its hard to understand what you intend. First i would recomend showing results with a material without the wireframe and high specular value to see modeling errors , then post a concept on what you are trying to achieve , because i dont know what you are asking exactly
@KarlP Your questions regarding edge flowing are a bit generic. The method of modeling something is subjective, you could start from a plane, or block out the entire shape and start detailing later. The first thing I notice about your model is that it is one piece; there is no reason to complicate things like that (unless you plan to print the model, but even in that case, I don't think it's necessary to have one single closed mesh). You should post some specifinc shape you are having trouble with, otherwise I don't think you will get much help here.
Is this the best way to approach a square inset with hard corners on a cylinder? Or are there any other methods/techniques that produce the same result? I just can't find the other comments with image/.gif solutions regarding about square insets on curved surfaces. Thanks in advance.
Is this the best way to approach a square inset with hard corners on a cylinder? Or are there any other methods/techniques that produce the same result? I just can't find the other comments with image/.gif solutions regarding about square insets on curved surfaces. Thanks in advance.
first page of this thread has the answer you seek , by Shepeiro.
Can you elaborate on why you view the single block solution I have as over complicated? Is it simply the challenge of meshing together fundamentally different shapes so why do that?
You've constructed the mesh much as you construct sentences: In a manner convoluted and obscure.
Ok, I have refocused and would still like your help,
I appreciate that this question has been asked a lot but the answers dont seem to go very far for me.
So this is my take on the "how do I harden this edge" question. The block out, trying to respect topology flow using a control loop before extruding
Additional support loops added with swift loop tool
Changes I made to the support loop flow beyond the edge I wanted hardened
Turbosmooth 1 itr
Turbosmooth 1 itr, high spec material
The questions 1. Have I finally understood "supporting the edge" and "use more geo"? 2. I used swift loop and ring connect but I would like better control of how tight my support loops are, any pointers on how I could better do that? 3. Any further suggestions?
You really shouldn't snip your posts in a thread about learning, @KarlP.
Yes, understood, but the comments were right so I reposted with a better focus, there was no value in the previous posts.
That's your assumption. Maybe the next person is in the same boat and by understanding your methodology they're able to also "better focus" by following your train of thought. Generally bad to snip as said, I mean half this page is just full-stops now ¯\_(ツ)_/¯
yeah, it would be a good way to have newbies etc see "what not to do" , now its only fullstops with no context ( wich make the replies also have no context , wich invalidates the time that people took to reply to you ).
Thats like those people that make dumb comments and then delete them because they are too afraid to "look bad" , when everyone knows they are jerks already :P
@KarlP To control how tight your support loops are, you simply move them, it's as simple as that. In your exercise, it's useless to push back and forth the loops to "conform with the flow", since they reside on a totally flat surface and don't contribute to the shape; and your mesh doesn't have to be all quads
Thanks @Joao Sapiro, Yeah I understand and actually that was part of the reason for snipping, it seemed like I was off to a very bad start and I didnt want any new potential commenters to write me off straight away, I guess it didnt achieve the desired affect. I do appreciate the time everyone takes to post here and you are right "Live and learn" :-)
@Ausonian Ok, I probably didnt use the best example as the surfaces are flat (also touched on by Perna), but if it were a curve I am on the right track?. Yeah I still have the "quads or die" mentality, though I have seen some really good tutorials about pentagons in forming a control loop. Thanks for the input Ausonian, I appreciate the guidance
1. I have no idea what "supporting the edge" means in this context or how you would qualify for understanding that. You'll have to be less vague. You have not understood "use more geo", as that refers to modeling curved surfaces, which are completely absent in your model.
2. Tools and results are two entirely different things. What does "better control" mean?
3. You write lengthy posts with lots of images yet they end up frustratingly vague and mystical, discussing high-flying abstract concepts when you ought instead adopt a more pragmatic approach. It seems you lack a concrete grasp of the concepts you touch on, such as "model flow", a term (among several) you seem to throw out more or less randomly, or with disproportionate priority. Your mesh construction seems to be following a bunch of hippie new-age concepts that you don't actually understand but somehow still feel compelled to adhere to.
There's nothing mystical or artistic to what you're trying to learn. It's a technical craft with absolute rules. You'll greatly benefit if you stop overthinking the mystical aspects. Currently both your posts and your meshes seem to have been constructed under the influence of cocaine. Slow down and start with the basics.
