Microphone Array Beamforming Stability via Auxiliary Weight Vector

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing microphone apparatuses face challenges in maintaining directional audio quality due to variations in microphone sensitivity and location, which affect the accuracy of beamforming algorithms like MVDR, leading to noise distortion and instability.

Innovation Solution

The proposed microphone apparatus employs a three-microphone array with a main beamformer and an auxiliary beamformer, where the auxiliary beamformer provides an independent weight vector to stabilize the steering vector determination, reducing noise and improving accuracy by using an auxiliary voice detector and adaptive filtering techniques.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If a standard MVDR beamformer is used with adaptive steering vector estimation, then the beamformer can adapt to varying source locations, but the determination of when microphone signals comprise the desired signal becomes problematic leading to noise distortion and instability

Engineering Contradiction:
Improveadaptive beamformingVSAvoidsignal stability
Core Design Contradiction:
Adaptability or versatilityVSReliability

Solution Approach 1:

The patent introduces an auxiliary beamformer as an intermediary component that processes microphone signals through a different computational path than the main beamformer. This auxiliary path provides a stable reference for steering vector determination, mediating between the adaptive requirements and stability concerns. The auxiliary beamformer's output is used to guide the main beamformer's adaptation without directly causing the instability that plagues standard adaptive approaches.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The beamforming system is segmented into two distinct processing paths: a main beamformer for primary signal processing and an auxiliary beamformer for stable steering vector estimation. This segmentation allows each component to specialize - the main beamformer handles adaptability while the auxiliary beamformer provides stability - resolving the contradiction between these two requirements.

Inventive Principle:
Principle #1Segmentation

2Device complexity

If microphone sensitivity variations and location deviations are not compensated, then the system is simpler, but the beamforming accuracy deteriorates due to finite production tolerances and aging

Engineering Contradiction:
Improvesystem complexityVSAvoidbeamforming accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The system performs self-calibration by using the auxiliary beamformer to automatically estimate and compensate for microphone sensitivity variations and location deviations. The auxiliary beamformer continuously monitors the actual system behavior and adjusts the steering vector estimates accordingly, enabling the system to self-correct for manufacturing tolerances and aging effects without external intervention.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The auxiliary beamformer provides continuous feedback about the actual microphone array performance to the main beamformer. This feedback loop allows the system to adapt to sensitivity variations and location deviations in real-time, maintaining beamforming accuracy despite manufacturing tolerances and component aging.

Inventive Principle:
Principle #23Feedback

3Measurement precision

If calibration techniques are applied to handle deviating microphone characteristics, then beamforming accuracy is maintained, but the system complexity and computational load increase

Engineering Contradiction:
Improvemicrophone calibration accuracyVSAvoidcalibration system complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The calibration function is merged with the normal beamforming operation by integrating the auxiliary beamformer into the existing signal processing pipeline. Rather than adding a separate calibration system, the auxiliary beamformer performs both calibration and signal processing functions simultaneously, reducing overall system complexity while maintaining accuracy.

Inventive Principle:
Principle #5Merging (Combining)

Data Source

PatentEP3675517B1Microphone apparatus and headset
Publication Date: 2021.10.20 GN AUDIO AS
  • EP3675517B1 patent drawingFigure 1
  • EP3675517B1 patent drawingFigure 2
  • EP3675517B1 patent drawingFigure 3

AI summary

The present invention relates to a microphone apparatus (30) with a main beamformer (31) that provides a main output audio signal (SM) as a beamformed signal by applying a main weight vector (BM) to a main input vector (MM). A main beamformer controller (32) repeatedly determines a main steering vector (dM) and adaptively determines the main weight vector (BM) in dependence on the main steering vector (dM) and the main input vector (MM) to increase the relative amount of voice sound (V) from the user (6) in the main output audio signal (SM). The microphone apparatus (30) further comprises an auxiliary beamformer (33) that provides an auxiliary beamformer signal (SF) as a beamformed signal by applying an auxiliary weight vector (BF) to an auxiliary input vector (MA) that is a subset of the main input vector (MM), and an auxiliary beamformer controller (34) that adaptively determines the auxiliary weight vector (BF) to increase the relative amount of voice sound (V) from the user (6) in the auxiliary beamformer signal (SF). The main beamformer controller (32) determines the main steering vector (dM) in dependence on the auxiliary weight vector (BF). This may enable the main beamformer controller (32) to utilize information derived independently of the steering vector (dM) and may thus improve stability and/or accuracy of the estimation of the steering vector (dM), and may further reduce the computation load for the main beamformer controller (32).