Once you've created your models, you're ready to photograph them. In the case of physical miniatures, this means positioning the model in a fixed location or on a movable mount, usually in front of a monochromatic background. The model is photographed, sometimes several times for different lighting or motion “passes.” If multiple models are required for one scene, such as a squad of fighter planes attacking a battleship, each model may be filmed separately, then composited, or pulled together, into one scene with the others, along with any extra passes that were required.

If the entire effects sequence is being created digitally, the process is somewhat different. CGI models are arranged in what is called a scene layout. This is equivalent to placing actors and scenery on a stage. Once in their positions, both the object and the camera can be moved in relation to each other, and to the world at large.

Digital Models

One advantage of digital models is that the visual effects artist has options regarding shooting. Each model can be filmed or rendered individually and composited later, as you would do with physical miniatures. The other choice is to set up an entire scene at once, including landscape, buildings, 3-D characters, clouds, trees, lighting, shadows, and so on. and then render the entire thing at the same time. Digitally rendering everything at the same time may seem simpler, but it takes much more computing power and limits your freedom when it comes to editing.

For example, if you're shooting a tractor-exploding sequence and realize the tractor should explode half a second earlier, you'd need to render the entire scene again, which might take hours. But rendering just the tractor, or even simply adjusting the existing explosion in the editing room, may only take a few minutes.

Digital Movement

If a digital scene involves movement (as compared to a still image), you'll most likely use a technique called keyframing. Remember how with stop motion the animator had to move the model in every frame? In computer animation, you only have to set up keyframes at certain points in a sequence. For example, let's say your script calls for an airplane to go into a barrel roll as it flies past the camera. For the first keyframe, you begin with the plane in an upright position, 100 meters away from the “camera object.” At a point five seconds later in the sequence, you move the plane to 100 meters on the other side of the camera, and you rotate it 360 degrees on the long axis, just as if it were doing the roll in real life. This is the second key-frame. The computer will calculate the change in position and axis between the two keyframes and then fill in the airplane's correct placement and attitude for every frame in between.

Of course, this is a highly simplistic example. To make this look at all real would require careful timing and adjustment to the movement curve. If you're just getting started, practice with very simple shapes that render quickly. Once you're confident in your skills, use the final model, but begin with simple test versions such as wireframe or animatics until the motion is perfect.

Another key element that you must consider is lighting. Whether real or digital, proper lighting can mean the difference between an imposing aircraft carrier and a bathtub toy. This is even more important if the effect will be added to existing footage shot on a stage or location.

Compositing

The final step is to put all the pieces, or elements, of your effect together. This is called compositing. Compositing can be thought of as layers of film stacked atop one another. If you're compositing your exploding-tractor footage, one layer has the background. The next has a moving stream. One has a farmhouse and another has just the smoke coming from the chimney. The topmost layers might be the tractor, then the explosion elements.

To keep all the layers clearly delineated from each other, each is defined by solid black areas called matting. The matte area can be used to define either the area that will be visible on that layer or the area that will be transparent, letting the layers behind show through. Each element layer and its accompanying matte layer can be moved, re-timed, color enhanced, brightened, dimmed, or adjusted in almost any way without altering the other layers.

Many 3-D software programs can create matte lines for you using alpha channels or other automatic settings. When using physical miniatures, mattes are often drawn by hand, or at least indicated manually, and then filled in by an editing computer later. To simplify this process, you would film your subject in front of a monochromatic background in a process traditionally called blue screening.

Blue Screens

When it comes to using blue screens, the truth is that the background screen doesn't have to be blue. Often it's green, or even orange. The strategy of using a single, vibrant color is to make it easier for the person (or computer) creating the matte to distinguish between what images remain and what needs to be matted out. So the choice of color is determined by the colors used in your scene. For example, if your evil alien has blue skin, you probably want to shoot him in front of a green screen. If the shot involves green grass and a blue sky, then orange might make things easier.

Television newscasts use a low-quality version of this in a process called chromakey. Controlled by a video mixing board, the camera is instructed to not “see” a specific color. A second image fed from a computer or even another camera is then keyed in to replace that color. If the weatherman is standing in front of a green wall, and the chromakey is set to that green, the weather map will appear behind him. If he makes the mistake of wearing the wrong color necktie, it might look as if he's being strangled by a low-pressure system.

Rendering in CGI

Since there's no film camera inside the digital world of your computer, the 3-D software program will provide an object that represents the camera. This can be pointed and moved just like a real camera. There will be settings for focus, zoom, and aperture, as well as special lenses and filters to match anything done in the real world. Once you're satisfied with your scene layout, there are several options for rendering:

• Wireframe or animatics let you quickly see the relationship of the objects to each other.

• Open GL renders a bit slower, but still fast enough for an easy check of color, shading, and shadows.

• Ray-tracing provides the most full realistic results.

As it renders, the computer saves the frames on a hard disk, either as individually numbered files, an image sequence, or a digital film file such as Quicktime or Windows Media. These can be replayed at a higher frame rate (usually the standard of twenty-four or thirty frames per second) to create the illusion of motion.

To more accurately portray the way the human eye perceives the behavior of objects on film, the software can add effects like motion blur, lens flare, and depth of field. Rising technology and falling prices have brought quality visual effects within reach of any filmmaker willing to take the time to learn the basics, and to experiment with just how much the finished product can be enhanced by visual effects.

Natural Effects

Besides film and camera behavior, natural effects are available on some software systems. Volumetric effects such as fog, clouds, and dust alter the way light passes to the camera. Particle effects can simulate rain or fire. To stay on schedule with a film containing several complex visual effects, an optical effects house will make use of a render farm, which ties several computers together, working on the same sequence simultaneously. For even larger films, studios will employ several effects companies, each assigned a different part of the film.