利用ADPCM进行网络环境下实时多点语音通信

Real-time Multi-point Speech Communication Using ADPCM in Computer Network

目前网络语音通信应用逐步增多,但效果还不甚理想。该文阐述了应用音频处理API函数、ADPCM编码调制技术和UDP协议进行网络环境下实时多点语音通信的方法,并给出其实现过程。实现机制为发送端利用低层音频服务的回调机制对音频数据块进行动态循环处理,防止了由于系统在写满语音输入缓冲区后,新、旧缓冲区交换时有一定的间隙而产生的输入数据泄漏。在传输之前采用ADPCM编码,有效地利用了现有带宽,传输后经缓冲解码再进行播放,尽可能地恢复原始语音信号。由于采用ADPCM编解码技术,实现了在各种网络环境下进行实时多点语音通信,效果较为理想。

At present,the applications of the speech communication in computer network have gradually increased,but its effects are not so satisfying.This paper presents a method of the real-time multi-point speech communication in network,which uses audio processing API function,ADPCM and UDP.The implementation of this method is also shown in this paper.The implementation mechanism is:the sender dynamically-circularly processes audio data blocks using the callback mechanism of low level audio service,and prevents input data from leaking for some intervals resulting from new buffer and old buffer switching after the speech input buffer is full.The use of ADPCM code before transmitting effectively utilizes present bandwidth.The speech received is decoded and then played after transmitting for recovering the original speech signal as possible as people can.As adopting the code-decode technology of ADPCM,the real-time multi-point speech communication in the various computer network is realized and its effects are satisfying.