I'm new in this forum and in subD's. I'm a max user so if I use some "max terms"... sorry. I've read all the posts in this thread and learned alot, from the examples and advices.
But I still have some doubts...
You mention that sometimes when you need to add more detail to a certain area but don't want to mess out other parts, you break that part (detach it... to a new object? to a clone?) and add detail... and then what? I tried that but in a circular object even if you place the new object EXACTLY in the point where it was... you clearly notice a seam!
Do you weld it back to place? But for that you have to deal with all the "loose ends"/edgeloops and such that you created...
I'm really in the dark here...
Is this a case of "more geometry"???
Any hint would be gold... thanks!!!
please post some pictures of what you mean.
I have no idea what you mean without pictures.
I think that on making the pictures I answered myself but anyway let me post them here and confirm my idea...
Image 1 is a typical case in wich you need more geometry, otherwise the tension between faces and the proximity generated when subdiving the mesh will cause lumps or straight edges...
Image 2 has a decent ammount of base geometry and the detail is localized. However, if I don't weld the 2 parts, there will be a visible seam...
So I guess my question is:
- Do you guys ALWAYS have to attach and weld everything after breaking it to add localized detail?
Or is it just in this cases (circular geometry)?
- This example could also be achieved through the use of "floating geometry, right?"
I think that on making the pictures I answered myself but anyway let me post them here and confirm my idea...
Image 1 is a typical case in wich you need more geometry, otherwise the tension between faces and the proximity generated when subdiving the mesh will cause lumps or straight edges...
Image 2 has a decent ammount of base geometry and the detail is localized. However, if I don't weld the 2 parts, there will be a visible seam...
So I guess my question is:
- Do you guys ALWAYS have to attach and weld everything after breaking it to add localized detail?
Or is it just in this cases (circular geometry)?
- This example could also be achieved through the use of "floating geometry, right?"
Your images are broken, so I might be missing what you're trying to say.
If I am breaking off an extra-detailed piece, like some sort of grille that is going to go on a simple box, I model it as a physically separate piece, not an integrated piece (think like a heating vent in your house, how it is a separate piece of metal that fits over the drywall).
With an integrated piece, you can typically terminate the extra edges needed for the high density grille, etc in flat surfaces, or by running them into symetry points. With a curved surface, though, you usually need to maintain the high density throughout the entirety of the curved surface or else you wind up with pinching on the curved surface.
With floating geometry, like a dimple in a flat surface, You simply model the dimple as a separate piece that floats above (think like a contact lens sitting on top of a sheet of paper). It doesn't look right from an extreme angle, but the normal/light maps you bake with it don't see it from that extreme angle, they see it from straight on, where it appears to be seamless.
I sent the images like this to illustrate the 2 ways I tried to solve the problem and it's steps... I never thought they could be confusing but that was my error... when you're doing something you end up expecting that everyone else sees what you see... sorry!
Thanks for your tips, they REALLy shed light into some of my doubts! And your example of the floating geometry has a contact lens is really vivid and cool! I could picture it in my mind with ease!
Removing those edges kills the pinching, but creates points of tension at these verts. Is there a way to get it completely smooth, or is it so minor that it won't show when I do baking?
Removing those edges kills the pinching, but creates points of tension at these verts. Is there a way to get it completely smooth, or is it so minor that it won't show when I do baking?
You could do it with a quadsphere cap. Take a cube, smooth it into an almost-sphere, and slap that on the end.
For polar pinching like that I tend to just add extra edge loops to reduce it to nothing because I am lazy. I also find having an n-gon rather than a bunch of tris reduces the pinching.
Thanks for the video, definitely helped with the part. But there's something in there that modo seems to own at? It relaxes that square-ish circle into a real circle. In WAY 1. Right after you cut out the sides and fill them in and then drag the a new loop inward.
thats a script that i have called perfect circle.
But you can do the same in max.
Select the verts that form the circle and apply the spherefy modifier and those points will make a circle
Thanks Bit. I'm new to modo and gonna try it out again. Only problem with Modo I see is no stack. Other than that it's a streamline 3D program.
Yep. There is no stack. That's one thing that max offers that is powerfull.
Modo's strenghts are the tools and the easy of use. Selecttions and stuff like that, are very powerfull.
please see the attached images & obj...i am trying to model this shape with a curvy surface & sharp corners...but if i bevel the corner edges for sharpness...the sharp edge continues throughout the middle of the object.if i try to terminate the corner edges a pinch appears...how do i maintain a curved surface with a sharp corner?
ok. its my time to ask questions
so how would you guys wold model this?
