Binaural filterbank hearing aid simulation V5

In this presentation, we describe the Binaural Filterbank Hearing Aid Simulation (BFHAS) which is a MATLAB-based package that simulates commonly available hearing aid signal processing. This package allows the user to process custom audio signals (e.g., noisy speech; music) with a suite of hearing aid algorithms. Linear or non-linear fitting prescription fitting formulae can be used for gain compensation. The hearing-aid algorithms include spectral-subtraction noise suppression, wide dynamic-range compression (WDRC), and frequency compression based on sinusoidal modeling of the high frequencies of the signal. Users can vary signal processing settings like number of bands; add/modify spectral subtraction; gains, knee points, attack and release times for WDRC; compression ratios and cutoffs for frequency compression etc. The simulation provides the processed hearing aid output that includes the compensation and hearing aid quality and intelligibility metrics for the input stimuli. Future updates will include the room-simulations and built-in noise generators to simulate different levels of signal degradation. We also provide a user guide and examples illustrating the use of different package modules. This package will be made available for download. [Work supported by NIH R01 DC012289 and by a grant from GN ReSound to University of Colorado.]