Abstract:

Although there are a number of electronic/acoustically based objective intelligibility rating methods (as opposed to subject based word score tests), each suffers from a number of limitations. The essentially nonlinear behavior of most sound systems can and does affect the accuracy of electronically based assessment methods, yet such effects receive little or no mention in the literature. The paper sets out to redress the balance and highlights, and discusses many of the practical limitations of carrying out such measurements. In particular, %Alcons, RASTI, and STI methods are reviewed. The paper describes how system nonlinearities, distortion, compression, digital signal processing, and bandwidth restrictions, as well as late acoustic reflections, can all affect and possibly corrupt the final result. The choice of test stimulus and method of analysis are also shown to affect the apparent result. [See NOISE-CON Proceedings for full paper.]