TABLE OF CONTENTS In any measurement, the signal processing has the potential to affect the measured response. This section contains a brief overview of signal processing for aspects that are relevant to the majority of sound and vibration measurements described in this chapter. It is restricted to filter analysis and does not cover measurement techniques using the Fast Fourier Transform for which many other texts are available (e.g. see Randall, 1987).
From the viewpoint of the analyser, the analogue input signal is a time-varying voltage; therefore the processing applies to both sound and vibration signals. The analogue input signal from sound or vibration transducers is continuous in time, but in the analogue-to-digital conversion it is sampled so that the subsequent analysis is carried out on discrete time data. For filter analysis, the digitized input signal passes through the filters and then the detector as shown in Fig. 3.2.
The effect of signal processing on the measurement of reverberation time is a specific issue that is discussed in Section 3.8.3.
Signals can be considered in two distinct groups: stationary and non-stationary.
'Stationary' is used to indicate no change over time, or no change during the measurement period. Stationary signals are commonly used in sound insulation measurements and fall into two categories: random and deterministic signals. For stationary random signals, the probability density function for the signal does not change with time. An example of a...
