Chiming in cause I'm not a butthurt party. @KarlP You should plan your model out so you don't end up with triangles or poles that negatively affect your smoothing. If you're done with your model and see a triangle but your sub-d looks perfect, then try and learn why it happened so you don't make that mistake again and move on, IMO. Sometimes using an 8sided cyl is better because it'll flow better with the rest of your model. Sometimes that magic number is 6, 7 or 13 depending on the surrounding GEO. Spacing of edges is only important if it gives shape, is placed on a curved surface or otherwise alters the result of the sub-d. If you have trouble following my train of thought I can do pics for you.
Irrelevant for everyone else: As for your attitude, and I say this with the most positive intention: You write like an intelligent person who has pride in that fact. Stop it, or you'll take twice as long learning even the simplest things. When perna gives it to you straight and without lube, take it like a man and move on. Getting hurt feels and stumbling on language is not conductive to progress and learning
1- Grow a thicker skin 2- Be more proactive and search the various pages of this thread starting from page 1 to get the answer you seek... 3- Dont act like a child when you dont understand basic concepts and blame it on the person that tried to help you , literally biting the hand that feeds you...
This right here. I still have a slight protection towards something I model or create in Substance Designer. I think everyone does, to a certain extent. I've learned that taking the advice of others to make my work easier and less messy has helped more than it hurts.
Many of us have known Per for a long time and while he can be blunt, I have never known him to be malevolent.
Lots of people give their time and advice freely on this forum, for which I might add, they charge a bunch of money for when they do professionally, and when they offer essentially the same piece of advice in the same thread maybe 100 times they are bound to get frustrated when people don't listen because it's not the advice they want to hear or because they think they know better. Per is no exception to this. I'm not sure people understand how incredibly frustrating it is to give sound advice over and over, only to have people ignore it because they don't agree, and then see that bad information propagate like wildfire across the internets by people who don't correctly understand the subject matter.
If you don't understand just read the damn thread over and over. This forum has the sum knowledge of arguably the best 3d artists in games over the past 18 years (almost) and this thread has over 140 pages of education that you would pay thousands for in a university. Reading and emulation was how people learnt before the internet. It's a technique that works.
The steps go something thusly:
Not understanding is fine.
Read the bloody thread until your eyes feel like peepy hamburger meat.
Try to find a solution yourself
If you still cant get the result you want or still don't understand move to step 5.
Ask questions and people will help.
People give their advice freely so listen to what they have to say.
TAKE THEIR ADVICE GRACEFULLY DAMMIT.
Don't argue with people who are more experienced and know better than you do.
Go back and try out the cool new techniques.
Be happy in the knowledge that you are a productive member of this fucking awesome community.
Keep practising until you are good enough to give advice to others.
You know what, everyone is right. I won't damage this thread by starting a flamewar.
And indeed, you're correct, I did enlist your advice many months ago, where you didn't fail to condescend to me then as well with the statement "chokra school of modeling."
But like I said, this thread deserves better than a flamewar.
1.) I never reposted because, in all honesty, I have a habit of doing that. That being said, 3d modeling is a hobby, and at the time my career got in the way.
2.) I still haven't found a solution to that part, but I'm biding my time working on other things so I can come back to it with a fresh mind. Nonetheless, here is my attempt
No fuck that Andy I'm Prince Tulip Pumpernickle and I deserve all the Peepy hamburgers I want! If I can't bevel you have to show me how but you can't say I'm wrong because that's wrong-shaming me. You just have to keep showing me over and over so when I finally do it right I can feel like it was my idea all along. THIS IS MY INTERNET!1
I agree with @metalliandy completely. This thread is, IMO, one of the best things on the internet, and it's value nearly incalculable. Hence why I'm not going to continue an argument.
Dunno, just thought those were to generally be avoided with Sub-D.
Simply not true, they're situational but are sometimes the better answer to solving a shape in sub-d.
Ok, so this whole thread is designed to teach people. If there's one single thing that perna and the others try to impress on people is to think and experiment for themselves. It's one thing to be taught that X is right or wrong for Y reasons, and another to hear that someone said X so it's always true.
Dunno, just thought those were to generally be avoided with Sub-D.
Simply not true, they're situational but are sometimes the better answer to solving a shape in sub-d.
Ok, so this whole thread is designed to teach people. If there's one single thing that perna and the others try to impress on people is to think and experiment for themselves. It's one thing to be taught that X is right or wrong for Y reasons, and another to hear that someone said X so it's always true.