Hi @Perna, I am a learner on here, I appreciate that you have 20+ years of experience in the industry but I do not. "The concepts" to me are only now becoming concrete, there has been lot of hard work from me to get this knowledge, everyone seems to have their own unique way of doing things and explaining them, sometimes these methods even appear, as a newbie, to be conflicting.
In response to point 1. Some definitions I have learned from across the sources I have used a. support edge - the edges added to control the hardness of the edge loop of polygons b. control loop - a loop of polygons that defines the way fill loops will run around the model. These loops typically have e-poles (5 edged verts) and n-poles (3 edged verts) which change the "flow of the topology" c. fill loops/fill edges - loops of polygons that add geometry for padding (for example between two control loops, I understand thats good practice) or to provide more verts/edges for further detail Please feel free to correct my understanding
In response to point 2. What I wanted better control on was the tightness of the support loops to the actual edge of the geometry. I could use swift loop tool but then I am placing by hand which is not consistent. I could use ring and then connect but again 90% on the tool slider for one loop of polys is not the same as 90% on the tool slider on another. The question was - How do I get better control over these tools so that I can consistently set my support edges to obtain uniform tightness, if thats what I want?
In response to point 3. Ok point taken, from here on, I will reduce my word count and wont post more than one image per post. It is comments like this that make learning this craft so difficult, "I need a more pragmatic approach", that comment along with most of what you actually say means 3/5th of nothing to me. Are you here responding to my posts to help me or to mock me? It seems the latter, particularly your penultimate sentence, I am not on cocaine and to put it bluntly, that is offensive and insulting. If you dont want to help or comment on the modeling techniques then dont. If I am as difficult to have here as you are making out then have me banned. I would like your help, I respect your experience and position in the community, but I dont need your help.
In response to the final paragraph The first piece of information that has actually been useful to me. I will slow it down and I think you are right possibly I need to be more technical about it all, which suites me, I am more of a numbers person. Thanks for the help in that regard.
I'm trying to model a belt but I've got no idea how to get the holes in and keeping my shading good. Any help would be appreciated.
@Sunray, have you seem this video https://www.youtube.com/watch?v=NKtncTvmQJc or similar? It has helped me. Also one thing that seems to be key is the number of sides to the hole, looks like you have 8 and that seems to be a good start. Finally, this connectivity should help, it spreads the edges evenly around the hole so that when you subdiv it averages their positions and produces a smoother result.
Hello! I wanted to make a star citizen environment for my portfolio but I came across a problem while I was modelling the door. Now most of the doors have curved edges and they are easy to make but when I was modelling the door whenever I used the inset, the curved edges began to collapse with each other but in the references, it shows no signs of them overlapping
I'm trying to model a belt but I've got no idea how to get the holes in and keeping my shading good. Any help would be appreciated.
@Sunray, have you seem this video https://www.youtube.com/watch?v=NKtncTvmQJc or similar? It has helped me. Also one thing that seems to be key is the number of sides to the hole, looks like you have 8 and that seems to be a good start. Finally, this connectivity should help, it spreads the edges evenly around the hole so that when you subdiv it averages their positions and produces a smoother result.
Thanks man I will give it a try when I continue working on the belt.
@DeathstrokeFTW - In the reference that's not all one piece of continuous geometry - geo is stuck into other geo. Inset might not be the only op you want to use - If I were doing this I'd use inset in some cases but also make my smallest and largest part then bridge the geo in between. That would give me more control over the end result.
@Sunray - If you look at the example from @perna you can see there's only a 4 sided hole, that's really all you need for this. You also need to contain that hole with supportive loops and keep the period of loops the same throughout the belt so it can bend uniformly. Making a flat hole (of any amount of sides) does not fully solve the problem you are having.
D'oh I should have thought about making multiple meshes and bridging them. Thanks for the nudge! But if anyone else has a more efficient way of handling that then please let me know xD
@TeriyakiStyle, thanks for the input. Particpating here is part of trying to make those decisions on what to believe. In regards to my answer to sunray, I will be more accurate in the future with any questions or explanations. I do know how subd works mathematically and why 8 works best.
Thanks again for taking the time to point me in a better direction.
