View Full Version : Traditional artist and non-traditional artist approach modeling differently

06 June 2002, 06:45 AM
This is sort of an extension of the "Does Drawing Help Modeling" thread.

I've been thinking about this a lot lately:

In the beginning when I first decided to get into 3D(about 3 years ago), box-modeling was all the rage, and it just seemed so unintuitive from my perspective of traditional art background. In fact, all the tutorials I read seemed unintuitive and written by people who weren't practicing traditional art. It was mostly mathematical and lacked true insight into the human anatomy. I just couldn't identify with all the non-traditional artist's approach to 3D, and I just felt stuck and frustrated. So, I uninstalled my 3D apps and went back to drawing and painting. The only time I touched any 3D was when I had to apply textures to levels for games with MAX.

Then later, NURBS modeling was all the rage. That didn't feel any closer to traditional art.

Now, years later, more and more traditional artists are getting into 3D, and as some of the 3D artists from the past have become better traditional artists, modeling approaches began to change.

My current frame of mind is this: traditional artists and non-traditional artists think differently, and have different approaches. The methods I've been reading about recently all reflect traditional art background, and I felt right at home learning from those methods.

Two of the "revelations" that changed everything for me:

The Human Head Resources website(now a dead link)- This guy opened my eyes as far as considering the anatomy, movement of the facial muscles while planning your models. His detailed explanation of the strengths and weaknesses of different modeling approaches, and his analyzation of other well-known modeller's wireframes really helped me understand what was important in a model.

thehobbitguy's tutorial- this was THE tutorial that changed my life. Prior to that, all other methods just didn't "sing" to me the way his did. After doing his tutorial, I learned all the important things that I needed in order to approach my own models.

And now, after some experimentation, I've developed my own method, which relies heavily on traditional art skills(but you can get around the drawing part if you had good photo-reference).

My steps(in this case, dealing with a human head):

1)Draw the head in 3/4 view(or use a photo). For this stuff I use Photoshop, since is't far easier to correct and line things up.

2)map out all of your edgeloops, and major divisions on the drawing. Keep experiementing until you have a set that best flows with how the facial wrinkles and muscles move in different expressions. Do AS MUCH planning as you can during this stage, as it will save you tons of trial/error later. You can even devise clever ways to avoid triangles and 5-sided polys at this point(if that's important to you).

3)draw a front and side view(or use photos) of your face, map in your edgeloops and divisions(need to match them up correctly)and use them as image planes.

4)using thehobbitguy's method: create a polygon in the shape of either your front of profile(up to you), then start doing some basic dividing by splitting edges(following your drawings exactly. Then pull out all the vertices out into 3D along only one axis(except the vetices that makes the contour of your head).

5)keep splitting/tweaking by following your drawings.

Now, this method might seem ridiculous or unnecessary for some people, but for me, at this point of my learning curve, is the mostly logical and practical method for a traditional artist, especially when the topology of your face needs to be anatomically correct for animation.

Any thoughts?

Iain McFadzen
06 June 2002, 08:10 AM
I don't really see how one technique is closer to "traditional" art (whatever that is) than any other. To assert that it's only recently that traditional artists have taken an interest in CG, or that it's only recently that CG artists have taken an interest in traditional art, is a generalisation bordering on the ridiculous.

For me box modeling is as close to scuplting as you can get, where you start with a cube and remove all the parts which don't look like your model. I've tried various other techniques but, like you say, this is the one which "sings" to me. TheHobbitGuy's technique just feels too mechanical. I rarely use orthos though, prefering to eyeball my models, so I miss out on the strongest advantages of that technique anyway.

But that's just me, I wouldn't presume to imagine it's the same for everyone.

I suspect your experiences of badly written box modeling turorials had more to do with badly written tutorials than any inadequacies in the technique itself.

06 June 2002, 10:21 AM
I also have to disagree, to me box modelling is much more like traditional art forms than any other 3d method.

