There are two important characteristics of vocals: fricative and plosive. These two kinds of sound environment noise cannot be created. As long as it is a human language, no matter which language has this feature, speech sensing technology can automatically sense the frequency component of the audio signal. Whether it is vocal, this technique can be applied to AGC (Automatic Gain Control).
Application scenario description:
There are many conference microphones in the conference room for picking up the speaker's voice. These conference microphones have the same brand and model. The position of the corresponding channel of the conference microphone in the sound reinforcement system is also the same, but the distance of the speaker from the microphone in use and the volume of the speaker is not the same. This result in different levels of conference microphones and the volume of speech amplification is large or small. The AGC (Automatic Gain Control) circuit can automatically compress large signals and automatically increase small signals.
The traditional AGC (automatic gain control) works:
The traditional automatic gain control is to preset a threshold. When the audio signal picked up by the microphone is lower than the threshold, the automatic gain control will automatically give the audio signal a gain boost, which is close to the threshold position; when the microphone picks up the audio signal is high At the threshold, the automatic gain control automatically attenuates the gain of the audio signal to bring it close to the threshold position, which basically ensures the constant value of the microphone input signal level.
Application of speech sense technology:
In the traditional application of the automatic gain control circuit, in addition to picking up the human voice, the conference microphone will pick up other sounds, such as flipping books, putting cups and keyboard tapping, which are mostly smaller than the threshold of automatic gain control. When the microphone picks up these sound signals, the automatic gain control circuit will work immediately, amplify these sound signals, and then amplify the sound to the speaker through the power amplifier, and the original small sound becomes a large environmental noise. This phenomenon is undesirable in conference room audio sound reinforcement systems. The biggest function of the speech detection algorithm is that the automatic gain control can distinguish the sound of human speech and the sound of non-human speech. Only the human speech sound is boosted or attenuated, and the environmental noise smaller than the automatic gain control threshold is not automatically increased. The control circuit performs amplification. In this way, the automatic gain control circuit with the speech sensing technology algorithm works more accurately and intelligently.
MAONO is an innovative designer and manufacturer of Lavalier, Podcasting, Wireless, Shotgun, Recording microphones and accessories for Smartphone, Camera and PC, etc.