Neural Feedback Cancellation for Hearing Aid Entrainment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Adaptive feedback cancellation algorithms in hearing devices suffer from performance degradation and audible artifacts, particularly in the presence of strongly self-correlated input signals like pitched speech and music, leading to reduced stable gain and entrainment artifacts.
Innovation Solution
Implementing a neural network-driven feedback cancellation system that identifies acoustic features and predicts target outputs to control feedback cancellation, trained offline and deployed on a hearing device processor, mitigating entrainment without altering the output sound quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If adaptive feedback cancellation algorithms are used in hearing devices, then feedback cancellation is achieved, but performance degradation and audible artifacts occur in the presence of strongly self-correlated input signals
Solution Approach 1:
The patent introduces an intermediary processing stage between the adaptive feedback canceller and the output. This intermediary uses neural network-based acoustic feature identification to detect self-correlated inputs and applies corrective processing to prevent entrainment artifacts while maintaining feedback cancellation effectiveness.
Solution Approach 2:
The system dynamically changes processing parameters based on acoustic feature analysis. When self-correlated signals are detected, the system adjusts feedback cancellation parameters to prevent entrainment, while maintaining normal operation for other signal types. This adaptive parameter adjustment resolves the contradiction by context-dependent optimization.
2Reliability
If signal processing systems process the output to reduce entrainment, then added stable gain is restored, but additional audible sound quality artifacts are introduced
Solution Approach 1:
The system performs preliminary acoustic feature analysis and takes preventive action before entrainment artifacts can manifest. By identifying self-correlated inputs early in the processing chain, the system adjusts feedback cancellation parameters proactively to prevent artifact generation, rather than attempting corrective processing after artifacts occur.
Solution Approach 2:
The system implements a feedback loop where acoustic feature identification continuously monitors the input signal characteristics and feeds this information back to the feedback cancellation processor. This closed-loop control enables real-time adaptation of cancellation parameters based on actual signal conditions, preventing entrainment while maintaining stability gain.
Data Source
AI summary
Disclosed herein, among other things, are apparatus and methods for neural network-driven feedback cancellation for hearing assistance devices. Various embodiments include a method of signal processing an input signal in a hearing assistance device to mitigate entrainment, the hearing assistance device including a receiver and a microphone. The method includes performing neural network processing to identify acoustic features in a plurality of audio signals and predict target outputs for the plurality of audio signals, and using the trained neural network to control acoustic feedback cancellation of the input signal.