The global stop-motion record Stop Motion Database

history

Dynamation: from Willis O'Brien to Ray Harryhausen

Stop motion

Corroborated 2 sources on file · Needs a primary source — official site, end credits, or a festival record.
Spotted an error? Suggest a correction →

Definition

Dynamation is the composite-photography process Ray Harryhausen developed out of Willis O'Brien's stop-motion effects work, letting a frame-animated miniature model appear to occupy the same shot as live-action actors and sets.

Overview

Ray Harryhausen was thirteen when he saw Willis O'Brien's King Kong (1933) and resolved to become a stop-motion animator; a family friend arranged an introduction, and O'Brien critiqued the boy's early models and pushed him toward formal training in sculpture and art direction. The mentorship became a working relationship on Mighty Joe Young (1949), where Harryhausen animated most of the effects shots under O'Brien's supervision; the film won O'Brien the Academy Award for Best Special Effects, and two decades later Harryhausen made The Valley of Gwangi (1969) as an explicit tribute to him.

O'Brien's method on King Kong sandwiched a miniature model between two painted glass plates to fake its presence among live actors -- accurate, but slow. Working on the low-budget The Beast from 20,000 Fathoms (1953), Harryhausen simplified this into what became Dynamation: live-action footage was rear-projected onto a screen, the animated model was positioned and shot frame by frame in front of it, and a sheet of glass with blacked-out sections (a split-screen matte) blocked the areas where actors needed to appear in the same frame; the film stock was then rewound and re-exposed through the inverse of that matte to complete the composite. Producer Charles Schneer coined the trademark name in 1957, and the film was not marketed under the Dynamation name until The 7th Voyage of Sinbad (1958).

Refined through variants publicised as SuperDynamation and Dynamation 90, the process let Harryhausen build sequences O'Brien's era could not attempt: the multi-skeleton swordfight in Jason and the Argonauts (1963), animated across four and a half months with dozens of tracked movements synchronised per frame, and the Medusa sequence in Clash of the Titans (1981), his final film. The technique's real legacy is procedural rather than proprietary -- it established rear-projection matte compositing as the standard way to marry model animation to a live-action plate, a problem every subsequent generation of creature and model animators, up through the digital rig-removal era, has had to solve in some form.

A Dynamation sequence began on paper, long before a model was built or a plate was shot. The exhibition Ray Harryhausen: Myths in Motion (The Arc, Winchester, 2026-27) shows the detailed drawings Harryhausen made to pitch ideas to film producers. It hangs them beside paintings and sketches by John Martin, Gustave Doré and Joseph Gandy, which the curators present as early influences on his art and storytelling. The pairing explains something about the composites themselves. A split-screen matte only works if the image is planned around it: where the creature stands, where the actor stands, and where the join between them falls. Martin's and Doré's tiny figures dwarfed by vast architecture and landscape are a ready-made grammar for that kind of staging. Harryhausen's pitch drawing was more than concept art. It fixed the composition the whole process would have to deliver, which is one reason his sequences read as designed shots rather than effects inserted into a scene.

Workflow

  1. Shoot or select the live-action background plate first, with actors performing to marks that indicate where the animated model will appear.
  2. Rear-project that plate onto a screen behind the animation stage.
  3. Position the model in front of the screen and animate it frame by frame, exposing the model against the projected background.
  4. Mask the areas of frame where a live actor must remain fully visible using a glass matte painted black over those regions, so the model exposure does not overwrite them.
  5. Rewind the film and re-expose it through the inverse of the matte, printing the live-action actor into the masked-out areas so model and actor share the final frame.

Worked examples

Related

Links