Making Of 'Judge Death'
Hi there, my name is Petr Nasirov. I'm a 24 year old character modeler from Belarus. "Judge Death" was created for the game art challenge Comicon'09 and it won 1st place. Before starting I spent about 2 days researching comic characters. When I saw Judge Death I instantly thought, "I want to model this character" because he looked interesting, unusual and fun.
Searching for references is one of the most essential stages. The more time you devote to studying the features of the character, the fewer mistakes you will make. At this stage you need to study all the detail of the character's structure and the design and material of his clothes, to figure out whether everything is functional and comfortable. Character creation is a complex multistage process. If you find a mistake near the end of the process it will take much more time to correct it. If worst comes to worst you will even have to modify the high and low-poly models, and the textures.
Regarding the modeling of the Judge Death, I took advantage of all the references I managed to find on the internet using the Google search engine. There is a great variety of designs of Judge Death, some of their elements appeal to me, some not. One poster became a reference for a hand as I liked the details and color. Another poster gave me an idea for the bird on the character's shoulder. The combination of Judge Death images helped me to create my own version of the character. The more reference I found the more detailed my version of the final character became.
Base Low-Poly Model
After studying the references that I had gathered, I got down to creating a base low-poly model. Primarily this stage is necessary to reveal the right proportions of the character and to define a general silhouette. At this stage I use the standard 3ds Max materials to set different colors to the model elements. This helps me to see not only the silhouette, but also the combination of the brightness of the different elements (Fig. 01). From this early point it is more convenient to work with a low-detailed model. You have to focus on ensuring the silhouette is correct.
After I am happy with the silhouette I proceed to the next stage: high-poly modeling. When preparing a model to export to ZBrush, I create a high-poly model with no detail in 3ds Max by means of the modifier Turbo Smooth (Fig.02). Pleats on clothes and folds on bodies, and small details like this should be made in ZBrush.
Export To Zbrush
When the high-poly model without details is ready, I delete the modifier Turbo Smooth and export the model in .obj format selecting the ZBrush Preset in 3ds Max 2009 (Fig. 03). For convenience I export the model to ZBrush in parts with the help of the command "Export selected", when the parts are symmetrical I delete the halves.
Next I create a high-poly model in ZBrush. Here I gather the model in the parts which it has been exported from 3ds Max. Each part is placed to a separate sub tool which makes it easier when high-poly modeling. (Fig. 04)
Usually I start working on the high-poly model from the 1st level of SubDivides and try not to proceed to the next level until I am sure the current level is detailed enough. This enables control of the mesh and smooth lines. (Fig.05)
When modeling I often use the Standard brush with no alpha. If I want to sharpen the edges or improve the small details then I add Alpha 39 (Fig. 06). When finishing the modeling I can modify some edges with the Pinch brush to achieve a higher level of sharpness (Fig. 07).
Also I like to work with the Clay Tubes Brush. This brush creates different organic objects with a complex friable surface.
When modeling I as often as possible turn to my references. For instance if I want to draw pleats on clothes I search for some photos, select the most interesting fragments and carry them to the model. The character becomes more alive when the modeler uses references.
Adding the colors on a high-poly model in ZBrush helps you to imagine the final result. If the model has UVW the Polypaint can be moved to the diffuse texture. To activate the Polypaint one need to carry out some simple actions: press the button Colorize on the tab Texture of the Tools panel. Then uncheck Zadd and Zsub on the upper panel of tools and press the button "Rgb". A color can be selected from the upper menu Color (Fig.08).
Final Low-Poly Model and Unwrap
When the high-poly model is completed, I create a low-poly model again but with the resemblance to the high-poly model by means of the program TopoGun. This program operates with high-poly objects and has an easy-to-use toolset to create low-poly models from the high-poly ones.
Each detail is created separately. Each high-poly element is opened in TopoGun and replaced by a low-poly one (Fig.09). Then all the parts are put together in 3ds Max where the low-poly model is being improved considering the limitations on the polygons number.
When unwrapping I use Unfold3D. In 3ds Max the model is divided into parts as it is before exporting to ZBrush, each part is separately imported to Unfold. Here I indicated the edges which will have the borders of UVW layouts, (orange lines on Fig.10) the program will then automatically unwrap the model. Then all the parts are put together in 3ds Max into one mesh and a general unwrap of the model is created.
In order to place all the layouts tightly one should start with the grouping of larger parts and then proceed to gradually cover the free zones with small parts. To save space the symmetrical parts are combined. (Fig.11)
A normal-map is the first texture I create. I like to render textures in 3ds Max. It allows manual correcting of the cage and after rendering I get a normal-map of high quality which almost does not need improving.
Like the previous stages we work with parts of model. We open the high-poly element and its low-poly version in the scene. Sometimes I apply the modifier push with a small value to a low-poly model to increase a low-poly in size when rendering the normal-maps.
When all the parts of the normal map are rendered the final general texture is created for the character and the normal-maps are tested. It would be better to test the character in the scene with the final lighting. To test the normal-maps quickly I use DirectX shader: Xoliulshader which can be found on the Internet.
Diffuse Map and Specular Map
When the normal-map is ready I proceed to create the textures for the ambient occlusion. It is necessary to supplement the shadows of the normal maps. This is the basis of the diffuse textures. For my character the ambient occlusion textures were created by means of baking the lighting from the high-poly model to low-poly, using the program xNormal. The lighting was baked for each part separately with the following settings: (Fig.12)
(Fig.13) This texture is laid in Photoshop with the Multiply mode on the layer, with the colors and facture to create shadows. The highlights are a result of the specular maps. I don't create them in the diffuse textures so the character has highlights only in places where the light falls. The base diffuse texture is created by mixing AO (Ambient occlusion) textures, color and facture (Fig.14)
The specular map is created from the diffuse textures by desaturating and increasing the contrast. On mat surfaces (for example fabric) the texture is made with less contrast, on the bright surfaces (metal) the contrast is increased.
Starting at this stage I improve the textures by thinking about the final scene render. (Fig.15)
I hope you've enjoyed this making of and perhaps you've even learnt a few tricks. If you have any questions, please don't hesitate to contact me via email. Thanks for reading!