Smart Hearing Device Voice Noise Distinction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional hearing aids fail to effectively distinguish between natural and non-natural languages, leading to poor performance in amplifying voices and canceling noise, as they do not account for real-time surrounding noise and lack precise vocal range recognition.
Innovation Solution
A smart hearing device system with dual microphones, one near the mouth and one farther away, analyzes sound data to differentiate between natural and non-natural languages using graph patterns, adjusting amplification and frequency settings based on user-specific hearing data to provide customized audio feedback.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional hearing aids amplify all sounds in the vocal range (500 Hz to 5000 Hz), then voice amplification is provided, but external noise cannot be distinguished from voices leading to poor noise cancellation performance
Solution Approach 1:
The patent segments the broad vocal range (500 Hz to 5000 Hz) into multiple sub-bands with different center frequencies and bandwidths. Each sub-band is processed independently to identify whether it contains voice or noise, allowing precise voice detection while canceling noise in other frequency regions.
Solution Approach 2:
The patent applies different processing characteristics to different frequency sub-bands. Each sub-band has customized filtering and amplification parameters tailored to its specific characteristics, enabling localized voice enhancement without amplifying noise in other regions.
2Productivity
If conventional hearing aids use simple vocal range classification, then processing is computationally simple, but voice highlighting and noise cancellation functions are not implemented well
Solution Approach 1:
The patent implements dynamic processing where the system continuously adapts to changing acoustic environments. Processing parameters such as amplification gains and noise cancellation strengths are dynamically adjusted based on real-time analysis of each sub-band's content, improving noise cancellation while managing computational complexity through efficient algorithms.
Solution Approach 2:
The patent replaces simple mechanical frequency filtering with sophisticated digital signal processing techniques. Digital algorithms analyze temporal and spectral characteristics of sound waves across multiple sub-bands, enabling intelligent distinction between voice and noise that goes beyond fixed mechanical filtering.
3Measurement precision
If conventional hearing aids do not analyze real-time surrounding noise, then device operation is simple, but voice amplification performance deteriorates due to inability to distinguish natural from non-natural language
Solution Approach 1:
The patent performs preliminary analysis of acoustic characteristics during device initialization and ongoing operation. By pre-characterizing the user's voice patterns and storing reference profiles, the system prepares voice identification templates in advance, enabling faster real-time comparison and distinction between natural language (voice) and non-natural language (noise) without excessive computational burden during active use.
Data Source
AI summary
The inventive concept relates to a smart hearing device for providing a control parameter and feedback for a natural language or a non-natural language determined by analyzing sound data, which includes a receiving unit that receives sound data of a voice signal and a noise signal from a first microphone and a second microphone being formed at one side, a determination unit that compares digital flow of the sound data with a previously stored graph pattern to determine a natural language or a non-natural language for the sound data, a processing unit that matches similar data for the determined natural language or non-natural language, based on a database including a natural language area and a non-natural language area, and a providing unit that provides a user with a one-sided sound converted by setting a control parameter in a natural language or a non-natural language specified according to the matched similar data.