Sign in to follow this  
v0dKA

Best Way to Model?

Recommended Posts

I've seen a couple different ways of modelling, and I'd like to know which is the most efficient. The method I'm most comfortable with is the one I've been introduced first. Basically, set the picture you want to model into 3D into the background of the modeller, find strategic vertices in the front view, and set their appropriate depths in the right view. Then you "connect the dots" to form faces. I learned this in a MilkShape tutorial, and it's my preferred method. In a blender tutorial, there was a slight variation to the last method. Basically the same thing, except all with bezier curves. I'm more comfortable with the first method, but I didn't find a way to do it in blender (plot vertices instead of bezier curves/polygons). So firstly, I'd like to find a way to plot only vertices in blender, and then connect them to form faces. For some reason, I found plenty of tutorials on this method for milkshape, but only one tutorial for blender. Is there a different preferred method for modelling in blender that I'm missing? Also, I keep finding tutorials about how to model from nothing in blender, but, again, only one on how to model from a reference picture. Am I getting the wrong idea about modelling in general here? How exactly do you model? Don't you use reference pictures? Or do you model from nothing, i.e., fire up blender and make shapes without any reference pictures?

Share this post


Link to post
Share on other sites
Depends on how well I know what I want to model. Fantasy things, such as spaceships or robots, I can usually whip up with minimal sketching. But if I want to model a complicated object I will try to sketch it out and get it "fixed" in my head before beginning. I usually use the so-called connect-the-dots method if I am modeling something that I want to have realistic proportions.

Otherwise, I can just fudge it. Really, it doesn't matter -- As long as the result is what you are looking for. There are various way to tackle the actual modeling itself, such as box modeling or spline/patches. I'd just use what comes natural to you and generates the best results. Don't hesitate to try new things and give them a chance before dismissing them as unusable.

Share this post


Link to post
Share on other sites
I'm not a modeller, but I see 'connect the dots' as a pretty slow way of doing things. I thought most modellers start of with some sort of primitive like a sphere, cube or cylinder and stretch and split faces to form whatever they want. This is the kind of method taught by Wings 3D.

Share this post


Link to post
Share on other sites
I don't know if Milkshape or Blender have the same tools available, but the way I work in Maya is to create a single polygon face in side view which is the outline of the head (this is just an example), it'll have as many verts as needed to get a properly detailed silhouette, but it's just a single flat face. The next step is to make a split through the face - for example, from nose ridge to the back of the head, following an approximation of the natural curves of the head. Once you have that first split, you select those inner verticies (i.e any that are not on the outer edge) and pull them out to where they should be in front view. This is an excellent way of quickly blocking out the shape and is very fast as well once you're used to it.

I don't like the sound of this connect the dots things...sounds cumbersome. For me modelling has to have a flow to it. If you think too much you get lost and slow down. I treat it like sculpting in a way. It should be intuitive.

Hope that helps.

Share this post


Link to post
Share on other sites
Guest Anonymous Poster
i'm a 3D STUDIO MAX User ...

to do the join the dots method in max you have to make a box first .. just so u can get into vertex editing mode.. :/

as for the best technique for modelling... thats the one that works best for you..

i try to learn many techniques..and hopefully master them..

the best way to make low poly heads is not the best way to make hi poly heads, and so on and so forth...

in max i use
spline modelling . box modelling . edge modelling . join the dot ....
don't be afraid to mess around with new techniques.. i would reccommend getting one technique down first though and then building on that...

...style

Share this post


Link to post
Share on other sites
usually the most efficient method is the one you're most comfortable with. That said, connect-the-dots is a horrible way to do. Yes, you can model using that... but I honestly feel sorry for you because you went and learnt how to model that way. It's harder to switch now.
This is actually the reason why I never liked Milkshape. I prefer to create my own triangles/quads on the fly or do box modelling.

Box modelling is where you first create a box then use all sorts of tools to change it into something (usually) completely different.

The way you model in say, metasequoia is closer to connect-the-dots as in milkshape, but not as tedious. Basically you create the triangles by clicking where the vertices go. (if you click an existing vertex, it will use that) So rather than first doing the points and then the faces, you'll do it all at the same time. It's a lot faster... especially in application as intuitive as metasequoia. But this "method" doesn't work so well in 4-window layout.

Share this post


Link to post
Share on other sites
Wow! So you guys make models right out of your head?

Call me weird, but I prefer connecting the dots [smile]. What if you had to model something complex, like a car, for instance? Wouldn't you want to put it in the background of your modeller and plot its points? Take this pic from a tutorial, for instance:



That specific tutorial can be found here.

Would something like that be done right out of your head? Out of a heavily manipulated cube?

Maybe I am just weird... What do you think?