Well I absolutely do try to experiment as much as possible. But I nonetheless thought quads were the most preferable polygon for subdivision, paramount to triangles.
So when are triangles okay? Do we still want to favor quads as a rule? What is the occasional triangle vs being sloppy?
Well I absolutely do try to experiment as much as possible. But I nonetheless thought quads were the most preferable polygon for subdivision, paramount to triangles.
So when are triangles okay? Do we still want to favor quads as a rule? What is the occasional triangle vs being sloppy?
I've been trying to remember the thread that had good visual examples and can't come up with anything, sorry. The "classic" example was capping the inside face the cylinder from a revolver, IIRC, but that can be done with quads just as easily. Personally, I terminate loops into triangles on flat surfaces often, or sometimes on curves if it isn't going to hurt the shape much and I don't have the time to go quad only simply for the sake of.
Well I absolutely do try to experiment as much as possible. But I nonetheless thought quads were the most preferable polygon for subdivision, paramount to triangles.
So when are triangles okay? Do we still want to favor quads as a rule? What is the occasional triangle vs being sloppy?
I've been trying to remember the thread that had good visual examples and can't come up with anything, sorry. The "classic" example was capping the inside face the cylinder from a revolver, IIRC, but that can be done with quads just as easily. Personally, I terminate loops into triangles on flat surfaces often, or sometimes on curves if it isn't going to hurt the shape much and I don't have the time to go quad only simply for the sake of.
You should post some examples, if you get a chance.
If you don't want to have collapsing chamfers on inset geometry then the right way to go about it is to do it the other way around. Do all the insets first from a rectangle shape, then chamfer. This keeps everything evenly sized. Order of operations is a big deal when it comes to chamfering.
Bonus tip: If you want variable width chamfers then just do insets first, break off the geo where you want different chamfer width second, chamfer, then reconnect.
If anyone wants examples let me know. Pls @ me so I'll see it,
Here is my try....! Extrude and then scaled by figure from x and y axis.
After extrude.... Is there way to get equal space (X & Y) by scale rectangle shape by dragging....?
Thanks Guys.
This problem occurs when your base geometries transforms are not frozen... Just freeze your transform before you apply the bevel and everything will be fine.
Hi, not gonna lie, i know this its pretty basic, but its driving me crazy.. (maybe i'm just obsessive) When i'm doing hard surface, in certain corners i get some pinchy in my shading due to the support loops. I dont know if i'm putting them wrong or if it is for the shading to look that way, since i cant really see the pinchy effect unless i use certain matcaps.. The "solution" i finded it's to add another edge loop, or to slide the rest loops only supporting the initial part.. (the last image its another model)
@DeathstrokeFTW I see nothing wrong with that topo; why you hate that? @Evidenz If the surface is flat (like your first example) , I see no reason to not relax the corner more, like this:
@DeathstrokeFTW I see nothing wrong with that topo; why you hate that?
I know its fine but I just think there could be a better way to make it look efficient.
if you obcess over small things like that , its the contrary to efficiency since it will literally make you waste your time trying to get something ( that will have no better visual impact ) to be perfect. Its good to search for clean ways to do stuff mind you, but if it looks good and clean, move onto next piece
It could be said that some of these loops are unnecessary as they don't contribute to the shapes: But it could also be that they're necessary for vertex painting and stuff like that. In the end I wouldn't bother to invest the time removing them unless needed.
This page is hilarious. Tris are bad, mockery, quad to quad is origin of whatever the hell... I'm balling over here. This thread is great but sometimes I think people just refuse to look, learn and listen to what people say.
For one, the whole triangles are bad thing is simply a newbies misinformation concept that is thrown way out of proportion. Let me put it like this, it's useful to plan out your geometry but it's almost inevitable that you'll eventually model a detail that was unaccounted for and requires more or less geometry to support its form. At this point most novices stop modelling, rip their hair out and then start over thinking that they've done something wrong. Instead you either want to find a way to re-route your geometry, simplify it or try and terminate geometry into a triangle dump somewhere on a flat or hidden area...
It could be said that some of these loops are unnecessary as they don't contribute to the shapes: But it could also be that they're necessary for vertex painting and stuff like that. In the end I wouldn't bother to invest the time removing them unless needed.
