Adaptive feedback cancellation in hearing aids with clipping in the feedback path

Area of Science:

• Audiology
• Signal Processing
• Acoustics

Background:

• Adaptive linear filtering is standard for hearing aid feedback cancellation.
• Nonlinearities in the feedback path can degrade the performance of these algorithms.
• The assumption of linearity is often violated in real-world hearing aid applications.

Purpose of the Study:

• To investigate the impact of feedback signal clipping on hearing aid feedback canceller performance.
• To evaluate the benefits of applying identical clipping to the cancellation signal to model feedback path nonlinearities.
• To determine the optimal strategy for handling signal clipping in feedback cancellation systems.

Main Methods:

• Simulated clipping of the feedback signal at the hearing aid microphone.
• Applied identical clipping to the cancellation signal to mimic feedback path nonlinearities.
• Measured changes in added stable gain and filter coefficient misadjustment under various input levels.

Main Results:

• Feedback signal clipping reduced the achievable added stable gain and increased misadjustment at high input levels.
• Clipping the cancellation signal mitigated these negative effects, allowing for higher gain and reduced misadjustment.
• The benefits of cancellation signal clipping diminished at the highest input signal levels.

Conclusions:

• Nonlinearities, specifically clipping, significantly affect hearing aid feedback canceller performance.
• Modeling feedback path nonlinearities via identical clipping of the cancellation signal can improve performance.
• Careful consideration of input signal levels is necessary when implementing clipping strategies for feedback cancellation.