Modeling - Helmet

[NanoGator]

I'd do it with Metanurbs. Start with a box, subdivide via metanurbs, then drag points around.. smooth shift.. bandsaw... etc.

[Longman]

Originally posted by NanoGator
I'd do it with Metanurbs. Start with a box, subdivide via metanurbs, then drag points around.. smooth shift.. bandsaw... etc.

Sub-D's = metanurbs

Correct?

[iatepaste]

Tried that. With metanurbs/sub-d's (same term), I can't get a truly spherical surface like it calls for. I'm giving it another whirl though.

[NanoGator]

Originally posted by iatepaste
Tried that. With metanurbs/sub-d's (same term), I can't get a truly spherical surface like it calls for. I'm giving it another whirl though.

Start with a metaform subivided box. Box turns into sphere when nurb'd. :)

[Longman]

What do you mean a truly spherical shape? I don't know of anything much more spherical than a tesselated ball with sub-d's turned on... ;)

[NanoGator]

Originally posted by Longman
What do you mean a truly spherical shape? I don't know of anything much more spherical than a tesselated ball with sub-d's turned on... ;)

Depends on the type of ball, really. The default sphere in LW has triangles at the very top, and they make for weird subdivision problems. An ideal sphere using metanurbs would be made with all quads. That's why I'm saying 'subdivide a box'.

[iatepaste]

Hey,

I'm not questioning your methods but hoping to learn something new so bear with me.

Even with a sub-d box, I always still get some "corner" areas. I don't mean 90 degress or anything but just slight smooth areas where there is not perfect curvature. Maybe I'm just too anal and no one would ever notice it.

I've come close to what I'm looking for by sub-d a sphere as mentioned but like you said, there always is that slight little abnormality that shows up. I haven't really played that much with a tess. sphere though so that's next on the list.

Thanks.

/devin

PS - I'll try to post some images. Maybe that'll help show what I'm futzing up.

[Cman]

When I first saw the thread appear, and your ref, I thought it was your final render! I was like, WOWOWOWOW!!
-LOL

Then I read your post. :)
Good luck! :thumbsup:

[dies-irae]

agreeing completely with cman...
you shouldn't trick us that way ;)

try using both subd and splines (subd a tesselated sphere for the main body and do the little parts by spline...)

[SplineGod]

Originally posted by iatepaste
Hi all,

I need to model a shape similiar to an old school motorcycle helmet or football helmet sans facemask. Below is a ref pict of what I'm talking about (minus the visor).

http://iatepaste.com/ganymede/Herb_And_Skills/ftknox/helmet.jpg

I've tried starting out with sub-d's (not true spherical shape and it gets messy), a booleaned shere, and even spline patches. So far it's a no go. I just can't get a clean shape, esp. after I try to give the item thickness.

I'd appreciate any insight. I've scoured the web and so far, nothing.
THX,
/devin
Something like that is pretty easy to do with splines. Heres an example of a helmet I created with splines. Its a lot faster to layout 39 point in 13 splines. I not only get the shape I want but I also get the polys flowing in just the right way that makes it easy to shift over to SubDs and get the rest of the details I need.

[iatepaste]

Well, I think I'm about to get it. I did a quick build and now want to go back and get a polished one. THANKS for all the help!!!

/d

[Kaiser_Sose]

Originally posted by SplineGod
Something like that is pretty easy to do with splines. Heres an example of a helmet I created with splines. Its a lot faster to layout 39 point in 13 splines. I not only get the shape I want but I also get the polys flowing in just the right way that makes it easy to shift over to SubDs and get the rest of the details I need.

Hey Larry, let me see if I got how you did this

You laid out some points in a arc fashio and placed them on 0 x

made another arc, smaller than the first one, and in back view move the to the right

did it again with even a smaller arc

selected the bottom points and made a spline

made one in the middle make two halves

made two cross splines

[wgreenlee1]

Here is mine...took longer to render than itdid to model...

[wgreenlee1]

wireframe shaded....

[SplineGod]

The way wgreenlee created the helmet will probably work fine for that particular helmet. I resort to splines in many instances because moving around 39 points (in my case) is much faster then moving around hundreds of points. I can see the form very quickly. Since I know how many polys will be in each patch I can easily predict the total number of polys. When the splines are in place I can see how the polys will flow. In my cast its very important because the flow determines HOW Im going to be able to smooth shift out the rudder on the helmet, the grooves, eye and mouth pieces. If I simply started with a sphere or other primitive I would have a much more difficult time doing these things. Im basically just using the splines as a layout tool.

