Hearing Prosthesis Sound Processing Using Confidence Measures
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional hearing prostheses, such as cochlear implants, face challenges in distinguishing desired sounds from unwanted noise, particularly in noisy environments, which hinders speech understanding.
Innovation Solution
A sound processing method that generates an electrical representation of sounds with desired and noise components, estimates noise components, calculates a confidence measure for the noise component, and processes the sound representation based on this measure to enhance noise reduction and channel selection for effective stimulation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional hearing prosthesis processes all input signals, then the device complexity is reduced, but the ability to distinguish desired sounds from noise deteriorates
Solution Approach 1:
The patent segments the audio spectrum into multiple frequency channels and processes each channel independently. This allows selective noise reduction in specific frequency ranges while preserving desired speech signals in other channels, thereby improving sound discrimination without requiring complex global processing of the entire spectrum.
Solution Approach 2:
The patent applies different processing strategies to different frequency channels based on local characteristics. By analyzing the signal-to-noise ratio in each channel and applying targeted noise reduction only where needed, the system improves overall sound quality while avoiding unnecessary processing complexity in channels with acceptable signal quality.
2Object-affected harmful factors
If noise reduction processing is applied to all frequency channels, then the noise reduction effect is improved, but the loss of information deteriorates due to potential speech signal attenuation
Solution Approach 1:
The patent dynamically adjusts the noise reduction gain for each frequency channel based on real-time signal-to-noise ratio estimates. The processing intensity varies dynamically across channels and time, applying stronger noise reduction only to channels with poor signal quality while preserving channels with good signal-to-noise ratios, thus reducing speech signal attenuation.
Solution Approach 2:
The patent changes processing parameters (gain, threshold, smoothing constants) based on the estimated signal characteristics in each frequency channel. By adapting parameters to local signal conditions, the system achieves effective noise reduction in noisy channels while maintaining speech signal integrity in clean channels.
3Measurement precision
If multiple noise component estimates are generated and combined, then the measurement precision is improved, but the device complexity increases
Solution Approach 1:
The patent uses feedback mechanisms where noise estimates from previous time frames inform current noise estimation. This temporal feedback allows the system to track noise characteristics over time and improve estimation accuracy without requiring multiple independent estimation algorithms running simultaneously, thus avoiding excessive complexity.
Data Source
AI summary
A method for processing sound that includes, generating one or more noise component estimates relating to an electrical representation of the sound and generating an associated confidence measure for the one or more noise component estimates. The method further comprises processing, based on the confidence measure, the sound.