Contribution
Scene-aware compensation strategies for hearing aids
* Presenting author
Abstract:
Single-channel noise reduction (SCNR) and wide dynamic range compression (WDRC) are commonly used in hearing aids to attenuate background noise and improve the audibility of a target speech signal. However, these processing stages can have opposing effects, as SCNR may unintentionally reduce soft speech components while WDRC can reduce the signal-to-noise ratio. A behavioral experiment with hearing-impaired participants was conducted to evaluate the perceptual preference of four different WDRC configurations with and without a deep neural network (DNN)-based SCNR stage. The tested WDRC systems included fast- and slow-acting compression and a scene-aware compressor that switched between fast- and slow-acting compression depending on the estimated presence of speech. Moreover, an ideal reference system was tested that applied fast-acting compression to the target signal while the background noise and reverberation was attenuated by 12 dB. The results showed that the DNN-based SCNR stage improved the preference for all WDRC systems. The ideal system was the most preferred, followed by scene-aware compression and slow-acting compression both with SCNR, whose preference scores were close to the ideal system. The results suggest that it is possible to increase the effective compression of the target speech through either ideal or scene-aware compression without compromising quality.