3d Modling Question

[OrbWeaver]

After that, getting a good likeness of anything is just art, so you should look for information on life drawing tutorials or something.

 

I don't think life drawing techniques would help in 3D modelling. Two-dimensional drawing requires looking at things in a different way - seeing the lighting and shading and 2D shapes, so you can draw them on flat paper as you see them rather than as you think they should be.

 

This is completely opposed to the technique of modelling, where you are working with 3D shapes with no reference to the 2D appearance.

[sparhawk]

Life drawing helps a lot, because in most cases you want to have a clear image of what you want to model. I strarted to learn drawing a bit just because of that.

[OrbWeaver]

Life drawing helps a lot, because in most cases you want to have a clear image of what you want to model. I strarted to learn drawing a bit just because of that.

 

Are you referring to drawing an idea you have in your head onto paper before modelling it? That is certainly valuable, but that is not what I understood by Domarius' reference to life drawing.

 

Based on the few art classes I took a long time ago, life drawing is the technique of looking at a scene you have in front of you, such as a bowl of fruit, and rendering it onto paper. This is quite difficult for many people (including me), because you have to project your 3D view of the scene onto a 2D image, which requires you to disassociate yourself from the objects and symbols in the real-life scene and draw exactly what your eye sees, i.e. areas of colour, darkness and light.

 

Modelling, by contrast, requires you to consider the 3D structure of the object and duplicate that, irrespective of how the particular view of the object appears to you, or how it is lit or shadowed.

[Springheel]

Life drawing will teach you about various muscle and tendon groups; what you can see through the skin in various poses and how the muscles bunch when compressed. Not to mention valuable observations of proportion, balance, etc. I would imagine most of those things are useful when modelling.

[OrbWeaver]

Life drawing will teach you about various muscle and tendon groups; what you can see through the skin in various poses and how the muscles bunch when compressed. Not to mention valuable observations of proportion, balance, etc. I would imagine most of those things are useful when modelling.

 

Oh, that sort of life drawing.

 

Yeah, that will definitely be useful for modelling characters etc.

[Hewer]

When drawing from life, you also have to get a good grasp of what your subjects 3D form is like in order to render it in 2D. From the many life drawing classes I've had, I know that moving your head 1/4 inch, or especially after you (or the model as the case may be) take a break, the entire composition can change dramatically according to your new point of view. Now the chin is 2 inches off the shoulder instead of almost touching it, or now you can see half of the orange instead of just the top of it. You have to have a really good grasp of your subject in 3 dimentions in order to adjust for these changes because you're never just copying exactly what you see (as in drawing off of a photograph), you're drawing what is physically there in 3 dimentions.

 

 

I hope I'm making sense here...

[FishFace]

In my experience, modelling talent is more important than artistic talent. As long as you are reasonably competent in the latter, you can produce reasonably competent work, as you can take good references and to an extent, push points around until it looks good. In that sense, the artistic sense of knowing when something looks shite and more importantly why it looks shite (and what in hell to do about that) comes in handy.

 

However, you can know exactly how to represent something in 3D, but if you don't have the modelling talent, there's no way in hell you can translate that into a model in any sensible time and with any sensible poly-count. (Note that with a high poly-count, not only does the model become unusable in games, but it becomes difficult to handle in the application, too.) For example, in a high-poly modelling situation, you might want to add a wrinkle to a face. It takes great skill, such as which I do not have (he said, from experience) to add one in such a way as to not a) make the model unweildy, make it look weird on other places and c) make the wrinkle look realistic in the first place.

In ones artistic imagination, it's quite trivial to simply push that bit in, and give this bit a tug, or on paper to darken an area, but the same operation will usually need to have been planned from the very beginning in the case of a 3D model.

[Domarius]

OrbWeaver I'm speaking from plenty of personal experience, not hypothesising. FishFace, your example of modelling talent being more important than artistic talent only pertains to you. Neither is more important thant the other. Someone with no artistic ability would see it the other way around.

 

Modelling is like drawing - there is some planning involved but you can always change things, especially the smaller things. Start rough, add detail.

 

Modelling is like drawing - knowing how to use your tools will only get you so far, after that it's about being an artist - having an understanding of the underlying muscle tissue and bone, how each body part should look at different angles - when you model you are constantly flicking the view around to see what things look like at certain angles, and on some it looks bad, so you fix it.

 

When I model, I only sketch a front view for a guide for proportions, no side view like the traditional way, because the rest I do by eye, by constantly rotating around and looking at it to see if it looks like a human body. I have tried many times using side views in conjunction with front view, and side views just get in the way.

[oDDity]

I don't sketch at all when doing models. I tend to work on the fly with no ridgid plan. Models are so pliable and easily fixed that I don't need to worry about making mistakes. I did do a lot of painting and drawing for many years before I ever touched a computer, but I wasn't anywhere near as good at it as I am at CG only after 3 years.

Modeliig is a bit of a misnomer here, since I'm as much a designer and texture artist as a modeler, as are most good modelers, so it really takes in all three arts of sculpture, painting and draughtmanship, so a background in 2d art defintiely helps in 2 of those areas, if not all three.

