or at least what is being rigged and what's being  skinned and animated and modeled is not the final   polycount version it'll be subdivided okay so  it's really important that the mesh is quadded   out because when you subdivide a quad you get four  polygons and it's very predictable so if you're   doing animation on something that's made with  quads and you subdivide it it'll look more smooth   triangles don't subdivide very well which in  game's case is fine because it's not ever going to   be subdivided but in the case of offline rendering  it's not going to be great it's going to start to   make the mesh messy it's going to add like pinches  and stuff and because you're not having this   perfect square that's subdivided then things like  your displacement map aren't going to move the   vertices in a predictable way that you're going  to want so it just causes problems so if triangles   are used on a mesh for offline rendering they just  hide them you know put them in the bottom of the   feed or between the fingers or behind the ears  or whatever something like that you definitely   don't want to have triangles on all the squishy  parts that move around so that's why we measure   triangles for game characters whereas for offline  rendering and you know like how high you know how   many polygons is your high poly model that that  type of stuff we just typically are counting   polygons which uh are usually quads so let's talk  about high poly models that is the beginning stage   of almost everything most of the time you're  going to spend as a model as making high poly   models one of the things about high poly models is  really the poly count doesn't matter what you need   from a high poly model is just enough polygons to  hold the detail that you need for that specific   model that you're making and no more and it needs  to be able to run in your machine in a reasonable   you know amount of time so that you can sculpt on  it and that's it so the number doesn't matter it   could be 5 million it could be 10 million you know  doesn't matter just make it as high as you need   don't go higher because it's just wasteful  it's going to make everything slower   so if you don't need tiny little pores then  you don't need it and if you have a jacket   or something you don't need you know 20 million  polygons high poly models are made for the shape   and for sculpting and that's it it doesn't have  anything to do with deformation because this is   about high poly models this is about a source  mesh that you will derive production meshes and   textures from one of the most common ways to make  final meshes that become these high poly models is   to use zebra mesher so we talked about manually  doing topology for production ready meshes when   optimal poly count is important well in high poly  models optimal poly count isn't important what's   important is the shape and the sculpting right  so z-remesher is made for this purpose it makes   geometry quickly that takes into account the shape  and does its best to evenly distribute polygons   on it okay last thing about high poly models is  subdivisions use subdivisions none of this super   high resolution dynamesh stuff you want to use  subdivisions so that you have the most control   it follows the principles that we talked about  before which is having topology that reinforces   the form that you're after helps enable you  to create clean polished forms and surfaces   more easily without the resistance  of having just that generic topology   being able to subdivide lets you get up and  start to create details and at any time you can   drop all the way down and work on primary forms  subdivisions can also serve as a proxy in zbrush   so the entire scene moves faster too so those  are the main differences in topology for high   poly models now we can look at more production  models all right now let's look at some examples   all right so some of you may know i recently  published my base meshes that i made for my own   work so that people that wanted to use those  or just buy a base mesh can get them and so   because of that i've got my own base meshes my own  topology that i spent a long time on so i thought   let's use it as an example for this so i'll just  briefly touch on a few of the principles that we   discussed and i'll show you how i implemented them  and then i'll show you a couple i guess unique-ish   things that i ended up doing to my bass  mesh and i can explain the reasons why   that might give you some idea on something that  you might want to do in your model so important   off the top to try and convey what i wanted to  accomplish with this base mesh is that it would be   enough to render in real time and look pretty good  but then also i can subdivide it and sculpt on it   it's my preferred way of working on organic  characters if i can i'd like to make the high   poly model or the subdivision level one the actual  real-time mesh that way they're connected so i'm   only doing the topology once so it's kind  of serving both purposes as the render mesh   and the base mesh so when i can do that i will and  that's what i'm doing with this this will be what   i subdivide and then work on so because of that  you know down here i'm making the polygons pretty   evenly spaced like a grid now this helps with  sculpting because you know just like pixels   in a canvas you want them to be similar size so  like if i sculpt veins and wrinkles then i have a   similar resolution over the whole hand as i  subdivide so same is true with the face though   i can't make it all even because just how faces  are pretty complicated you can see like because   of deformation it makes more sense for me to have  more loops around the eye and the mouth so we're   starting to get rectangles you know rather than  like square polygons which is okay but it's just   something that is worth noting i think that's fine  but that's why i have some more faces than i need   but in general we're getting to a space now where  real-time meshes don't have to be you know super   cheap so i still don't think this is very high  but this definitely could render and you know if   i showed you without any lighting whatsoever it  still looks pretty smooth you know