@TeriyakiStyle The rule is number or faces at the point of placement +4 - reference: this forum, wirrex sept 2015. Reason: the goal is to produce quads using this formula/method you can support the hole geometry with a ring of quads.
Subdivision: a vertex of a quad face is split to create a new quad face. The centre of the face is the position of the original vertex. The newly created vertices are position around this centre via a weighted average with the surrounding vertices of the orignal quads in the unsubdivided mesh. Why do we want quad? Because quads will create quads when subdivided. Also quads deform well in animations due to the two degrees of freedom.
Feel free to correct me, apologies for the language used, it is wordy.
@TeriyakiStyle The rule is number or faces at the point of placement +4 - reference: this forum, wirrex sept 2015. Reason: the goal is to produce quads using this formula/method you can support the hole geometry with a ring of quads.
Subdivision: a vertex of a quad face is split to create a new quad face. The centre of the face is the position of the original vertex. The newly created vertices are position around this centre via a weighted average with the surrounding vertices of the orignal quads in the unsubdivided mesh. Why do we want quad? Because quads will create quads when subdivided. Also quads deform well in animations due to the two degrees of freedom.
Feel free to correct me, apologies for the language used, it is wordy.
Tris will also create quads when subdivided...the rule "quad only" when its spewed out without even trying to understand it will only do harm instead of helping. I think that what you mean by the 8 sided thingy is that when you subdivide its in the power of 4 ? altho thats tottally irrelevant when modeling .
The hole with quads can be done with 4, 5 , 6 sides, its irrelevant, but damn i have no idea what you are saying. i get a headache just reading what you write.
I seriously recomend you work first on the way of how you communicate before attempting to ask for help because i guarantee you that nobody will understand what you try to say and simply dismiss you as a troll or a weirdo.
@Ausonian Ok, I probably didnt use the best example as the surfaces are flat (also touched on by Perna), but if it were a curve I am on the right track?. Yeah I still have the "quads or die" mentality, though I have seen some really good tutorials about pentagons in forming a control loop. Thanks for the input Ausonian, I appreciate the guidance
Depends on what you are modeling, only practice can improve you, not theory; post something you really are having difficult with, so someone can help you. Lastly, tris aren't evil, either
I have no clue why I'm getting freaked by doing this kind of topo. I did two versions of the door (difference is just the size). Its in one geo though.
thats the problem xD I think im doing it wrong and in an inefficient way but both of them look ok. Its just that im doing star citizen-ish stuff and trying to do what they did without making mistakes. I just got to have a small environment done and well polished so I can get a job I'm a little over the edge today... Ill just go get some sleep
@DeathstrokeFTW To avoid the collapsing corners, I do as this: Instead of insetting all the way, I make a very small inset (or one with 0 as value), select the corner edges, scale them uniformily, and then scale the entire face.
DeathstrokeFTW I personally prefer the one on the right, I like the uniform curves more than the converging curves. Seems more like the sort of design choice artists in Star Citizen would make.
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,
Admiral, you don't need a lot of sides. Use sub-division modeling instead:
Yeah dude except not if the other side of the block doesn't match up with the cylinder such that its topology acts as a supporting/controlling edge.
What do you propose we do then, huh?
You know, you treated KarlP like a jerk, and then a trivially identifiable problem with this graphic flew right over your head.
Maybe there's a solution to this graphic that doesn't involve more geometry, but it's not obvious to me. So instead of just talking, why don't you walk the walk and show us how it's done?
My opinion? This is an edge case (no pun intended) that doesn't generalize to modeling in practice, which involves vastly less ideal topology than that of this example.
@KarlP, Perna is being a dick. Keep practicing, don't let anyone put you down.
An @Perna, instead of condescending with statements like "use sub-d," why don't you think before you post a virtually useless graphic?
Admiral, you don't need a lot of sides. Use sub-division modeling instead:
Yeah dude except not if the other side of the block doesn't match up with the cylinder such that its topology acts as a supporting/controlling edge.
What do you propose we do then, huh?
You know, you treated KarlP like a jerk, and then a trivially identifiable problem with this graphic flew right over your head.
Maybe there's a solution to this graphic that doesn't involve more geometry, but it's not obvious to me. So instead of just talking, why don't you walk the walk and show us how it's done?
My opinion? This is an edge case (no pun intended) that doesn't generalize to modeling in practice, which involves vastly less ideal topology than that of this example.
