Clinical critical band estimation.

A modified version of the method of loudness summation, developed for clinical critical band estimation, is presented. The stimuli are noise bands centred around 1 kHz. The standard procedure includes 1) determination of the "sensitivity curve": the loudness difference between broad band noise and narrow band noise as a function of the level. For clinical use this function is necessary for evaluation of the level with the most rapid growth of loudness with bandwidth, i.e. the optimum level for sharp determination of the critical bandwidth. 2) Determination of the critical band function, i.e. the difference in sound pressure level required for equal loudness of the test noise band and the reference as a function of the bandwidth. This determination is performed at a level with maximum growth of loudness, evaluated by the sensitivity curve. 3) Critical band estimation by a mathematical model, described in another work (Bonding et al., 1978) from the data obtained. The method is analysed regarding feasibility and reproducibility.