for the most   part the head looks smooth which is an important  part the arms look pretty smooth uh the fingers   look okay so it doesn't look stand outish you know  it doesn't look like really polygonal so that's   good you can see little jaggies here if you really  look but for the most part um this is my kind of   minimum for uh most things now i could also use  this for some kind of pre-rendered i mean i could   easily just smooth this once and get four times as  many and now all of a sudden this is going to be   totally fine more than i need so depending  on my purpose i can do that so i want it to   be as versatile as possible okay so other than  the amount of polygons and the size you can see   the general flow here is pretty easy right this  is all horizontal because she can bend back and   forth same with the elbow you can see it's already  getting a little bit of the bend in there now on   the joints that bend like this that just bend in  one direction pretty easy way to think of it is   just you just want to make it just like a grid  just like a bendy straw you know you want to have   some edges and as it bends you want to be able to  have more edges so that when it does bend it still   looks smooth if you didn't have that many edges  then all of a sudden when it's starting to bend   you know it starts to look sharper and that's not  what you want when you see these facets right here   even now it's getting a little pointed on the  elbow but i think that's fine it is an elbow   i even have an extra loop here to surround the  elbow bone so that when she does a full extension   hopefully this shape can still be supported the  chest you may notice isn't as round as it could be   that's fine with me because i did make this base  mesh universal this is the same topology for my   male shape as my female shape that we're  looking at right here and i don't plan on   ever doing a character that is 100 topless or  anything like that that would be a custom model   you know the fact that she would be  wearing some kind of costumer clothing   i could always rip the arms off if she was fully  covered so i'm not even using this body mesh   other than maybe a shape to make clothing on top  of so that's not really a problem at all all right   so now i'm just going to point out a few you know  somewhat unique things that i did to my bass mesh   and i can give you a couple reasons why i did  that first we can just start with the head is   if you notice i have some stars here stars are  vertices that have five edges coming from them   uh typically it's indicative of a direction change  it's just a lot of polygons and so there's like   you know the flow is kind of turning so you do  need to have stars for the most part if you're   going to have you know a mesh that has quads this  has quads everywhere game meshes are going to be   triangles right so that's not a problem if you  use triangles but because this mesh is meant to   be versatile when i'm sculpting on it i don't  want triangles and if i'm doing subdivisions on   render for displacement maps i don't really want  triangles so why not do quads it's also easier to   make selections and uv map and stuff like that but  here you can see these are all direction changes   here's some more stars these are all direction  changes uh without using triangles so back to   her neck so yeah you can see some stars here now  this is because i have this head split like this   i wanted to be able to have the head with enough  room to have neck holes and stuff if i wanted to   detach the head and still have a little bit of  the chest so i prefer doing it this way rather   than having concentric circles go down lower and  lower and lower and also the shoulder right here   has its own lines so that she can raise her arm  right and then this gets to compress and then   this stretches out right here on the underside  so i'm getting the lines here for deformation   i'm getting the topology here that's gonna flow  from her chest over to the top of her deltoid   again when she lifts her arm up hopefully it keeps  that shape that way you can keep the bump here and   you can even see on the underside i actually  just used it to reinforce the armpit shape   so as she moves around hopefully this will still  be able to deform the shoulders are a big point of   deformation and stretching in characters so really  a lot goes on the technical animator being able to   have like supporting bones and good skin waiting  and stuff like that but this is just a way that   i reinforced it with some extra geo and kept the  flow going without using triangles so you can see   here if you pay attention that you can see like  some more stars star right here start right here   this isn't necessarily a star but it's a direction  change so let's take a look at some of the stars   on the face so with these stars or poles uh a vert  with five edges it can cause some tension in the   mesh visibly actually when you're smoothing it  you know it'll just feel sharp so it's not going   to be as smooth as like this area right here  when it's subdivided it's going to be super   smooth not a problem at all but you know stars  like this could and when you're sculpting it   it's not as easy to get the exact shape that you  want so when making faces it's best practice to   strategically place where you're putting these  stars so i'll explain why i chose where i chose   this is just a direction change not really a  star but you can see like i wanted a forehead   with lines like this so that it can compress so  she can like raise her eyebrows and furrow them so   it's best if the forehead is a grid in my opinion  and then so i just have this you know flow come   around so that it can meet up with the eyes which  are concentric circles concentric rings which is   good same with the mouth the mouth and the eye  are both very similar in that they are rings   going outward so it can do like the squashing  and stretching and stuff and it can look around   the nose is really just about holding the shape  there's some deformation here right here you   can see the flow come around so instead of the  concentric rings from the mouth we end here and   now we're coming up over the nose so this is