Internet Phone Conference Mixing Technology Research

Along with computer network technology, communications technology continues to evolve at the same time, the network construction of the conference call is also in high gear. Now, conference calling technology is moving towards multi-media communications session-oriented IP conference call can be depending on the direction of the system, using the network is currently the fastest growing Internet. The telephone conference system is the use of multimedia communications technology, network terminals and other equipment, in the off-site or multiple locations through the IP network between the transmission channel with the meeting as a means of communication.This is done in each session, first select a President of the Conference to control the voice. This greatly limits the exchanges between the participants, has brought a lot of inconvenience to the meeting. But in audio frequency module, mixing technique made a telephone conference to resolve this problem. So that more participants will no longer be able to apply for a voice to speak at the same time, and all the members attending the meeting can simultaneously receives a different message from the speakers, to each Council member be able to say what purpose.This article first to the VOIP principle of work, several kinds of basic application patterns as well as the key technologies which involved in the course of communications has carried on the brief introduction. Then introduced the two traditional mixing schemes and optimized, further improve the mixing processing speed. Next, analysis of several existing mixing algorithm, mixing the quality is the key problem of overflow and noise. In order to resolve the problem effectively, this paper proposes an adaptive reduced spectrum mixing algorithm, which not only in the adaptive weighting algorithm control mixing, on the basis of the overflow problems also introduced to reduce the noise of thought ,right speech signal Fourier transform, from the noisy speech power spectrum by subtracting the noise power spectrum, and thus receive a more pure voice frequency spectrum to reduce the noise, and then Fourier transform them with anti-mixing weights.We calculate linear relationship between consuming time and mix channel of each sample point according to the sampling points total running time. The result shows that time consumption of improved method increases lower and lower with mixing Road. It indicates that speed of improved algorithm increased dramatically under many mix roads. then it evaluates the ability of understanding and clarity by an objective PESQ evaluation method. Through PESQ score, it shows that sound quality is more clear, natural, smooth. It is closer to "face to face" meeting.