Abstract :
A mathematical solution is given for the design of a quantiser suitable for quantising, with minimum mean-square error, a signal with an exponential amplitude probability-density function. This is particularly relevant to speech, the probability-density function of which approximates to the exponential case.