like  if you raise your cheeks and stuff you can create   these shapes so you have to have these polygons  and able to make some of these shapes if you're   going to do blend shapes and things like that so  it's important that you have the edge flow echo   the facial shapes or the direction a face can  slide around on top of that it's extra geo here   that's maintaining the shape of the head looking  good while it's rendered in all angles so it's   serving dual purposes so where i place these poles  right here and right here and right here and right   here now i put them here because i heard a good  tip a while ago that if you have to put a pole   somewhere on a face it's good to put it in an area  that's that doesn't deform that much or already is   somewhat hard you know by that i mean like a  bone or something so here we have this temple   doesn't move that much either i think it's  a perfect place for a direction change   this just trying to find a space in between like  like i said like here's a section that deforms   right and then here is a section that deforms  this is like what would make the crow's feet   and like if you're doing a big smile and then we  have the jaw which can open so this area right   here can swing open so it's a good place to put it  because you have these different sections that are   kind of moving in different ways in their in their  major directions but then also i kind of think of   a place right here is kind of a good spot because  not it's really just floating around i'm not doing   a lot of crazy detail right there so for me it  made sense to put it right there right here i'm   thinking maybe kind of like the orbital bone again  like we have the squashy eyelids around the eyes   and then we have like a cheekbone and it's coming  in different directions so i feel like this is a   fine place to put it if there's a little bit of  a corner on it i think that's okay most of this   stuff by the way not that perceivable this is just  the super like geeky stuff that you'll get into   once you're once you've made several base meshes  and you're just trying to get the perfect mesh   something i think is probably worth  pointing out and calling out here is   this kind of structure here i wouldn't do this if  i was doing a mesh purely for sculpting purposes   and i wouldn't do this if it was for offline  rendering only i really only added this here   for facial animation at this resolution if i were  to use this as a real-time mesh and that's to get   the squash of the brow so that we can move the  brows in this direction and get this to to kind of   compress and make a wrinkle right here other than  that i would have just like had the lines come up   you know that would be the simplest way but i  put it there again you can see some stars or   some poles and those i feel like are the perfect  place to hide these because there's going to be   eyebrows right here so you can almost never see  them when they're placed right here i've seen   them placed in other places if they're too high  you know might be a problem if they're up here   so i always try to put them right here if i can  same with the ones down here again like the nose   isn't moving too much it's kind of like a a solid  structure right there so that's why i put those   there as well so you can see the the line flowing  underneath the structure of the neck uh the ear   the topology if you if you notice some poles  again and what that is kind of doing is making a   kind of a section here so the ears it's like  its own island and that's how i model it so   that i can have topology that supports the  structure of the ear the ear doesn't animate   so it's purely for the form of it so really what  i'm trying to do is stitch together these sections   that i know have to be a certain way you know  and then i'm getting like even flow in between   and then i'm trying to stitch them together and  finding ways to like route edge flow if i need   to if i have some extra edges or something but the  ear kind of a good example of of edge flow that's   purely for the model and the structure of it and  not for deformation at all all right so this is   my base mesh and uh some of the reasons why i did  some of the things i did now let's take a look at   a couple of examples of meshes that i think are  not too good okay so here is the demo head from   zbrush this comes with zbrush it comes with a  couple base mesh tools that you can use to get   started so i thought we could look in here and see  if we could pick some things that could be better   overall it's not too bad i don't think um  the main things that stand out to me are the   the neck i mean this one kind of stands out uh  obviously they put the poles in different places   which is fine but there's like a the direction of  the flow you know is is not supporting the shape   that much it's being stretched out here to conform  to where they want you know essentially to conform   to this part probably so i think that could be  better it's actually kind of going against the the   flow of the tendon of the neck and the muscles so  i would do that a little bit differently you can   see we do have some concentric rings here around  the mouth which is good to see this is the best   way to do a mouth i might go a little bit more but  this is uh this is not so bad uh then we have this   star here that actually has like one two three  four five six seven edges i think that's way   too much stress on this one vert so you can  even see like the flow how it takes a sharp   turn and the difference between this face  and this face you know it's really big so   i think i would probably add some more topology  here and change the flow there's even a change in   direction here on the bridge of the nose i assume  to support the bridge of the nose but i think   we could still do that by having like the edge  rings going over the nose it would also help with   animation and deformation like if you're doing  uh snarls raising the nose squashing that kind   of thing like intense frowns and stuff right now  we don't really have topology that would reinforce   shapes like that it looks like it's built more  for the static shape than it is for if it deformed   even right