Acoustic Elimination Technology Overview

Acoustic Elimination Technology Overview

Echo Cancellation Technology – Introduction

In instant messaging applications, it is necessary to carry out two or more real-time voice communications. In high demand applications, external speaker playback is usually used. This will inevitably result in an echo, that is, after one party speaks, passes through the other speaker. Play it, and then it is collected by the other Mic and sent back to yourself (as shown in the right figure). If you do not deal with the echo, it will affect the quality of the call and the user experience, and more serious will form a shock, resulting in howling.

Echo cancellation means that after Mic collects the sound, the sound played by the local speaker is eliminated from the sound data collected by Mic so that the sound recorded by Mic only has the voice of the local user.

The traditional echo cancellation is based on hardware. DSP processing chips are integrated on the hardware circuit. For example, fixed telephones and mobile phones commonly used by us all have special echo cancellation processing circuits. However, there are technical difficulties in using software to implement echo cancellation, including technical difficulties. The most widely used QQ super voice in China is the adoption of foreign GIPS technology.

Acoustic Echo Cancellation Technique - Basic Principle The principle of echo cancellation is based on the correlation between the loudspeaker signal and the multi-path echo generated by it, establishing a speech model of the far-end signal, using it to estimate the echo, and continuously modifying the filter The coefficient makes the estimate closer to the real echo. Then, the echo estimate is subtracted from the microphone's input signal to achieve echo cancellation.

About Noise Suppression (ANC) The principle of the second-generation noise suppression technology developed by Tenveo is to perform spectral analysis on the digital signal sampled in real time so that the intensity and spectral distribution of the background noise response can be analyzed. Then a filter can be designed based on this model. When someone speaks, they do signal analysis at the same time. Based on the analysis, the ANC can analyze the speaker's spectrum. According to the background noise and the speaker's spectrum, the filter makes real-time changes according to the comparison of the two signals. When the spectrum of the sound passes, the frequency spectrum of the background noise is suppressed, and its energy is reduced, for example, by 15 to 20 decibels, and it is apparent that the noise suppression effect can be experienced.