Easily until I hit the receiver piece, and then much more slowly
edit: actually that looks like it would be crazy fun to model. Once you solve the tricky bit it all looks pretty easy, unless you want to model that embossed detail in.
In LW I'd spline it out, then make poly patches. This one I just made a poly cylinder, removed 270 deg of it then added the divot, made a spline to guide it and shaped it. Like 4 minutes.
Kind of a weird shape but the steps to make this is pretty simple and short.
I'm using 3ds max but any app should have similar tools to do this..
1. create two objects, one with a sharp corner and another with the desired curve.. you'll want to maintain the same proportions, edge distances and number of edges..
2. Join the objects and perform a bridge operation. Keep it only at only 1 segment.
3. Chamfer the edge that was created between the two objects with multiple edges and a large chamfer distance.. that way it smoothes out the transition.
4. Delete the outer most edge on the sharp corner.
5. Apply Symmetry to the model.. and there you go. You can vary this up quite a bit by changing the distances between the objects and how much you chamfer the edge.
There might be some guys that can make this is zbrush or mudbox.
Like the dude that made Batman over at gameartisans for comicon.
He used zbrush for his hardsurface modeling
What's his flow? I always find it a pain in the ass to do hard surface on Zbrush. It's just not efficient to me. Half the time I find myself making hi res alphas to just to create grebble indents. Just creating the hires HR in all Max then detailing in Zbrush saves time for me.
This is question for the Max users out there. How do you setup for Sub Div modelling?
I was showing a friend how to model a shape (the shapes not important) and he questioned why I had a TurboSmooth in my stack. He then pointed out that you get the same results by turning NURMS smoothing on within the edit poly. I now use this method, as I can't see any reason to use TurboSmooth, but through browsing this thread I'v noticed alot of people using it.
Any reason why I would use a modifier for the smooth instead of NURMS?
perna: nice solution.. a bit simpler and easier to edit later on. How dare you one up me! :P
Bad Spleen: As far as I know Turbosmooth and the NURMS smoothing in Edit Poly are the same but different routes to take when smoothing your model. They both use the NURMS method to smooth the object. The main difference is that turbosmooth, since its a modifier allows you to apply its effect to multiple objects, groups etc. simultaneously through instancing and referencing... while the Edit Poly option only applies to the one object.
So for instance, if you have several objects in your scene and they are all being subdivided using the edit poly NURMS, you would have to go into each object to change their settings, but with Turbosmooth you can instance the effect and change one objects modifier to change all of the others at the same time...
There's also Meshsmooth which has additional features like different Sub-D methods and weighted verts and can it can keep faces convex unlike Turbosmooth. It's only drawback is that its slower. Basically we have Turbosmooth because users wanted the same speed as the Edit Poly version but with the convenience of a modifier. So they stripped Meshsmooth of its added qualities and created the Turbosmooth modifier.
Another benefit of using the modifier is that you can add smoothing after other modifiers without having to smooth the base object first... handy if you have a long stack of modifiers and don't want to deal with the additional geometry until later.
You can also instance your turbosmooth modifiers across different objects and toggle them on/off, all at once.
And perhaps the display is actually faster... back in previous versions Editable Poly objects were slow to draw, but turbosmooth converted them to Mesh instead.
Dunno how it goes nowadays though, as pieces of Max are completely rewritten all the time...
If you want your viewport to render even faster and you know (absolutely) that you are done modeling you can convert your model to a Editable Mesh after you smooth it out. Editable Mesh is for some reason much easier for the computer to handle than Edit Poly. When I have a huge scene with hundreds of objects I usually do this after backing up everything in a separate file, its amazing what a difference it makes.
nah i need an hard corner there and 90" between lower surface, but without that bug, but when there are 3 edges it destroys my lower surface (i could separate this but isnt there any trick that will solve this?
this isnt sharp ant the other one with 3angles smooths badly
Emesh renders faster, however there's no need to back up or to wait until you are done modeling. I'm not sure why you suggest that? You can convert back and forth between emesh and epoly without any problems
The only reason I would suggest that is because i usually use a large modifier stack when i'm modeling, mostly as a history as well as using deformers and such. It's always nice to be able to go back and forth and so I usually backup the model in case I actually need to go back and change something quickly without having to remodel from the current version. Its just about saving time for me
I am serious but you know i wanted to avoid this 3 edges goin for all the lower surface, coz this all is bend in letter C, when i got 3 edges goin threw all the surface it gives me an hard edge everywhere. ok never mind just separated the outer element from the lower surface.
I can't decipher your message. Just do like shown in my image and you'll be fine.
