CGSociety

{digg}

CGSociety Tutorial - Alessandro Baldasseroni
24 January 2008

Working at Blur Studio on the Hellgate: London cinematic, Alessandro Baldasseroni created the Hunter model. He tells all on CGSociety.

lessandro Baldasseroni is one of the most prolific artists in the industry, submitting works since the original Character Modeling d’artiste title, and the original ELEMENTAL Autodesk book. Alessandro submitted the Hellgate: London ‘Hunter’ image to Ballistic Publishing’s ‘EXOTIQUE 3’ last year. The Hunter model was made for the Hellgate:London cinematic, made by Blur Studio Inc. The Hellgate: London game was created and developed by Flagship Studio.

After my time in the production deadline, I decided to make an illustration for my own education, using some additional models I had done previously. I tried to stay as close as possible to the given references. This was started as a three-quarter angle sketch depicting the main proportions. Not having a heavily detailed sketch is good for me, since there is room to add more personal touches. I feel this pushes the artist in me to figure out visual and mechanic detailing solutions, while keeping the overall feeling unaltered.

I immediately recognized three kinds of elements I had to deal with. A generic male body covered with a dark grey undersuit; a dark blue light rubber armour and a metallic light blue heavy armour. My modeling workflow was basically bounded to these priorities. The figure sets the proportions, and then it is covered with armor plates.

Due to some other tight deadlines, and also because the body has not that very specific anatomic definition, I took a generic male model and started changing his proportions with a free form deformation box and soft selections. I left the facial features and most of the muscle definition undefined, because those parts were supposed to be covered by the armour. The Soft Selection is an extremely valuable tool for quick tweaking of proportions. Just be sure to flag also ‘edge distance’ with an appropriate value into its rollout to have a very localized control of the fall-off.

As you can notice in the stack figure, I also make use of the Symmetry Modifier during the modeling process, and TurboSmooth with two levels of subdivision, constantly at the top of the stack. Assigning a short cut to the ‘show end result on/off’ button you can easily model at step zero and immediately see the overall smoothed result by pressing a button (I use the space bar, for example).

Once I’m satisfied with the general proportions of the body under the armor, it’s time to begin covering it with the metal plaques. Of course, the dark blue ones first, since they’re the closest to the body, then the light blue ones.

No special techniques here. Once again the modeling is in subdivision with the very same stack as above. I usually start with a single quad, then Extrude edges all around, trying to stay close in volumes and shapes to the given reference. In a lot of the meshes, you need to make sure the volumes and shapes are solid, keeping an eye to not go too far from the underneath of the body.

When I’m satisfied with the general volume, I just try to develop a good quad topology of the main volumes. Of course the topology is done to take a count of the main cuts and holes, but the rest can be easily done with a normal map or a bump map. You need to decide how much detail is worth modeling and what can be put into a bump map.

I usually adopt a ‘rule of thumb’ that means if something looks tricky to carve or extrude into an existing geometry, I usually put it into a Normal Map. When you make something for production you don’t usually have time to model everything, so you need to set priorities.

As you can see in the picture a lot of cuts are not currently modeled, they are done with a bump map. So basically all the modeling followed this workflow, most of the detailing as you can see has been left to my imagination, which was a lot of fun! After modeling all the armor I modeled some folds in subdivision, just in those areas where they were more visible like the middle of the arms and the back of the knees.

The presentation renderings are done in mental ray. The material is a simple mental ray SSS Fast Skin material while the lighting a couple of photometric area lights and a back omni light with ‘final gather’ on.

Basically when texturing, I divide the mesh up by material. That means that each material shares a unique texture sheet with very few exceptions. Every piece of mesh needs to be UVedited to achieve this. I find the use of pelt mapping very convenient for complex shapes like these. After the UVediting, I usually render a template of the UV (after subdividing the mesh at Step One) so that I have a base to paint over the diffuse texture.

The diffuse texture is a mix of photorealistic metal textures and painted rust/dirt. I used a free brushes collection made by and courtesy of

These brushes are simply great to create variations of dirt and rust, and having the UV template in an underlying layer helps me to place the dirt, scratches, decals and everything else exactly where I want it.

The material used to create the light blue heavy armor is similar in many ways to the leather under suit and some other chrome parts. Basically it’s a Max Blend material. The idea behind is to blend two different materials with the same diffuse map but different specular properties. This gives the appearance that the scratched areas of the armor has a shiny metal underneath.

A totally different material can be chosen to share a different metal diffuse map as the underlying metal, but in this case I was comfortable enough with the same diffuse texture. To blend the two different specular materials, I used a grey scale mask where the darkest parts would have shown the metal material underneath.

Here again to paint the mask, I used some custom jagged brushes to get a natural feeling of random scratches. Specular maps are derived from the diffuse map. I usually put a Hue/Saturation regulation level on top of the diffuse map, bringing the Saturation slide to zero so that the picture results in gray scale. At this point, I put another Brightness/Contrast regulation level on top to regulate the intensity of the specular parts.

