Toward bias minimization in acoustic feedback cancellation systems

Area of Science:

• Acoustics
• Signal Processing
• Hearing Aid Technology

Background:

• Acoustic feedback is a common problem in hearing aids, reducing sound quality.
• Existing feedback cancellation systems can introduce bias, affecting performance.
• Human psychoacoustics indicates tolerance to specific phase variations.

Purpose of the Study:

• To propose a novel technique for bias suppression in acoustic feedback cancellation systems.
• To utilize time-varying all-pass filters to achieve bias suppression.
• To analyze the impact of filter parameters on steady-state error.

Main Methods:

• Implementing all-pass filters with time-varying poles in the forward path of a hearing aid.
• Deriving algorithms for time-varying pole locations based on bias analysis.
• Examining practical factors like filter order, number of varying poles, and pole standard deviation.
• Conducting simulations using a hearing aid model.

Main Results:

• The proposed method achieves bias suppression through constant magnitude and varying phase frequency response.
• Analysis reveals the dependence of steady-state error on filter parameters.
• Strategies for pole variation are introduced and evaluated.
• Simulated results support the effectiveness of the proposed technique.

Conclusions:

• The novel technique effectively suppresses bias in acoustic feedback cancellation systems.
• Time-varying all-pass filters offer a promising approach for hearing aid audio enhancement.
• The study provides practical insights into optimizing filter parameters for improved performance.