They are necessary :P I still am not finished with the wall and the floor, most of them help me place a few small details and Normal map decals(making a trimsheet after im done here)
@DeathstrokeFTW I see nothing wrong with that topo; why you hate that? @Evidenz If the surface is flat (like your first example) , I see no reason to not relax the corner more, like this:
thats the second method i usually use.. i just want to learn a correct way, since every tutorial i've seen on hard surfaces, they hardly do the alternatives i listed, to the contrary they just put the edges really close and since its a video i really dont know if there its pinchy or not.. god i sound like an obsessed moron and i know those pinchy shading artifacts will be barely visible in a bake. but since im trying to learn good topology
So when are triangles okay? Do we still want to favor quads as a rule? What is the occasional triangle vs being sloppy?
There are no rules.
As long as the object looks correct when it's subdivided then move on.
Sometimes people get told to avoid triangles when they're new to SubD modelling because triangles don't subdivide as neatly as Quads. Additionally, when you're just starting out you're unlikely to be trying to model some crazy shape that actually requires triangles as part of it.
No one really cares about sloppy-looking topology either. The only time people will care if the wireframe looks pretty is when its part of the end result (i.e. if you're doing a wireframe render.) Hell, some of the topology I use when high poly modelling is hilariously disgusting to look at; but that's fine because it works, and it was quick/easy to do.
As long as the model actually looks correct and is optimized appropriate use whatever geometry you want.
@Dan Powell - Right on all points except the messy topology - If you're ever passing off meshes to other team members or revisiting your mesh at a later date you want good topology.
I'd hope this would go without saying honestly. If you giving anything to anyone then it needs to be in a state where they can work on it. Obviously don't purposely make your topology a mess, but don't waste time making it beautiful either unless it's going to be seen.
The reason you do a wireframe render is to show your good topology skills haha.
I see "Good topology skills" as making something that looks correct when it's rendered. Doesn't have to be aesthetically-pleasing topology, it just has to work - and if possible try not to make it an absolute mess (but don't get caught up in making it perfect as it'll slow you down.)
I see pretty-looking topology is a waste of time unless you're doing a wireframe render for something like a product advert (where wireframe renders are sexy), or unless you're just trying to add breakdowns for a portfolio piece.
Well, about just going with whatever topology as long as it works is kinda of a tradeoff imo. If you build your topology correct it might be easier for you to just retopologize stuff and to easily make changes.
And yeah, by correct I'm not saying only quads, this is not really correct.
Good topology is not what you truly strive for, but something that comes from having a good workflow. Having a non-destructive workflow is what you ultimately want to achieve as much as possible. It is just that messy topo generally comes from a lot of manual editing when compared to using things like working in the modifier stack.
i just want stuff to be easily editable on my end when artists send me their HP. If it has messy topology it makes stuff harder to edit, but im not requiring full quad models, heck i very usually have ngons roaming free on my models and they smooth nicely.
Can I get some help with this shape? I've tried cutting into the mesh, extruding and flattening (left) and then using half a cylinder as a boolean (right), but it still doesn't look right and I'm sure there has to be a better way.
Can I get some help with this shape? I've tried cutting into the mesh, extruding and flattening (left) and then using half a cylinder as a boolean (right), but it still doesn't look right and I'm sure there has to be a better way.
I am almost certain that this particular case was talked about like 50 times in this thread
Replies
@KarlP You should plan your model out so you don't end up with triangles or poles that negatively affect your smoothing. If you're done with your model and see a triangle but your sub-d looks perfect, then try and learn why it happened so you don't make that mistake again and move on, IMO.
Sometimes using an 8sided cyl is better because it'll flow better with the rest of your model. Sometimes that magic number is 6, 7 or 13 depending on the surrounding GEO.
Spacing of edges is only important if it gives shape, is placed on a curved surface or otherwise alters the result of the sub-d.
If you have trouble following my train of thought I can do pics for you.
Irrelevant for everyone else:
As for your attitude, and I say this with the most positive intention: You write like an intelligent person who has pride in that fact. Stop it, or you'll take twice as long learning even the simplest things. When perna gives it to you straight and without lube, take it like a man and move on. Getting hurt feels and stumbling on language is not conductive to progress and learning
(I'm probably spewing garbage)
Lots of people give their time and advice freely on this forum, for which I might add, they charge a bunch of money for when they do professionally, and when they offer essentially the same piece of advice in the same thread maybe 100 times they are bound to get frustrated when people don't listen because it's not the advice they want to hear or because they think they know better. Per is no exception to this.
