A note on acoustic decay measurements

Abstract Two sources of experimental errors in acoustic decay measurements are investigated: smoothing produced by the averaging device and “ringing” of the bandpass filter. Both linear and exponential averaging are considered. It is concluded that if it is only the average slope of the decay curve which is of interest, it suffices that the integration time of the detector is less than one quarter of the reverberation time of the system under test (with linear averaging), that the “reverberation time” of the detector is less than the half of the reverberation time of the system (with exponential averaging), and that the product of the filter bandwidth and the reverberation time of the system is at least 16. The requirements are much stronger if the important initial part of the decay is needed, however.