Comparison of room impulse response measurement methods

Abstract

One of the most important measures of a system such as a room is impulse response. Almost all characteristics of that system can be calculated directly from the impulse response, i.e. in room acoustics numerous objective parameters such as reverberation time, early/late ratios and RASTI can all be obtained from it. In room acoustics, due to large reverberation times, longer impulse responses are needed and a large dynamic is desired. The impulse response of a system is defined as the output when a perfect pulse, or delta function, is applied to it. Hence, the simplest of all techniques is just to apply a short duration pulse to the room and then measure its response. Other broadband signals can be used, with various processing, to calculate the impulse response. One example is the use of a chirp, a short duration sine sweep. This approach has more energy output than a pulse, but requires post-processing. If one stretches out the chirp so that it is a continuous repeating sweep, even more energy is output, and this requires similar post-processing. Finally, a pseudorandom noise, or a maximum-length sequence signal, with a Fast Hadamard transform can be used, to calculate the impulse response. For all of these techniques to be valid, the system under study must be linear and time invariant (except with the pulse method where time invariant is not a problem). This paper compares each of these approaches, and describes the strengths and weaknesses of them when applied to room acoustics