Nonlinear Acoustic Echo Suppression via Distortion Function Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing acoustic echo control methods fail to effectively suppress nonlinear acoustic echoes in speech applications, leading to distorted signals and reduced voice recognition accuracy in hands-free telephony due to the assumption of a strictly linear acoustic echo path, which is not valid for low-cost audio systems or high playback levels.
Innovation Solution
A method for residual echo suppression that generates a nonlinear signal by applying a distortion function to the original reference signal, calculates residual echo power for both linear and nonlinear components, and applies a room model to estimate power associated with each signal, using a weighted sum of weighted powers to achieve combined echo power estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If acoustic echo control methods assume a strictly linear acoustic echo path, then the processing is simpler and works for ideal conditions, but nonlinear distortion from loudspeaker or amplifier cannot be compensated and echo artifacts remain in the output signal
Solution Approach 1:
The patent segments the acoustic echo path into linear and nonlinear components. The linear component is handled by traditional AEC methods while the nonlinear component is processed separately using distortion function estimation. This segmentation allows each component to be compensated appropriately without overwhelming complexity
Solution Approach 2:
The patent performs preliminary estimation of the distortion function using training data before actual echo suppression. By pre-characterizing the nonlinear behavior of the loudspeaker/amplifier, the system prepares compensation parameters in advance, reducing real-time processing complexity while maintaining reliability
2Ease of manufacture
If traditional AEC and RES methods are used, then the system works for linear echo paths, but echo artifacts remain in processed output signal when nonlinear distortion is present
Solution Approach 1:
The patent introduces a distortion function estimation module as an intermediary between the traditional AEC/RES system and the final output. This intermediary component models the nonlinear behavior and enables compensation without requiring a complete redesign of the existing AEC infrastructure
Solution Approach 2:
The patent changes the parameter representation by estimating distortion function parameters (such as polynomial coefficients) that characterize the nonlinear behavior. By working with these estimated parameters, the system can compensate for nonlinear distortion while maintaining compatibility with traditional linear processing pipelines
Data Source
AI summary
A method for residual echo suppression is provided. Embodiments may include receiving an original reference signal and applying a distortion function to the original reference signal to generate a second signal. Embodiments may include generating a non-linear signal from the distortion function that does not include linear components of the original reference signal. Embodiments may also include calculating a residual echo power of a linear component and a non-linear component, wherein the linear component is based upon the original reference signal and the non-linear component is based upon the non-linear signal. Embodiments may further include applying a room model to each of the original reference signal and the non-linear signal and estimating a power associated with the original reference signal and the non-linear signal. Embodiments may include calculating a combined echo power estimate as a weighted sum of a weighted original reference signal power and a weighted non-linear signal power.