To me the esscence of traditional art is its trial-and-error nature, you can sit down with a piece of clay / blank sheet and create something with no planning at all (ok you may have rough ideas about what you're going to do) this is esentially the same as box modelling, you can start with a single 6 sided cube and create anything you want. ( it may not turn out very good however)

I think when you have to sit down and plan out your model/sculpture from the start you are losing the esscence of art, it becomes more of a craft (for both 3d and traditional art), but this is more something for the "is 3d art" debate.

06 June 2002, 12:30 PM
Iain- I think you might be right about that--all the tutorials I read back then were horrible. It was either very primitive, convoluted, or just plain laughable(in their inaccuracies). But nowadays, I see much better written tutorials popping up everywhere. It's a good time for newbies to get in now. :)

I didn't mean to say that only recently traditional artists became interested in 3D, or vice versa. All I meant was that it happens at a faster rate and more often now compared to back then, and that change has made an impact in the way 3D modeling is done(and of course, technology is another factor).

teacup- I think the reason why it feels more traditional to me is because I'm spending more time actually drawing instead of playing with polygons. Having a perfected drawing with all the polygon divisions already planned out feels like I've already accomplished the task by 50%. I mean, what I end up with basically looks like a wireframe, except it's all drawn by hand. Then to follow that drawing when modeling is just a breeze(relatively speaking). Drawing and painting to me feels like second nature, so it's easier for me to plan that way than to do it with vertices, edges, and faces.

I think I prefer thehobbitguy's method because I can get the overall look of a model very accurately and quickly with an established profile or front shape, and only need to worry about pulling vertices in one direction, where as in box modeling, I'd have to tweak in all 3 axis's, and no defined shape was established in the first place.

Maybe my case is unique because I'm such a weirdo. :D But, I definitely didn't post this to say that box modeling is bad in anyway. God knows there are plenty of kickass 3D artists that use box modeling. Just for me, my brain hurts less when I use a different way. :p

06 June 2002, 05:46 PM
interesting, as a fellow 'art student getting into 3d' I found the best method, for me anyway, was to create the model vertex by vertex, face by face(very simplified description of the method). I know this is a much less efficient method, but it helped me learn where things fit in the 3d world, I dunno why.

Iain McFadzen
06 June 2002, 06:35 PM
There is nothing wrong with that technique at all, you just have to go with whatever makes sense to you. As long as you end up with a good model nobody is going to give a damn how you modelled it.

It is certainly true that as 3D apps develop and the tools are refined they are becoming much more intuitive and user-freindly. The artists has to think less about the softaware and more about the art itself, which I guess, in retrospect, is kinda what Lunarique was saying in the first place.

06 June 2002, 06:48 PM
"As long as you end up with a good model nobody is going to give a damn how you modelled it. "

still working on that one too. :):)

06 June 2002, 01:39 AM
I think I was lucking coming from both art and technical backgrounds, my mother is sculpture/potter and my father is a draughtsman/painter. So I have already had both 3d (sculpture) and 2d ingrained on me.

I could see that coming from a drawing/painting background your method would be good, infact its quite similar to the way you model in autocad, always working on flat planes and drawing in the detail.

06 June 2002, 11:20 AM
modeled this after photos. built poly by poly by adding points to the mesh . works well for me . Btw iīm in 3d now since 2 years, before that i had a 14 year drawing and painting backgroundhead ( :hmm:

06 June 2002, 02:26 AM
Hey, that's pretty cool.

On one of the projects I worked on when I was art directing at a game company, one of the modellers built a model using the same way. I just provided a front and side drawing of the character, and he just plotted points all over and then connected the polys.

When you plot the points, do you just follow the contours of features, or do you have the divisions already planned out in your photo?

06 June 2002, 09:20 AM
i actually draw the points on the living model so i donīt have to do it later . i use an eyeliner for that itīs way easyer to remove than any magic marker.
this is where i started.

06 June 2002, 11:31 AM
Hehehe. That looked pretty funny.

Would it have been easier if you actually drew in all the edges too?

06 June 2002, 12:17 PM
i will sure do this with the next ones iīll do. this was my first head ever so i didnīt know exactly where to put the points.

