Multistage Echo Control System for High Speech-to-Echo Ratios
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional acoustic echo cancellers in mobile devices struggle to effectively cancel echo signals with speech-to-echo ratios greater than 30 dB, leading to unrecognized or misrecognized user speech in voice recognition systems due to the dominance of echo signals over user speech.
Innovation Solution
A multistage echo control system (MS-ECS) that selectively employs pre-echo control filters, adaptive echo cancellation, and dual-channel echo suppression to attenuate, cancel, and suppress both linear and nonlinear echo signals, improving voice recognition performance by reducing residual echo.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a conventional acoustic echo canceller is used, then the device structure remains simple, but it cannot cancel echo signals with speech-to-echo ratios greater than 30 dB, causing user speech to be masked by echo
Solution Approach 1:
The patent divides the echo cancellation process into three distinct stages: pre-echo control (filtering nonlinear echo), adaptive echo cancellation (canceling linear echo), and dual-channel echo suppression (suppressing residual echo). Each stage handles specific types of echo, allowing the system to effectively cancel high-level echo signals while maintaining manageable complexity through functional segmentation.
Solution Approach 2:
The system dynamically selects and configures which echo control functions to apply based on the detected echo level. When echo exceeds 30 dB, all three functions are activated; when echo is below 30 dB, only necessary functions are applied. This dynamic adaptation allows the system to maintain high reliability across varying echo conditions while optimizing device complexity.
2Measurement precision
If multiple echo control functions are applied to handle high echo levels, then voice recognition accuracy improves, but computational complexity and processing time increase
Solution Approach 1:
The system dynamically adjusts the application of echo control functions based on real-time echo level detection. When echo exceeds 30 dB, all three functions (pre-echo control, adaptive echo cancellation, and dual-channel echo suppression) are activated to maximize voice recognition accuracy. When echo is below 30 dB, the system reduces processing by applying only necessary functions, thereby managing computational complexity while maintaining sufficient recognition accuracy.
Solution Approach 2:
The pre-echo control function is applied first to filter nonlinear echo components before the adaptive echo cancellation processes linear echo. This preliminary filtering action simplifies subsequent processing by removing difficult-to-cancel nonlinear components early, reducing the overall computational burden while improving final voice recognition accuracy.
3Object-generated harmful factors
If echo suppression is aggressively applied, then residual echo is reduced, but user speech quality may be degraded
Solution Approach 1:
Each echo control function targets specific types of echo with specialized processing: pre-echo control handles nonlinear echo components, adaptive echo cancellation addresses linear echo, and dual-channel echo suppression targets residual echo. By applying locally-optimized processing to different echo types rather than uniform suppression, the system effectively reduces residual echo while preserving user speech quality.
Solution Approach 2:
The system continuously monitors the echo level and dynamically adjusts the application and parameters of echo control functions based on feedback from the acoustic environment. This feedback mechanism ensures that echo suppression is applied appropriately to reduce residual echo while automatically adjusting to prevent degradation of user speech quality under varying conditions.
Data Source
AI summary
A method for controlling echo of a voice recognition system includes receiving at least two audio input signals corresponding to sound sensed by at least two microphones in a physical space. A first audio input signal of the at least two audio input signals is received on a primary channel, and each remaining audio input signal is received through a respective secondary channel. The method includes selecting, by a processor, based on an echo power level of a speaker in the physical space, a subset of echo control functions (ECFs) from among a plurality of ECFs of a multistage echo control system. Each ECF modifies the at least two audio input signals to reduce echo. The method includes generating a corresponding number of audio output signals by processing the signals received on the primary and secondary channels through the selected subset of ECFs, and outputting the audio output signals.