I'm not sure people understand how incredibly frustrating it is to give sound advice over and over, only to have people ignore it because they don't agree, and then see that bad information propagate like wildfire across the internets by people who don't correctly understand the subject matter.
If you don't understand just read the damn thread over and over. This forum has the sum knowledge of arguably the best 3d artists in games over the past 18 years (almost) and this thread has over 140 pages of education that you would pay thousands for in a university. Reading and emulation was how people learnt before the internet. It's a technique that works.
The steps go something thusly:
And indeed, you're correct, I did enlist your advice many months ago, where you didn't fail to condescend to me then as well with the statement "chokra school of modeling."
But like I said, this thread deserves better than a flamewar.
1.) I never reposted because, in all honesty, I have a habit of doing that. That being said, 3d modeling is a hobby, and at the time my career got in the way.
2.) I still haven't found a solution to that part, but I'm biding my time working on other things so I can come back to it with a fresh mind. Nonetheless, here is my attempt
Here's the end result:
Ok, so this whole thread is designed to teach people. If there's one single thing that perna and the others try to impress on people is to think and experiment for themselves. It's one thing to be taught that X is right or wrong for Y reasons, and another to hear that someone said X so it's always true.
So when are triangles okay? Do we still want to favor quads as a rule? What is the occasional triangle vs being sloppy?
How the edges are parallel going, when I am trying extrude and inner extrude.... at end its overlapped....?
Is there any other way to do...?
Thanks.
there are good reasons when and why to avoid triangles
but quadonly just for the sake of quadonly is a shit dogma
Extrude and then scaled by figure from x and y axis.
After extrude.... Is there way to get equal space (X & Y) by scale rectangle shape by dragging....?
Thanks Guys.
Thanks.
When i'm doing hard surface, in certain corners i get some pinchy in my shading due to the support loops.
I dont know if i'm putting them wrong or if it is for the shading to look that way, since i cant really see the pinchy effect unless i use certain matcaps..
The "solution" i finded it's to add another edge loop, or to slide the rest loops only supporting the initial part..
(the last image its another model)
I see nothing wrong with that topo; why you hate that?
@Evidenz
If the surface is flat (like your first example) , I see no reason to not relax the corner more, like this:
It could be said that some of these loops are unnecessary as they don't contribute to the shapes:
But it could also be that they're necessary for vertex painting and stuff like that. In the end I wouldn't bother to invest the time removing them unless needed.
For one, the whole triangles are bad thing is simply a newbies misinformation concept that is thrown way out of proportion. Let me put it like this, it's useful to plan out your geometry but it's almost inevitable that you'll eventually model a detail that was unaccounted for and requires more or less geometry to support its form. At this point most novices stop modelling, rip their hair out and then start over thinking that they've done something wrong. Instead you either want to find a way to re-route your geometry, simplify it or try and terminate geometry into a triangle dump somewhere on a flat or hidden area...
god i sound like an obsessed moron and i know those pinchy shading artifacts will be barely visible in a bake. but since im trying to learn good topology
There are no rules.
As long as the object looks correct when it's subdivided then move on.
Sometimes people get told to avoid triangles when they're new to SubD modelling because triangles don't subdivide as neatly as Quads. Additionally, when you're just starting out you're unlikely to be trying to model some crazy shape that actually requires triangles as part of it.
No one really cares about sloppy-looking topology either. The only time people will care if the wireframe looks pretty is when its part of the end result (i.e. if you're doing a wireframe render.) Hell, some of the topology I use when high poly modelling is hilariously disgusting to look at; but that's fine because it works, and it was quick/easy to do.
As long as the model actually looks correct and is optimized appropriate use whatever geometry you want.
I'd hope this would go without saying honestly. If you giving anything to anyone then it needs to be in a state where they can work on it. Obviously don't purposely make your topology a mess, but don't waste time making it beautiful either unless it's going to be seen.
I see "Good topology skills" as making something that looks correct when it's rendered. Doesn't have to be aesthetically-pleasing topology, it just has to work - and if possible try not to make it an absolute mess (but don't get caught up in making it perfect as it'll slow you down.)
I see pretty-looking topology is a waste of time unless you're doing a wireframe render for something like a product advert (where wireframe renders are sexy), or unless you're just trying to add breakdowns for a portfolio piece.
And yeah, by correct I'm not saying only quads, this is not really correct.
Can I get some help with this shape? I've tried cutting into the mesh, extruding and flattening (left) and then using half a cylinder as a boolean (right), but it still doesn't look right and I'm sure there has to be a better way.