Subjective Evaluation of Reverberation Times of Sound Fields with Non-Exponential Decays

Two different techniques for the evaluation of reverberation time are compared for the subjective reverberance in an auditorium. The first is an existing technique where the reverberation time is obtained from the best straightline fit to the measured decay curve that has been smoothed by the Schroeder integrated impulse method. The second is a proposed technique which accounts for the 50-ms average response time of the human ear and the 3-dB just-detectable loudness change of the ear, and is so-called subjective evaluation of reverberation time: As a result, the total acoustic energies in successive 50-ms intervals of the decay are calculated, and considering the decay of the total energy, the reverberation time is calculated by interpolation. The direct sound is included in the first 50-ms interval, and the 3-dB just-detectable loudness change is included for the first and its adjacent 50-ms intervals. For exponential or near-exponential decays at the associated audience seats in the auditorium, the reverberation times produced by both techniques are close to each other and correlate well to the impression of reverberance from a subjective test at these seats. The first technique is thus sufficient to obtain a reverberation time which can describe the extent of reverberation sensed by the ear. However, for highly non-exponential decays at other seats, the reverberation times evaluated by both techniques can be significantly different, and those from the second technique correlate better to the results of the subjective test. The reverberation time calculation must thus incorporate the perception limits of the ear where the first 10 dB of the decay is affected most by the limits. This paper provides an understanding of the evaluation of reverberation time for subjective reverberance under different decay conditions.