This demonstration presents processed audio files that aim to improve the cochlear response of a hearing-impaired person (HI) to that of a normal-hearing person (NH). The NH and HI responses are simulated using a biophysical model of human cochlear processing [1]. The adopted sound stimuli are sentences from a Dutch speech corpus [2]. The HI_unprocessed response shows how a typical sensorineural hearing loss affects the response along different cochlear center frequencies. Especially for higher frequencies, the HI cochlear response (black dashed) is up to 7 dB lower than the NH cochlear response.
The HI_CoNNear, NH_NALRP [3] and HI_NAL_NL2 [4] show cochlear response simulations of the hearing impaired listener after the speech was processed by a hearing-aid algorithm. Cochlear responses for two reference processing schemes (red, blue) are shown along with our novel neural-network-based CoNNear algorithm for augmented hearing (green).
Whereas NALRP is a linear processing scheme , both CoNNear and NAL_NL2 are nonlinear processing schemes. CoNNear uses an entirely different method for sound processing that is based on a closed-loop neural-network model [5] and is fundamentally different from the hearing-aid processing performed in NAL_NL2. However, the sound quality of our processing is equally good (or even better). Our CoNNear method reached its objective to return the HI_unprocessed response as close as possible towards the reference NH response. This demonstration shows that CoNNear-based methods for hearing-aid signal processing are a viable route to further explore.
[1] Verhulst, S., Altoe, A., & Vasilkov, V. (2018). Computational modeling of the human auditory periphery: Auditory-nerve responses, evoked potentials and hearing loss. Hearing research, 360, 55-75.
[2] Luts, H., Jansen, S., Dreschler, W., & Wouters, J. (2014). Development and normative data for the Flemish/Dutch Matrix test.
[3] Byrne D, Parkinson A, Newall P. Hearing aid gain and frequency response requirements for the severely/profoundly hearing impaired. Ear Hear. 1990 Feb;11(1):40-9
[4] Keidser, G., Dillon, H., Flax, M., Ching, T., & Brewer, S. (2011). The NAL-NL2 prescription procedure. Audiology research, 1(1), e24.
[5] Drakopoulos, F., & Verhulst, S. (2023). A neural-network framework for the design of individualised hearing-loss compensation. IEEE/ACM Transactions on Audio, Speech, and Language Processing.