Marcel Gorri did a breakdown of his final diploma project at Think Tank created with Maya, ZBrush, and V-Ray.
My name is Marcel Gorri, I am a 3D Modeler, Texturer and Technical Artist from Brazil. I recently graduated from Think Tank Training Centre, specializing in creating Environments for Films. In Brazil, I got my bachelor degree in Computer Engineering and I worked in software development after completion. However, at a specific point in my career, I decided to switch fields and find a job that could honor my creative and artistic side. I’ve always been fascinated by the creation of entire scenes and environments and once I realized this was possible as a career in movies and games, I decided to pursue a career in the 3D field. I started studying 3D in Brazil and after some time I got into Think Tank, and now here I am.
Wonderbook – Book of Spells
Wonderbook – Book of Spells is my final project for the full diploma program at Think Tank, focusing on Environments for Films. For such project, both my mentor and I decided that whatever concept I would choose, it should be a great concept containing good variety of materials (reflective, matte, SSS), both hard surface and organic elements, large and small artifacts (furniture objects / walls and door panels), hero items (complex objects with different materials to be displayed in turntables), different colors, and finally, interesting lighting. I spent a good amount of time searching for concepts with these elements until I finally found Wonderbook – Book of Spells by Atomhawk, which besides being exceptionally beautiful contained all those features and also was in a theme that I like a lot: a medieval look and feel with lots of stone and wood.
Since the beginning, one of the main goals for this project was to be as close to the original concept as possible. I wanted to focus on developing my critical eye, specifically the ability to translate what I’m seeing into what I’m creating in 3D, something that was heavily put into practice while creating this environment. The only item in the original concept that was left out of scope was the knight armor which, in order to reach the quality level we were pushing, would require more time than I had. In fact, it could be a whole different final project by itself. Because I chose the approach of matching the concept, I didn’t have to create much or combine different references. My mood board was mainly to understand the structure of the objects in the scene as well as their materials. Other than that, the original concept was always my main visual reference.
The project started with the basic block out in order to define the main shapes and location of the objects in the scene. Since the goal was to match the original concept as much as possible, I didn’t have to imagine how to fill the environment, or how to create a mood. Everything was already there in the concept, I just had to translate it into 3D. This might look easier at first, but on the other hand, if something wasn’t fairly close to what was depicted in the concept I had to adjust it again and again until it was close enough to the original. For that, I took great care to match the perspective of the scene with the perspective of the concept, because everything else would depend on this initial setup.
Having the correct perspective set in my scene in Maya, I started to block out the initial shapes to define the building structure like the walls, floor, ceiling and windows as well as the objects in it (e.g. the wooden cabinet and shelves, the globe, the blackboard and the books) until I was able to represent them all correctly in the scene. Once this basic block out was done, the workflow adopted was to continuously refine the scene. So the basic block out evolved into a more refined form. Then some basic materials and basic lights were added in order to have the initial feeling of the scene.
Having established the block out baseline, I went from a block out phase to a detailed modeling with proper topology phase. The concept art, although wonderful, brought some technical challenges. Some of its areas were not so clear like details of the foot and base of the globe and the structure of the balcony in the upper right area of the room. For those cases, I had to go through some refinement cycles trying to discover what would make sense for a shape that was barely visible in the original image. I had to take into consideration not only what would fit the shape in the concept but how such objects and structures work in the real world. This indeed turned out to be a challenge but also a great opportunity for me to exercise and develop my critical eye and problem-solving skills for what wasn’t so clear in the concept and to create something that would work in the 3D and real worlds.
Another important step in the modeling phase was sculpting all the wall and floor parts in the scene. I always wanted to sculpt surfaces like that, it was one of the reasons I liked this concept, so it was a good opportunity to exercise this skill. After sculpting, I generated displacement maps in ZBrush and then applied them to simple flat planes using V-Ray displacements.
