WARNING! This Making of Contains Nudity.

Software Used :

ZBrush

The original idea for this image came out of the references I’d collected. Whenever I start a new artwork, I do a lot of research and then I start the modeling (Fig.01).

Fig.01

With these references, I blocked the model with a simple base and I did some primary tests with the composition of the model (Fig.02).

Fig.02

As the idea matured, I really needed some specific references for the towel on the head and how its drapery would work. So I took some pictures of a relative of mine doing the exact pose with a towel (Fig.03).

Fig.03

I used the same real-life reference for the towel in the model’s hand (Fig.04).

Fig.04

Picture 1. The end result could look for example like this. Notice that this is a stylized head which is not exactly realistic. For example the eyes are too big.

Human head is one of the most difficult subjects to model believably. The structure and form of a human face is so familiar to everyone that even a small fault in the 3D model is easy to notice.

This tutorial covers the 3D modeling of a human head in 3D Studio MAX using polygons. The head is modeled roughly using polygons. The head will get its final form when Meshsmooth modifier is added to it. When the Meshsmooth modifier is added to it the face becomes a so called subdivision surface. Subdivision surface means a surface which is created by dividing the original surface into smaller polygons. Meshmooth modifier does exactly this. This way the rough surface can be made rounder and smoother. The end results is a head similar to the one in picture 1. Notice that this is a stylized head which is not fully realistic. For example the eyes are too big and the ears have been simplified. However, I believe that this model could very well work as a basis for realistic geometry of a human head.

Topology, Edge loops and Poles

Picture 2. In the picture above you can see the polygon construction of the model. By seeing the construction of the model you can conclude how well the model is suited for animation.

Picture 3. The picture above is rendered so that you can see the Isolines of the Meshsmooth modifier.

In picture 2 and 3 you can see the topology and edge loops of the model. By seeing the structure of the model you can conclude how well the model is suited for animation. Edge loop means a row of edges that goes through successive four sided vertices. Four sided vertex is a vertex where four edges meet. In my opinion, in an ideal model that is made for animation edge loops are parallel or perpendicular to the movements of the muscles.

There are almost only four sided polygons (quads) in this model which is good if the model is going to be animated. One should always aim to construct the model from four sided polygons when modeling a character. There are naturally three sided and five sided poles in this model. A pole is a vertex where less or more than four edges meet. Pole ends an edge loop and might sometimes cause trouble in animation. If you want to know more about the subject I recommend reading Bay Raitt’s instructions about 3D modeling and edge loops. Another great resource where to find more information about edge loops, head modeling and 3DS MAX is www.cgsociety.com.

3D Studio MAX Tools

Picture 4. Low polygon version of the human head. This is how the complete 3d model will look like before the Meshsmooth modifier is added.

Before starting the 3D modeling one should be familiar with the following concepts and 3DS MAX tools:

• Editable poly
  • working in sub-object modes
  • copying polygons
  • weld
  • cut
  • extrude
  • create
• Symmetry modifier
• Meshmooth modifier
• Basics about materials

Reference Pictures for the 3D Head

Picture 5. Reference pictures provide a lot of help when modeling a human head.

It’s best to use real photographs as reference pictures when modeling a human head but if real photos are not available you can download pictures below.

Create reference pictures in 3D Studio MAX according to picture 5. Change Left view to Right view. Create two plane objects whose size is 512 x 512 (the same as the pixel dimensions of the pictures. It’s relevant that the proportion between width and height is the same as in the pictures) and add textures to them. Set the Self-Illumination of the materials to 100 (material will become self illuminating and therefore is easier to see).

• Human head reference picture: right
• Human head reference picture: front

Prepare the Views

Turn on Smooth + Highlights (F3) and turn off grid (g) in all view ports. Set the width and height segments to one in both plane objects. The head is considerably easier to to create when we can model according to the reference pictures in the background.

Page 2: 3D Modeling a Mouth in 3DS MAX

Software Used :

ZBrush, Maya

Introduction :

I am a big fan of military topics. Since I had already done a soldier, this time I wanted to try a World War II German general.

Refrence :

Because this character was based on the real world look-and-feel, I didn’t draw any concepts for it. Therefore, I needed a ton of references instead. I found an old man reference from 3d.sk which I decided to use for my character; the reason was simply that it was a perfect reference to practice my ZBrush sculpting skills with and would also fit my character pretty well (Fig.01).

Fig.01

Modeling :

I created a plane in ZBrush with the reference image as the texture. Then, I started to model the head from a sphere. It’s much easier to get the right proportion if you have an image plane (Fig.02).

