Audio signal nonlinear post-processing method and system capable of realizing acoustic echo cancellation

Abstract

The invention discloses an audio signal nonlinear post-processing method and system capable of realizing acoustic echo cancellation. The system includes a non-linear processor (NLP) (104) for receiving the following signals and taking at least two signals as inputs: an upcoming far-end signal and a plurality of capture-end signals; the NLP (104) first For each frequency band, one or more coherence measures between the received signals are calculated, and based on the one or more coherence measures, a corresponding suppression factor for each frequency band is derived. The NLP (104) also applies the aforementioned suppression factor to one of the capture signals in order to substantially cancel the echo of the capture signal.

Description

technical field [0001] The present invention relates generally to methods and systems for echo cancellation in telecommunication systems, and more particularly to a method and system for canceling residual echoes of an error signal by non-linear post-processing of the error signal. Background technique [0002] Voice quality is an important concern for phone system providers. As customer demands for voice quality continue to increase, it is critical that phone system providers continue to improve. Echo is a delayed manifestation of the original transmission, and if the delay is long, it can seriously disrupt the speaker. If the short-range round-trip delay is less than 20 milliseconds, the talker cannot distinguish echoes and sidetones in the handset of the telephone. However, for long-range communications (such as satellite communications), the remotely generated echo signal often exhibits a significant delay. In addition, the speech and channel coding necessary in digit...

AI Technical Summary

Problems solved by technology

The problem, however, is that after this first stage there is usually still some audible echo in the residual error signal

Images