After the modeling phase, I had to layout the UVs in order to be able to texture the environment. At this stage, keeping a correct and consistent texel density is very important. In order to achieve that I had to consider different factors, such as camera planning, specifically, the distance that each object would be from the render cameras, render time optimization and compelling visual results. The render cameras movement must be carefully planned, otherwise, a good balance between allocated resources and a reasonable render time cannot be achieved. This is a problem that might appear later on in the production of the environment when a good amount of effort has already been dedicated. If you leave all your textures in 4k resolution for instance, even for objects that are not close to the camera, excessive resources are going to be allocated (texture files loaded in RAM memory) causing slow render times or even crashing the rendering process.
Once done with the UVing phase, the next step in the texturing process was to gather real-world references for the materials of the objects in the scene and finding good tileable images to be used as an initial base for the textures. With such images at hand, I could work with procedurals, masks, and brushes to create the targeted visual, adding primary, secondary and tertiary details in order to make the textures look more realistic.
One particular challenge in the texturing phase was creating the map texture for the globe. It might be difficult to create a seamless texture on a spherical object, especially in one where the texture has so much significant information like a map with meaningful shapes, letters, and numbers. Again, it took me some iterations to refine this texture until I got to a point where I felt satisfied with the result.
Creating the textures in programs such as Substance Painter, Mari and Photoshop is just one step in the whole process of creating the visuals for the objects. After that, I had to create shaders in V-Ray with the proper settings to recreate those materials in real life. After exporting all the maps from the texture process such as diffuse, specular, glossiness, and bump, I connected them to V-Ray materials, set properties and went through an interactive process of rendering, checked the results, refined either the textures or the shaders, and then rendered again until reaching good results. Some shaders were not created with maps exported from the aforementioned programs but rather with masks and procedural images combined in a node-based shader network, – for example, glass materials.
The textures are not an isolated step, they work in tandem with the lights and shaders, so I also had to adjust and refine the lighting rig for the scene. The concept has very interesting light effects but some of them are not necessarily physically accurate, so I had to take this into consideration as well.
In the light rig, I have some lights that are really important, with big effects like the moonlight coming from outside of the windows and reaching the wooden cabinet area. I also added other lights with very specific shapes (adding masks to their color fields) so I could specify the shadows to exactly reflect what was shown in the concept. Besides that, I had some other smaller lights positioned in very specific places to generate color and reflection effects like the light coming from the left corridor area in the floor level, and also some shining reflections in objects like the globe and some jars.
All these effects had to be combined in the end with the fog and god rays in the scene. In order to do that, I had a separate scene file focusing only on the creation of a V-Ray environment fog for the scene. After creating and positioning the environment fog in this scene, I applied a dark material to the rest of the environment so the rendered frames would contain only the fog/god rays information to be combined later with the regular frames from the environment in Nuke. There I also added some of the noise effects, color and contrast adjustments to the composition. If I had created the environment fog in the original scene file, the render time would be much longer and it would be much harder to do fine adjustments.
Another interesting touch to the scene was the creation of the visual effects for the candle flames and the magic smoke coming out of the cauldron. For both of them, I used the same technique, which was to create a 2D fluid in Maya adjusting its shapes/particles, colors, and movements until it looked like a flame or a special magic smoke. After that, I rendered a large number of frames, approximately a thousand, which was a pretty fast process using the nCache mechanism that Maya provides. I assigned different sequences of frames to cards (simple planes) using masks to drive the opacity and then obtaining the animated flame effect. Having a big set of images to choose from, I could use 3 or 4 different sets from the same rendered frame sequence and still have different results for each one of them. Since each card had one image in each specific frame, it didn’t consume extra resources in the overall render process at all. In fact, it was lighter than rendering particles together in the same scene.
In the end, the biggest challenge I faced was the fact that the concept provided me with a very good looking environment but also an incredible amount of objects to create. The biggest challenge was just the sheer amount of effort and determination required to complete it. At the same time, it was a great opportunity to learn. This project allowed me to grow my skills in all areas of creating a 3D environment from understanding the concept and shading to rendering, and compositing. This allowed me to develop my critical eye and granted me more experience growing confidence in my skills and getting me ready and excited for new challenges ahead.
If you found this article interesting, below we are listing a couple of related Unity Store Assets that may be useful for you.