Abstract
The borderland between perfectly periodic and a-periodic sounds is explored, using trains of sharp pulses as a test signal. Jitter, i.e. the r.m.s. deviation from the average interval proves to be a meaningful measure of the imperfection of periodicity. Measurement of jitter in a bowed string are presented. Limits for the perception of pitch and imperfect periodicity are discussed in terms of relative jitter. Finally, an experiment is described showing that the hearing system is able to perceive jitter, without having aces to the components of the sound spectrum.