Multi-channel voice mixing method and system

Abstract

The invention discloses a multi-channel voice mixing method and system. The method comprises the steps of: S1, performing decoding, packet loss concealment and jitter control on network data packets input by each voice channel to obtain a smooth audio data stream input by each voice channel; S2, performing continuous detection and voice energy calculation on the audio data stream input by each voice channel, selecting voice channels participating in current voice mixing according to the continuity and voice energy of the audio data stream input by each voice channel and the historical voice mixing information of each voice channel, and updating the historical voice mixing information of each voice channel; S3, completing a voice mixing operation based on an adaptive normalized voice mixingalgorithm according to the audio data streams input by the voice channels participating in current voice mixing to obtain voice mixing data; and S4, outputting the voice mixing data to each voice channel. The method and the system can realize high-quality voice mixing.

Description

technical field [0001] The invention relates to the technical field of communication. More specifically, it relates to a method and system for mixing multiple voices. Background technique [0002] In recent years, with the progress and development of VOIP technology, there are more and more demands for voice conferencing. Voice conferencing can realize simultaneous voice communication of multiple users, and has broad application prospects in the communication field. Generally speaking, there are two schemes for voice conferencing: a centralized sound mixing scheme and a distributed sound mixing scheme. The so-called centralized audio mixing solution is based on the server-client architecture technology. All clients are connected to the server, and the server performs unified audio mixing, and then distributes to each client through the language channel in the network to realize multi-person voice communication. The so-called distributed audio mixing scheme means that the ...

AI Technical Summary

Problems solved by technology

The problem is: the superimposed signal is easy to overflow, once the overflow occurs, it needs to be truncated, that is, the value exceeding the upper limit is set as the upper limit value, and the value exceeding the lower limit is set as the lower limit value

This kind of processing will obviously destroy the waveform of the original signal, and will introduce noise, which sounds very noisy, the voice fluctuates, and sometimes there will be harsh popping sounds, which seriously affect the subjective experience of the call.

And as the number of callers participating in the mixing increases (clients and voice channels also increase at the same time), this overflow frequency will become more frequent, eventually resulting in the inability to distinguish valid voices at all

Especially for voice conferencing based on VOIP technology, due to the inherent jitter of the network itself, the sound is prone to packet loss and discontinuity. If it is not processed, it will often lead to a worse sound mixing effect

Therefore, there is an urgent need to design a high-quality sound mixing method. However, in fact, even if the best sound mixing algorithm is used, the sound mixing effect may not be ideal in practical applications. This is obviously not the case. The problem of the mixing algorithm itself, but the problem of the mixing strategy (scheme), involves the overall design and optimization of many aspects

Images