Physical theatre performer Rob van Vuuren gets some tips from seminar-leader Vincent Truter during our introductory motion capture seminar . The sphere at the back is the Polhemus dome that generates the magnetic field from which the sensors on Rob's body read their position.

We ended up using a combination of key frame and performance animation. The hands, face and feet were key frame animated, while the bodies were performance animated. Next time I think we’d try to capture the hands at the same time using sensor-gloves.

We used a Polhemus Ultratrak motion capture suite in combination with Kaydara’s Motionbuilder to capture our body animation. The Ultratrak, courtesy of SABC2, is the only of its kind in South Africa, but is quite a primitive suite compared with Polhemus’s current systems or the Vicon optical systems. It runs off a 386 processor and works well in a calibrated area of 4m x 2m x 4m (x,y,z) though we were not that experienced in the calibration method and could probably have increased this somewhat. The suit is attached to the computer by a tail of cables that is quite cumbersome for performance.

For certain actions the small acting area does create problems. Actions involving traveling over a distance (such as when Byrth turns and runs to catch the fish in his net) were tricky and required the stitching together of separate takes to create the final motion. We could also only accommodate one character at a time. This made co-coordinating motions between characters quite tricky, but not impossible.

The Ultratrak system uses a spherical magnetic field generator to create an energized area. This area is then calibrated by recording the readings of the sensors at specific grid points inside the space. The calibration of the space needs to be done at regular intervals to ensure accuracy, especially if the layout of electronic goods in the room changes.


The Polhemus sphere generate a magnetic field that can be read within a calibrated area (the grid area).
The movements of a performer wearing sensors within this magnetic field can be recorded
and this motion data applied to digital puppets.

For a recording session, the actor, wearing the sensor suit, is placed in the center of the space in a T-stance and their proportions are recorded and translated onto the virtual puppet. The sensors are assigned to joints on the virtual puppet so that, by moving the sensors attached to the suit, the actor animates the joints of the virtual puppet.

We used 13 sensors on the body at the heels, knees, pelvis, chest, shoulders, elbows, palms and head. We had a shortage of sensors and would have like to have tried more for the spine, hips and props.