here putting this star right here isn't  so good because we want to have this shape right   here of the nasolabial fold so that when the face  does this you can get that crease you can get that   compression of the forms of the cheek and  everything you know i don't know is this helping   but you know what i'm saying so i don't think  that's a great place for a star and i wouldn't   want to do a vert that has so many edges anyways  definitely want to loosen that up and just get the   the flow a little bit nicer like this and probably  just cross the nose uh the ears okay doesn't have   a lot of the internal structure supported probably  do that in sculpt though but yeah so that's this   mesh then let's take a look at one that i think  is worse all right so this is the julie mesh   from zbrush also comes free i think this has  subdivision so it's probably you know you probably   wouldn't want to start with this one i mean you  could always dynamesh it but you can see here   how the ear is just totally botched you can tell  that the topology was just made simply and the   ear was tried to was like attempted to be built  out of that so in this kind of case you'd have to   subdivide a lot in zbrush to get enough topology  in here to make the ear shapes and you'd have to   fight the topology a bit to like tuck it and make  those convex forms so by not having the topology   like reinforce the shape be built custom for an  ear it's gonna be a lot more difficult to sculpt   and model and obviously you lose it completely  unless you have a very high subdivision level   so i think that's no bueno and then we can see  the mouth here has some problems first right   but but also we we don't have those concentric  rings to make like a kiss face or like a big smile   right here the edge flow just comes off this way  and it would really wouldn't support like you   know this kind of cheek shape you wouldn't  be able to get the puffy cheeks of a smile   making like a kissy face would be almost  impossible so this is an example of topology   that deformation would not be good it definitely  was not made with deformation in mind and it   looks like it was made probably on for high  poly models it was probably made so that you   could sculpt a static shape but i still think you  could make topology for this purpose a lot better   if the structure of the mesh supported the shape  it was trying to be more like in the ears case and   like in the noses case you know by not having  the nostril like reinforced with edge loops   then you have to sculpt everything and you're  going to be fighting against this grid like shape   rather than an actual like circular shape that's  reinforced with the topology itself all right so   there's a couple examples of topology i think are  not so good and while we're on the subject i want   to recommend a couple books if you want to get  more deep into topology for animation and of faces   there's the book stop staring by jason osipa which  i've recommended before i've read that book and   i like it a lot not only does it talk about good  techniques and principles for edge flow for facial   animation but also it talks about modeling a face  uh modeling a bass mesh for a face so i recommend   that book to any character artist starting out or  someone that's interested in making characters for   facial animation another book that you can get  which is a somewhat famous uh ebook is this one   the art of moving points by brian tyndall so in  this book it's a you know it's a digital book like   i say but um it's even got some videos let's see  if we'll see a video here oh here we go so here's   a topology laid out you can see these um sections  the concentric rings that we're talking about and   they describe the sections and what they do and  why you'd want to build them that way so great   reference this book is really good was recommended  to me by an animator and you can see uh because   it's a digital book there's even some videos which  i can't play right now but trust me there's videos   all right i want to take a moment to thank  this video sponsor skillshare skillshare is   an awesome online learning resource with tons  of videos on all different kinds of subjects so   if you're interested in learning stuff and you've  got an interest in anything that you want to know   more about skillshare is a great place to check it  out i watch all kinds of different videos that are   you know that are somewhat related to my interest  somewhat not i'm just curious and i like to uh   look into stuff one of those things is this logo  design with aaron draplin i've seen his other   videos on the site and even though i don't you  know think i'm a graphic designer i love hearing   about graphic design and his passion i also  love the way the videos are made but there's   a ton of different subjects on skillshare that  you could check out for yourself so really you   could type in anything that you want and and see  if there's something that you're interested in   to learn more about the first thousand people  to click the link down below in my description   get a one month free trial to skillshare so you  can check this out for yourself take a few classes   and learn something for free oh cinematic color  grading dope so that is as much about topology   as i could think to put in this  video for now so thank you for   watching up to this point i hope you learned  something something that you could apply to   your own models i'll have an art related video  coming out pretty soon so stay tuned for that but   thank you for watching and go out there and make  some cool models until next time peace out you

Channel: J Hill

Views: 38,529

Rating: 4.9779005 out of 5

Keywords: 3d modeling, topology, basemesh, animation, modeling, 3d tutorial, 3d character, character art, face topology, cg art, game character, game art, cinematic model, cinematic character, modeling tutorial, 3d modeling tutorial, how to model, maya character, zbrush character, wireframe, sculpting, how to make a character, quad draw, topogun, simpligon, real-time art, real time art, real time model, ue4, unreal engine 5, blender tutorial

Id: 6Kt0gW3_kio

Length: 40min 33sec (2433 seconds)

Published: Fri Aug 13 2021