Jump to content

Model Topology - Common Solutions


Recommended Posts

I have been interested lately in modelling topology solutions.  The solutions are pretty software agnostic as good topology is pretty universal regardless of the software package being used.  Also given that there are many ways to solve a particular situation, I have been interested in as many solutions as possible and have wondered if others have the same interests.

 

 I found this site which I think is just an awesome resource.

 

As I look at this site, it does make me wonder if Topology should be its own sub-forum in the modeling section.  You present a problem involving a particularly challenging basic shape and you just post a solution that shows great topology (good  edge flow, no complex poles, evenly spaced quads, etc).  As Core4D moves to embrace other software packages there is no need to present a tutorial, rather just the rough initial shape and then the solution --- or multiple levels of solutions:  one level for hard surface models that won't be sub-divided but will be UV'd, another for hard surface models that will be UV'd and sub-divided and a third for character models that will be UV'd, sub-divided and deformed.

 

Most of us know the modeling commands but what most of us lack is the thought process behind the topology solution.  That information only comes from the masters of modeling who probably have grown their techniques over years of refinement. 

 

My hope is that over time, there is always access to that level of good source of reference information  - whether it be from the "Topology Guides' site or within Core4D itself.  

 

Just a thought.

 

Dave

 

 

Sorry...but I simply do not have enough faith to be an atheist.

Link to comment

For quite some time I've been thinking about some kind of Wiki page about general stuff regarding 3D...like you mentioned, topology for example and so on. There is plenty to cover and I need to figure out how would that page look like. 

Link to comment

I'm sure I have read before about a topology reference paper or 2, and a couple of books over the years, but the names of any of them escape me right now of course, and none of them were free ! 🙂  That one you have found looks excellent, and there were examples in that I hadn't seen so clearly explained before, and it IS free, so it certainly gets my recommends !

 

But I tend not to think topology should have its own forum section on the Core, as the 2 categories we have, HS and Sub-div almost always feature topology advice in the answers, some of which is strategizing and solving specific topo issues and so that seems an adequate place for any specific topo discussion to live. Also I rather doubt whether there are enough people interested in advanced topology theory to justify its own section ! We're a dying breed, I'm told, and we (the Core) are usually being asked for less categories not more !! 🙂 

 

But I can quite see how it might work in a Wiki sort of format (good thoughts, Igor!), and would happily contribute in my own small way to such a thing as time permits...

 

CBR

Link to comment
14 hours ago, Cerbera said:

Also I rather doubt whether there are enough people interested in advanced topology theory to justify its own section ! We're a dying breed, I'm told, and we (the Core) are usually being asked for less categories not more !! 🙂 

Cerbera,

 

Good points....as usual.  But I do find your comment rather startling given how good topology just makes everything easier (UV, deformation, and can actually speed up the rendering process).   Bad topology always shows itself in the final render (phong errors, unplanned bumps or pinching, etc) so if anyone wants to improve their art, they cannot get away from learning good topology.  

 

Now,  I will admit it takes work, thought and a lot of practice...so maybe people are lazy but there is no "make great art" button so ultimately you need to do the work.  For me it is not the extra effort in modeling that is the barrier (the tools are designed to support topology creation after all) but rather the knowledge of which topological solution works best OVERALL to the entire model.  I emphasize "overall", because I do find situations where I think I have a good topological solution to a certain section of a model but then as I continue modeling realize that I need to go back and do certain sections all over again.  Experience (and a bit of planning) is the only way to avoid that and that is why I am doing web-searches for sites and techniques on topology.  

 

So getting back to your statement that good topologists are a dying breed, I would "hope" that you are wrong.  3D modeling is a skill that people can earn a living with so saying that those who do it well and with a high level of professional precision have been dwindling in number over time is like saying "Good Doctors are a dying breed" or "Good teachers are a dying breed".  

 

Honestly, I think you vastly underestimate the value you bring to the 3D community with your exceptional modeling skills.  

 

Dave

Sorry...but I simply do not have enough faith to be an atheist.

Link to comment

Times are changing Dave, things need to be done faster rather than slower. Renderers these days can cope with bad topology, even with bad UVs, so having perfect topology is really not necessary these days, depending on what you do of course. For animation topology will always matter I guess but for hard surface stuff that doesn't have to bend its really not necessary. Tri-Planar and Curvature maps can solve almost all your problems with bad UVs. I've seen a looot of bad topologies that works out just fine in the projects and no one really cares when they do.

Link to comment
2 hours ago, 3D-Pangel said:

So getting back to your statement that good topologists are a dying breed, I would "hope" that you are wrong.  3D modeling is a skill that people can earn a living with so saying that those who do it well and with a high level of professional precision have been dwindling in number over time is like saying "Good Doctors are a dying breed" or "Good teachers are a dying breed".  

 

Do you remember thatched rooves ? There used to be Haymakers and Thatchers everywhere, and a lot of those were very highly skilled. And then roof tiles (booles, re-meshers and the volume builder) came along, and that was pretty much the end of that ! 🙂 There are still Thatchers, but now only about 3 of them, doing all the highly specialised thatching that remains, which isn't much !

