3D art is becoming more and more prevalent in our every day lives: For example, most of Ikea’s advertisements are 3D renders, not real objects. Opening a dialogue and increasing 3D literacy with “the uninitiated” will lead to understanding, appreciation, and interest in the art. Creating accessible and easy to follow documents or videos can only benefit the 3D industry.
It seems that every year, the games released look better and better, but rarely does the hardware, the consoles games are played on, improve. Even a PC user will only upgrade every so often. Without a more powerful system, how do games look better and better all the time?
The answer, is optimization. As new generations of hardware hit the market, the game development industry is bombarded by new “next gen” tools to take advantage of this new power. As time goes on, tricks are discovered and shared, creating ways for artists to optimize their props. These shortcuts allow artists to push the envelope on what is possible, as they are always applying past lessons to new projects.
While readily available technology and artist skill will always play a a factor in the visual aspect of games, they are not the sole causes for the evolution of game graphics.
7) Final Assembly in Game Engine
After the model is created, UVed, and the texture and other appropriate maps are created, it needs to be assembled in the Game Engine, which serves as a framework for the game world, and allows artists, programmers, and animators to put their work together in to a full game. In the game engine’s level editor, a 3D modeler will place the props they’ve made around the level, apply the appropriate maps and textures, and use lighting tools to set the mood of the level, bringing the game world to life.
While parts of games may be procedurally generated, such as minecraft, or may be sculpted in something like Zbrush or mudbox, rather than modeled using polygonal modeling, this basic workflow is used across the industry to create assets and props for games.
Normal mapping, as discussed in a previous post, is a way of applying normal information from a high poly mesh to a low poly mesh, to give it more detail. But what are normals? How do they work?
Normals, or surface normals, simply put are the direction that the surface faces. Each side of a cube has a unique normal, as they all face in different directions. Since high poly models are much more detailed than low poly models, their have much richer normal information. By seeing where this normal information differs, 3D modeling software can figure out where to apply new normals to a surface (even if it is not it’s natural, default normal), and where to leave it the same. While this does not affect the actual shape of the model, it will affect the way light reacts when hitting that model, since the way it’s faces are pointing, or sections of its faces are pointing, has been changed by the normal map in the computers eye.
Using imagine manipulation software, such as Photoshop, a 3D artist can import a UV template, and create a texture for the 3D model. The 3D model already has shape, but no color, and there is no information about the material it is made from. Adding texture to the model will help to give it character, inform the viewer of it’s materials and will give it a more pleasing and complete look, but a texture alone does not make for a complete material.
6) Creating Normal, Specular, and other maps.
At this point, our low poly model has been created, UVed, and textured, but color alone does not sell the illusion of reality. All surfaces have micro details that are all but impossible to capture in a low poly model on their own, as well as different levels of reflectiveness, also known as specularity.
Normal Maps, which allow micro-amounts of detail and smoothing from the high poly model to be projected on to the low poly model via UV space, are a common way to “fake” detail on to a model. Since every part of the model is represented both in 2D UV space, and 3D XYZ space, these details can be “projected” on to the low poly model to give the illusion of detail.
Similarly, specular maps, as well as maps that can control emissiveness, displacement, and other details can be created to add more realism and detail to the model.
3) High Poly Model
Although the order in which the High Poly and Low Poly models are created is interchangeable depending on the workflow (creating an extremely detailed model and removing detail from it, or creating a simple model and adding detail to it), it’s exact place in the workflow depends often on personal preference. As the name suggests, a High Poly model is created with a large amount of detail, which would be undesirable for use in a real-time rendering situation, but can be beneficial in several other ways, as explained below.
4) UV Mapping
UV mapping is one of the most important parts of the 3D workflow. It involves taking a 3D model that exists in 3D “XYZ coordinate” space, and unwrapping and unfolding it so that it sits perfectly in 2D “UV coordinate” space. The easiest way to think of this process is backwards origami. By unfolding the model and laying it flat, the 3D artist creates a 2D image that can be drawn on. Since every polygon, edge and vertex are laid out in this 2D space, each section of the image is also associated with a location in 3D space.
1) Gather Reference
3D artists rarely create any art from scratch. Normally, reference material is either gathered, created by a concept artist, or both. At the very least, a 3D artist will have a “visual guide” to work with in order to create an appropriately styled piece of art. Reference is important to plan out what is to be created, since many elements of the model may be important to the player, signifying hidden items, the path forward, or explosive or hazardous materials.
2) Low Poly Model
A low poly model is a model created using a relatively low number of polygons (Though standards change each generation of consoles) made to be rendered in real time in the game engine. Since rendering becomes more and more difficult for each polygon added to the model, extraneous details are excluded from the low poly model.