A nonlinear regression approach to the evaluation of reverberation times using Schroeder’s integrated impulse response method

Abstract

Reverberation decay curves can be obtained by backward integration of room impulse responses [M. R. Schroeder, J. Acoust. Soc. Am. 37, 409–412 (1965)]. The evaluation of reverberation times is often achieved by a regression line fitting the reverberation decay curves. However, the successful application of this method requires either a careful choice of the integration limit or estimation of the mean‐square value of background noise where background noise is present in the room impulse responses to be evaluated. In the present paper, an alternative method for evaluating reverberation times from Schroeder’s decay curves using a nonlinear iterative regression approach is proposed. The regression process is based on a nonlinear curve model using the generalized least‐square error principle rather than a linear model as used in the linear regression. The present paper will describe the principle of this approach and discuss the advantages and disadvantages involved. Comparison of results obtained using this approach and alternative methods will also be presented.