Motion Capture Systems
The term motion capture encompasses a range of systems for recording body motion quantitatively. They can be divided into two broad categories. In this chapter, we’ll get an overview of the main kinds of motion capture system used in music research.
The first category is camera-based systems. These derive motion data from video. They can be further sub-divided into marker-based and markerless systems.
In marker-based motion capture systems, reflective markers are put on the subject, and video is recorded from multiple angles. Three-dimensional positions are then triangulated, meaning that they are calculated from each of the video recordings.
Markerless motion capture systems are also being developed. Here it is not necessary to wear any markers on the body. The software is able to track distinctive features in the subject, and derive relative motion from the video. This motion data can be mapped onto an existing framework, such as a 3D model of the human body, hand or face. Markerless systems can be based on regular video cameras, but it is also common to use depth-sensitive cameras. Such cameras perceive depth using a pattern of near-infrared light.
The second category, sensor-based systems, derive more detailed motion data from a smaller number of particular points.
Among these are acoustic systems; which work by echolocation. They consist of a device which emits periodic acoustic signals, and a receiver. They work over long distances, however can be affected by temperature, atmospheric pressure, and humidity.
There are also electromagnetic tracking systems, which consist of a transmitter, and sensors. The transmitter emits low-frequency magnetic fields, and a sensor placed on the moving body measures the strength of each of those fields. From data on the field’s strength in each axis, you can calculate the sensor’s position and orientation.
A third kind of motion capture uses small inertial trackers, such as gyroscopes, accelerometers and magnetometers. These track the orientation, angular velocity, rate of angular change, and acceleration. Such sensors are good at detecting relative motion changes, but struggle more with finding the absolute position or orientation in space.
Different motion capture systems are better suited to different environments. When choosing what kind of system to use, a researcher must take into consideration: size of setup, complexity, cost, and whether she wants to collect qualitative or quantitative data.
To do marker-based or electromagnetic motion capture, you need a small dedicated lab. These systems can collect very detailed data on how individuals or small groups move. They give us detailed numerical data about a small number of points, from which concrete analyses can be produced. However, our analysis is limited to the specific points that we choose. It also means putting musicians and listeners in artificial environments, sometimes wearing heavy or awkward equipment.
Regular video is easy, fast, and low-cost. Video retains the expressive nuance of original performance, which sensors alone can’t. With infrared cameras, we can also capture motion in dark environments, like a nightclub. We can see how large groups move in their everyday settings. However, plain video of real-world situations is complex and difficult to interpret with certainty. Different researchers can derive diverse qualitative data about movements from the same video, depending on their aims and observations.
In the next video, Alexander takes us to the lab at the University of Oslo, and tells us more about how researchers choose what motion capture systems to use.
© University of Oslo