Segreto: You don't have to collapse the stack, just make sure the last entry on the stack converts to emesh, like shell or turn to editable mesh. The result is you have the speed advantages of emesh, retain your stack, and can still model in epoly.
ahh i didn't realize that you still got the speed advantage doing that, i'll keep that in mind thanks
Turbosmooth turns to editable mesh too. Just set it to zero divisions and toggle it on and off- you'll see that even the viewport shading is different.
Hey guys got another bit of a snag here. Basically I'm wondering if I should leave the edgeloops I've highlighted as is, or can I remove the blue lines somehow (which really don't need to be there) while retaining the red ones and having them connect to the other edges that reside below the blue highlighted ones?
Much appreciated GCMP! So is it safe to say that I can never really overdo it with edges when going the sub-d approach? I'm sure we have to be efficient with it, but at the same time, seems I can go place as many edges where need be.
No problemo.
Your mostly limited by the maximum your hardware can handle when it comes to high poly models unlike low poly where it's how capable the game engine software is.
Thanks for the feedback Perna and GCMP, and yeah I'll definitely need those green lines as well! From the camera view I took a print screen of, it's hard to tell the shape of this thing, but it's a lil' odd at that corner, but I'm sure the reworked loops should work fine without any problems.
I've seriously got to dig myself out of that low-poly method of thinking. It's gone for the most part, still there at times though when I'm doing this kind of stuff.
Replies
There's lots of ways to make that shape. In this video i show 2.
The pic posted earlier uses WAY 1, but i think WAY2 is nicer.
http://vimeo.com/6143806
i guess the video hasn't finished uploading.
here is the direct link
http://dl.getdropbox.com/u/93793/Modo_Sight_Top.avi
please post some pictures of what you mean.
I have no idea what you mean without pictures.
I think that on making the pictures I answered myself but anyway let me post them here and confirm my idea...
Image 1 is a typical case in wich you need more geometry, otherwise the tension between faces and the proximity generated when subdiving the mesh will cause lumps or straight edges...
Image 2 has a decent ammount of base geometry and the detail is localized. However, if I don't weld the 2 parts, there will be a visible seam...
So I guess my question is:
- Do you guys ALWAYS have to attach and weld everything after breaking it to add localized detail?
Or is it just in this cases (circular geometry)?
- This example could also be achieved through the use of "floating geometry, right?"
[edit]Fixed images, you had https://dl-web.getdropbox.com/get/subD_01.jpg?w=1cbb558e instead of https://dl-web.getdropbox.com/get/subD_01.jpg
[editedit] No that is still wrong. Looks like you're coping the wrong link entirely.
Your images are broken, so I might be missing what you're trying to say.
If I am breaking off an extra-detailed piece, like some sort of grille that is going to go on a simple box, I model it as a physically separate piece, not an integrated piece (think like a heating vent in your house, how it is a separate piece of metal that fits over the drywall).
With an integrated piece, you can typically terminate the extra edges needed for the high density grille, etc in flat surfaces, or by running them into symetry points. With a curved surface, though, you usually need to maintain the high density throughout the entirety of the curved surface or else you wind up with pinching on the curved surface.
With floating geometry, like a dimple in a flat surface, You simply model the dimple as a separate piece that floats above (think like a contact lens sitting on top of a sheet of paper). It doesn't look right from an extreme angle, but the normal/light maps you bake with it don't see it from that extreme angle, they see it from straight on, where it appears to be seamless.
I sent the images like this to illustrate the 2 ways I tried to solve the problem and it's steps... I never thought they could be confusing but that was my error... when you're doing something you end up expecting that everyone else sees what you see... sorry!
Thanks for your tips, they REALLy shed light into some of my doubts! And your example of the floating geometry has a contact lens is really vivid and cool! I could picture it in my mind with ease!
Thanks for your time
Removing those edges kills the pinching, but creates points of tension at these verts. Is there a way to get it completely smooth, or is it so minor that it won't show when I do baking?
You could do it with a quadsphere cap. Take a cube, smooth it into an almost-sphere, and slap that on the end.
For polar pinching like that I tend to just add extra edge loops to reduce it to nothing because I am lazy. I also find having an n-gon rather than a bunch of tris reduces the pinching.
Thanks for the video, definitely helped with the part. But there's something in there that modo seems to own at? It relaxes that square-ish circle into a real circle. In WAY 1. Right after you cut out the sides and fill them in and then drag the a new loop inward.
But you can do the same in max.
Select the verts that form the circle and apply the spherefy modifier and those points will make a circle
Thanks Bit. I'm new to modo and gonna try it out again. Only problem with Modo I see is no stack. Other than that it's a streamline 3D program.
