Interactive fitting of multiple algorithms implemented in the same digital hearing aid

Abstract

The main purpose of this study is to examine the applicability of an adaptive and interactive optimization strategy to fine-tune three hearing-aid algorithms simultaneously: dynamic compression, temporal signal enhancement, and noise reduction. The optimal combination of these three algorithms was determined by a multidirectional pattern search with an adaptive step size. Additionally, we applied a round-robin procedure to validate the results of the optimization procedure. For both procedures the listeners were asked to compare two consecutive, differently processed sentences in continuous and fluctuating background noises on speech intelligibility. Ten hearing-impaired and four normal-hearing subjects participated. The reliability and consistency of the multidirectional pattern search was low, especially for the fluctuating noise condition. The results of the round-robin procedure did not correspond closely with the pattern search results. These findings suggest that the current implementation of a multidirectional optimization procedure has not yet proven to be applicable for the necessary individual fine tuning of complex signal processing strategies, when the objective is to maximize speech intelligibility.