That's well said. And it's true of almost everything in 3D. In fact, I'd go a step further and say that asking "Why?" is a good start, but also try it for yourself. Does having a triangle on your mesh melt your normal map or something? No? Then you've just discovered that the advice you were given is wrong.
Or, at least, it comes with qualifiers that you weren't given. Like, animation or deformation concerns.
Actually it depends. Tris are bad because of the way they smooth; They tend to create bulges, like NGons on subdivided meshes... otherwise on planar geo it's totally fine and it's a time saver...
these "issues" are fully visible at least with shinny materials like blinn or phong.
If you're using Maya create a platonic solid of icosahedron.
Smooth it 3x using linear which will keep the triangles.
Create a smaller sphere and then shrink wrap the smoothed icosahedron to regain a spherical shape as it will have some verts protruding after a linear smooth.
Then extrude hexagonal shapes several times to get the steps.
ok nm that gets unpredictable results after going more in depth with the process. Some of the polys lose symmetry. Try a geodesic script from creative crash?
I unfortunately have to step out and can not try this but it looks promising: http://trevorius.com/scrapbook/uncategorized/geodesic-sphere-with-ui/
EDIT ok that still doesn't really work either.keeping symmetry between polys is nigh impossible.
I created an icosahedron, scaled all the verts down very close to a dense sphere, shrink wrapped it then smoothed.
Scaled the new icosahedron up to where the shallowest verts barely stuck out of the sphere. Hand picked a group that had matching elevation above the sphere scaled them down to match. Shrink wrap smooth again.
Repeated this process two more times and STILL got uneven shapes.
As a half-hearted suggestion check this site out? http://www.ziptiedomes.com/geodesic-dome-calculators/6v-geodesic-dome-calculator.htm
This one should be easy for you guys. Is this geo okay, or is there some better way of doing it that I'm not seeing? It looks alright when it's smoothed, but my OCD is telling me to get rid of that 5-sided polygon... I feel like I'm missing something really freakin' obvious here.
This one should be easy for you guys. Is this geo okay, or is there some better way of doing it that I'm not seeing? It looks alright when it's smoothed, but my OCD is telling me to get rid of that 5-sided polygon... I feel like I'm missing something really freakin' obvious here.
If it sits, it fits!
Basically, if it gives you the desired smoothing well than.. !
I actually use 5 sided polygons in a lot of places in my models. As long as it smooths as expected or the issue is very small just move on to more important stuff
If you are determined and really want to only have quads on that mesh, there are several solutions. This is a quick one.
You have to move vertices in order to have an even distribution of geometry, or at least a gradual shift.
You can use of average vertex or edit edge flow to automate the process, but you'll have to delete some faces and possibly draw temporary faces in order to make the program understand exactly what you want to blend and in which direction.
Back to the geodesic dome problem. I modeled out a hex with a raised center and then extruded some platforms and then duplicated it out into a square array.
I tried to use a couple of non-linear bends to make it into a dome but it's pinching one of the axes.
I figure this is the way to do it though if only I could figure out the right way to deform it.
Here's an OBJ if anyone wants to have a go https://drive.google.com/file/d/0B6QMsNRqTwb7R09DTWtnNlkwTXM/view?usp=sharing
That's well said. And it's true of almost everything in 3D. In fact, I'd go a step further and say that asking "Why?" is a good start, but also try it for yourself. Does having a triangle on your mesh melt your normal map or something? No? Then you've just discovered that the advice you were given is wrong.
Or, at least, it comes with qualifiers that you weren't given. Like, animation or deformation concerns.
Exactly, incomplete advice can be just as harmful as bad advice, especially if repeated ad infinitum without understanding.
Actually it depends. Tris are bad because of the way they smooth; They tend to create bulges, like NGons on subdivided meshes... otherwise on planar geo it's totally fine and it's a time saver...
these "issues" are fully visible at least with shinny materials like blinn or phong.
