Dual-Microphone Echo Suppression via Power Level Difference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing acoustic echo cancellation methods in speech telecommunications face limitations, particularly in hands-free configurations with low near-end to echo ratios, leading to high distortions and poor communication quality due to residual echo and ambient noise interference.
Innovation Solution
A dual-microphone echo suppression method that exploits the power level difference between primary and secondary microphone signals to enhance echo cancellation, using adaptive filtering and echo postfiltering techniques, with double-talk detection and power level difference-based control to optimize echo suppression gains.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-generated harmful factors
If acoustic echo cancellation algorithms (adaptive filter + echo postfilter) are used, then echo suppression is improved, but speech quality deteriorates during double-talk periods due to high distortions
Solution Approach 1:
The patent segments the echo cancellation task into two independent processing paths: one for echo-only periods (using adaptive filter + postfilter) and another for double-talk periods (using alternative processing). This segmentation allows optimized handling of each scenario separately, preventing speech distortion during double-talk while maintaining echo suppression during echo-only periods.
Solution Approach 2:
The patent implements dynamic scenario classification that adapts the processing mode based on real-time detection of echo-only versus double-talk periods. The system dynamically switches between different processing strategies: aggressive echo suppression during echo-only periods and speech-preserving processing during double-talk periods, thereby resolving the contradiction between echo suppression and speech quality.
2Reliability
If double-talk detectors and noise only detectors are used to limit disturbance impact, then AEC performance is improved, but device complexity increases
Solution Approach 1:
The patent extracts and utilizes readily available signal energy information from the microphone path and loudspeaker path to perform scenario classification, rather than implementing complex double-talk detectors. By taking out and processing only the essential energy comparison, the system achieves reliable AEC performance with minimal added complexity.
3Manufacturing precision
If multi-microphone configurations are used to improve speech quality in low near-end to echo ratio scenarios, then echo cancellation performance is improved, but device complexity increases
Solution Approach 1:
The patent makes the existing dual-microphone hardware (originally designed for other purposes) serve the additional function of echo scenario classification. The primary and secondary microphones, already present in the device, are universally utilized for both their original functions and for detecting echo-only versus double-talk periods through energy comparison, thereby improving echo cancellation without requiring additional hardware.
Data Source
AI summary
A method for processing an audio signal is described comprising receiving a first audio signal via a first receiving path comprising a first microphone; receiving a second audio signal via a second receiving path comprising a second microphone; and performing echo suppression of the first audio signal based on the first audio signal and the second audio signal.