Systems and methods for in-ear microphone voice signal processing

By extracting features from in-ear microphone speech signals using a data-driven model on a computing device and applying time-varying correction gain for filtering, the problem of limited frequency bandwidth of in-ear microphones in noisy environments is solved, thereby improving signal quality and intelligibility.

CN122181007APending Publication Date: 2026-06-09EERS GLOBAL TECHNOLOGY CO
View PDF 1 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN Β· China
Patent Type
Applications(China)
Current Assignee / Owner
EERS GLOBAL TECHNOLOGY CO
Filing Date
2024-06-13
Publication Date
2026-06-09

AI Technical Summary

Technical Problem

In the prior art, in-ear microphones have limited frequency bandwidth for capturing voice signals in noisy environments, resulting in reduced signal quality and intelligibility. Furthermore, existing bandwidth extension technologies are complex and expensive.

Method used

By using a data-driven model on a computing device to extract features from in-ear microphone speech signals, predicting and applying time-varying correction gain for filtering, bandwidth extension is achieved and speech signal quality is improved.

Benefits of technology

It adapts to voice content in real time, improving the signal quality and intelligibility of in-ear microphones in noisy environments and simplifying the bandwidth expansion process.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122181007A_ABST
    Figure CN122181007A_ABST
Patent Text Reader

Abstract

A method is provided that includes obtaining a first in-ear microphone (IEM) voice signal and a reference voice signal; decomposing the voice signal into signal components spanning frequency bands; determining, for each frequency band, a time-varying true value correction gain based on a difference between corresponding first features of the signal components; determining second features of the first IEM voice signal; configuring a model to predict a time-varying correction gain of an estimated true value correction gain from the second features; obtaining a second IEM voice signal; decomposing the second IEM voice signal into signal components spanning the frequency bands; determining third features of the second IEM voice signal that match the second features; executing the model for predicting the correction gain based on the third features; applying the correction gain to the second IEM voice signal to obtain a processed IEM voice signal; and outputting the processed IEM voice signal.
Need to check novelty before this filing date? Find Prior Art