Microphone-Based Acoustic Echo Cancellation Without External Reference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing acoustic echo cancellation technologies rely on external reference signals, which become ineffective in environments with nonlinearities or unknown audio sources, and fail when external reference signals are unavailable, leading to difficulties in isolating desired audio in noisy conditions.
Innovation Solution
The implementation of microphone-based acoustic echo cancellation, where one microphone signal is used to cancel another, utilizing adaptive filters to determine noise/echo cancellation coefficients and buffer them for later application, allowing for echo and noise cancellation without relying on external reference signals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If external reference signals are used for acoustic echo cancellation, then echo cancellation performance is improved in controlled environments, but the system fails when external reference signals are unavailable or in environments with nonlinearities
Solution Approach 1:
The system uses its own microphone signals as reference inputs for the adaptive filter, eliminating dependency on external reference signals. The microphone signals inherently contain the echo path information needed for cancellation, allowing the system to serve itself and adapt to any acoustic environment without external assistance
Solution Approach 2:
Instead of using external reference signals to cancel echo, the patent inverts the approach by using the microphone signals that contain the echo as the reference. This reversal allows the system to extract and cancel echo components from the same signals that carry the echo information, making the system self-sufficient and adaptable
2Measurement precision
If traditional acoustic echo cancellation is used, then desired audio can be isolated in quiet environments, but the system becomes ineffective in noisy environments with background noise or unknown audio sources
Solution Approach 1:
The system continuously updates the adaptive filter coefficients using the error signal (difference between actual and desired output) and the reference microphone signals. This feedback mechanism allows the filter to learn and adapt to changing acoustic conditions, including background noise characteristics, thereby maintaining precise audio isolation in dynamic environments
Solution Approach 2:
The adaptive filter coefficients are dynamically adjusted in real-time based on the incoming microphone signals and acoustic conditions. This dynamic adaptation enables the system to maintain optimal echo cancellation performance across varying noise levels and acoustic environments, rather than relying on fixed parameters
3Measurement precision
If external reference signals are required for echo cancellation, then cancellation accuracy is maintained in controlled settings, but system complexity increases and operation becomes more difficult in uncontrolled environments
Solution Approach 1:
The system eliminates the need for external reference signal inputs by using its own microphone array signals as references. This self-service approach simplifies deployment and operation in uncontrolled environments, as the system can autonomously extract and utilize the necessary echo path information from its existing signal inputs without requiring additional external equipment or signal sources
Data Source
AI summary
A multi-microphone device that can perform acoustic echo cancellation (AEC) without an external reference signal. The device uses the audio data from one of its microphones as a reference for purposes of AEC and acoustic noise cancellation (ANC). The device determines filter coefficients for an adaptive filter for ANC when cancelling one microphone signal from another microphone's signal. Those filter coefficients are buffered and delayed and then used for AEC operations cancelling one microphone signal from another microphone's signal. When desired audio (such as a wakeword, speech, or the like) is detected, the device may freeze the coefficients for purposes of performing AEC until the desired audio is complete. The device may then continue adapting and using the coefficients.