Share this post


Link to post
Share on other sites
Guest Anonymous Poster
depends...
reference material is used by most... even if it is just their own quick sketches scanned in..
photo's are used as references for photo-realism ..
if your doing fantasy, sci-fi etcetera that just isnt possible.. there are no photo's of made up things?...

i personally use reference pictures.. sometimes just to get the base proportions right, and then i do my own slant once the foundation is there..

it's a style thing overall..

and as far as join the dots go.. it is not the best way to model quickly, though it is a good way to make sure you fit in details such as noses and eyes on the face.. i wouldnt use it for making a car mind.. that could be done way quicker extruding and cuttin poly's (box modelling) ... if i was going hi-poly car i would think about patches as they can do supersmooth curves that aren't to hard on the processor, thing is patches are messy once u want to go into angular details.. there are pro's and cons to each method... some are quicker, some more accurate.. more often then not i mix and match techniques on the same model

i would reccommend u continue using reference pictures as they are a good help starting.. even pro's use em.. though they don't rely on them as a blueprint in general, more a suggestion..

hope i have been helpful rather then patronising... :/

Share this post


Link to post
Share on other sites
You know how to sculpt an elephant, right? You start with a granite block, and you cut away anything that doesn't look like an elephant!

I actually do this: start with a big box or other simple shape, and cut out anything that doesn't look like what I want, using CSG. Sometimes, I build the pieces I want to cut out using CSG in turn.

At the end, I end up with a very low-poly outline of the shape I want. This can work as the coarsest LOD, and/or collision mesh. Then I start subdividing edges or faces and using soft selection to push/pull vertices around to make the shapes rounder and better matching to whatever I'm trying to accomplish.

I'm not a great modeler, but I thought I'd post this method, as it seems pretty different from what others have posted in this thread.

Share this post


Link to post
Share on other sites
I guess it also depends if you're trying to create an exact reproduction of something.

If you wanted a precise model of that car that had to be EXACT, it would be silly to not use a reference and place everything precisely according to the reference.

But if you don't mind a bit of creative liscense, you can just take an idea and run with it. For a recent project at Uni I made a car basically from a box just based on a perspective painting of one. (http://plang.lawngnome.org/car if you're interested). More fun that way.

If join the dots works for you then go for it, but as some one pointed out - since that method becomes harder to use when dealing with fantasy or made up things, try some box modelling too to compliment your current skillset.

Every approach has a different pros and cons and is most suited to different situations I guess.

Share this post


Link to post
Share on other sites
Quote:
Original post by hplus0603
You know how to sculpt an elephant, right? You start with a granite block, and you cut away anything that doesn't look like an elephant!


Hmmm, that sort of goes against how people think. People think positively - what makes up the elephant? I doubt the thought "What doesn't make up an elephant?" crosses their minds much [smile]. But then again, it's just me. And I'm known for utilizing rather... unorthodox... techniques....

Anyways, seems my thinking pattern goes a little against the "normal" people's techniques. I'll most certaintly give try these new techniques a try (once I figure out how to #!%#@ use blender!!)

Say, anyone know any alternatives to their $60 manual? Sorta short on money....

Share this post


Link to post
Share on other sites
As has been said, go with what your comfortable with, although you method certainly isn't the most efficient way to do it. ;)

I prefer to model from edge extrusion (MAX user here). Normally I just sketch out a spline for the foot, extrude that upwards and scale the top verts to create the shape of the foot/shoe, extrude up to the hips (Scaling on the way), do some welding and rotating at the hips, mirror and merge, then work with the torso from there, cleaning up verts here and there as I go.

Most other people I've seen have been horrified with my technique, but it gets results at roughly the same speed as box modelling and it's what I'm most confortable with. Personally I've never liked box modelling, as it takes me longer to try and remove the 'boxy' look than it does to actually model in all of the edges from the start.

Quote:
Wow! So you guys make models right out of your head?


I can if I really want, but I find it a lot better to model from quick sketches because it's a lot easier to spot proportion issues on a 2D drawing than it is in 3D, for me at least. Otherwise I seem to spend the last 1/4 of my modelling time trying to fix up small shoulders, large hips and monkey-length arms.

Share this post


Link to post
Share on other sites
Point modelling is really quite unefficient.

Box modelling is the most used technique, you should be able to google for tuts, there should be lots.

The problem with dots is that you will need the images in the backround of the object you're modelling. You can live with that if you're modeling cars and such but it should be problematic if you want to model something out of your head (Fantasy/SciFi stuff or whatever you can't find blueprints for).

Quote:
Call me weird, but I prefer connecting the dots . What if you had to model something complex, like a car, for instance? Wouldn't you want to put it in the background of your modeller and plot its points? Take this pic from a tutorial, for instance:

Bad tutorial, you would usually do cars with splines/patches, having the lines flow correctly in all views, in that tutorial you just get a flat 2d side of the car. I'd google for better tutorials.

Box modeling is not necessarily the best choice for cars and other objects with smooth round surfaces, you'll be better off with splines/patches there. For anything else, box modeling should be the fastest and easiest way to go.

Share this post


Link to post
Share on other sites
check this site!: http://www.onnovanbraam.com/
You'll find some tutorials on creating cars&stuff but you'll also find a huge database of blueprints(cars, planes, even humans).

Hope this helps..

Share this post


Link to post
Share on other sites
I am no 3d Artist, but I find that the best results are achieved by using the same technics used when drawing, that is, set your reference planes, use a sphere for the head, a cylinder for the torso, arms and legs, shape those basic objects into a character with point edit and the help of your reference planes.

I find the connect the dots technic to usually generate models that are highly detailed on the side view but poorly detailed on the front side due to its "cookie cutter" nature.

but thats just me. [smile]

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this