As I noted in February (Part 3), reverberation time can be measured using one of two methods. In the first, the room or space under acoustic scrutiny is allowed to reach a steady state condition with a pink noise test signal. The pink noise signal is switched on for a few seconds and then suddenly switched off, and the resultant sound decay is recorded and measured. The pink noise employed is usually a broadband signal covering the range from at least 80Hz to 10kHz. This is then filtered into octave or 1/3-octave bands during the calculation process.
Alternatively, filtered bands of noise can be used. This technique is sometimes useful under either noisy conditions or where very low reverberation times have to be measured. As established last time, although reverberation time is defined as the time it takes for the sound to decay or reduce by 60dB, in practice, this is not what happens these days. The first 5dB of the decay is disregarded and then the decay is measured over either a 20dB or 30dB window to yield either the T20 or T30 values, extrapolated as though they had been measured over the complete 60dB decay period.
Although this dramatically reduces the signal-to-noise ratio required to make a reliable measurement, it is often forgotten that a much larger “noise-free” decay in the sound amplitude is required because the noise floor can affect the result. In practice, the base of the decay curve must be at least 10dB above the ambient noise level. This, plus the 5dB at the top of the decay curve that is to be ignored or computationally thrown away, means that, for a T30 measurement (i.e., 30dB decay measurement), the acoustic signal must be at least 45dB over the background noise [5 + 30 + 10 = 45dB] or, in the case of a T20 measurement, 5 + 20 + 10 = 35dB. Figure 1 summarizes the situation for a T20 measurement.
Achieving a 45dB signal-to-noise ratio using a broadband signal can, in many situations, demand very high sound signal levels, particularly when the spectral content of the background noise is taken into account. For example, let’s assume that we are measuring in a room where the background noise is reasonably well controlled and is, say, 40dBA. Thus, at the measurement locations, the sound level must be at least 85dBA, which is quite loud.