I've corrected the links... sorry about that! I'm new to Dropbox...
Yep. There is no stack. That's one thing that max offers that is powerfull.
Modo's strenghts are the tools and the easy of use. Selecttions and stuff like that, are very powerfull.
i don't understand very much what youare asking
you want curved surfaces with sharp corners?
Is there a reference you're working off? or some concept? I just can't imagine anything that would look like that. I'd love to see an example though.
so how would you guys wold model this?
Eu modelava-a com a mao nos tomates, it's my favourite way to model!
olololol grande vitor!!! por falar em tomates, tou aqui com uma comixaozita. olol
Easily until I hit the receiver piece, and then much more slowly
edit: actually that looks like it would be crazy fun to model. Once you solve the tricky bit it all looks pretty easy, unless you want to model that embossed detail in.
I would use polygons.
http://www.3dtotal.com/domwar04_timelapse/
Like the dude that made Batman over at gameartisans for comicon.
He used zbrush for his hardsurface modeling
Kind of a weird shape but the steps to make this is pretty simple and short.
I'm using 3ds max but any app should have similar tools to do this..
1. create two objects, one with a sharp corner and another with the desired curve.. you'll want to maintain the same proportions, edge distances and number of edges..
2. Join the objects and perform a bridge operation. Keep it only at only 1 segment.
3. Chamfer the edge that was created between the two objects with multiple edges and a large chamfer distance.. that way it smoothes out the transition.
4. Delete the outer most edge on the sharp corner.
5. Apply Symmetry to the model.. and there you go. You can vary this up quite a bit by changing the distances between the objects and how much you chamfer the edge.
What's his flow? I always find it a pain in the ass to do hard surface on Zbrush. It's just not efficient to me. Half the time I find myself making hi res alphas to just to create grebble indents. Just creating the hires HR in all Max then detailing in Zbrush saves time for me.
First up, excellent thread, very useful.
This is question for the Max users out there. How do you setup for Sub Div modelling?
I was showing a friend how to model a shape (the shapes not important) and he questioned why I had a TurboSmooth in my stack. He then pointed out that you get the same results by turning NURMS smoothing on within the edit poly. I now use this method, as I can't see any reason to use TurboSmooth, but through browsing this thread I'v noticed alot of people using it.
Any reason why I would use a modifier for the smooth instead of NURMS?
- Adam
Bad Spleen: As far as I know Turbosmooth and the NURMS smoothing in Edit Poly are the same but different routes to take when smoothing your model. They both use the NURMS method to smooth the object. The main difference is that turbosmooth, since its a modifier allows you to apply its effect to multiple objects, groups etc. simultaneously through instancing and referencing... while the Edit Poly option only applies to the one object.
So for instance, if you have several objects in your scene and they are all being subdivided using the edit poly NURMS, you would have to go into each object to change their settings, but with Turbosmooth you can instance the effect and change one objects modifier to change all of the others at the same time...
There's also Meshsmooth which has additional features like different Sub-D methods and weighted verts and can it can keep faces convex unlike Turbosmooth. It's only drawback is that its slower. Basically we have Turbosmooth because users wanted the same speed as the Edit Poly version but with the convenience of a modifier. So they stripped Meshsmooth of its added qualities and created the Turbosmooth modifier.
Another benefit of using the modifier is that you can add smoothing after other modifiers without having to smooth the base object first... handy if you have a long stack of modifiers and don't want to deal with the additional geometry until later.
And perhaps the display is actually faster... back in previous versions Editable Poly objects were slow to draw, but turbosmooth converted them to Mesh instead.
Dunno how it goes nowadays though, as pieces of Max are completely rewritten all the time...
guys how to solve this? cand figure that out now
this isnt sharp
The only reason I would suggest that is because i usually use a large modifier stack when i'm modeling, mostly as a history as well as using deformers and such. It's always nice to be able to go back and forth and so I usually backup the model in case I actually need to go back and change something quickly without having to remodel from the current version. Its just about saving time for me
ahh i didn't realize that you still got the speed advantage doing that, i'll keep that in mind thanks
Solution for Hatred
and G3L
RED lines are new lines.
Your mostly limited by the maximum your hardware can handle when it comes to high poly models unlike low poly where it's how capable the game engine software is.
perna agreed I love autodesk:thumbup:
I've seriously got to dig myself out of that low-poly method of thinking. It's gone for the most part, still there at times though when I'm doing this kind of stuff.
Help?
My bad, it was late, honest :poly122:
Glad you got it fixed Hatred!
ZacD - why don't you just have a go at it?