Making Of 'Mr Coutinho'

Hi, my name is Wesclei Barbosa, and I'm from brazil, and currently work for Tv Globo, a television channel over here. I want to show you how I created my image "Mr. Coutinho" and to try and show a simple workflow between 3ds max and ZBrush, to get good characters.

About the image

In this project I wanted to try some techniques that I have seen in some CG forums. One of which is to project a photo onto models using ZBrush, in order to get your textures, the other is to get mesh details from it.

After looking around I choose this guy, because he has a lot of detail in his face and what better way to test this techniques.


For this guy I've used some photos of an old guy, that I have in my "resource folder" (it's a good idea that you have a lot of photos of all kinds of people whether they are old, young or children etc...) this kind of references always is very helpful.

So here is our guy.


The modeling process that I've used is poly-by-poly for the face and box modeling for the rest...there are a lot of tutorials of how you can do it, so I will just show some wires of my process.

Ok, now we can export our model so that we can play with it in Zbrush.

Right some observation is needed here, we don't require the UVW map from max, as we will be using the projection from ZBrush. Why, I hear you ask. I mentioned before, that we will project photos onto our mesh in ZBrush, so we can use the group UV to automatically map our model, like on this image.

After we have applied the UV, we can start to project our textures onto the model. For this, we will use the 'Projection Master' in ZBrush. Using this we can, paint textures, copy areas like the clone stamp in Photoshop, saturate or desaturate our textures and create some deformation, using tools like Deco brush.

Now, what we have to do is choose the tool Plane 3d, and increase your resolution to get more points on the mesh, doing this we get more control points to deform the plane and the texture. Select the alpha for the texture, choose the DragRect stroke, leave the material default and on texture choose your photo. On the Z intensity parameter put 0 to intensity. Now you can apply the photo to the mesh.

You probably wont get a good result on the first projection (like in the ear area), but now you can edit the mesh of the plane, like a simple object in ZBrush, so using move tool, you can adjust the texture on your model.

By doing this whole process again for the other angles, you can get this result.

Ok, now we have the colour/diffuse map for the model. What we have to do now is to create a mask based on the intensity of this map. It's a great way to get information about the bump/displace of the real photo.

Now we can use some tricks in Zbrush. Using the commands inflate and smooth, on the deformation menu, we can apply a "bump" on the model and with this technique you get a great base to start to deform your model with the traditional tools of Zbrush,
inflate brush, projection master etc.

This is the model after some adjusts.

Now we can export the bump/displace to use in 3dsmax.

Textures & Shaders

For the render I've used Mental ray in 3dsmax. The reason for this is because I like to use the Fast skin shader, as it's a good shader and with it, I have a lot of control over maps.

For the final shader, I've painted these maps.

Here are the parameters that I used on the shader.

But the values can be different for others models, because they use the scale of the object, so you have to make some render tests to get a better result. one tip is try some values on the scale factor parameter.

The shader for eye is very simple, just a phong shader with a high value of specular, the secret here is the way that it was modeled, two separated meshes, one for iris and the other the cornea. I've used two different materials, for the cornea I've used a raytrace material with a fallof on refraction parameter and a high value of specular too.

With this and some tricks on reflection, I get this result.


The light process is very simple, I've used two MR area light, one omni for backlight and a spot for the key light. So I created one sphere with visible to camera turn off and used it like my environment. To bounce the light to the model, I turned on the final gather parameter.
To get better reflections, I created two planes in front of the eyes.


For the render process, I've used MR sample pixels in 16 to improve the quality of anti-aliasing, and to the post process I rendered a pass of ambient occlusion and Z-depth information.

Post Process

In Photoshop, I played with levels and colour balance to enhance the result, so I used the ambient occlusion pass with a blend of multiply to improve the shadows on the model.
To finish it, I use the information on the Z-depth pass to apply a lens blur with this values.

Then I applied a film grain to add some texture to the image

So, here's the final image.

Fetching comments...

Post a comment