Research And Implementation Of Multimedia Interactive Software Based On Android

With the development of digital image processing technology and streaming media transmission technology, people look forward to a wide range of information exchange. In recent years, the development of mobile internet and smart phones is very fast. It makes the development of application based on smart phones possible. In this paper, I design and implement multimedia interactive software based on Android. You can send instant messages, make voice calls and chat with video. The multimedia interactive software consists of two parts, including server and client. The software uses SIP as the control signaling. By the way, it transports audio and video data via RTP. Audio is coded by G711and video is coded by H.264.The main work and innovations are as follows:(1) In order to lay the foundation for the development, I study the key technologies, including audio and video synchronization, SIP protocol, XML parsing, Android NDK development and H.264intra prediction coding technique.(2) Design a general framework for the multimedia interactive software, analyze the software function modules. According to the function requirements of software, this paper designs the function modules are modular and gives method to achieve development.(3) For the purpose to achieve SIP signaling transmission and media transmission, the paper ports open source libraries such as Mjsip and Jlibrtp to software.(4) For the jitter and lip not synchronous which caused by network packet loss and shuffle, this paper presents two cache mechanism to eliminate jitter, and proposes a method for audio and video synchronization. By setting cache, it achieves media internal synchronization. When it detects audio packet loss, it inserts comfort noise based on timestamp and main stream synchronization control mode. This method can improve the audio and video synchronization result which has little influence to audio quality.(5) In order to improve video coding efficiency, this paper optimizes and ports open source library OpenCore to software. First, the algorithm determines image flatness and chooses a corresponding intra prediction mode to predict the image. Second, for intra4×4prediction, this paper achieves a fast mode algorithm, reducing the computation amount. Then, I transplants the library based on Android NDK. It improves video coding efficiency without affecting the view.After the test, the software is stable, easy to operate and able to complete its intended function.