Fig.02

After I’d roughly done the sculpting, I exported a high-poly OBJ file to Maya and started to do re-topology. This is a very important step if you want your model to be used for real production. Modeling is only the first step of the production process; there is a lot of work after that, such as unwrapping, rigging, animation, etc. If you can keep your wireframe clean and appropriate, it will make everyone’s life much easier. I used a Maya plugin called “Nex”, which is a very nice tool to do re-topology with; using this plugin will keep your polygons in quads and evenly-spaced easily (Fig.03).

Fig.03

After completing the re-topology process, I exported another OBJ file and went back to ZBrush. With clean edge-loops, this time I was finally able to work on the details. There’s one thing I’d like to point out at this point: each area of the face has a particular skin texture. It’s important not to use the same texture everywhere because it will look unrealistic (Fig .04).

Fig.04

With this model my goal was to learn how to use normal mapping for a low poly character, a technique that was new for me in that moment. The model Arkeon Sannath it’s from a game called Confrontation.

Modelling:

Probably cause I’m more used to working in high poly than in low poly, I decided to started modelling the high poly version first. Compared to other models, this part was a bit difficult, basically due to the small size of the reference that I’m working from. The real figure is only a few centimetres high, so it was difficult to see all the details. The modelling pose of the arms was really strange, but it was the closest to the original pose to the real reference. The big shoulders didn’t allow for much arm movement, so I decided to model it this way to avoid problems.

I modelled most of the parts as separated meshes, and doing nothing really special with it. After all the poly modelling done, I only had to add some chamfer edges in parts to keep the hard edges when turbo smoothing.

Once I had the high poly version finished, I cloned it and started to delete some edges and lots of the bolts and bumps to keep the poly count as low as possible. Really I didn’t have any poly limit, so I just tried to make it with less possible polys as possible, without losing too much of the quality. At the end I finished with 7887 tris for both the character and the sword.

Unwrapping:

Once the character was modelled, It was time to unwrap the low poly version. First I started by unwrapping all the pieces separately, some using cylindrical mapping, but the majority using pelt mapping. To save work I just unwrapped one side of the symmetrical parts (legs and arms) and then mirrored them to have the other side cloned and unwrapped.

When I had all the pieces, I attached all of them in a single mesh, and added a new unwrap modifier (really in two meshes, one for the sword and one for the character). Then I just had to scale all the pieces and place them trying to have the maximum space occupied, as in the image. Once finished, I collapsed the unwrap modifier and detached the pieces as they were before attaching them, because I wanted to create the normal maps separately for each piece, as I thought it would be easier that way.

Normal Mapping:

To learn how to do the normal mapping I had read the tutorial that comes with 3dsmax. They explained it really well so I’m not going to explain it too, but there’s a very important thing. In the tutorial it said to change the “Normal map space” from tangent to world. If you’re doing a character, don’t do that change, or all the normal mapping information would be wrong when moving the character. Leave it in tangent, and it will work well.

Modelling:

Once I had all the normal maps done, I attached the low poly pieces so I only had two objects, the character and the sword. I also collapsed all the normal maps, creating just two maps. Before starting to texture, I did another render to texture. I added a white skylight to the scene and a plane to make the ground, and added a flat grey material to all the objects. With this I did a lighting map of the sword and the character with render to texture. I used this map as a top layer of the textures in Photoshop, set on multiply and 30% opacity, to make the texture look less flat and to help with the painting.

To create the textures I used some base layers, some masks and some self made brushes. For example, with the metallic texture that I have used on most parts, I started with a metallic base layer, nothing special, just a very clean metal texture, that I duplicated and moved to make it fill all the texture size. Then I used a very oxidized metal texture with a mask to add the oxide spots. I also created two brushes from mask images to add some scratches that I used in different layers with different opacities and blending modes. The same with the mud and the blood, I used combinations of masked layers and self-made brushes. I’ve also added some hue/saturation and brightness/contrast layers with masks applied on them to do the darker bolt colour or the gilded parts, and finally, I added some layers with dark or white colours to enhance the shape of the modelling in the texture.

I created 2 textures, the colour texture and a specular/reflection texture, that was the same colour texture in black and white and with a few changes in some layers to make it look more as I wanted.