06 June 2002, 01:33 PM
wyatt, that's the most ingenious thing I've ever seen.... I admire that intuition... and will have to grab the eyeliner dot idea for the next time I try a head. Thanks!

06 June 2002, 02:52 AM
Actually, this reminds me of when they showed behind the scenes for making of Terminator2. They had whatshisname being scanned, with all the lines drawn on his face. They just spun him around and scanned his head, and then used the data in some 3D proggy.

Iain McFadzen
06 June 2002, 08:00 AM
The only problem I can see with that technique is that you will be working with a fairly arbitrary set of points. Good deformations are heavily dependent on edge topology, and IMO good mesh topology needs to be built into the model from the outset. Nice flowing edges also make rigging a head much more straight-forward, which is important if you want to avoid falling out with guys further down the pipeline who have to deal with your models.

It also looks like it takes all the fun out of modeling, but that's a personal thing and completely irrelevant in a production environment.

06 June 2002, 12:23 PM
Interesting...I'm relatively new to 3D, but have always drawn and painted, and I find the vertex by vertex technique far more analagous to drawing. The box technique seems more like an engineering approach and I find it all but impossible - although, strictly speaking, I s'pose it is more analagous to sculpture.

Mr Wyatt - that is outrageous and ingenious!! If you see lots of people wandering about with pen marks all over their faces, you'll know you live near me.

06 June 2002, 01:19 PM
oh you mean scarey freaky xfiles guy?

07 July 2002, 02:35 PM
Im consider myself a traditional guy and box bomelling seems the closest to drawing/sculpting. I hate working with bitmaps in the background because it removes the fun out of the modelling process for me. Kinda like when you trace a drawing or rotoscope. Then again i'm just newbie so what do i know :D

07 July 2002, 03:01 PM
well not much it seems, at least about the filmindustry where i hope to get to work in as soon as i finished my studies at university. what i mean is not that you know nothing and that i disrespect you but there are in fact issues in 3d that need god reference to make it believable. as for visual-fx you have to replicate the realworld so good that nobody notices. you canīt do that from scratch, and if you could you wouldnīt have the time for it, so you use reference and so the modeling goes quicker. the funpart in making a good ( but fast modeled) default-grey model look real comes afterwards while texturing, lighting and animating the bastard. modeling is only one step to the final result, but itīs the first step and therefor a very important one. It all depends on what you want at the end. if one wants a realistic head or something like that, then saying that using reference takes the fun out of modeling is a bit, ... letīs say : " narrowminded"
by the way visual-fx like in fightclub would have been impossible without using a big pile of reference-pics.
look here (


Iain McFadzen
07 July 2002, 03:28 PM
Yeah that is true to a certain extent, but you can use references without necessarily tracing a pic in the viewport or drawing dots all over someone's head.

07 July 2002, 02:51 AM
Well, for me, my models has to look like my paintings, so it's kind of important to use image planes and model that way. I would say that the entire 3D production thing is a lot less fun to me than coming up with the story, writing the screenplay, doing concept designs..etc.

If all 3D artist did thing in the "fun" way, production will grind to a halt. :p

I suppose it's all about perspectives. For me, bringing the screenplay to life on screen is the most important thing--but that's from the perspective of a creator.

For some, doing the modeling/animating/lighting..etc is the most important thing--from the perspective of a visual artist.

07 July 2002, 05:40 AM
Originally posted by MrWyatt
i actually draw the points on the living model so i donīt have to do it later . i use an eyeliner for that itīs way easyer to remove than any magic marker.
this is where i started.

