| With the advancement of technology and the development of VoIP technology,high-speed,low-latency,and non-stuck Internet voice calls are becoming the daily needs of netizens.Moreover,the sudden epidemic in 2020 highlights the importance of online meetings.With the empowerment of science and technology,instant messaging software and online conference systems have developed rapidly,allowing enterprises and users to get rid of the constraints of distance and communicate more conveniently.However,existing software often relies on acquaintances to socialize,causing work and life affairs mixed and work efficiency reduced.At the same time,there are problems such as complicated operation of the login call function,inconvenience for users,expensive traffic,and lack of Internet signals in some cases.Based on these problems,considering the high penetration rate of Wi-Fi and Android smartphones,this thesis studies the multi-player voice communication system based on Android in wireless LANs.The system is specially developed for small and micro enterprises to meet and discuss,which can join the meeting without logging in.In the meanwhile,it is possible to make free calls with the help of Wi-Fi,making communication inside small and micro enterprises more convenient,cheaper and more efficient.The main work of this thesis is as follows:1.The overall architecture design of the multiplayer voice communication system is accomplished,and the functions are realized such as registration and login,creating conference,joining conference,friend management and voice call.The system uses a three-tier C/S architecture,specifically client,server,and database,the client provides an interactive interface to the user;the server processes the messages sent by the client;and the database uses MySQL to add,delete,modify and check data information.Among them,joining the conference function does not require login and registration,which is convenient for users to use;creating the conference function is performed after the user logs in,and the function of adding friends is applied for frequently voice calls.2.A complete set of processing procedures for the voice transmission process is designed and realized,including audio collection,echo cancellation,audio compression coding,network transmission,jitter buffer,mixing and audio playback.Among them,the audio collection uses the Audio Record class provided by the Android system to obtain the sound recorded by the audio hardware;the echo cancellation is accomplished by the 3A algorithm;the audio compression codec module uses the industry's commonly Opus to compress the voice data;the network transmission uses The RTP protocol with the packet sequence number and the RTCP protocol that provides feedback control;and Audio playback is performed by the Audio Track class.3.Aiming at the problem of delay jitter caused by unstable network transmission,jitter buffer technology is adopted to establish a dynamic buffer at the receiving end to enable smooth playback of audio data and effectively reduce voice jitter on the basis of Net RTQ of WebRTC.In terms of the mixing problem of the multiplayer voice communication system,the thesis studied several commonly used mixing algorithms,and proposed an adaptive weighted mixing solution for the three loudest audio streams in the conference.After comparison experiments with mixing algorithms such as the additive clamping method,the situation of speech overflow is obviously suppressed,and the audio mixing effect is good.4.Functional and non-functional tests are conducted on the multiplayer voice communication system.The results show that the multiplayer voice communication system is running normally,and various functional tests are normal.Meanwhile,the voice call quality(MOS)is improved from 3.17 to 3.89 after optimization,and the packet loss rate is reduced from 6.64% to 1.15% through optimization.The optimized voice quality is better,achieving the expected performance. |
PDF Full Text Request