From a galloping horse and a man in a striped suit to creatures that act live on set: 150 years of turning movement into numbers.
People first studied movement with photographs, then traced it by hand. Psychologists showed that a few moving dots are enough to see a person. Gait labs, games and then films built systems that turn those dots into 3D skeletons, and actors learned to perform creatures. Today capture works outdoors, underwater, and with no suit at all.
Sinbad: Beyond the Veil of Mists, Pentafour, Chennai and Los Angeles
2004
Feature film shot entirely on a capture stage
The Polar Express, Sony Pictures Imageworks and Vicon
2014
India's first photorealistic mocap feature
Kochadaiiyaan
June 1878Photographing and tracing motion
1878 – 1972
Photographing and tracing motion
Muybridge and Marey take movement apart with cameras. Marey's black suit with white lines is the first marker suit. Fleischer and Disney trace real performers onto drawn characters.
1878
June 1878
A galloping horse, frozen in 12 photos
Eadweard Muybridge, for Leland StanfordPalo Alto, California, USA
Muybridge lined up a row of cameras beside a racetrack. Threads across the track fired the shutters one after another as the horse ran past, catching each phase of its stride.
Why it mattered. It proved that a camera could take movement apart, and showed that all four hooves leave the ground at once.
Marey, a doctor who studied how animals move, built a chronophotographic gun that took 12 pictures a second, all from one viewpoint, to study birds in flight.
Why it mattered. Measuring motion from many timed pictures is exactly what a mocap camera still does, only faster.
Marey dressed a man in black with thin white stripes and dots along his arms and legs, then photographed him walking and running in front of a black wall. Only the white lines showed, so each photo became a moving stick figure.
Why it mattered. Black suit, bright marks, dark background: it is the idea behind every optical marker suit today.
Fleischer projected film of a real person, one frame at a time, onto glass and traced it onto paper. Dave Fleischer, dressed as a clown, was the reference for Koko the Clown in Out of the Inkwell.
Why it mattered. Copying a real performer's movement into a drawn character is the goal of motion capture, done by hand.
After Fleischer's patent ran out, Disney animators filmed live actors and used rotoscoping as reference for Snow White and the Prince, to make their movement look human.
Why it mattered. It showed that audiences read real human motion as more believable, a big reason mocap exists.
Johansson shows a few moving lights read as a person. Labs and gait clinics build camera systems that triangulate markers, and games and adverts start using them.
1973
A person made of 10 to 12 lights
Gunnar JohanssonUppsala, Sweden
Johansson fixed small lights to an actor's main joints and filmed them moving in the dark. When the dots were still, viewers saw a random pattern. The moment they moved, people instantly saw someone walking, running or dancing.
Why it mattered. It proved that a few moving points hold most of what we read in body movement, which is why a few dozen markers can capture a performance.
Carol Ginsberg and Delle Maxwell, MITCambridge, Massachusetts, USA
A performer wore a suit with LEDs on the joints. Two cameras with special light sensors each reported the 2D position of every LED, and a computer combined them into 3D points that drove a stick figure live.
Why it mattered. It was triangulation from two cameras driving a digital character, the core of optical mocap.
Vicon, Oxford Metrics (Julian Morris and team)Oxford, England, UK
Vicon's early camera systems were built for clinical gait analysis: measuring how children with cerebral palsy walk before and after surgery. Hospitals and universities helped set standard ways of placing markers.
Why it mattered. Medicine came first. The marker sets and methods that films use grew out of gait labs.
For the advert Brilliance, the team filmed a real actress with reference points marked on her body, and a computer tracked those points to move a chrome computer-graphics robot.
Why it mattered. It was one of the first times captured human motion drove a CG character on television.
For the scene where people walk behind an airport X-ray scanner, the team recorded the actors with motion capture to animate glowing CG skeletons. Accounts differ on how much of the capture data ended up in the final shots.
Why it mattered. It is often named as an early feature-film use of motion capture.
Sega's arcade games Virtua Fighter and Virtua Fighter 2 used motion capture to animate their 3D fighters. Soon many studios built their own capture stages for games.
Why it mattered. Games became one of the biggest users of mocap, and still are.
Mocap drives main characters. Gollum makes it acting, The Polar Express shoots a whole film on a stage, and Avatar adds head-mounted cameras and a virtual camera.
1999
A main character from mocap
Ahmed Best and Industrial Light & MagicLondon, UK and San Rafael, USA
Star Wars: Episode I featured Jar Jar Binks, the first main character in a feature film created with motion capture from an actor's performance.
Why it mattered. Mocap moved from background extras to a leading role.
Pentafour Software (later Pentamedia Graphics)Chennai (then Madras), India, and Los Angeles, USA
Sinbad: Beyond the Veil of Mists was the first feature-length computer-animated film made with motion capture. Actors were captured in Los Angeles in 1997, and a couple of hundred animators in Madras built the film.