[SplineGod]

Heres more or less the order that I created things.
I made a couple of template polys that gave me the general form of the helmet. I picked points off those template polys and created the splines. This keeps me from having to weld all the time and insures I follow the form I laid out. I tune the flow last and make sure that I get patches that work.

[wgreenlee1]

Here is another one ...more along the lines of Larrys....
This time PozExtender.

[wgreenlee1]

Heres the Modeler Screenshot....
Sorry about the compression...going from 1600X1200 down to 800X600 leaves something to be desired...

[Primus]

Man I loved Dave Steven's Rocketeer :thumbsup: Nice take Spline God :cool:

[SplineGod]

It looks like a knights helmet of some kind. :)
I dont like the extender method because you really cant see the final form like you can with splines before starting to model. You also end up dealing with a lot more points still. Its too much like point by point modeling. It takes much longer to finally work out a smooth model.

[wgreenlee1]

Originally posted by SplineGod
It looks like a knights helmet of some kind. :)
I dont like the extender method because you really cant see the final form like you can with splines before starting to model. You also end up dealing with a lot more points still. Its too much like point by point modeling. It takes much longer to finally work out a smooth model.

Well maybe for you it takes longer....for others perhaps not...
Splines are too mechanical for my taste there is no "feeling" I get from splines.
They seem like wire or boards where extender tools feel like clay or dough....
They have thier purpose but for true form and style extend tools seem more friendly too me:cool:

[SplineGod]

Originally posted by wgreenlee1
Well maybe for you it takes longer....for others perhaps not...
Splines are too mechanical for my taste there is no "feeling" I get from splines.
They seem like wire or boards where extender tools feel like clay or dough....
They have thier purpose but for true form and style extend tools seem more friendly too me:cool:
It DOES take longer. Its simple math. Moving around 100 points will take longer then moving around 30 points. The more points you have to move around the more localized the changes become. Splines create a mesh that is the average of the curves. You use a very small amount of points to create and control a larger amount of points. Point by point modeling was the ONLY organic way to model in Lightwave before there were SubDs of any kind. I know how to model that way which is why I dont do it anymore because Lightwave now has far better tools and methods. I also used to use splines for 90% of my organic modeling simply because that was the only quick way to do any organic modeling before Subpatches came to Lightwave. When Subpatches showed up I relied less on splines and more on subpatches for certain things. More and more tools have shown up that make it easier to control the flow for example. Before spin quads you had to manually kill the polys and recreate them to achieve the same effect. For me, Lightwave has so many powerful tools that point by point or edge by edge modeling is just not necessary anymore.
As far as "feeling" goes. Thats in the artist. Points, polys and splines have no feeling of their own. How they are formed it a reflection of the artists feelings and skill. There are times when having a clay like way of modeling is an advantage. There are also times when having more precise ways of modeling is an advantage. Modeling a helmet to look exactly like some kind of reference is a situation where Im more inclined to model precisely as opposed to freeform. Splines are great for that.
Building things point by point or edge by edge is still IMO a very slow and tedious way to model when there are better ways to go about it. I dont know of any traditional artists who paint or sculpt a very detailed eye, then another very detailed eye and then slowly grow the rest of the head that way. Its a great way to make a lot of errors. Those errors become magnified quickly and just increases the amount of tweaking.
Let me put it another way. Let say someone knows how to box model, point by point model and spline model all equally as well.
Spline modeling is generally quicker simply because it uses the least amount of data. The truth of the matter is to know when to use each method. I wouldnt, for example, use spline to make a hand or fingers. I would use point by point for detailing small local areas but not as a general modeling method. :)

[wgreenlee1]

Well..we shall agree to differant methods but I still dont see where you are coming up with this moving 100 points around and point by point and edge by edge stuff.
The helmet I created..the second one took all about 15 serious moves and a few tweaks here and there it was done to that point.
I have all my extend tools hot-keyed and the rest are move,rotate,stretch and so on.
And splines having less data I still dont understand either.Sure the cage does but you still have to patch it and that creates data also.....or perhaps I'm missing something here?

