Echo Suppression Using Constant Crosstalk Coefficient
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing echo suppression technologies struggle to effectively remove echoes with large distortion in the echo path, especially in environments with high-level noise, due to the complexity of echo crosstalk estimation and sensitivity to speech detection errors.
Innovation Solution
The method employs a constant crosstalk coefficient to estimate and subtract echo components from the near-end signal, using the output of an echo canceller and microphone signal, which simplifies the computation and reduces noise interference, even in conditions with high distortion and noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If complex echo crosstalk estimation is used to remove echoes with large distortion, then echo suppression performance is improved, but computational complexity and sensitivity to speech detection errors increase
Solution Approach 1:
The patent extracts only the essential information needed for echo suppression by using a constant crosstalk coefficient instead of complex real-time estimation. This coefficient captures the main echo characteristics without requiring complex computational processes, thereby reducing device complexity while maintaining echo suppression effectiveness.
Solution Approach 2:
The patent performs preliminary action by pre-determining the crosstalk coefficient based on echo path characteristics before actual echo suppression is needed. This preliminary determination allows the system to avoid complex real-time estimation during operation, reducing computational complexity while maintaining reliable echo suppression performance.
2Reliability
If complex echo crosstalk estimation is used to remove echoes with large distortion, then echo suppression performance is improved, but sensitivity to speech detection errors increases
Solution Approach 1:
The patent extracts only the essential echo characteristics by using a constant crosstalk coefficient, eliminating the need for complex real-time speech detection and estimation. This approach reduces sensitivity to speech detection errors while maintaining effective echo suppression, as the constant coefficient does not require continuous adjustment based on fluctuating speech conditions.
Solution Approach 2:
The patent performs preliminary determination of the crosstalk coefficient based on echo path characteristics before operation. This preliminary action eliminates the need for continuous speech detection during echo suppression, thereby reducing sensitivity to speech detection errors while maintaining reliable echo removal performance.
3Device complexity
If constant crosstalk coefficient is used to estimate and subtract echo components, then computational complexity is reduced, but echo suppression accuracy may be affected
Solution Approach 1:
The patent changes the parameter approach from complex real-time estimation to a constant coefficient. This parameter change simplifies computation significantly while maintaining sufficient echo suppression accuracy for practical applications, as the constant coefficient captures the essential echo characteristics without requiring continuous adaptation.
Solution Approach 2:
The patent performs preliminary determination of the crosstalk coefficient based on echo path characteristics before operation. This preliminary action allows the system to use a constant coefficient during operation, reducing computational complexity while maintaining adequate echo suppression accuracy for hands-free telephone applications.
Data Source
AI summary
Coefficient generator generates a crosstalk coefficient that is a predetermined value and that is used to calculate the amount of crosstalk of an echo. Converter uses either the output signal of a sound pickup device or the signal obtained by subtracting the output signal of an echo canceller from the output signal of the sound pickup device as a first signal, corrects the first signal based on the crosstalk coefficient generated in coefficient generator, and produces a near-end signal obtained by removing the echo from the first signal.