Hearing Noise Prediction for Low-Latency Speech Enhancement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing hearing devices face challenges in providing satisfactory noise cancellation and speech enhancement, with a need for improved methods to effectively reduce background noise and enhance speech.
Innovation Solution
A hearing device equipped with a set of input transducers, a processor, and a receiver that utilizes a neural network to predict future noise, allowing for noise estimation and subtraction from the input signal, thereby enhancing sound quality and reducing latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If traditional noise cancellation methods are used, then noise reduction can be achieved, but speech enhancement remains unsatisfactory and processing latency increases
Solution Approach 1:
The neural network predicts future noise levels before they fully occur, allowing the system to prepare noise cancellation in advance. The processor applies the neural network to predict future noise based on current and past noise levels, enabling proactive rather than reactive noise cancellation, which reduces the perceived latency in the noise reduction process.
Solution Approach 2:
Traditional noise cancellation methods are replaced with a neural network-based predictive system. The patent substitutes conventional signal processing mechanics with machine learning-based prediction, where the neural network analyzes patterns in transducer inputs to forecast noise levels, enabling more accurate and faster noise cancellation compared to traditional mechanical or algorithmic approaches.
2Object-affected harmful factors
If traditional noise cancellation methods are used, then noise reduction can be achieved, but speech enhancement remains unsatisfactory
Solution Approach 1:
The system continuously monitors transducer inputs and uses the neural network's predictions to dynamically adjust noise cancellation parameters. The processor compares predicted future noise with actual noise levels and refines the cancellation approach in real-time, creating a feedback loop that continuously improves speech enhancement quality while maintaining effective noise reduction.
Solution Approach 2:
The neural network predicts future noise parameters (such as noise level, frequency characteristics, and temporal patterns) based on current and historical data. By anticipating changes in noise parameters before they occur, the system can pre-adjust cancellation parameters to maintain optimal speech enhancement quality across varying acoustic conditions, rather than reacting after noise changes have already degraded speech quality.
Data Source
AI summary
A hearing device includes: a set of input transducers configured to provide a transducer input, the set of input transducers comprising a first input transducer; a processor configured to process the transducer input to obtain a first input and a second input, and to provide an electrical output signal; and a receiver configured to provide an audio output signal based on the electrical output signal; wherein the processor is configured to: apply a neural network to the second input for provision of a second output, wherein the second output is a prediction of future noise; obtain a noise estimate based on the second output; and subtract the noise estimate from a first input magnitude of the first input for provision of a first output; and wherein the electrical output signal is based on the first output.