Microphone Coherence Filtering for Wind Noise Reduction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing audio systems struggle to effectively reduce wind noise in signals generated by microphones, particularly in environments with air turbulence, which degrades audio quality.

Innovation Solution

An audio system processes signals from multiple microphones to determine coherence between them, applying a filter based on spectral gain and a band-pass filter to reduce wind noise by filtering out uncorrelated noise components.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If wind noise reduction filtering is applied to microphone signals, then audio quality is improved, but signal processing complexity increases

Engineering Contradiction:
Improveaudio qualityVSAvoidsignal processing complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent replaces complex mechanical wind noise reduction mechanisms with a computational approach using coherence-based spectral filtering. Instead of using physical barriers or mechanical filters, the system uses digital signal processing to analyze the coherence between multiple microphone signals and selectively filter wind noise frequencies, thereby reducing hardware complexity while maintaining audio quality.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent changes the parameter of noise filtering from fixed-frequency filtering to dynamic coherence-based filtering. By calculating the coherence between signals from multiple microphones across different frequency bins and time frames, the system adaptively identifies and filters wind noise components, improving audio quality without requiring overly complex fixed-structure filters.

Inventive Principle:
Principle #35Parameter changes

2Object-affected harmful factors

If coherence-based filtering is used to reduce wind noise, then wind noise reduction effectiveness is improved, but computational requirements increase

Engineering Contradiction:
Improvewind noise reduction effectivenessVSAvoidcomputational energy consumption
Core Design Contradiction:
Object-affected harmful factorsVSUse of energy by moving object

Solution Approach 1:

The patent applies partial coherence-based filtering by selectively processing only those frequency bins and time frames where wind noise is detected through coherence analysis. Instead of uniformly processing all signal components, the system identifies regions with high coherence indicative of wind noise and applies filtering only to those specific portions, reducing overall computational energy consumption while maintaining effective wind noise reduction.

Inventive Principle:
Principle #16Partial or excessive action

3Measurement precision

If multiple microphones are used for coherence analysis, then wind noise filtering accuracy is improved, but device complexity increases

Engineering Contradiction:
Improvewind noise filtering accuracyVSAvoidmicrophone array complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent introduces coherence as an intermediary metric to bridge the signals from multiple microphones. By calculating the coherence between microphone signals across frequency bins and time frames, the system creates a unified measure of correlation that simplifies the comparison and filtering process, thereby improving filtering accuracy without proportionally increasing device complexity.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP3968659B1Systems and methods for reducing wind noise
Publication Date: 2025.09.24 META PLATFORMS TECHNOLOGIES LLC
  • EP3968659B1 patent drawingFigure 1
  • EP3968659B1 patent drawingFigure 2
  • EP3968659B1 patent drawingFigure 3

AI summary

The disclosure is generally directed to a system for reducing wind noise. A system includes one or more processors coupled to a non-transitory computer-readable storage medium having instructions encoded thereon that, when executed by the one or more processors, cause the one or more processors to obtain signals respectively generated from two or more microphones during a time period, the signals representing acoustic energy detected by the two or more microphones during the time period, determine a coherence between the signals, and determine a filter based on the coherence. The filter is configured to reduce wind noise in one or more of the signals.