Bone-Air Audio Signal Mapping for Clearer Speech Output
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio signal acquisition methods using bone and air conduction microphones face challenges in ensuring intelligibility and noise reduction, with bone conduction signals losing important information and air conduction signals having high noise levels.
Innovation Solution
A system utilizing a trained machine learning model to map bone conduction audio signals to equivalent air conduction data, determining a target set of equivalent air conduction data that indicates semantic content, and outputting a target audio signal with improved fidelity and reduced noise levels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If bone conduction microphone is used to acquire audio signal, then speech signal can be obtained, but important information is lost
Solution Approach 1:
The patent combines bone conduction audio signals and air conduction audio signals through signal processing to generate a target audio signal. The bone conduction signal provides clear speech information while the air conduction signal supplements frequency components, merging their advantages to reduce information loss and improve fidelity.
Solution Approach 2:
The patent introduces a machine learning model as an intermediary to map bone conduction data to equivalent air conduction data. This intermediary transforms the limited bone conduction signal into a more complete audio representation by learning the mapping relationship between the two signal types.
2Loss of information
If air conduction microphone is used to acquire audio signal, then complete speech information is obtained, but noise level increases
Solution Approach 1:
The patent extracts useful speech information from the air conduction audio signal while separating it from noise components. By using the machine learning model to identify and extract relevant speech features, the system retains complete speech information while removing harmful noise factors.
Solution Approach 2:
The patent uses the noise-affected air conduction signal in a way that converts its harmful noise characteristics into beneficial information. The machine learning model learns to identify speech patterns even in noisy conditions, transforming the noisy air conduction signal into a clean target audio signal that combines the advantages of both sensing methods.
3Device complexity
If only bone conduction audio signal is used, then device complexity is reduced, but audio quality deteriorates
Solution Approach 1:
The patent makes the audio system multi-functional by enabling it to process both bone conduction and air conduction signals through a unified machine learning framework. The same system can operate with either signal type or both together, providing audio quality enhancement without significantly increasing device complexity.
Data Source
AI summary
The method for audio signal generation may include obtaining a bone conduction audio signal and an air conduction audio signal. The method may also include obtaining a trained machine learning model that provides a mapping relationship between a set of bone conduction data derived from a specific bone conduction audio signal and one or more sets of equivalent air conduction data derived from a specific equivalent air conduction audio signal. The method may also include determining a target set of equivalent air conduction data corresponding to the bone conduction audio signal using the trained machine learning model based on the bone conduction audio signal and the air conduction audio signal. The method may further include causing an audio signal output device to output a target audio signal representing the speech of the user based on the target set of equivalent air conduction data.