 |
 |
 |
 |
 |
 |
HOME |
PRODUCTION |
TEAM |
CRITIQUE |
CRITIQUE2 |
ROLE |
Motion Capture Research According to the ‘White Paper’ released in 1995 by Scoot Dyer, Jeff Martin, and John Zulauf, motion capture “involves measuring an object’s position and orientation in physical space, then recording that information in a computer-usable form. Objects of interest include human and non-human bodies, facial expressions, camera or light position, and other elements in a scene.” [1] Research in motion capture (mocap) is increasing due to the strength of new software and hardware technology becoming available in the market. Although most people identify motion capture used specfically for the purpose of animation, this is not always the case. For example at Ohio State University, a motion capture lab has been designed to capture date from motion capture in order to study its uses for film, animation, performance, and biology. According to their website, “the lab consists of a 14 camera VICON8i system located in a 3000 sq. foot space, and uses VICON’s iQ2.5 for data collection and cleaning as well as Kaydara’s MotionBuilder 7.5 for attaching skeletons, motion cleaning, and retargeting of data.” [2] A human or animal or sometimes an inanimate object is used for data collection from a motion capture system. According to Lisa Marie Naugle, a dancer and Assistant Professor at UC of Irvine breakdown the procedure for motion capture is as follows: [3] -studio set-up (she usually uses 8 to 24 cameras for multiple captures)  Motion Capture Lab at VSR courtesy of http://www.digital-humans.org/mocap.htm There are various methods and configurations of motion capture. Richard Widgery from Kinetic Impulse describes the various types of motion capture: [4] Magnetic Motion Capture Systems There are field generators that are placed in a room which then magnetic sensors are placed on the person's / objects body which detect the individual positions and orientation in relation to the generator. Magnetic Sensor
Advantages: the position and rotation can be fairly accurate. The orientation in space can be identified. The information gathered can be used in realtime which can make it useful for live broadcast. The system is also less expensive than optical. Disadvantages: when the distance increases, the data becomes less accurate and noisy. The system overall is not as good as optical and it is prone to interference from metal that could be embedded in the floors or ceiling. The movement of the actor is limited since the system is directly connected to the computer.
Marker
 Courtesy of digital-humans.org/mocap.htm The data obtained from motion capture is not always accurate because of inaccuracies with the technology when it comes to capturing the right type of data. Unconscious body movement and stretching of skin can produce false data. That is why a clean up process is required after capturing the data to make sure what is used is as accurate as possible. Advantages: the performer is free to move around since there are no cables attached to the body. There can also be more performers on stage. The data captured is clean and detailed. Disadvantages: Since it is an optical system, inaccuracy could occur from light interference. Depending on the movement, the "reflective dots" can be blocked which can cause data loss or occlusion. Since there must be so many "reflective dots" placed on the suit, it can be uncomfortable for the actor to wear. The system can be very expensive and can range from $150000 to 250000.
In a Mechanical system, the performer would wear a metal skeleton suit (usually hooked to the performer's back). When the performer starts moving, the skeleton or "exo-skeleton" moves as well and sensors are triggered in each joint. There is also different types of Mechanical systems from gloves to arms or even animal shapes like Monkeys which can good for key framing. Advantages: the system is not prone to any interference from light or magnetic fields Disadvantage: Since the system is placed on the performer, there is no sense of "ground" or jumping is not recognized. The system must be constantly calibrated. The data is not alway reliable.
According to Professor Jessica Hodgins at Carnegie Mellon University, “A common problem with real time motion capture is the presence of noise. If you plot the motion data you will notice that the curves aren’t always smooth, there are some unsettledness or jerkiness. There are many filtering systems and techniques to eliminate the “bad data”. During offline capture, one can go back to the data and “clean it up” before applying it to the computer character.”[5] As new methods are being developed the data will begin to become more accurate and functional and its uses will spread to other fields.
Courtesy of Vicon.com References:
|
|