@KarlP, Perna is being a dick. Keep practicing, don't let anyone put you down.
An @Perna, instead of condescending with statements like "use sub-d," why don't you think before you post a virtually useless graphic?
the ammount of entiltement on your post is amazing. Its basically saying "you wrote something that i didnt even bother to properly research\try on my own because im used to beeing spoonfed knowledge , so its your obligation to do it for me since i think that my time is more precious than yours"
Joao Sapiro said: the ammount of entiltement on your post is amazing. Its basically saying "you wrote something that i didnt even bother to properly research\try on my own because im used to beeing spoonfed knowledge , so its your obligation to do it for me since i think that my time is more precious than yours"
man , you suck.
First of all, no, I don't at all expect to be spoonfed knowledge, but you can strawman me all you want if it makes you feel better. No, By using that graphic, in whatever context he did, the onus is on Perna to understand and, more importantly, identify to beginners its limitations.
Second, WooOOOoooOOOooo strong words from someone who couldn't be bothered to even consider the problems with the graphic I noted. Solve the problem I proposed without simply using more geometry and I'll shutup.
I'll admit the following: I demonized @Perna perhaps more than he deserves, but nonetheless the amount of condescension his words carry is palpable.
Additionally, I'm not saying the graphic isn't useless, but it WON'T work in every circumstance. Which I see was indeed stated. Maybe I jumped the gun here, but nonetheless I saw a lot of undue condescension the last few pages as well.
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...
Replies
...
First i would recomend showing results with a material without the wireframe and high specular value to see modeling errors , then post a concept on what you are trying to achieve , because i dont know what you are asking exactly
Your questions regarding edge flowing are a bit generic. The method of modeling something is subjective, you could start from a plane, or block out the entire shape and start detailing later. The first thing I notice about your model is that it is one piece; there is no reason to complicate things like that (unless you plan to print the model, but even in that case, I don't think it's necessary to have one single closed mesh). You should post some specifinc shape you are having trouble with, otherwise I don't think you will get much help here.
I appreciate that this question has been asked a lot but the answers dont seem to go very far for me.
So this is my take on the "how do I harden this edge" question.
The block out, trying to respect topology flow using a control loop before extruding
Additional support loops added with swift loop tool
Changes I made to the support loop flow beyond the edge I wanted hardened
Turbosmooth 1 itr
Turbosmooth 1 itr, high spec material
The questions
1. Have I finally understood "supporting the edge" and "use more geo"?
2. I used swift loop and ring connect but I would like better control of how tight my support loops are, any pointers on how I could better do that?
3. Any further suggestions?
Thanks all
Yes, understood, but the comments were right so I reposted with a better focus, there was no value in the previous posts.
Thats like those people that make dumb comments and then delete them because they are too afraid to "look bad" , when everyone knows they are jerks already :P
Live and learn !
To control how tight your support loops are, you simply move them, it's as simple as that.
In your exercise, it's useless to push back and forth the loops to "conform with the flow", since they reside on a totally flat surface and don't contribute to the shape; and your mesh doesn't have to be all quads
@Ausonian
Ok, I probably didnt use the best example as the surfaces are flat (also touched on by Perna), but if it were a curve I am on the right track?. Yeah I still have the "quads or die" mentality, though I have seen some really good tutorials about pentagons in forming a control loop. Thanks for the input Ausonian, I appreciate the guidance
Hi @Perna, I am a learner on here, I appreciate that you have 20+ years of experience in the industry but I do not. "The concepts" to me are only now becoming concrete, there has been lot of hard work from me to get this knowledge, everyone seems to have their own unique way of doing things and explaining them, sometimes these methods even appear, as a newbie, to be conflicting.
In response to point 1.
Some definitions I have learned from across the sources I have used
a. support edge - the edges added to control the hardness of the edge loop of polygons
b. control loop - a loop of polygons that defines the way fill loops will run around the model. These loops typically have e-poles (5 edged verts) and n-poles (3 edged verts) which change the "flow of the topology"
c. fill loops/fill edges - loops of polygons that add geometry for padding (for example between two control loops, I understand thats good practice) or to provide more verts/edges for further detail
Please feel free to correct my understanding
In response to point 2.
