Binaural pre-processing for contralateral sound field attenuation and improved speech-in-noise recognition

Area of Science:

• Auditory Neuroscience
• Signal Processing
• Acoustics

Background:

• Speech comprehension in noisy environments is a significant challenge.
• Current hearing aid technologies struggle to isolate target speech from competing sounds.
• Spatial audio processing offers potential solutions for improving speech intelligibility.

Purpose of the Study:

• To develop and evaluate a novel algorithm for enhancing speech perception in free-field conditions.
• To attenuate the sound field contralateral to the target ear using linear subtraction of the weighted contralateral stimulus.
• To assess the effectiveness of this method in improving signal-to-noise ratio and speech reception thresholds.

Main Methods:

• Mathematical justification for weighting the contralateral stimulus based on head-related transfer functions (HRTFs).
• Implementation of the algorithm in the frequency domain.
• Technical and experimental evaluation with normal-hearing listeners in simulated free-field conditions.

Main Results:

• Substantial improvements in signal-to-noise ratio (up to 30 dB) and objective intelligibility.
• Significant improvements in speech reception thresholds (SRTs) for sentences in noise (up to 10 dB).
• Minimal impact on virtual sound-source lateralization, with performance dependent on HRTF weighting parameters.

Conclusions:

• Contralateral HRTF-weighted subtraction effectively enhances speech intelligibility in challenging acoustic environments.
• The algorithm shows promise for improving the listening experience for users of binaural hearing devices.
• Further research can optimize the azimuth range for weight calculation to maximize benefits.