From lab images to the movie screen
Computer-generated imagery, usually shortened to CGI, means pictures or moving images produced by a computer, then combined with or used instead of filmed material. By the 1990s it was a standard part of big-budget film, but it took roughly two decades to get there. The progress came from several places at once: university labs, specialized effects companies, television studios and the research groups of large firms. There was no single starting point.
This article follows the commercial and cinematic path. The people who made experimental films are covered in the profile of animation pioneers, and the underlying techniques are described in the history of computer graphics, so here the focus is on how the images reached the screen.
The 1970s: small appearances
In the early 1970s, computers were mainly used to process or manipulate images in ways that filmmakers could not easily do by hand. The 1973 film Westworld is often cited for using digital processing to simulate the pixelated view of a robot character. The effect was made by scanning and processing frames rather than by generating a three-dimensional world, but it was one of the first times audiences saw computer work credited in a feature film.
A sequel, Futureworld (1976), included computer-generated images of a hand and a face, based on work done at the University of Utah. In 1977, Star Wars included a short sequence of wireframe graphics showing a planned attack route. These were small moments, and the main effects in these films were still physical models, optical compositing and painted backdrops. Even so, they demonstrated that computers could contribute to films.
Studios and labs take shape
During the late 1970s and early 1980s, a number of groups began to specialize in the field. Names often mentioned include Information International Inc., Digital Productions, Robert Abel and Associates, MAGI and the Computer Graphics Lab at the New York Institute of Technology. In addition, the film company Lucasfilm formed a computer division in the late 1970s, which developed software and hardware and later became a separate company, Pixar, in the mid-1980s.
These groups differed in approach. Some focused on rendering, the process of calculating how a modeled scene would look. Others built tools for artists, or sold hardware. The common feature was that they usually combined researchers with filmmakers, and that they were often short of both computing power and money. Producing a single frame could take a long time on the machines of that era, and a film might need many thousands of frames. These constraints are why the earliest uses tended to be short sequences.
Early 1980s: the first extended sequences
Two films from 1982 are often named as milestones. Tron placed live actors inside an environment that was largely computer-generated, with several companies contributing different sequences. It was ambitious, though much of the film’s look relied on traditional techniques combined with the computer work. Star Trek II: The Wrath of Khan included a sequence, produced by the Lucasfilm computer group, showing a planet transformed by a fictional device. It is often cited as an early example of a fully computer-generated shot in a feature film, and it used techniques for modeling natural shapes such as terrain.
In 1984, The Last Starfighter used computer-generated spacecraft in many scenes, instead of physical models. In 1985, a knight formed from a stained-glass window in Young Sherlock Holmes is commonly described as an early photorealistic character made with computer graphics. Those descriptions are simplified, since earlier work existed, but they mark a point at which the effects started to blend into live action.
Television, commercials and music videos
CGI did not appear only in films. Television stations adopted flying logos and animated title sequences in the late 1970s and 1980s, and advertisers used computer images to attract attention. Music videos also made use of the new look. Because these projects were shorter and had smaller budgets than a feature film, they gave studios a way to pay for equipment and practice.
The 1990s: effects become routine
By the end of the 1980s and into the 1990s, speed and software had improved. The Abyss (1989) included a water creature that interacted with live actors, and Terminator 2 (1991) used a liquid metal character. In 1993, Jurassic Park combined computer-generated dinosaurs with physical puppets and was widely seen as a turning point. In 1995, Toy Story became widely recognized as the first feature film made entirely with computer animation, a subject covered in the history of computer animation.
At this stage, the production process looked quite different from the early days. Large teams of artists modeled, textured, lit and animated characters, and compositing software merged the results with filmed footage. Hardware improved as workstations and later personal computers became capable of rendering at reasonable speeds.
Why it is a team story
It is tempting to give the story a single hero, but each milestone depended on many contributors: software authors, artists, producers and the researchers who developed the mathematics. Techniques were shared at academic conferences and sometimes moved with people between employers, which is part of why similar capabilities appeared at several places around the same time.
CGI has also not replaced older methods. Practical effects, miniatures and painting are still used alongside it, and the best results often combine methods. What the early years established was that computers could be one tool among several in image-making, and that the cost and quality of that tool would keep improving.