What I wanted better control on was the tightness of the support loops to the actual edge of the geometry. I could use swift loop tool but then I am placing by hand which is not consistent. I could use ring and then connect but again 90% on the tool slider for one loop of polys is not the same as 90% on the tool slider on another.
The question was - How do I get better control over these tools so that I can consistently set my support edges to obtain uniform tightness, if thats what I want?
In response to point 3.
Ok point taken, from here on, I will reduce my word count and wont post more than one image per post.
It is comments like this that make learning this craft so difficult, "I need a more pragmatic approach", that comment along with most of what you actually say means 3/5th of nothing to me. Are you here responding to my posts to help me or to mock me? It seems the latter, particularly your penultimate sentence, I am not on cocaine and to put it bluntly, that is offensive and insulting.
If you dont want to help or comment on the modeling techniques then dont. If I am as difficult to have here as you are making out then have me banned. I would like your help, I respect your experience and position in the community, but I dont need your help.
In response to the final paragraph
The first piece of information that has actually been useful to me. I will slow it down and I think you are right possibly I need to be more technical about it all, which suites me, I am more of a numbers person. Thanks for the help in that regard.
Also one thing that seems to be key is the number of sides to the hole, looks like you have 8 and that seems to be a good start.
Finally, this connectivity should help, it spreads the edges evenly around the hole so that when you subdiv it averages their positions and produces a smoother result.
I wanted to make a star citizen environment for my portfolio but I came across a problem while I was modelling the door. Now most of the doors have curved edges and they are easy to make but when I was modelling the door whenever I used the inset, the curved edges began to collapse with each other but in the references, it shows no signs of them overlapping
I should have thought about making multiple meshes and bridging them. Thanks for the nudge!
But if anyone else has a more efficient way of handling that then please let me know xD
Particpating here is part of trying to make those decisions on what to believe.
In regards to my answer to sunray, I will be more accurate in the future with any questions or explanations. I do know how subd works mathematically and why 8 works best.
Thanks again for taking the time to point me in a better direction.
The rule is number or faces at the point of placement +4 - reference: this forum, wirrex sept 2015.
Reason: the goal is to produce quads using this formula/method you can support the hole geometry with a ring of quads.
Subdivision: a vertex of a quad face is split to create a new quad face. The centre of the face is the position of the original vertex. The newly created vertices are position around this centre via a weighted average with the surrounding vertices of the orignal quads in the unsubdivided mesh.
Why do we want quad? Because quads will create quads when subdivided. Also quads deform well in animations due to the two degrees of freedom.
Feel free to correct me, apologies for the language used, it is wordy.
The hole with quads can be done with 4, 5 , 6 sides, its irrelevant, but damn i have no idea what you are saying. i get a headache just reading what you write.
I seriously recomend you work first on the way of how you communicate before attempting to ask for help because i guarantee you that nobody will understand what you try to say and simply dismiss you as a troll or a weirdo.
Lastly, tris aren't evil, either
Its in one geo though.
I'm a little over the edge today... Ill just go get some sleep
To avoid the collapsing corners, I do as this:
Instead of insetting all the way, I make a very small inset (or one with 0 as value), select the corner edges, scale them uniformily, and then scale the entire face.
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,
What do you propose we do then, huh?
You know, you treated KarlP like a jerk, and then a trivially identifiable problem with this graphic flew right over your head.
Maybe there's a solution to this graphic that doesn't involve more geometry, but it's not obvious to me. So instead of just talking, why don't you walk the walk and show us how it's done?
My opinion? This is an edge case (no pun intended) that doesn't generalize to modeling in practice, which involves vastly less ideal topology than that of this example.
@KarlP, Perna is being a dick. Keep practicing, don't let anyone put you down.
An @Perna, instead of condescending with statements like "use sub-d," why don't you think before you post a virtually useless graphic?
man , you suck.
Second, WooOOOoooOOOooo strong words from someone who couldn't be bothered to even consider the problems with the graphic I noted. Solve the problem I proposed without simply using more geometry and I'll shutup.
I'll admit the following: I demonized @Perna perhaps more than he deserves, but nonetheless the amount of condescension his words carry is palpable.
Additionally, I'm not saying the graphic isn't useless, but it WON'T work in every circumstance. Which I see was indeed stated. Maybe I jumped the gun here, but nonetheless I saw a lot of undue condescension the last few pages as well.
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...