 

I fear the same may happen to topologists, and have seen signs that this may be the case for some years now... the inherent laziness of humans, and our overall unchecked vaulting ambition (simultaneously our best and worst attribute as a species !) to be doing everything faster (but not necessarily better) may make our decline inevitable... 

 

This ultimately may mean I have to make money elsewhere, but with the noble goal of trying to remain in work you actually like (!) I will doggedly hold on as long as I feasibly can, and try and be one of the good ones that can remain ! And if that isn't a possibility then I will continue to enjoy topological problems in my own time, and there will presumably be a convention every year where the worlds only 5 remaining topo guys meet in a tiny local village hall and reminisce about the good old days ! There would probably be themed beermats🙂

 

2 hours ago, 3D-Pangel said:

Bad topology always shows itself in the final render (phong errors, unplanned bumps or pinching, etc) so if anyone wants to improve their art, they cannot get away from learning good topology.  

 

That's not quite right.  It doesn't always show itself in render by quite a long shot, especially if SDS is not involved ! Your topology can literally be anything you like on flat surfaces - all tris, hundreds of complex poles, or even 1 big ngon (from which it would be easy to generate a cohesive perimeter loop for that face / section, which in a lot of cases will make it render flawlessly.  However a big disadvantage remains in that some of the modelling and selection tools will work much more unpredictably (or worst case not at all!), and unhelpfully for the rest of the build because they are all built to work with sensible poly flow ! 

 

And as Igor points out, curvature maps and tri-planar mapping can solve a great deal of texturing problems in the real world - I would be the first person to admit there is no point in UVing a rock OR particularly caring what its topology is doing !! If one of those is made of 100% triangles that can actually be better (in terms of even displacement and polygon efficiency) and there are cases where topology genuinely doesn't matter at all (3D printing jumps to mind), or matters considerably less. I also predict that as remeshing tools become better and better the areas in which hand-made topology is crucial become ever smaller and more specialised.

 

So, you know - I've made my bed, and I want to keep lying in it, but I do kind of accept it might be on fire ! And because I spent 20 years building this particular bed, I'll put up with a little smoke, and keep the fire extinguisher of diversity nearby ! 🙂

 

With the hope of ending this post on a more positive point, there will always be 'artisans' - people who do things nicely and properly for the sake of their own peace of mind, people who WANT to take the time to know ALL the things and take a great deal of joy in the process. Some people still write musical scores by hand on paper ! Some people still make beer 'ye olde waye'  and some people still thatch rooves, and there are enough humans now that every specialisation, no matter how introspective or 'niche' can have a place in the world, and find other people who share that interest.  And of course it has never been easier to connect with like-minded folk !

 

CBR

 

Link to comment

I too fear topological space wizardy and professional pointpushers are a scarcity these days, kinda like a unicorn, almost mythical creatures that no one is sure even exist these days;) but in seriousness it does seem like this art form is becoming less common. Saying that it's almost a double edged sword. I've been doing some modelling work for a German studio lately,  hence my untimely absence of late, and they  were almost taken back by how neat and clean my meshes were in comparison to some of the crap they were used to having to work with  which as freelancers can only go in our favour.

Link to comment
On 6/28/2021 at 1:45 AM, VECTOR said:

I too fear topological space wizardy and professional pointpushers are a scarcity these days, kinda like a unicorn, almost mythical creatures that no one is sure even exist these days;) but in seriousness it does seem like this art form is becoming less common. Saying that it's almost a double edged sword. I've been doing some modelling work for a German studio lately,  hence my untimely absence of late, and they  were almost taken back by how neat and clean my meshes were in comparison to some of the crap they were used to having to work with  which as freelancers can only go in our favour.

Now, to provide a more positive outlook, were the people at the German studio upset that your models had such a clean topology?  Were there complaints that you took too long or put in effort that was not needed?  Honestly, I think there is a natural attraction to quality provided you know what quality looks like.  What I fear is that with the rapid proliferation of CG everywhere, the average person does NOT know what a quality model looks like.  So it is not because people no longer care about perfect topology, they just don't know what it is or why it makes a difference.  


So the argument should be "does it make a difference" rather than topological space wizards being a dying breed.  If it does make a difference, you will therefore always stand out amongst the crowd.  Your work will continue to be recognized even tangentially as Vector found out from the German studio. 

 

"Yeah, the models we get from Vector or Cerbera just always hold up in all situations -- they are easier to texture (eg. tri-planar projections just come out better because the quads are just so evenly spaced on all sides), they deform well, and we never get any rendering errors.  Can't say that with models we get from Joe Moron who graduated from the Triangle Academy."

 

Honestly, if modeling quality wasn't important why would there be Checkmate certified models -- which are usually sold at a premium.  I would imagine that true "top of the pyramid" professionals at ILM, Weta, etc...etc..etc...know and demand perfect topology because their multi-million dollar projects cannot afford any mishaps from shoddy models.

 

So, to sum up, you are NOT a dying breed.  You just have uneducated non-professional customers.   You can either educated them yourselves (which can be painful) or they will be educated by the problems they encounter by accepting sub-standard work (which is even MORE painful).    

