Quality Aspects of Digital Hearing Aid Algorithms

Abstract:
For several modern signal-processing schemes for hearing aids (e.g. single- or multi-channel noise reduction schemes, dynamic compression, dereverberation) technical measures are used during the development process for performance prediction that are either inappropriate or can only approximately predict the success for hearing impaired (and also normal hearing) people. In this context, not only the speech intelligibility, but also the subjective signal quality is of importance. In this study we present and compare subjective and objective assessment methods. Particularly, the objective quality measure PEMO-Q that is based on modern auditory models is evaluated for different hearing aid algorithms. This approach appears to be significantly superior to the classic SNR-based approaches in some quality aspects.