[OrbWeaver]

I guess you could say that 3D art skill is modelling skill multiplied by artistic skill. It doesn't matter if one of them is mediocre as long as the other is high, but if one is zero then your work will be crap no matter what.

[FishFace]

I'm certainly crap at texturing. It doesn't help that Blender isn't too hot on the texture front - I found C4Ds method so much more intuitive, because it was uniform and extensible in a uniform manner. But the real problem is that I suck at creating the textures for, e.g. a face. No matter how good your model is, it'll always look CG unless you get a decent texture on the thing.

[Unskilledlaborer]

I have been working from a tutorial and have seen a few people's Blender work, faces and heads and human stuff like that . many of them are using VERTEX PAINTING(if that's correct term). Is this compatible with doom. they have some good stuff. I know that most games use skinned objects (skinned with a 2D paining) so i was just wondering what is the standard.

 

what are some things i should know about when i start a model?

 

ODDITY do you put in small details like nose holes and stuff like that?

 

how many FACES (term?) are you limited too in games like doom ?

[Unskilledlaborer]

that screen shot is my work just so you know!!!

 

not that anyone cares but i am proud of my little head thing...wife doesn't care so I tell you all...

[sparhawk]

It looks much better then my head attempts.

[OrbWeaver]

You can use vertex colours in Doom 3 if you use Goofos' export script to export to a .ASE. I don't know whether the LWO exporter works or not.

[oDDity]

I never try to paint face textures, I always use real photos as a base an modify them.

As for vertex painting, forget it - that's equivilent to dipping a rat in paint and getting it to roll around on a canvas. It's useless for anything short of terrain blending.

You need a paint program like photoshop or GIMP (it's free) to paint the textures after making a UV map.

You also need to generate normal maps. This is not the place for an impromtu tutorial.

 

ODDITY do you put in small details like nose holes and stuff like that?

 

how many FACES (term?) are you limited too in games like doom ?

 

NO you don't need details when using normal maps. IN fact, details tend to mess up the normal map and make the model look worse.

 

My characters range from 3500 to 5000 triangles for doom, but then they're a lot more complex, and all multipart meshes than the simple one-piece Doom 3 monsters.

[gleeful]

My characters range from 3500 to 5000 triangles for doom, but then they're a lot more complex, and all multipart meshes than the simple one-piece Doom 3 monsters.

 

are you serious?

 

i've made streetlamps with more triangles than that!

 

kind regards

gleeful

[oDDity]

Are you serious, becasue if you are you're going to have to alter them. 1000 triangles is pushing it for a street lamp. You're obviously not making full strategic use of alpha maps and normal maps to fake detail, but are rather modeling it in.

[gleeful]

Are you serious, becasue if you are you're going to have to alter them. 1000 triangles is pushing it for a street lamp. You're obviously not making full strategic use of alpha maps and normal maps to fake detail, but are rather modeling it in.

 

i'm trying to cut back on polygons, i really am.

 

it's those darn round forms. spheres and circles (also cushions on couches btw) - haven't found a way to fake them yet. any they really explode the polygon-count.

 

as long as it's square i'm doing fine.

 

kind regards

gleeful

[OrbWeaver]

Remember that Doom 3 automatically smooths models. For round shapes you can get away with 8 or 9 sides, or even as few as 5 if you don't really see the end point.

[oDDity]

IT doesn't auto smooth them, you set the smoothing angle in your 3d app.

THe problem is always the edges or rounded objects, they always look like shit. Especially something like a plate, if you want to make a plate look like a smooth circle from a top view, you really need a hell of a lot of polygons, amd no amout of smoothing will help it. You could use an alpha map to make it look perfectly circular, but then you can't give it any depth.

[OrbWeaver]

IT doesn't auto smooth them, you set the smoothing angle in your 3d app.

 

It does if you use ASE models. For LWO things may be different.

 

THe problem is always the edges or rounded objects, they always look like shit. Especially something like a plate, if you want to make a plate look like a smooth circle from a top view, you really need a hell of a lot of polygons, amd no amout of smoothing will help it. You could use an alpha map to make it look perfectly circular, but then you can't give it any depth.

 

True. If the model is actually meant to be round when viewed by the player, then you need about 32 sides to make it look convincing.

[sparhawk]

Smoothing is a vertex feature, and doesn't have to do with the angles. It also only defines how the light is calculated between three vertices. If it is not set to smoothing, then the face is a flat colour, while with smoothing it is, well, smoothed.

[Fingernail]

yes but you need to set those properties in your 3D app.

 

Lightwave works on angles, Max works on "smoothing groups", which do not correctly import to D3, so it's best to a) use lightwave and don't bother with setting smoothing groups, just unweld points you want to have flat edges.

[sparhawk]

There is no such thing as a smoothing angle on vertices. This is a toggle. You either set it on or off, and that's it. You can look up the OpenGL standard to check it out yourself if you don't believe me.