I think you need to re-read the bolded part of Warren's post.
If you are determined and really want to only have quads on that mesh, there are several solutions. This is a quick one.
You have to move vertices in order to have an even distribution of geometry, or at least a gradual shift.
You can use of average vertex or edit edge flow to automate the process, but you'll have to delete some faces and possibly draw temporary faces in order to make the program understand exactly what you want to blend and in which direction.
I've been doing some sub-d stuff for archvis recently and this thread has been invaluable for reference from all the contributors.
I'm wondering though, have many people been using stuff like Marius Silaghi's openSubdiv or Max's 2015 onwards, Blender etc, openSubdiv to alleviate some of the extra workload managing support loops for mostly linear edge treatments?
For stuff like the above I'd probably just use that approach, though for a soft furnishing I was just modelling I quite quickly went back to DIY loops to get the exact form I needed from the smoothing.
I'm not sure I'd give myself a headache over managing linear edge treatments when I can just use these newer approaches.
So, since probooleans give me headaches and some people say not to use them 'cause of all the trouble they can cause, what is the best way(s) to cut perfect holes in basic shapes like cylinders, squares, even spheres? I've looked at youtube videos, but I feel members here might have faster ways.
So, since probooleans give me headaches and some people say not to use them 'cause of all the trouble they can cause, what is the best way(s) to cut perfect holes in basic shapes like cylinders, squares, even spheres? I've looked at youtube videos, but I feel members here might have faster ways.
Use a regularize script, or a spherize command to circle your selection(there are bunch online). IMO booleans are gold but not for things such as a perfect hole. To make a hole in a cylinder just inset it's cylinder cap, if you want mini holes, just do a double quad inset where you want them, hope this quick reply help...
It seems like a good idea to use these approaches.
As a Maya user, I do not have any of this fancy modeling stuff. But I like to retopologize intersection of tubes with quad draw, it's still faster than doing it the traditional way.
Gee, I feel like I post a ton of silly questions here... Oh well! I have yet another problem. This area is supposed to be smooth, but the shading looks weird and lumpy, probably because of the poles. Could someone show how I should make the topology to avoid these issues? The shading problem really doesn't come across very well on a picture, but it shows very well when a lightsource is moving across the model.
Yeah that's not the best place for poles. Try to keep the lines parallels on the side.
i try to think how the model is machined irl, and basically, it's simple shapes with booleans operations. So think booleans with support loops (even if you don't really model it with booleans operations)
1- make cylinder
2- select edge ring
3- connect
4- select the new loop
5- chamfer (from 1 to 2 edges) so you have now 2 lines in the right place
6- select the 2 new loops
7- chamfer both with the same value
8- select both poly rings
9- extrude inwards with the same value
The shape is super-simple. Let's assume the reference is facing the Z axis.
You have a profile with visible finger indents, that you extruded along Z axis.
Cut the shape along its Z symmetry, now you have a triple profile with finger indents.
Leave the central one alone and move the extreme ones back along the X axis.
ok, i'm stuck, and i cant for the love of good not figure this shape out!
The Ref. I have is really bad, and if i had a better one, i'd probably could do a better job.
1. I'm overthinking it?
2. how's the shape going?
Please help me
You're overthinking it though. just by seeing how the handle catch the highlight you can pretty determine if this or that area is "concave", i mean I like to say "do it as you see it". This is why blockout is so important. If i can give you a special advice : visualize indents from all the views.
Probably overthinking it, got to this point, not really happy, but atleast i figured out how to get to that point. Thank you guys for the tips. If you guys have time, i would be very helpfull to see a block in!
How would you guys model the area where the "hole" kinda goes over the rounded edge? I want to both maintain the roundness of the hole, and keep a consistent curve along the edge of the grip.
How would you guys model the area where the "hole" kinda goes over the rounded edge? I want to both maintain the roundness of the hole, and keep a consistent curve along the edge of the grip.
You need to use enough geometry.
