Echo cancelling method based on self-adaptive time delay estimation

Abstract

The invention provides an echo cancelling method based on self-adaptive time delay estimation. The echo cancelling method is used for cancelling an echo between a speaking end and a listening end of a communication system, the speaking end generates a speech signal and transmits the generated speech signal to the listening end; an echo signal which can return to the speaking end is generated between the speaking end and the listening end. The echo cancelling method is characterized by comprising the following steps of: (a) acquiring the echo signal to be used as an actual echo signal; (b) generating an analogue echo signal according to the echo signal; and (c) subtracting the actual echo signal from the analogue echo signal to reduce the echo electric level for returning the echo signal of the speaking end. The method not only can effectively cancel the echo containing Gaussian noise in a communication line, but also can well cancel echo containing impulse noise.

Description

technical field [0001] The invention relates to the communication field, in particular to an echo cancellation method based on robust adaptive time delay estimation for magnetic resonance imaging. Background technique [0002] Echoes generally refer to signals that arrive at a given point by other means than the normal path. At this point, the signal has sufficient magnitude and delay that it can be perceived as different from a signal carried by a normal path. Human hearing can perceive sound echoes with a certain intensity and a delay of more than 50ms. Voice communication is one of the most basic telecommunication services, and echo problems are often encountered in voice communication. [0003] In the traditional public switched telephone network (Public Switched Telephone Network, PSTN), since the impedance in the 2 / 4 wire conversion circuit cannot completely match the impedance of the telephone terminal, the input signal of the 2 / 4 wire conversion circuit will be par...

AI Technical Summary

Problems solved by technology

Among the many methods proposed so far, the noise is mostly described by the Gaussian model. However, many practical studies have shown that the noise in the actual communication line system is often impulsive, and the classical algorithm based on the Gaussian noise model is not robust to this. The performance degrades and even cannot be used

Images