UC Berkeley Main Campus
Department of Art Practice, Center for New Media, Fall 2004
CGI is three-dimensional, but the computer screen is two-dimensional. The Maya interface offers many tools to see and manipulate all three dimensions, giving users a sense of control over all three dimensions with a flat interface. First off, Maya can show one scene in it's 3D space from many different points of view at the same time. Each view is generated by a separate virtual camera. Often, Maya users switch between cameras to see all sides of a scene, or even work with three or four cameras at the same time. One important element which is always present in Maya is time, which is represented in the Keyframe bar at the bottom of the screen. There is only one time per scene, and it is measured in frames. A typical time line would run from frame 0 to frame 300, and that time line would represent an animation duration of 10 seconds at 3 frames per second.
The Maya interface further consists of four layers of menus, the Modeling, the Animation, the Dynamics and the Rendering Layer. These four layers of menus reflect different stages in the production of an animation, and they offer approptiate tools and commands for each stage. Users cycle between these layers using the pulldown menu on the far left of the top bar of the interface, or by using Function keys F2 through F5.
It is important to learn the CGI terminology which helps you describe what you want to do in a way which registers with CGI tools and procedures. If you know what a tool or a procedure is called, you can find it with the "Help > Find Menu Item" command or in the expansive Maya Online Manual. Our course offers a CGI terminology page, and we also define key terms below.
The math involved in 3D graphics includes vector geometry, trigonometry, and stereometry. The most basic CG unit is a point in space with three values, x, y, z determining its location on the three axes of 3D-space, x (width), y (height), and z (depth).
Three such points in space define three corners, three edges, a normal and a facet. Facets are stored as a set of three points in space with three values each. 3D objects are composed of many facets, which configure the representation of an object in space. Since they contain many corners, they are all called polygons. Extremely complex polygons are referred to as meshes. Meshes are used to describe scanned objects and contain between 100,000 and several billion polygons.
Michelangelo's David for example is a mesh, which was scanned from the original sculpture by Marc Levoy and his team. The resulting model contains two billion polygons. Much like pixels in high resolution images, many polygons eventually blend to the illusion of a smooth continuous surface.
Few objects in our experience conform to the geometric clarity of polygons. Many are based on continuous, curved shapes defined by many points. To match our experience, CG programs allow for the design of objects based on curves, which stand in for polygons with very many points.
Non-Uniform Rational B-Splines
The term spline originates from shipbuilding, where the planks used for the ship's hull need to be bent into exact spline forms. The curvature is defined both by the resilience of the wood and the position of the ducks which hold the bent wood in place. Thus, the spline is defined exactly with four points and a constant. Similar curves are also produced by suspended strings such as laundry lines. Such curves for example have served Spanish architect Antonio Gaudì as an inspiration.
Primitives
Primitives are regular geometric bodies which have exact mathematical definitions. In Maya, the definitions of these bodies are readily available in Make>Nurbs>Primitives. Their parameters, such as height or radius can be adjusted by the user.
Deconstruction into Primitives
The first step in CG model construction is the deconstruction of the object we want to model into its primitive shapes. This deconstrcution is an abstraction from the real object we seek to represent. It is the process of reducing a representation to its subjective essence. A successful abstraction only consists of what seems essential to the author of the abstraction. This process is associated with the cubist and the abstract expressionist tradition in painting and sculpture, but has its roots in African traditional sculpture. For more images on African traditional sculpture, please visit this site:
http://www.sas.upenn.edu/African_Studies/Sculpture/menu_Sculpt.html
The head by Raymond Duchamp-Villon shows an integrated abstraction of all facial elements to an expressive whole. Such a head is geometrically much simpler than any realistic head. At the same time, it is very effective at communicating a sentiment.
The process of abstraction is essential to CG, because it lowers computational loads and can increase the expressive value of our models. Compare for example, Picasso's Demoiselles d'Avignon with a classic CG game still from Castlevania. The methods for simplification have not changed in 100 years. However, the expressive agenda is radically different.
From an initial image (photograph or drawing) of an object or person of your choice, make a matching static model using only primitives, polygons. extrusions from polygons curves, and extrusions, lofts and revolutions of these curves. Prepare to justify your choices for each part of your model. Try to match proportions and expressions closely. Use no more than 53 geometric shapes. The emphasis should be on expression, not on matching the initial image: show what the photograph feels like to you. There is a classic example here of stylized runners by Kazuma Morino. Present both your initial image (in jpg), and a truntable playblast no longer than 400 frames on a memory key in a folder named as follows: firstinitial_lastname_firstinitial_lastname_a02. This assignment is due this Wednesday.
An example of a chair maya (.ma) file with a turntable animation setup in maya is downloadable here.
