Low-Latency Hearing Aid Signal Processing With Neural Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing hearing aids suffer from high processing delays that affect the quality and timing of audio signal processing, which can impair the effectiveness of hearing compensation and user experience.
Innovation Solution
A hearing aid design that utilizes a neural network-based encoder and decoder to convert audio signals between domains, optimizing the processing to achieve low-latency signal compensation, allowing for high-frequency resolution and time alignment of audio signals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional audio signal processing methods are used in hearing aids, then processing accuracy can be maintained, but processing delay increases significantly
Solution Approach 1:
The patent replaces traditional mechanical signal processing methods (filter banks, FFT-based frequency domain processing) with a neural network-based system. The neural network processes audio signals directly in the time domain, eliminating the need for complex mathematical transformations and reducing processing delay while maintaining or improving processing accuracy through learned features.
Solution Approach 2:
The patent changes the fundamental processing parameters by operating directly in the time domain rather than transforming to the frequency domain. This parameter change from frequency-domain processing to time-domain neural network processing reduces the computational complexity and processing delay while maintaining signal fidelity and processing accuracy.
2Measurement precision
If high-frequency resolution processing is implemented, then audio quality improves, but processing time increases
Solution Approach 1:
The patent substitutes traditional frequency-domain analysis methods (which require long FFT windows for high frequency resolution) with a neural network that achieves high-frequency resolution through learned time-domain features. The neural network can identify frequency patterns and characteristics directly from time-domain signals without requiring extensive windowing or transformation, thus achieving high frequency resolution with minimal processing time.
3Reliability
If complex signal processing algorithms are used to compensate for hearing impairment, then compensation effectiveness improves, but device complexity increases
Solution Approach 1:
The patent replaces complex multi-stage signal processing algorithms (including frequency domain filtering, gain application, and time domain synthesis) with a unified neural network model. The neural network learns the complex transformations needed for effective hearing compensation directly from training data, consolidating multiple processing stages into a single integrated system that achieves comparable or superior compensation effectiveness with reduced algorithmic complexity.
Data Source
AI summary
Disclosed herein are embodiments of a hearing aid including at least one input unit for providing at least one stream of samples of an electric input signal in a first domain, at least one encoder configured to convert said at least one stream of samples of the electric input signal in the first domain to at least one stream of samples of the electric input signal in a second domain, and a processing unit configured to process said at least one electric input signal in the second domain. The at least one encoder is configured to convert a first number of samples from said at least one stream of samples of the electric input signal in the first domain to a second number of samples in said at least one stream of samples of the electric input signal in the second domain.