Having both regulation levels in a folder and saving the .psd document, allows me to quickly change parameters for fine-tuning of the specular. It’s also possible to drag/copy this specular folder on top of a different diffuse map, so maintaining the same specular values.

The leather material is just a simple Max material with quite a high specularity. Cables are chromed material reflecting a HDRI map and glows are just self-illuminated standard materials. The render is done in Brazil with a simple spot light and a very low intensity global illumination.

« Some sample materials

Already having a monster model I did some time back for the game itself, the intention was to put both the Hunter and the monster (Karnagor) into a nice composition showing both of them in a sort of relationship, moreover everything should have some hints that the scene was in London. The monster model was made in subdivision and most of his detail is by normal mapping (sculpting done in ZBrush).

I asked to an artist friend of mine; Antonio Mossucca if he was interested in helping me with the composition. I briefly explained to him what I wanted to achieve and he came out with these nice sketches, which are pretty close in terms of composition to the final picture.

The only element I was missing was the Big Ben model, so I decided to make a quite low poly version as it was supposed to be a background element. I was missing also a sort of weapon and a ground terrain. For the weapon I wanted something used by snipers with a scope, not really futuristic. I preferred something adapted roughly to be used as a flamethrower.

So when all the desired elements were textured and put into the scene, it was time for lighting and rendering! Let’s say up front, I wasn’t after anything photorealistic, but giving instead a painterly feel. To achieve this, especially when dealing with 3D images the post process done in a 2D application is crucial.

It’s often a waste of time to spend in realistic and complicated shaders. Most of their properties will be washed out during the post process work, so it’s better to be plain with materials and lighting.

The basic lighting setup consists of a main spot light (better with no color), a back omni light, and a few omni lights for the glowing parts here and there. I didn’t use any global illumination, just a plain Scanline render because another pass of ambient occlusion will be composited later. Then it’s time for heavy post processing!

In Photoshop, I usually start by settling on a color scheme. The base render did not have much of a dominant color on purpose. I prefer to have much more control in post, and I love flexibility. In this case I wanted some warm tones, so I started painting the background clouds with brown/red tones; very low in saturation.

Put the ambient occlusion layer in Multiply mode; Opacity of the level at around 20 per cent. I also put a Hue/Saturation level connected to this layer in order to shift the Hue of the figures to brow/reddish tones, matching the background light. As you can see, the rendered Big Ben has been put in a layer behind the main figures and color shifted as well. Smoke has also been painted in the foreground with a soft edged brush with low opacity.

Time to add some fire here and there on another layer! I made an extensive use of some real fire pictures collected on a black background. Sometime back I bought a reference CD called Pyromania. They’re pretty easy to composite in Hard Light mode so the flame can be extracted. I also painted a yellow highlight in Soft Light mode over the leg of the monster, because I expect some light of the fire to shimmer over it.

I wanted some smoke to move in the direction of the composition. To paint this you need a round hard-edged brush first, for mid-tones and shadows, then blend the ‘blob’ together with the Smudge tool leaving the hard edges in the areas where the light source is supposed to hit. The burn tool at this point helps to enhance the brightest parts of the smoke. I introduced some painted sparks close to the fire and some debris floating around in the air too. This gives a ‘chaotic’ feeling to the image.

Time for a color balance. Things were beginning to look a little flat, so I introduced some bluish tones in the background and the shadows by using the Color Balance tool (ctrl-B). I also noticed that at high resolutions, some textures were going to lose their definition. A quick trick to reduce this effect is to put a grayscale high resolution sample of a similar texture in Overlay mode over the interested areas.

In the final image, I made the tones even warmer and saturated by adding an orange Soft Light layer. Some fog has been added in the background to give the scene a little more depth. The lamp in the background is a paintover of a real london lamp. I also added some blood painted blotches on the Monster and brightened some areas of the Hunter to make him pop up more.

My name is Alessandro Baldasseroni 34 years old, born and raised in Milan. I’m currently employed at Venice based Blur Studio Inc. (CA), in charge of character modeling. I’ve been working in the industry for seven years, starting as a generalist at Milestone, a leading Italian game development company.

Later on, I started freelancing for various worldwide popular studios, from Flagship Studio, to MassiveBlack and now Blur Studio. In October 2007, Blur decided to hire me full time. I still love to make digital illustrations in my spare time and my artworks have been featured as covers and articles in many international 3D magazine publications and Ballistic books.

Related links:

Society of Digital Artists

CGWorkshops

CGTODs

Forums

WIKI

Members

CGPortfolio

CGChallenge

CGForum

BSW

CGGallery

Features

CGJobs

CGS Members

Ballistic Publishing

The Society

Learn

Discover

Create & Share

Shop

Social