But try this on a cat huh? ;)

07 July 2002, 08:19 AM
well cats are different i think. cats, unlike humans, are a completely other species and therefor easyer to fake, even when everybody knows cats the human eye still stylises what it takes to find a cat convincing, so it may be possible to build a cg-cat where every human goes: " wow, what a realistic cat, thought it was a real one. cg? you must be kidding." and so on, but if cats could speak they would probably say something like : " gosh this cg-cat sucks big time. "
i think we can easily model everything so that it looks convincing to us, everything but humans, because we are humans and that is the species we know best and recognize the tyniest mistake. ok what i really wanted to say is : you donīt have to paint dots on a terrified cat to model it realistic a bunch of photos even from different cats should do the trick. all i say is that i do have fun modeling human heads like i do and iīm fast too, thatīs important for me because i want to do visual-fx someday after my studies, and i know that you donīt get a job when youīre
1. slow
2. inacurate
sorry if iīm such a chatterbox but sometimes i feel missunderstood.:p

07 July 2002, 04:10 AM
It makes me remember of an image a teacher have shown to us at school. It was in the 3D historic lesson and he showed us what they did on Terminator 2 for the T1000. The actor (cant remember his name) was running naked (for the execption of his boxer) in the street and his body was painted with all the wire frame!! I found it pretty funny at the time but there was worst. Remember Tron from Disney! The 3D is all a big .txt!!! All writen! The make a render a every 2 week because it was too long! They were writing the position of each point manually! Must have been so painful!!! (that's for the kind of huge ship that I cant remember the name. . .). So I think THAT was inintuitive!!;)

07 July 2004, 06:54 AM
I dont think its really so important to compare the workflow of modeling in 3d with classic arts workflows like painting, sculpting etc.
As far as I see it the most important thing is that the workflow is as intuitive and pleasant as possible. If it is it can be picked up by anyone and will be nice and logical to work with.
I think 3d modeling interfaces and workflows still have a long way to go for that.
So far I liked the box modeling aproach better because its more fun than the others and I want 3d modeling to be fun and not just a pain in the ass.
As painting can be fun to some and sculpting can be fun to others.

07 July 2004, 08:06 PM
In my view, the closest modeling way from traditionnal arts is box modeling. Starting from a simple cube, I detail and detail and detail the mesh until I get the good shape. This step by step way of doing comes from traditionnal skills.

That works well at drawing or sculpting, but their is an extra parameter in 3d (supposing the model is going to be animated) : the topology. The HobbitGuy method (which I do not like at all) is an extrem, all the topology is planned before the model is shaped. Another extrem would be to shape without caring abotu topology, and then draw edges loops. I think we have more or less a foot in each extrem.

Here is the way I model : First, as I have a little idea of how my model will be, I place many helpers I need. A rough of skull made of a sphere, two ears sized, ... With extruded planes, I place to main lines (eg : sterno-cleido-mastodian muscle of the neck). Then I start from a cube, and begin to model generous shapes. If I do not like the Hobbit Guy method, that is because I fell so confusing to have a detailled mesh (with to much vertices) unshaped. I can not manage to place all first and move the vertices then (that is the most time consuming method to me, and the result won't be great).

Materials : I have drawn on a sheet of paper with different colors a front view of a human face with the main topology edges orientations, fromsome threads readings online and from my own experience. I have some photographs of face from fashion magazines.

About having a refs in 3d views, I do not really use it. My first models were made without it (I though it wasn't "gentle" to model using them). But yeah, it prevents time comsuming to have a face profile in the side view. What is interesting me is working on the model, shaping. I love the shape, how the shape evolves . So I prefer turn around the model until I spot something I could improve (proportion or character style).

Tell me how you model, and I will say you what kind of artist you are.
Yeah, it could be a good question about the traditionnal artist / CG artist issue. I would add/say first that the word artist is really hard to use in french. It makes references to the ability of the human to create something beatiful... but what is beautiful?... is their something which is beatiful to everyone? ... Seldom are the cases this word is right used.

Talking about technic...
Being an artist (enlgish speaking :) ), I like to switch from art to technic and from echnic to art. I like to sculpt, and MrWyatt was presenting a well known method to get a real model into 3D. I am studying a cheep/inHouse method to scan real models in to 3D using a webcam, a lazer and a turning table. From a given rotation angle, you get a range of images from your model. Then, thanks to a more or less complicated script/software package, the images are analyzed to build the mesh of your model. The only problem is that mesh is made on million of polys, and is crappy on some areas (nostrils, ears, eyes). I am working on making a melScript for Maya to "topologize" this heavy poly model....

