Echo Cancellation Device for Hands-Free Voice Systems

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Miniature hands-free voice communication systems face challenges in effectively eliminating nonlinear echo components, which often damage the near-end voice and degrade duplex performance, due to limitations in acoustic design and phase consistency of transmitters.

Innovation Solution

The solution involves using an echo elimination device and method that utilizes amplitude matching and voice probability estimation to separate voice from echo, employing an array echo elimination unit, self-adaptive echo elimination unit, and residual echo elimination unit, which do not require strict phase consistency of transmitters, and dynamically adjust filtering intensity based on voice probability.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Object-affected harmful factors

If nonlinear echo component elimination is performed to suppress echo, then echo is reduced, but near-end voice is damaged and duplex performance degrades

Engineering Contradiction:
ImproveechoVSAvoidduplex performance
Core Design Contradiction:
Object-affected harmful factorsVSReliability

Solution Approach 1:

The patent applies local quality by differentiating between linear and nonlinear echo components and applying different elimination strategies. The linear echo component is eliminated using self-adaptive filtering, while the nonlinear echo component is handled through residual echo filtering that specifically targets nonlinear distortion without damaging near-end voice. This localized approach to different echo types resolves the contradiction by protecting voice quality while suppressing echo.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent employs dynamic adjustment of filtering intensity based on voice probability estimation. The residual echo filtering module dynamically adapts its operation to the current acoustic environment, adjusting the degree of echo suppression applied. This dynamic approach allows the system to maintain high duplex performance while effectively suppressing echo, as the filtering strength is optimized in real-time rather than being fixed.

Inventive Principle:
Principle #15Dynamics

2Measurement precision

If array signal processing with transmitter array is used to achieve full-duplex call, then echo extraction and voice separation are improved, but transmitter phase consistency requirements increase and acoustic design becomes strictly required

Engineering Contradiction:
Improvevoice separation accuracyVSAvoidacoustic design requirements
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the echo elimination process into distinct functional modules: array echo elimination unit, self-adaptive echo elimination unit, and residual echo elimination unit. Each module handles a specific aspect of echo suppression, allowing the system to achieve effective voice separation without requiring strict phase consistency across all transmitters. This segmentation reduces the overall acoustic design complexity while maintaining measurement precision.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces an intermediary approach by using a primary transmitter and auxiliary transmitter configuration where the auxiliary transmitter(s) provide additional echo information that is processed separately. This intermediary structure allows the system to extract echo components without requiring the main primary transmitter to meet stringent phase consistency requirements, thereby reducing device complexity while maintaining voice separation accuracy.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Reliability

If self-adaptive echo filtering is applied to eliminate linear echo component, then linear echo is eliminated without damaging near-end voice, but nonlinear echo component remains and may damage receiver

Engineering Contradiction:
Improvenear-end voice protectionVSAvoidnonlinear echo component
Core Design Contradiction:
ReliabilityVSObject-affected harmful factors

Solution Approach 1:

The patent implements continuity of useful action by cascading multiple echo elimination units in sequence. The array echo elimination unit continuously processes signals to remove linear echo components, followed by the self-adaptive echo elimination unit that continues the elimination process, and finally the residual echo elimination unit that addresses nonlinear components. This continuous multi-stage approach ensures both near-end voice protection and comprehensive echo suppression, including nonlinear components that could damage the receiver.

Inventive Principle:
Principle #20Continuity of useful action

Data Source

PatentEP2858338B1Echo cancellation device and method for small-scale hands-free voice communication system
Publication Date: 2017.08.23 GOERTEK INC
  • EP2858338B1 patent drawingFigure 1~3
  • EP2858338B1 patent drawingFigure 4(a)~4(b)
  • EP2858338B1 patent drawingFigure 4(c)~5(a)

AI summary

The present invention discloses an echo elimination device and method for a miniature hands-free voice communication system. The system comprises a receiver, a primary transmitter and an auxiliary transmitter, a distance from the primary transmitter to the receiver being greater than that from the auxiliary transmitter to the receiver. The device comprises an array echo elimination unit, a self-adaptive echo elimination unit and a residual echo elimination unit, which are structurally cascaded in turn. The array echo elimination unit, with inputs being a signal of the primary transmitter and a signal of the auxiliary transmitter, performs array filtering to obtain one path of output signals; the self-adaptive echo elimination unit, with the input signals being a signal of the receiver, the output signal of the array echo elimination unit and a signal of the auxiliary transmitter, performs self-adaptive filtering to obtain two paths of output signals; the residual echo elimination unit, with the input signals being the two paths of output signals of the self-adaptive echo elimination unit, performs voice probability estimation and echo matching to obtain an echo-eliminated voice signal. Thus, the duplex performance can be enhanced, and the phase consistency of the transmitters is not strictly required.