Squeal determining method and squeal determining device

Abstract

The invention relates to a squeal determining method and a squeal determining device. The method comprises the following steps: acquiring the current frame of a received audio signal; determining N sampling frequency points in the current frame according to preset sampling frequencies, wherein each of the N sampling frequency points corresponds to one frequency, and N is a positive integer greater than or equal to 2; determining the energy corresponding to each sampling frequency point; and determining squeal sampling frequency points in the current frame according to the energy corresponding to each sampling frequency point. According to the technical scheme, all potential squeal frequency points in an audio signal can be found comprehensively in the frequency dimension, no squeal sampling frequency point is missed, then, the continuity is analyzed on a possible squeal frequency point from the time domain, the duration of a high-energy signal of an audio frame where the squeal frequency point is located is judged to avoid misjudgment of the audio frame, and thus, comprehensive, accurate and fast detection of squeal in the audio signal is ensured.

Description

technical field [0001] The present disclosure relates to the technical field of howling, and in particular, to a method and device for determining howling. Background technique [0002] At present, with the increase of network bandwidth and the popularization of mobile Internet, more and more users install various real-time voice communication applications on the terminal, but in some specific cases (such as two mobile phones are close to each other) the voice When talking, there will be a sharp howling (a kind of feedback sound), which greatly affects the hearing effect. Among them, the reason for the "howling" is that a closed loop is formed in the two directions of the voice communication sending and receiving, and the total gain of the loop is greater than 1, which causes the signal to be continuously amplified on the loop. The frequency points are concentrated, thus forming a large monophonic signal, which is very harsh. From the causes of howling, it can be seen that...

AI Technical Summary

Problems solved by technology

[0002] At present, with the increase of network bandwidth and the popularization of mobile Internet, more and more users install various real-time voice communication applications on the terminal, but in some specific cases (such as the distance between two mobile phones is relatively close) When talking, there will be a sharp howling (one of the feedback sounds), which will greatly affect the hearing effect

Among them, the reason for "howling" is that a closed loop is formed in the two directions of voice communication, and the total gain of the loop is greater than 1, which leads to the continuous amplification of the signal on the loop, and in some specific conditions determined by the device. Concentrate on the frequency points to form a larger monotone signal, which is very harsh

It can be seen from the causes of the howling that in order to suppress the howling well, it is necessary to accurately find the frequency of the howling, but the prior art cannot accurately determine the frequency of the howling in the audio signal, especially Yes, in actual equipment, the howling sound is often discontinuous, but periodically intermittent, and different equipment, because the position and structure of the microphone and the speaker are different, therefore, the howling sound The frequency is also different, even the same equipment, under different initial excitation conditions, the howling frequency it produces is also different, which further increases the difficulty of determining the howling frequency

Images