Voice Quality Enhancement Technology For Mobile VoIP

Now Internet is gradually changing into the information infrastructure of global commercial applications. With the rapid expanding of the number of the Internet users, their information formats required from the Internet are changing from text and data to multimedia, such as combination of voice, image and video. As the wireless network gradually matured, extensive use of wireless networks also makes VoIP more flexible. Mobile VoIP is the technology of using the wireless Internet for voice transmission. As one of the key function of communication, the quality of voice communication over IP is very important. The environment of wireless network is not stable, so the voice quality of mobile VoIP is worse than wire Internet VoIP. The research on voice quality enhancement of mobile VoIP is important and has practical value.IP network is originally designed for data application and its implementation is correspondently simple. It provides best effort service. This design is a weakness for voice communication over IP network because the characteristics of the real-time data are not considered. When current IP network is used for real-time voice communication, the requirement of voice communication such as real-time, delay and jitter will not be satisfied. Now the study of voice quality enhancement of mobile VoIP goes into two general categories:network-centric approach and end system-based approach. Network-centric approach should configure routers to support quality of service (QoS) and end system-based approach employ control techniques such as congestion control, packet loss concealment, forward error control and jitter buffer control. End system-based approach is a hot research area because they are independent for the evolution of the underlying network technologies.Because of unstable of wireless network and best effort service of Internet, mobile VoIP meets some challenges such as packet loss, packet delay and delay jitter. Voice communication is a latency-sensitive business, large delay and delay jitter will affect the interaction and comfort of voice communication. At same time, packet loss also decreases the voice quality. Our main contributions include the following areas. Firstly, by analysis the protocol of mobile VoIP related we proposed a profile for mobile VoIP. This profile also provides a useful limited for the following study.Secondly, by analysis current forward error control technique, this paper proposed an adaptive forward error correction technique for mobile VoIP. In the end-to-end voice communication networks, the packet losses caused by network congestion and wireless channel errors reduce the efficiency of the forward error correction scheme. Especially for mobile VoIP, congestion and wireless channel status changed very fast. Adaptive forward error correction (FEC) technique can adaptive adjusting the FEC block according to the network status.Thirdly, this paper proposed GWSOLA algorithm by introduce the gain control into the standard WSOLA technique. GWSOLA algorithm could adjust the level of the voice segment for overlap and add in order to maintain the voice signal level consistent. Because packet loss concealment method by using WSOLA algorithm is Independent to the voice coding algorithm, this method is more suitable for multi-party IP conference system and can be used for any voice coding algorithm. This paper also proposed a voice signification transmission detection (VST) method. VST detection is a powerful method that improves the performance of GWSOLA when the location of lost frame is at VST section.Finally, this paper proposed an adaptive network congestion control method integrate with enhanced NLMS predictor of transmission jitter for mobile VoIP, this method can adaptive adjust network congestion according to the network status. When the networks are congestion, the network delay and packet loss rate will increase rapidly. It will seriously affect the voice quality. Our method is dynamically adjusted the sending rate of the source voice according to network status. It can alleviate network congestion and improve voice quality.The simulation result shows our methods can improve the voice quality of mobile VoIP and have practical value for commercial applications.