[SplineGod]

Sure you can freeform model pretty quickly in Lightwave. Nobody will dispute that. Its not hard to bang out a helmet. For me Ive NEVER been in a situtation when doing work at a studio where I was allowed to just bang out something that I wanted from my own head. You almost always have model sheets for character, props, vehicles and so forth. You typically have to match them pretty closely. Many times you have to match them exactly.
That being said it will take longer to extend an edge and tweak every vertex to match some kind of contour vertically AND horizontally as youre growing your geometry. Lets say you have 10 points across your edge and you need to have 10 sections of polys up the rest of the model. Thats 100 vertices you will have to tweak. With splines I dont have to do that. I can create the contours with splines with a minimum about of points. The problem with modeling with SubDs is twofold:
1. When you bandsaw or knife thru your subpatch polys is that you really arent cutting thru the subpatch object but the control surface (the actual polys). This means that as you make cuts the geometry becomes flat and faceted unless you tweak your points out to get rid of the flat areas. This wouldnt be a problem if you could cut the implicit surface rather then the control surface.
2. The subpatch object (implicit surface) is only an approximation to the actual control surface (the polys). This means you have to add more cuts to get the shape of the implicit surface to more closely match the shape of the control surface. This means more tweaking. Spline patching allows you to easily get around those issues because the polys follow the shape of the spline much more exactly. That means less tweaking.
Heres an example of what I mean. To create this same shape using Subpatches and extending edges will take MUCH more time to do. It literally took several seconds to create this shape with splines, patch it and then turn on SuBDs. To do the same by extending edges and then tweaking will take minutes. Keep in mind that I know how to model point by point and poly by poly....I did it that way for years. I was around when metaform came to Lightwave and then metanurbs. Ive used and know pretty much every modeling technique in Lightwave. Ive simply learned from LOTS of trial and error over a period of years when to use which tools and when to use other tools and methods. I dont JUST use splines. I find that too many people disregard various other tools and methods (not saying you do that). The more you know the better off you are. Ive sat next to people who modeled pretty much only point by point and were excellent artists and could do some nice work. The majority of the time I could get things out much faster. Anyways, Im just sharing what I know and experience Ive aquired doing this stuff since before Lightwave came to the PC or Mac. :)

[iatepaste]

Looks like quite a debate is going on now but I just wanted to post my final solution to the problem.

I tried using a tessalated sphere but in the end, I ended up having geometry errors in the render even though it didn't show up in OpenGL. This occurred primarily after I tried to give the model thickness.

I ended up using splines. I do, however, still have frustrations when it comes to tight areas w/ higher poly counts (i.e. around the cutout area). To avoid some of the pinching I sacrificed the shape a bit by using bandglue to minimize on polys. I'll def. be taking advantage of splines a bit more in the future.

Thx again for all the help.

http://iatepaste.com/ganymede/Herb_And_Skills/ftknox/helmet_2.jpg

BTW, I was looking to build something similiar to a helmet so that's why the cutout area is so small here.

[Kaiser_Sose]

What about using one of those thickner plugins on the mirrored helmet to avoid those problems

[iatepaste]

I did. The dirty mesh was an issue with the tesselated sphere after using such a plug-in. I used Thickener.p on the spline based one and it worked perfect.

[SplineGod]

Originally posted by iatepaste
I did. The dirty mesh was an issue with the tesselated sphere after using such a plug-in. I used Thickener.p on the spline based one and it worked perfect.
I dont use thickener because its easy to do without.
You simply select your polys, hit the f key to flip them. Hit the c key to copy them. Hit the f key to flip them again. Smooth shift them slightly. Hit the v key to paste the polys. M key to merge.
Thats it.

[wgreenlee1]

But the edges wont line up properly will they?
Plus can you thicken up to 500mm like that Larry?

[SplineGod]

Originally posted by wgreenlee1
But the edges wont line up properly will they?
Plus can you thicken up to 500mm like that Larry?
The edges line up perfectly. I flip and copy the original geometry.
This will end up being my inside facing surface. I then flip and smooth shift the geometry. This creates the outside part of the helmet plus the edge. Last I paste in the geometry I copied and then merge. Its very quick.

[CGTalk Moderation]

This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.