Abstract
This work looks at using coherent averaging to measure transducer responses with high precision without using an anechoic chamber. The use of coherent averaging in building acoustics is familiar to those who use deterministic signals (e.g.. MLS) as the basis for their measurements. The primary concern is to achieve a sufficient signal to noise ratio so that results relate to the system being measured rather than unrelated noises. This technique requires conditions to be unchanging with time but coherent averaging can be used to advantage in situations where conditions are purposefully rendered time-varying. It is possible to select or reject contributions to a measurement by choosing some transmission paths to be time invariant and making others – ones we wish to remove from the measurement - varying. In previous work we have shown that by rotating a loudspeaker-microphone couple in a highly reverberant room we can measure the loudspeaker (or microphone) transfer function with the same accuracy achievable in an anechoic chamber by using this technique to suppress the reverberation in the room. In this more recent work we consider whether any particular deterministic signal – e.g. A log ‘chirp’ – is more advantageous for this application and whether the availability of the new soundfield-type microphones offers an improved way for making such measurements.