When holes and curves intersect, the smoothing interpolation is altered, and will be stopped by irregular lines.
You must make sure that your basic shape looks similar enough to the end curve wherever the intersection happens.
It's in the title of the thread for a reason
Topology could be cleaner but I'm a practical guy, if it smooths right it can be ugly IMO
Not quite, I can tell at a glance that some parts have a wrong position on the Z axis (too far or too close) not representing the actual shape.
Also the topology you chose for the triangle shaped hole won't help when it comes the time to bend it (because it's placed on a curved surface, as you can teel observing the handle base)
I just found this amazing site filled of great artist! Maybe someone could help me with a tedious task I wanted to achieve!
I am trying to model a sphere with a subtracted star but I want to make it with the proper way! Using quads and right stuff as you all know!
I am making this for personal achieving and studying!
I take this challenge to learn a bit more of 3D modelling!
Take a look at the attached images so you can see what I have done!
You'll get the Idea of what I want!
I thinks this is harder than I thing so this is why I am asking here for some tips!
At the moment I can't get the points of the star sharp, I am getting rounded ones! and I do not know how to get it!
Hope you can throw me some tips!
Thanks in advance!!!
Oh! I am using C4D for modelling!
hey thanks! Im gonna give it a try! Kind of complicated to me!
That wrapping idea sounds interesting!
Also the main idea is to have the interior of the star smooth as the sphere surface!
Thanks in advance!
For this I would recommend working from a higher poly sphere. SubD is not gonna work with your current mesh. All the unevenly distributed edges are breaking curvature.
Oh Gosh!! Closer!!!!!! Thanks ZacD @ZacD! That is a really close solution! I need it to be sub D but star interior like sphere surface! the problem i am facing now is the edges! I need to have them with great bevel! and keeping the star points sharp aswell! Gonna try to upload some screenshots soon! Thanks in advance to everyone elping me! Really appreciate it!
Pura Vida!
Replies
Or, at least, it comes with qualifiers that you weren't given. Like, animation or deformation concerns.
these "issues" are fully visible at least with shinny materials like blinn or phong.
http://i.imgur.com/bwATJDJ.jpg
Smooth it 3x using linear which will keep the triangles.
Create a smaller sphere and then shrink wrap the smoothed icosahedron to regain a spherical shape as it will have some verts protruding after a linear smooth.
Then extrude hexagonal shapes several times to get the steps.
I unfortunately have to step out and can not try this but it looks promising:
http://trevorius.com/scrapbook/uncategorized/geodesic-sphere-with-ui/
EDIT ok that still doesn't really work either.keeping symmetry between polys is nigh impossible.
I created an icosahedron, scaled all the verts down very close to a dense sphere, shrink wrapped it then smoothed.
Scaled the new icosahedron up to where the shallowest verts barely stuck out of the sphere. Hand picked a group that had matching elevation above the sphere scaled them down to match. Shrink wrap smooth again.
Repeated this process two more times and STILL got uneven shapes.
As a half-hearted suggestion check this site out?
http://www.ziptiedomes.com/geodesic-dome-calculators/6v-geodesic-dome-calculator.htm
Basically, if it gives you the desired smoothing well than.. !
You have to move vertices in order to have an even distribution of geometry, or at least a gradual shift.
You can use of average vertex or edit edge flow to automate the process, but you'll have to delete some faces and possibly draw temporary faces in order to make the program understand exactly what you want to blend and in which direction.
I tried to use a couple of non-linear bends to make it into a dome but it's pinching one of the axes.
I figure this is the way to do it though if only I could figure out the right way to deform it.
Here's an OBJ if anyone wants to have a go
https://drive.google.com/file/d/0B6QMsNRqTwb7R09DTWtnNlkwTXM/view?usp=sharing
The subject has been extensively covered elsewhere.