07 July 2004, 06:40 PM
Well, I have to say that you need to know whats there and what connects in order to make it believable in a model. Even if you so call "draw from the imagination" your really drawing from memory of certain images youve seen beforehand (so your the reference library). There is no shape in existance that doesnt exist in the real world. One thing that helps me alot is to put primitives as guidlines in the model so that I know exactly where to wrap a polygone around a 3d form. Think of it as using construction shapes in traditional media. It really helps a lot to do so.

07 July 2004, 01:22 AM
I just read the first post from Luna. And i think I work out the same way as you are. There are little different in between. However, I don't quite agree that traditional and non traditional atrist approach differently.

Its more base on his or hers habbit of modeling. Personally, I more concentrade on how the topology is formed. So i work in a little bit on together:

1. draw out the topology of what i think it should go on another layer of the reference picture.

2. follow the layout to in 3D spaces. Continue working to form a better topology. By that time the shape of my 3D model is probably off.

3. go back and follow the topology to redraw the new layout on top of my reference.

4. Rotoscope my new drawn topology in 3D space.

This is how i model human heads. I start with a box most of the time, but in the middle of the process, i'll cut out parts (nose, eye, ear...) to model it separately.

ps. Sometimes i feel guilty on the rotoscoping part.

07 July 2004, 03:41 AM
hehe, interesting topic.


Spline modeling can be very intuitive. as far as I find.

you 'draw' the shape you want using spline in 3D space first, and then worry about connecting them to create surface.

that way you get the shape you want instantly..unlike box modeling.


07 July 2004, 04:56 AM
I agree about the spline modeling. Splines are easy to layout quickly and they define form and flow with very little data. I use splines quite a bit myself in Lightwave where theyre geared towards laying out polys. From there I turn on SubDs and continue on.

07 July 2004, 10:00 PM
I suppose it's all about perspectives. For me, bringing the screenplay to life on screen is the most important thing--but that's from the perspective of a creator.

For some, doing the modeling/animating/lighting..etc is the most important thing--from the perspective of a visual artist. Interesting thread, there are definately differences between those groups.

My main use of 3D is as a storyteller, but that doesn't force me into doing tight sketchs and tracing them in 3D. I'd much rather have a good script, loose and effective sketchs/thumbnails and then tigthen them up in 3D. I use references all the time, but (as per Iain's point) I don't put them into the viewport and trace them.

I started doing computer animation as a sideline at my traditional Bauhaus style art school (OCA). Once spline modelling hit 3ds, just after max1 came out, I felt more comfortable doing some of the in 3D -- even more so once Bay Raitt's modelling style got popular and well explained (his video, with Irfan's and Matt Clarks were real eye openers into that technique). Now I'll do loose sketchs and tighten them up in 3D -- on the horizon maybe concepting in Zbrush would work (for characters) :).

A friend of mine who went through art school at the same time, but didn't start to do computer animation until the last year and a half, is bound to a '2D then translate' school of thought.

It's just a question of the tools that you can be comfortable with creatively. If your (self)education has your creative bound up in 2D stage then a very technical 3D process would be fine. If 3D was a large part of your workflow while you were being (self)educated about the storytelling process then you probably want more creative freedom at that stage... even my animatics tend to rough 3D once past the thumbnail stage.

It'll be interesting to see how things go with people who've had access to 3D tools since they were kids... Something like ZB2 to sketch characters would be great in the hands of a talented 10 year old, and that person who would be thinking 3D naturally for their whole career.

07 July 2004, 06:37 AM
Something like ZB2 to sketch characters would be great in the hands of a talented 10 year old, and that person who would be thinking 3D naturally for their whole career.
Some of the recent concept stuff I have seen from Glen Southern is amazing! It is such a cool concept/ look (2.5D sketches)

07 July 2004, 05:46 PM
So true, that material is just excellent.

Even the Zsphere aspect can be interesting; what Pixolator did here ( ) is pretty expressive/playful.

CGTalk Moderation
01 January 2006, 10:00 AM
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.