Motion capture, commonly abbreviated to mocap, is a technique for recording movement from a real performer or physical subject and converting that movement into digital data that can be analysed, edited or applied to a computer generated character, virtual object or other digital representation. In animation and visual effects, motion capture is often used to obtain human body movement, facial performance or other physical actions that would be time consuming to reproduce manually.
A performer wears markers, sensors or a specialised capture suit, or is recorded by cameras that track identifiable features. The resulting data describes positions, rotations, trajectories or other movement information over time. Motion capture does not automatically create a finished animation. Captured data normally requires cleaning, retargeting, editing, timing adjustments and artistic refinement. Its value lies in providing a detailed starting point based on observed human performance. The technology is used in film, television, games, virtual reality, biomechanics, research, training and other fields.
Optical motion capture systems use cameras to observe a performer and determine the movement of markers or identifiable points. Marker based systems commonly place reflective or otherwise trackable markers on the performer, while markerless systems use computer vision to estimate body landmarks directly from images. Multiple cameras allow three dimensional positions to be reconstructed through triangulation or related computational methods. The quality of the capture depends on camera placement, calibration, lighting, occlusion, marker visibility and tracking algorithms.
Optical systems can capture complex movement with high spatial resolution, but occlusion can cause problems when one body part hides a marker from the cameras. Markerless approaches can reduce the need for specialised suits but introduce their own challenges involving image quality, clothing, lighting, body overlap and inference. The resulting motion data can be highly detailed, but it remains dependent on the capabilities and assumptions of the capture system.
Inertial motion capture uses sensors containing combinations of accelerometers, gyroscopes and sometimes magnetometers to measure orientation and movement. These sensors are typically attached to body segments through a suit or straps. Inertial systems can operate without external cameras, making them useful in environments where optical capture is impractical. They can also allow performers to move through larger spaces without the same camera visibility requirements.
However, inertial sensors can experience drift, in which small orientation errors accumulate over time, and magnetic interference can affect systems that use magnetometers. Hybrid systems combine different sensor types or use external references to improve reliability. The appropriate capture method depends on the movement being recorded, required accuracy, environment, budget, portability and intended application. No single motion capture technology is ideal for every production.
Facial motion capture is a specialised area that records detailed facial movement for digital characters. Systems may use head mounted cameras, facial markers, optical tracking, depth cameras or machine learning based computer vision. Facial capture can record movement around the eyes, eyebrows, cheeks, mouth and jaw and then transfer or interpret those movements for a digital face.
The relationship between a performer's real facial anatomy and a stylised or nonhuman digital character is not necessarily direct, so retargeting and artistic adjustment are often required. Facial capture can preserve subtle performance details that would be difficult to reproduce manually, including small changes in expression and timing. However, the capture data does not inherently determine the final artistic performance. Animators and facial artists may adjust exaggeration, timing and deformation to ensure that the character communicates effectively and remains consistent with the project's visual style.
Retargeting is the process of adapting captured movement from the performer or source skeleton to the rig and proportions of a digital character. This is necessary because performers and digital characters rarely have identical anatomy. A person's arm length, joint placement, spine structure and range of motion may differ substantially from those of a creature, stylised character or digital double. Retargeting systems use relationships between corresponding joints and controls to transfer motion while compensating for these differences.
The result may still require manual correction. A captured foot movement may cause unwanted penetration through the floor, or a human shoulder motion may deform a creature's body incorrectly. Animators therefore clean the data, adjust constraints, correct contacts and modify poses. Motion capture is best understood as a source of movement information rather than a complete replacement for animation expertise.
Motion capture is particularly prominent in video games because interactive characters require large amounts of movement for walking, running, combat, climbing, jumping and other actions. Capturing performers can provide realistic motion efficiently, after which the data can be processed into animation libraries and blended at runtime. Film productions also use motion capture for digital characters and creatures, although workflows vary considerably.
Performance capture is a broader term sometimes used when body movement, facial performance and voice or other aspects of acting are captured together. The terminology should not be treated as perfectly standardised across the industry. Some productions use specialised capture stages with controlled cameras, while others use portable systems for location based work. The technique can also be combined with keyframe animation, allowing animators to retain captured realism while deliberately changing poses, timing or expression.
Motion capture has important limitations. Captured movement may contain noise, tracking errors, occlusions, unwanted performer motion or artefacts caused by equipment. A performer may also move differently when wearing equipment or working within the physical limitations of a capture stage. Physical human movement does not automatically translate convincingly to a creature with different proportions or to a stylised character whose design requires exaggerated motion. Capture can also reproduce incidental movements that are undesirable for the final performance. Cleaning and editing may therefore require substantial time. Furthermore, motion capture records what a performer did under specific conditions rather than automatically determining what the character should do in the narrative context. Artistic direction remains necessary to choose, modify and integrate the captured performance.
Motion capture has expanded beyond entertainment into scientific and technical applications. Biomechanics researchers can use movement data to study gait, posture and athletic technique, while medical and rehabilitation applications can analyse movement patterns. Sports technology can record body mechanics, and virtual reality systems can use tracking to map a user's physical actions into a digital environment. Industrial and ergonomic studies can investigate how people interact with tools and workspaces.
In entertainment, the continuing development of computer vision, real time tracking and machine learning is making capture more portable and increasingly capable of operating without traditional marker suits. Nevertheless, the fundamental principle remains consistent: physical movement is measured, represented as digital data and subsequently interpreted or applied within another system. Motion capture is therefore a bridge between physical performance and digital animation rather than a self sufficient substitute for artistic animation.
