Influence of the occlusion effect over the prediction-error feedback cancellation system in hearing aids.

This work presents a theoretical analysis of the prediction-error method-based adaptive feedback canceller in hearing aid applications. The studied scene takes into account the occlusion effect caused by the partial or complete closing of the ventilation opening. Such a situation may occur in high gain applications to avoid undesired whistling. Deterministic recursive equations and steady-state conditions were derived for the mean weight behaviour of the predictor and the adaptive filter. The expected theoretical predictions were compared to Monte Carlo simulations, showing very accurate agreement. The simulation results suggest the steady-state performance of this feedback canceller is not affected by the occlusion effect, however the occlusion is still perceived, being annoying to the user.