 

So you know modelling...what you need to learn now is marketing. 

 

Effective marketing is a huge piece of anyone success.  Fortunately, successful marketing is easier when you have a great product or service to sell and your models are a great product and your modeling capabilities are a great service.

 

Dave

 

 

Sorry...but I simply do not have enough faith to be an atheist.

Link to comment
9 minutes ago, 3D-Pangel said:

Now, to provide a more positive outlook, were the people at the German studio upset that your models had such a clean topology?  Were there complaints that you took too long or put in effort that was not needed?  Honestly, I think there is a natural attraction to quality provided you know what quality looks like.  What I fear is that with the rapid proliferation of CG everywhere, the average person does NOT know what a quality model looks like.  So it is not because people no longer care about perfect topology, they just don't know what it is or why it makes a difference.  


So the argument should be "does it make a difference" rather than topological space wizards being a dying breed.  If it does make a difference, you will therefore always stand out amongst the crowd.  Your work will continue to be recognized even tangentially as Vector found out from the German studio. 

 

"Yeah, the models we get from Vector or Cerbera just always hold up in all situations -- they are easier to texture (eg. tri-planar projections just come out better because the quads are just so evenly spaced on all sides), they deform well, and we never get any rendering errors.  Can't say that with models we get from Joe Moron who graduated from the Triangle Academy."

 

Honestly, if modeling quality wasn't important why would there be Checkmate certified models -- which are usually sold at a premium.  I would imagine that true "top of the pyramid" professionals at ILM, Weta, etc...etc..etc...know and demand perfect topology because their multi-million dollar projects cannot afford any mishaps from shoddy models.

 

So, to sum up, you are NOT a dying breed.  You just have uneducated non-professional customers.   You can either educated them yourselves (which can be painful) or they will be educated by the problems they encounter by accepting sub-standard work (which is even MORE painful).    

 

So you know modelling...what you need to learn now is marketing. 

 

Effective marketing is a huge piece of anyone success.  Fortunately, successful marketing is easier when you have a great product or service to sell and your models are a great product and your modeling capabilities are a great service.

 

Dave

 

 

I think alot of it comes down to time and money vs quality.  Alot of customers want it quick and cheap and usually that doesn't correlate with a  quality piece of work. Some also just don't care about the topology aslong as the final render is okay, this is a different case with animation ofcourse. I've seen some of the models on turbo squid that are checkmate certified and am honestly stunned how they managed to achieve it because the quality of some have been aweful. This also goes for some of the work I've seen from some of these artist working at big studios, the models have been amazing but the topo full of triangles, ngons and generally just not particularly clean or neat. Modelling quality definitely matters, half the issue these days is people don't want to spend the time it takes to achieve said skills to what you would consider an "Elite" Level, so although good topologist may not be technically "dying out" the quality definitely seems to be slipping, and people as anal about clean meshes as @Cerbera and I are most definitely a scarcity and not the norm, which is why customers react the way they do when receiving the models.

Link to comment
30 minutes ago, VECTOR said:

I think alot of it comes down to time and money vs quality.  Alot of customers want it quick and cheap and usually that doesn't correlate with a  quality piece of work. Some also just don't care about the topology aslong as the final render is okay,

That's why I never worked per hour but per model when I was freelancing. That way I was able to deliver good quality, triangle free models and I never had issue with that. Working per hour, exactly due to that reason always felt wrong and it simply sucks as there is always pressure of making sure you don't overdo it when not necessary. On the other hand, putting so much effort for something that in the end doesn't actually matter, makes no sense, simply because then you are wasting your time, simply because  no one will care about your all quads nice topology models, especially when you know in the final shot that's hidden anyway, so in the end, you lose, not really the customer, especially if he doesn't ask you to have topo which is all quad. 

 

On the other hand, triangle topology has it place and its use so its not really fair to say its wrong. In some cases optimization is much more important then topology, especially when you know that in the render topology really doesn't matter at all, not even 1%. So people these days are thought to be more efficient while preserving good end quality. Imagine making rocks being all quads just because...it just makes no sense. Bending objects are different story so I would say at this point, every technique has its place and none is wrong or right. What matter is that it works, looks good and its optimized for the task.

Link to comment
1 hour ago, Igor said:

Imagine making rocks being all quads just because...it just makes no sense. Bending objects are different story so I would say at this point, every technique has its place and none is wrong or right. What matter is that it works, looks good and its optimized for the task.

Everything comes down to how the model is to be used.  If it is a background rock, I 100% agree.  But if it is going to be used in an extreme close-up with height maps applied, or even it is to be made out of marble and you want some SSS applied, then a fully triangulated model with complex poles and n-gons may give you some trouble. 

 

So the marketing of models made to the highest possible standards should be "The most trouble free models no matter how they are used or rendered that money can buy".  Again, you may not want that guarantee in all situations because it will make no sense, but there will always be cases were people are willing to pay for it. 

 

Dave

Sorry...but I simply do not have enough faith to be an atheist.

Link to comment
×
×
  • Create New...

Copyright Core 4D © 2023 Powered by Invision Community