For example,
http://forums.autodesk.com/t5/modeling/hexagon-ball/td-p/4021084
The reference picture just happens to not show the pentagons, but they must be there
Anyway, starting from a Geosphere, smooth it and add a sculpt deformer to make it spheric
EDIT: Concerning Sculpt Deformer, I'm using Max...is there a version of that tool in Max?
Your question was interesting to me as well so thank you + i guess spherize modifier will do the job in Max.
Exactly, incomplete advice can be just as harmful as bad advice, especially if repeated ad infinitum without understanding.
I think you need to re-read the bolded part of Warren's post.
Or another way like this,
I'm wondering though, have many people been using stuff like Marius Silaghi's openSubdiv or Max's 2015 onwards, Blender etc, openSubdiv to alleviate some of the extra workload managing support loops for mostly linear edge treatments?
For stuff like the above I'd probably just use that approach, though for a soft furnishing I was just modelling I quite quickly went back to DIY loops to get the exact form I needed from the smoothing.
I'm not sure I'd give myself a headache over managing linear edge treatments when I can just use these newer approaches.
Use a regularize script, or a spherize command to circle your selection(there are bunch online). IMO booleans are gold but not for things such as a perfect hole. To make a hole in a cylinder just inset it's cylinder cap, if you want mini holes, just do a double quad inset where you want them, hope this quick reply help...
As a Maya user, I do not have any of this fancy modeling stuff. But I like to retopologize intersection of tubes with quad draw, it's still faster than doing it the traditional way.
i try to think how the model is machined irl, and basically, it's simple shapes with booleans operations. So think booleans with support loops (even if you don't really model it with booleans operations)
In Max, how do I make these spaces equal in distance? Image here: http://i.imgur.com/FX6DTtb.jpg
I'd build half of it and use the symmetry modifier.
2- select edge ring
3- connect
4- select the new loop
5- chamfer (from 1 to 2 edges) so you have now 2 lines in the right place
6- select the 2 new loops
7- chamfer both with the same value
8- select both poly rings
9- extrude inwards with the same value
The Ref. I have is really bad, and if i had a better one, i'd probably could do a better job.
1. I'm overthinking it?
2. how's the shape going?
Please help me
You have a profile with visible finger indents, that you extruded along Z axis.
Cut the shape along its Z symmetry, now you have a triple profile with finger indents.
Leave the central one alone and move the extreme ones back along the X axis.
You're overthinking it though. just by seeing how the handle catch the highlight you can pretty determine if this or that area is "concave", i mean I like to say "do it as you see it". This is why blockout is so important. If i can give you a special advice : visualize indents from all the views.
bEst regards
You can make it rounder, sharper, etc. Just move the vertices around.
Thank you
Tack S
When holes and curves intersect, the smoothing interpolation is altered, and will be stopped by irregular lines.
You must make sure that your basic shape looks similar enough to the end curve wherever the intersection happens.
It's in the title of the thread for a reason
Topology could be cleaner but I'm a practical guy, if it smooths right it can be ugly IMO
Also the topology you chose for the triangle shaped hole won't help when it comes the time to bend it (because it's placed on a curved surface, as you can teel observing the handle base)
I just found this amazing site filled of great artist! Maybe someone could help me with a tedious task I wanted to achieve!
I am trying to model a sphere with a subtracted star but I want to make it with the proper way! Using quads and right stuff as you all know!
I am making this for personal achieving and studying!
I take this challenge to learn a bit more of 3D modelling!
Take a look at the attached images so you can see what I have done!
You'll get the Idea of what I want!
I thinks this is harder than I thing so this is why I am asking here for some tips!
At the moment I can't get the points of the star sharp, I am getting rounded ones! and I do not know how to get it!
Hope you can throw me some tips!
Thanks in advance!!!
Oh! I am using C4D for modelling!
That wrapping idea sounds interesting!
Also the main idea is to have the interior of the star smooth as the sphere surface!
Thanks in advance!
Thanks in advance!
If it needs to be smoother, you can subdivide the dodecahedron more.
Pura Vida!