Rigging:

As I just wanted this character to test the normal mapping, and I only wanted him to be posed and not to be animated, I used a simple biped to rig it, and the skin modifier to skin it. I only added a few bones with splineIk to move the cloak. If it was for an animated character I would probably use bones for all the character, but in this case, the biped was ok for most of it

Rendering:

The final renders were done with Brazil. I used an hdri skylight and an area light for the main light source

First Steps: workspace and references

Before we get started modeling, the first thing we want to do is make our workspace more efficient. The way I like to do this is to simply split my view down the center, putting the resulting left viewport in front view (numpad 1) and the right viewport in side view (numpad 3). If you’re not familiar with how to split your view, please reference this short video tutorial: http://www.cgcookie.com/articles/2008/12/16/blender-tip-splitting-the-workspace. This allows us to work from both sides of the model at the same time without having to switch our view constantly. It also gives us more views of the model to help with accuracy and proportion.

Now that we have our workspace setup, let’s go ahead and bring in our background image for reference. Today we are going to use a simple, rough drawing of mine that has a front and side view. Anytime you are working from references (which should be almost always!) try to get as many angles as possible. This is particularly important when we work from photo references. Here is the drawing:

To place the reference into the background of your workspace:

1. Go to View > Background Image
2. Click Use and Load to navigate to your image.

Do this with both viewports. The next step is to adjust the X and Y positions to line up your image, it’s best to align the center of the head from both views with the Central Axis point (where each of the three axis’ meet.)

Modeling: mirroring and structure

With our workspace and references set up it’s time to start modeling. The first thing we want to do is add a mirror modifier to the default cube so that we only have to work on one side of the model; anything we do will be mirrored across the central axis. But, before we do that, we need to add a central loop of vertices to our cube, along with deleting one half. This way we don’t mirror our cube on top of itself. You can do this by:

1. Going into Edit Mode with Tab
2. Hit Control + R to activate the Loop Cut tool.
3. Cut a new loop, vertically, along the cube by clicking the MMB when you see the purple line with your mouse hovered over the cube.
4. From the Front View, make sure everything is deselected with A and then select the left-most vertices and hit X > Delete Vertices

The last thing we need to do before we start modeling is adding a mirror modifier for symmetry:

1. Go to the Edit Buttons (F9) and click Add Modifier > Mirror
2. Click Do Clipping

We are now ready to really get down to business!

Modeling: edgeloop structure

The single most important thing to remember while modeling a character head is the structure of your mesh. This is referred to as “topology.” Edgeloops, or continuous lines or circles of edges, are the primary concern with topology. Proper edgeloops allow your model to deform well during animation; they also make tweaking and detailing your model much easier!

To get started:

1. Select the back side of the cube
2. X > Delete Vertices

We do this because we want to work from a single face. What we are going to be doing is laying out a series of edgeloops to map out the structure we want for the mesh. Let’s start at the chin by moving our remaining face with G to line up with the reference from both view.

Due to the variations in our drawing it is going to be necessary to compensate between the views from time to time.

What we are now going to do is use the Extrude tool to lay out our loops. To do this:

1. Select the two outside vertices with Shift + RMB
2. Hit E > Only Edges to extrude.

Extruding will automatically place you into grab mode, which allows you to place the newly created edge where you want it. In this case, along the jaw bone. You can use Rotate, Scale and Grab to help you position the edges. When you’re done you should have something like this:

We can continue using this same technique to get the following for the top of the head:

As you can see, we are starting to define the structure of the mesh and the shape of the head, much as a traditional artist would use reference lines to sketch out a head.

Before we go too much further, we need to go ahead and map out the eye, as it is one of the most important areas of the head, and it’s topology is essential the rest of the mesh. To do this we are going to add a circle from the Front View:

1. From the Front View, left click in the center of the eye to position the 3D Cursor
2. Hit Spacebar > Add > Mesh > Circle
3. Use 8 Vertices and a Radius of 0.500

Next you want to use your translate tools (grab, rotate, scale) to postion each of the vertices to fit the shape of the eye socket:

Now with everything selected (A):

1. Hit E > Only Edges
2. Then immediately hit S to scale in.

Use this same technique for around the nose and the mouth:

That’s it for the structure, this will then allow us to connect all the areas and not have to worry so much about getting the topology right as we have just laid out the major areas.

Modeling: filling in the gaps

With the topology laid out, it’s now just a matter of filling in the gaps (essentially). This means our job has just gotten a lot easier. Using just the Extrude tool, Translate tools, Fill tool, and the Loopcut tool we can very quickly complete rest of the head.

Lets start with the nose:

1. Select the inside edge of the eye socket
2. Hit E > Only Edges
3. Move into position at the bridge of the nose.



Read More