Why it mattered. An Indian studio led the first full-length mocap film, years before Hollywood's big ones.
Final Fantasy: The Spirits Within aimed for photorealistic human characters, animated with motion capture. It was the first widely released film made this way.
Why it mattered. It showed how hard it is to make digital humans feel alive, even with real motion.
Andy Serkis and Weta DigitalWellington, New Zealand
Serkis acted Gollum on set with the other actors, then again in a marker suit on a capture stage. In The Two Towers his body motion streamed live onto the CG Gollum; animators used his face and voice as the guide for Gollum's face.
Why it mattered. It turned mocap into acting. People began to talk about performance capture and the actor behind the creature.
Robert Zemeckis, Sony Pictures Imageworks and ViconCulver City, California, USA
The Polar Express used 72 cameras around a capture area about 10 feet (3 m) square. Up to four actors were captured together, bodies and faces: each wore about 48 body markers and 152 small face markers.
Why it mattered. It was the first film shot entirely on a motion-capture stage, and showed face capture's promise and problems.
Andy Serkis and Weta DigitalWellington, New Zealand
For King Kong, Serkis studied gorillas at London Zoo and in Rwanda. He had 135 small markers glued to his face, about two hours of make-up a day, to drive the ape's expressions.
Why it mattered. Face markers mapped a human face onto a very different creature face, a big step in retargeting.
James Cameron, Lightstorm and Weta DigitalLos Angeles, USA and Wellington, New Zealand
For Avatar, actors wore skull caps with a tiny camera on a boom in front of the face, instead of hundreds of glued markers. Cameron used a virtual camera to see the Na'vi in Pandora live, on a capture stage described as six times larger than any before.
Why it mattered. The head-mounted camera and virtual camera became standard tools of performance capture.
Two light sensors in 1983 grew to dozens of cameras on big film stages (SIGGRAPH, AWN, Wikipedia).
1983 Graphical Marionette (MIT): 2 cameras
2004 The Polar Express: 72 cameras
2012 Kochadaiiyaan shoot at Pinewood: 48 cameras
2011 – today
Capture anywhere
Capture moves outdoors and underwater, game engines make it real time, and machine learning finds poses in ordinary video and phones.
2011
August 2011
Apes captured outdoors
Weta Digital, with Andy Serkis as CaesarVancouver, Canada and San Francisco, USA
Rise of the Planet of the Apes took performance capture out of the studio onto real locations, including a battle on a Golden Gate Bridge set. A camera that could see the capture markers in daylight made it possible.
Why it mattered. Mocap no longer needed a dark stage; creatures could act beside live actors on real sets.
Soundarya Rajinikanth, with RajinikanthChennai, India, and Pinewood Studios, London, UK
Kochadaiiyaan was captured at the Centroid lab at Pinewood Studios with 48 cameras, and Rajinikanth performed his triple role in a marker suit. Big Indian effects films such as Ra.One (2011) and Baahubali (2015) also leaned on VFX, though published accounts say little about how much mocap each used.
Why it mattered. It was billed as India's first photorealistic motion-capture feature.
Ninja Theory, Epic Games, Cubic Motion and 3LateralAnaheim, California, USA (SIGGRAPH)
On stage at SIGGRAPH, actress Melina Juergens performed as Senua from the game Hellblade. Her body and face were captured and rendered live in a game engine, then cut into a finished scene in minutes. It won Best Real-Time Graphics and Interactivity.
Why it mattered. Game engines made real-time performance capture good enough to be close to final.
Zhe Cao, Yaser Sheikh and team, Carnegie Mellon UniversityPittsburgh, USA
OpenPose was the first open-source system to find the body keypoints of many people in ordinary video in real time, using deep learning and part affinity fields.
Why it mattered. It started the markerless wave: pose from video, no suit.
The iPhone X's TrueDepth camera projects over 30,000 invisible infrared dots onto the face. Apple's ARKit turns them into 52 blend-shape weights, such as jaw open and brow up, many times a second.
Why it mattered. Facial capture that once needed a studio became something anyone could try at home.
A study compared OpenPose with several video cameras against a marker system. About 80% of joint errors were under 30 mm, but some were over 40 mm. Markers still win when millimetres matter.
Why it mattered. It measured the trade-off: markerless is easy and cheap, optical is precise.
James Cameron, Lightstorm and Wētā FXManhattan Beach Studios, California, USA
For Avatar: The Way of Water, the team captured actors underwater in a huge tank built at Manhattan Beach Studios (reports of its size vary). Air bubbles act as tiny wiggling mirrors that fool the cameras, so actors and crew held their breath while performing.
Why it mattered. Capture moved into places once thought impossible for optical systems.