Wireless Acoustic Sensor Network Audio Enhancement

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing audio enhancement systems using wireless acoustic sensor networks face challenges in improving audio quality, especially in environments with non-stationary noise, due to lack of perfect synchronization and reliance on pre-defined source models.

Innovation Solution

Implementing a distributed source model adaptation using Non-negative Matrix Factorization (NMF) in a wireless acoustic sensor network, where each terminal adapts and shares audio source models to enhance audio quality without requiring an exact source model, allowing for real-time processing and adaptation to changing noise and speaker environments.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Object-affected harmful factors

If noise reduction methods are used based on spectral subtraction with distant microphones, then background noise can be reduced, but the quality is limited for non-stationary noise due to lack of perfect synchronization

Engineering Contradiction:
Improvebackground noiseVSAvoidnoise reduction quality
Core Design Contradiction:
Object-affected harmful factorsVSReliability

Solution Approach 1:

The patent implements dynamic source model adaptation where the system continuously updates source models based on current audio signals from multiple microphones. This allows the noise reduction algorithm to adapt to non-stationary noise conditions in real-time, resolving the contradiction between noise reduction capability and reliability for changing noise environments.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system employs feedback mechanisms where the processed audio signals from multiple microphones are used to refine and update source models iteratively. This feedback loop enables the system to improve its noise reduction performance continuously, addressing the synchronization issues that limit reliability in wireless acoustic sensor networks.

Inventive Principle:
Principle #23Feedback

2Reliability

If microphone array signal processing is used, then audio quality can be improved, but perfect synchronization of acoustic sensors is required which is not the case for wireless acoustic sensor networks

Engineering Contradiction:
Improveaudio qualityVSAvoidsynchronization requirement
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent enables each terminal in the wireless acoustic sensor network to independently adapt and share source models with others. This self-service approach allows the system to achieve microphone array processing benefits without requiring centralized synchronization control, as each node autonomously updates its source models based on local measurements and communicates with peers.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The system uses dynamic source model adaptation where synchronization requirements are continuously adjusted based on current audio conditions. The source models are updated in real-time to account for varying acoustic environments and terminal positions, eliminating the need for perfect synchronization while maintaining high audio quality.

Inventive Principle:
Principle #15Dynamics

3Reliability

If supervised NMF is used with pre-defined source models, then better performances are achieved, but the system cannot adapt to unknown sources in real-time

Engineering Contradiction:
Improvenoise reduction performanceVSAvoidadaptation to unknown sources
Core Design Contradiction:
ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent implements dynamic source model adaptation where the system transitions from static pre-defined models to continuously updating models. The source models are refined in real-time based on actual audio signals from multiple microphones, enabling the system to adapt to unknown sources while maintaining the performance benefits of supervised NMF.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system employs feedback mechanisms where processed audio signals are used to continuously refine source models. This feedback loop enables the system to learn and adapt to unknown sources over time, combining the reliability of pre-defined models with the versatility to handle novel acoustic environments.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP2888736B1Audio enhancement system
Publication Date: 2019.06.26 HUAWEI TECH CO LTD
  • EP2888736B1 patent drawingFigure 1
  • EP2888736B1 patent drawingFigure 2
  • EP2888736B1 patent drawingFigure 3

AI summary

The invention relates to an audio enhancement system (200), comprising at least two terminals (201a, 201b), each one comprising at least one acoustic sensor and processing means, wherein the acoustic sensors of the at least two terminals (201a, 201b) are wirelessly coupled with respect to each other forming an acoustic sensor network, each terminal (201a) being configured to provide information of at least one audio source model of the terminal (201a) to at least one of the other terminals (201b), wherein information of an audio source model of a terminal (201a) describes an audio characteristic of at least one audio source (205) impacting on the at least one acoustic sensor of the terminal (201a), and wherein the processing means of the at least two terminals (201a, 201b) are configured to perform audio enhancement processing based on the information of the audio source models of the at least two terminals (201a, 201b).