An Object Oriented Animation Engine

By Yatin S. ShelkeIntroduction to Animation Engines

So far, we've covered many of the core elements of an animation, including the Graphic Object, Animation Object, Camera, Frustum, Layers and Layer Folders. This article (and the next) will bring them all together to design .

Video games and Flash are examples of animation engines. An animation engine is the code in the animation application that drives the animation. The animation engine has many responsibilities:

Taking drawing data (animation objects, layer arrangements) and events as input Delegating the rendering tasks to animation objects Making sure that drawing is done in the order specified by layers and layer folders Scene graph management Collision detection (in games) Satisfying real time drawing requirements ( if any) Managing memory for animation objects, layers and layer folders.

A good animation engine should be able to handle all kinds of drawing and animation implementations without knowing how they are implemented. A good animation engine will not know whether the rendering is done with OpenGL or Direct3D. It will not know which platform-specific functions are used for keeping time; it shouldn't even know how memory is allocated. It must deal only with abstractions that are not platform or library dependent. It can then cater to a wide variety of animation applications on a variety of platforms. Animation engine design therefore is one of the ultimate challenges for an Object Oriented designer/programmer.

Animation Object examples

Let us build some animation objects based on the discussions thus far. The objects discussed in this article are:

An animation object that will move the camera (AnimCamera). An animation object that will draw a Graphic Object (class AnimFace), built on a simple Graphic Object, namely a rectangular GraphicFace. An animation object for moving a Graphic Object (class AnimTranslate). It will cause the Graphic Object in AnimFace to move.Camera based on OpenGL API

Before discussing the class AnimCamera, let's see how the camera works in a graphics API such as OpenGL (this discussion is based on the C API for OpenGL). Imagine a scene that contains several objects. Each object is drawn at a specific location in the virtual world. Let's call the virtual world coordinates "world coordinates" and use the prefix "world_". These world coordinates are mapped onto the screen to "screen coordinates" using prefix "screen_". When I use a triplet (x,y,z) for coordinates, that means world coordinates (since the real world has 3 dimensions). If the coordinates are described as a pair (x,y), these are screen coordinates, since screens are flat.

By default, the scene is drawn with the viewer at the origin of the virtual world's 3D Cartesian coordinate system, i.e. at (world_x,world_y,world_z)

View the Original article