MIMO Acoustic Echo Cancellation Through Direct–Residual Path Modeling

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional MIMO acoustic echo cancellation methods face challenges with high cross-correlation and time-varying loudspeaker rendering systems, leading to degraded sound reproduction quality and non-unique system identification.

Innovation Solution

The proposed acoustic processing device employs a dual-path model, using pre-defined MIMO FIR filters for direct echo cancellation and an adaptive MIMO FIR filter for residual echo cancellation, with adjustable weighting coefficients optimized through block-recursive least squares and GFDAF algorithms, to handle highly correlated loudspeaker signals and time-varying rendering algorithms.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If conventional approaches use nonlinear distortions, decorrelation filters, or phase modulation filter banks to reduce cross-correlation between tweeter signals, then acoustic echo cancellation performance is improved, but sound reproduction quality is degraded

Engineering Contradiction:
Improveacoustic echo cancellation performanceVSAvoidsound reproduction quality
Core Design Contradiction:
ReliabilityVSObject-affected harmful factors

Solution Approach 1:

The patent segments the echo path into direct path and residual path components. The direct path is handled by pre-defined MIMO FIR filters that model the acoustic transfer function from loudspeakers to microphones, while the residual path handles remaining echoes. This segmentation allows accurate echo modeling without degrading sound quality through aggressive decorrelation.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces pre-defined MIMO FIR filters as intermediaries to model the acoustic transfer function. These filters act as a mediator between the loudspeaker signals and the echo cancellation process, providing an accurate representation of the acoustic path without requiring nonlinear distortions or aggressive decorrelation that would harm sound quality.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Adaptability or versatility

If the time-varying rendering algorithm is used to adapt to different acoustic environments, then adaptability is improved, but system identification becomes non-unique and convergence is slowed

Engineering Contradiction:
Improveadaptability to acoustic environmentsVSAvoidconvergence speed
Core Design Contradiction:
Adaptability or versatilityVSProductivity

Solution Approach 1:

The patent performs preliminary action by pre-defining MIMO FIR filters that model the acoustic transfer function before the adaptive processing begins. These pre-defined filters capture the essential acoustic characteristics in advance, providing a solid foundation for echo cancellation that speeds up convergence and maintains accuracy even when the rendering algorithm changes over time.

Inventive Principle:
Principle #10Preliminary action

3Reliability

If the dual-path model with pre-defined and adaptive MIMO FIR filters is used, then echo reduction robustness is improved, but device complexity increases

Engineering Contradiction:
Improveecho reduction robustnessVSAvoidprocessing architecture complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the echo cancellation into two parallel paths: a direct path using pre-defined MIMO FIR filters for modeling acoustic transfer, and a residual path using adaptive MIMO FIR filters for handling remaining echoes. This segmentation distributes the processing complexity across two specialized paths rather than requiring a single complex adaptive system, improving robustness while managing complexity through functional decomposition.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP4205376B1Acoustic processing device for MIMO acoustic echo cancellation
Publication Date: 2025.10.22 HUAWEI TECH CO LTD
  • EP4205376B1 patent drawingFigure 1
  • EP4205376B1 patent drawingFigure 2
  • EP4205376B1 patent drawingFigure 3a

AI summary

An acoustic processing device (300) for performing MIMO acoustic echo cancellation is disclosed. The acoustic processing device (300) comprises a first signal reception unit (303) adapted to receive a plurality of loudspeaker signals and a second signal reception unit (305) adapted to receive a plurality of microphone signals. Moreover, the acoustic processing device (300) comprises a processing circuitry (301) adapted to enable echo reduction, the processing circuitry (301) being configured to determine for each microphone signal an estimated echo signal, wherein the estimated echo signal comprises an estimated direct echo signal and an estimated residual echo signal. The processing circuitry (301) is further configured to determine a respective echo reduced microphone signal based on the respective microphone signal and the estimated echo signal.