In-Ear Audio Signal Compensation Using Machine Learning

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In-ear microphones struggle to capture clear audio signals due to sound attenuation through the Eustachian tube, resulting in muffled and less recognizable sound compared to traditional air-emitted audio signals, especially in noisy environments.

Innovation Solution

A method using a machine learning model, trained with both in-ear and outer audio signals, to convert in-ear audio signals into compensated audio signals that mimic air-emitted audio signals, employing convolutional neural networks or deep neural networks to enhance speech quality and recognition.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Object-affected harmful factors

If in-ear microphones are used to collect sounds from the ear canal, then the wireless earphone can function in noisy environments, but the high frequency sound is attenuated and the audio appears muffled

Engineering Contradiction:
Improveambient noise interferenceVSAvoidaudio signal quality
Core Design Contradiction:
Object-affected harmful factorsVSMeasurement precision

Solution Approach 1:

The patent introduces an intermediary processing system consisting of multiple microphones (in-ear and outer microphones) and signal processing algorithms. The outer microphones capture ambient sound while the in-ear microphone captures the user's voice through the ear canal. These signals are processed together to compensate for the attenuation and improve audio quality, resolving the contradiction between noise rejection and signal fidelity.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent applies parameter changes by using equalization filters to modify the frequency response of the captured audio signal. The system adjusts amplitude and phase parameters across different frequency ranges to compensate for the high-frequency attenuation caused by the ear canal transmission, thereby restoring natural audio quality while maintaining noise rejection capabilities.

Inventive Principle:
Principle #35Parameter changes

2Device complexity

If traditional audio signal processing is used for in-ear microphones, then the device structure remains simple, but the audio signal remains muffled and less recognizable

Engineering Contradiction:
Improvesignal processing structureVSAvoidspeech recognition accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent implements preliminary action by capturing reference audio signals from outer microphones before processing the in-ear microphone signal. These reference signals contain ambient noise characteristics that are used in advance to create noise profiles and compensation filters, which are then applied to the in-ear signal to enhance speech recognition accuracy without significantly increasing device complexity.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent transitions from single-dimensional signal processing to multi-dimensional processing by incorporating both temporal and spectral analysis. The system processes audio signals in the time domain for noise estimation and in the frequency domain for spectral compensation, adding dimensional depth to the processing approach and significantly improving speech recognition without proportionally increasing hardware complexity.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Data Source

PatentUS10848855B2Method, electronic device and recording medium for compensating in-ear audio signal
Publication Date: 2020.11.24 HTC CORP
  • US10848855B2 patent drawing
  • US10848855B2 patent drawing
  • US10848855B2 patent drawing

AI summary

A method, an electronic device, and a recording medium for compensating an in-ear audio signal are provided. The method is applicable to the electronic device having a processor. In the method, an in-ear audio signal transmitted through an inner ear when a user speaks is captured by using an in-ear microphone, and an outer audio signal transmitted through air when the user speaks is captured by using an outer microphone in a training stage. Then, a machine learning model of audio signals is established for an objective function and is trained by using the in-ear audio signal and the outer audio signal. Finally, the in-ear audio signal captured by the in-ear microphone is converted into a compensated audio signal by using the trained machine learning model in an online stage, and the compensated audio signal is output.