All kinematic studies of human motion employ measurement techniques which introduce noise into displacement data. Commonly, the data, as time related functions, are differentiated to produce velocity and acceleration information. Unfortunately, differentiation amplifies the noise present to such an extent that additional signal treatment is essential.

The study was conducted to compare film generated acceleration curves with the analog acceleration curves of single segment movements. An instrumented segment was manually manipulated by the hand to produce analog records of the angular displacement and acceleration of the segment. Simultaneous filming of the segment produced synchronized displacement data. Acceleration functions were determined from these data using “finite difference”, “Chebyshev least squares polynomials” and “digital filtering followed by finite difference” techniques. Digital filtering combined with a first order finite difference technique produced acceleration data very closely approximating the analog signals. The other two techniques were clearly inferior. Data are provided to enable the reader to evaluate his own differentiation procedures.