Hearing Aid Signal Processing with Dual Source Separation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Hearing aids often experience unsatisfactory results due to interference signal amplification and freezing of source separation, where the coefficient set fails to adapt, leading to incomplete signal separation.
Innovation Solution
A method and apparatus for processing input signals in hearing aids that involve continuous determination and comparison of correlations using multiple coefficient sets, with the second source separation resetting and adapting based on the first, and periodic storage and recall of coefficient sets to prevent freezing and optimize separation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a single source separation is continuously adapted, then the coefficient set adapts to the acoustic situation, but the system may freeze in a local minimum and fail to achieve complete separation
Solution Approach 1:
A second source separation is continuously reset to prevent freezing, ensuring it remains ready to replace the first source separation if freezing is detected. This preliminary preparation allows rapid recovery from local minima without requiring complete reinitialization.
Solution Approach 2:
The system continuously compares the correlation values from the first and second source separations. When the second source separation achieves a smaller correlation (indicating better separation), this feedback triggers a replacement of the first coefficient set, enabling the system to escape local minima and achieve complete separation.
2Reliability
If a second source separation is continuously reset to avoid freezing, then freezing is prevented, but the second source separation has only minimally adapted its coefficient set and may provide unsatisfactory results
Solution Approach 1:
The second source separation is continuously reset with the current coefficient set from the first source separation, rather than with random initial values. This preliminary action preserves the adaptation progress while preventing freezing, allowing the second source separation to quickly converge to a better solution without starting from scratch.
3Reliability
If the coefficient set of the main source separation is replaced by the additional source separation, then freezing is corrected, but the main source separation delivers unsatisfactory results due to inadequate adaptation
Solution Approach 1:
The system continuously monitors and compares the correlation values from both source separations. Only when the second source separation achieves a smaller correlation (demonstrating superior performance) does the system replace the first coefficient set. This feedback mechanism ensures that replacement occurs only when it will improve signal separation quality.
Solution Approach 2:
The system dynamically switches between the first and second source separations based on real-time performance comparison. This dynamic approach allows the system to adapt to changing acoustic situations and escape local minima while maintaining high signal separation quality through continuous optimization.
Data Source
AI summary
Processing two input signals in a hearing aid with a first source separation and a second source separation is provided. A comparison of a first correlation with a second correlation and with a change in the first coefficient set as a function of a second coefficient set when the second correlation is smaller than the first correlation, with the second coefficient set being changed as a function of the first coefficient set when the second source separation is reset.