AGC automatic gain control of WEBRTC

When the loudness of the voice to adjust the need for voice automatic gain (AGC) algorithm processing , voice chat will use this algorithm.

The simplest hard gain processing is to multiply all audio samples by a gain factor, which is also equivalent to multiplying the gain factor at each frequency in the frequency domain, but because the perception of all frequencies of the human hearing is not linear, is followed by the equal loudness curve, resulting in such processing, it sounds like the frequency of the increase, some frequency weakened , resulting in amplified speech distortion.

To make the frequency of the entire band sound "the same", the gain must be made at the loudness level, rather than in the frequency domain, i.e. the frequency of the speech is weighted by the equal loudness curve, not by a fixed gain factor.

By some visible, the automatic gain processing of speech can be roughly divided into two parts:

(1) Determination of the loudness gain factor.

(2) The loudness gain factor is mapped to the equal loudness curve, and the gain weights of the final frequencies are determined.