Abstract: A method for adjusting an adaptive exponential filter, especially for noise reduction. The input of the filter is supplied with a series of time-discrete input signals containing both data and noise. An estimate of the standard deviation of the filter input signal has been determined prior to the execution of the adaptive exponential filter method. The method including: measuring the deviation between a current filter input signal to the filter and the prior sample period's output signal from the filter; determining the proportion of the deviation to the estimate of the standard deviation; and adjusting the discrete time constant of the filter in response to the proportion to provide a filter output signal at the output of the filter.