A 3D breakdown for 'The Barbarian'
Senior 3D artist Ivo Diependaal takes us through how he made July 2015's winning gallery image in ZBrush
This tutorial will give a overview of my work process on this project and focus on a couple of key points to bring this work to life. There will be something in here for all skill levels but a basic understanding of the used software packages is required.
Reference is key!
Searching for reference images is the very first thing I do when starting any new project. These can be inspiring works of art from other artists to anatomy studies or specific images from nature related to your topic. Any other source that inspires you or makes an impression is also good; the important point here is that it informs your work. Pinterest is a very easy to use and helpful website that allows you to dive into someone else's library of images and specific collections, and from there create your own collection. You can also easily create ref sheets with Kuadro or Picasa.
Sculpting the character
In ZBrush I modeled a very rough body shape; I started with the body because I knew I wanted to make a chunky, muscular character. It helps to start off with the things that are a definite given – for this project a big muscular guy. At this stage I concentrated on the torso area which I found helped to give me a better overall feel for the character. I mainly use the Move and Clay Buildup brushes, but I find that the closer I get to the final look the less pressure or intensity I need to use. The choices made in the torso design give cues for shaping the rest of the character and inform the look of the face.
I started sculpting the face as soon as the rough proportion of the body where done.
I always try to sculpt my character with an open mouth, this makes it a lot easier to make expressions and other changes to the face and jawline later on. Besides that it is less trouble to make an open mouth shape into a closed one than the other way around. Here it is exaggerated because I knew I wanted to have him with his mouth wide open eventually.
The sculpting is straight forward, and tackled in numerous tutorials, so I will not go over it too much depth:
- First – the basic structures of cheek bones and nose bridge were sculpted. Basically all the hard points on the face, where the skull shape is most visible (cheekbone, forehead, chin, and bridge of the nose).
- Second – sculpting the large skin/soft areas around the mouth, sides of the nose and so on. This is important to do in strokes that follow the muscle structure of the face; this will give you a strong foundation for the rest of the character.
- The primary details are setup with the DamStandard and secondary ones are the directionality strokes over the whole model. Than a thin layer of final compression wrinkles. Filling up the spaces with the Standard, Inflate and StandardGravity brush to add weight to the bigger wrinkles.
- Finally different skin reliefs are added with alpha brushes including nose pores, extra detail under the eyes and on the forehead.
Note that there is only real detail in the face; this is to balance out the work, and lead the viewer's eye to the face. Because of the proportions of the character, the overall body mass already draws the main focus; therefore other things are done to draw the viewer to the face.
- A bit more detail in the face than the body and clothing
- New color introduced in or around the face (face paint, and white hairs)
- Adding the fog of his breath
- Face placed in between two of the same object types (armor)
It should be noted that often detail is not that important, especially if the basic anatomy, posture or design construct is lacking. Many are too quick to go into all the details before the foundation is there. In the end this will not make it better, but more likely mask the underlying problem and keep you from progressing in your work.
The posing was first done with the ZBrush mannequin. This makes testing poses quick and easy to check and adjust. When satisfied with the initial pose you can add the mannequin as a SubTool to your character. Then go in to Transpose Master and replicate the mannequin pose you made.
It was very important that the pose emphasized the size and width of the character and that the viewer's eye would be guided over the image to points of interest in a natural manner. In the example below I have drawn out different guide lines to point this out.
The basic character color maps were painted in ZBrush using different colors with a Spray brush, and later adding a bit of cavity color, and a cavity mask in the initial shader. Most of the props were textured in MARI; this is a lot easier, because subdividing to get a certain texture resolution is not needed there.
Different maps were created from the base character color map by changing the hue, saturation, brightness, and contrast in Photoshop. I also used this to paint some of the extra details such as the vein detail for the sub-dermal, and a cavity map multiplied over the base reflection map. These skin maps where all hand painted, to get the look and feel I wanted.
The shading basics
The shading was done in Maya with V-Ray. Initially I had two setups, one with SS2 shader, and one with the new V-Ray skin shader. I ended up liking the look I got with the fast SS2 shader more. The networks would be the same, but with a slightly different result in the falloffs and Specular highlights. The skin material was started off with the fast SS2 shader piped into the Blend material as the main shader and a V-Ray material as the first coat for the sharper Specular highlights. After that come all the shaders that are not part of the main skin, but really layers on top. Such as overall dirt bump and spec, and a second layer of specific dirt for the arms and shoulders. Followed by the face paint, that is basically a blue V-Ray shader, and in the mask an alpha map of where the paint should be. I also put that texture in the bump of the paint, to elevate it from the skin. Finally, I also put the Blend material in a V-Ray bump shader, and added a slight extra bump based on cavity, besides the standard normal maps for small details.
For the metal I just used a color and reflection map created from photo textures.
With specific chip marks, dull and reflecting surface areas. IOR set to around 8 to get a nice metal like reflection in conjunction with the GGX BRDF type. I always use the color map as an extra bit of bump, besides my ZBrush Normal map to just get extra details in there. The shader was used as base layer in a blend material again, so that a layer of fine dirt in the non reflecting parts could be added, with different highlight values, and bump.
The final render and composite
For the render I created a three point lighting setup as a start. Later I added to these with a few fill lights for the eye area, and a top light for the hair. After some tweaking I noticed that under-lighting the overall image worked much better when later comping it in Photoshop for the result I wanted.
The goal was to create the final look in comp without doing things you couldn't do for animation in NUKE or other composite software. Keep this in mind when rendering a turntable and easily be able to recreate the same look. Basic passes were rendered Beauty, Diffuse, Reflection, GI, SS, Specular, and Lighting. I took the ones I needed in to Photoshop and created the final image with the use of different blend modes.
It gave me pretty good results and reduced the time I needed to tweak lights and test render on my slow machine. So using this method trying to get as much out of the renders I could in Photoshop probably saved me a lot of time in the end.
I hope you learned something from this overview. If you have any questions feel free to ask them. Best of luck on your current and future projects!