Echo cancellation method, device and equipment and storage medium

Abstract

The embodiment of the invention discloses an echo cancellation method, device and equipment and a storage medium, wherein the method comprises the steps: determining a far-end frequency domain signalcorresponding to a far-end sound signal at a first moment, and determining an echo estimation frequency domain signal according to a first coefficient of a self-adaptive filter at the first moment andthe far-end frequency domain signal; performing echo cancellation on a near-end sound signal at the first moment according to the echo estimation frequency domain signal to obtain a residual frequency domain signal; determining a first adjustment factor; determining a control factor; updating the first coefficient according to the control factor and the first adjustment factor to obtain a secondcoefficient of the self-adaptive filter at a second moment so as to perform echo cancellation on the near-end sound signal at the second moment according to the second coefficient, wherein the secondmoment is the next moment of the first moment; and outputting an echo cancellation signal according to the filtering state and the residual frequency domain signal. By adopting the embodiment of the invention, the viability and robustness of the self-adaptive filter can be improved, and the accuracy of echo cancellation signal output is improved.

Description

technical field [0001] The present application relates to the communication field, and in particular to an echo cancellation method, device, equipment and storage medium. Background technique [0002] In modern multimedia communication systems, no matter in the field of voice communication or human-computer interaction, in order to improve the communication quality, the echo in the collected sound signal must be eliminated or suppressed. However, traditional echo cancellation methods often completely rely on the accurate detection of the state of the echo sound field (double talk, echo path changes, etc.). Since the adaptive filter (Adaptive Filter, AF) itself has no robustness guarantee measures, the parameter characteristics of its output are not accurate, so there is a greater risk of misalignment in the adaptive filter coefficients, and the robustness Reliability cannot be guaranteed. At the same time, in the echo cancellation method using multiple filters, the calcula...

AI Technical Summary

Problems solved by technology

However, the traditional echo cancellation methods often rely entirely on the accurate detection of the state of the echo sound field (double talk, echo path changes, etc.)

Since the adaptive filter (Adaptive Filter, AF) itself has no robustness guarantee measures, the parameter characteristics of its output are not accura...

Images