Limitations caused by distortion in room impulse response measurements by swept sine technique

The significance of a room impulse response implies the requirement that its measurement should have a high level of accuracy in certain applications. One of the common problems in a measurement process is nonlinearity leading to the distortion of a room impulse response. Limitations caused by the distortion in room impulse response measurements by swept sine technique are analyzed here by the simulations and measurements. For the investigation, both linear and exponential swept sines are used as an excitation signal. In the simulations, this signal is modified by the nonlinearity model in the time domain with or without memory. On the other hand, the distortion in measurements is achieved either by applying the nonlinearity model or by using higher excitation level and a loudspeaker with a highly nonlinear characteristic. The results show that the most of distortion energy is located in the non-causal part of the extracted impulse response. In this way, the distortion products can be separated from the linear response. However, a part of the distortion energy remains in the causal part of room impulse response. As a consequence, in the case where distortion is present, the dynamic range of a measured response can be limited in a similar way as in maximum length sequence technique, although the saturation level (maximum dynamic range) is higher for swept sine technique. Thus, swept sine technique is also vulnerable to a certain extent to distortion that limits the quality of measured impulse response.The significance of a room impulse response implies the requirement that its measurement should have a high level of accuracy in certain applications. One of the common problems in a measurement process is nonlinearity leading to the distortion of a room impulse response. Limitations caused by the distortion in room impulse response measurements by swept sine technique are analyzed here by the simulations and measurements. For the investigation, both linear and exponential swept sines are used as an excitation signal. In the simulations, this signal is modified by the nonlinearity model in the time domain with or without memory. On the other hand, the distortion in measurements is achieved either by applying the nonlinearity model or by using higher excitation level and a loudspeaker with a highly nonlinear characteristic. The results show that the most of distortion energy is located in the non-causal part of the extracted impulse response. In this way, the distortion products can be separated from the linear response. However, a part of the distortion energy remains in the causal part of room impulse response. As a consequence, in the case where distortion is present, the dynamic range of a measured response can be limited in a similar way as in maximum length sequence technique, although the saturation level (maximum dynamic range) is higher for swept sine technique. Thus, swept sine technique is also vulnerable to a certain extent to distortion that limits the quality of measured impulse response