Why Photorealistic People Can Be So Costly to CGI
There is no universal “single most expensive” CGI task: a film’s hardest shot depends on its scale, realism, and how many shots must be made. But photorealistic digital humans are among the most demanding effects to create. Viewers know exactly how a face should move, so even tiny mismatches can make a computer-generated person feel wrong.
Every Detail Has to Work Together
A digital human is more than a detailed model. Artists and technicians must reproduce skin’s subtle translucence, the movement of facial muscles, the wet reflections in eyes, and the way hair shifts. The face also has to perform: a smile or glance must convey the intended emotion, not merely resemble the actor in a still image.
These elements interact. Skin that looks convincing under one light may look waxy under another. A perfectly modeled face can still fail if its expression is mistimed, its gaze misses the other character, or its movement lacks the small irregularities of a living person. Matching a real actor’s performance often means carefully tracking expressions and then refining the result by hand.
Why One Shot Can Take So Long
The work grows with every close-up, lighting change, and adjustment. A revision to the face may require the team to rerender the shot—the computer’s process of calculating the final image—then check that the skin, eyes, hair, and shadows still fit together. If the character appears in many shots, the effect must remain consistent across different angles and performances.
Other effects, especially water, fire, crowds, or large-scale destruction, can also demand enormous effort. The practical answer is less a permanent champion than a recurring challenge: the most expensive CGI is often the effect that must fool the viewer at close range, for long enough to carry the story.
