新型MPEG2码流合成系统的设计与实现

根据数字电视发展的要求,设计实现了一种新型的MPEG2码流合成数字电视发端系统,可将视音频的信源编码与复用集成为一体.本文讨论了视音频打包与复用的关键技术,包括时序匹配,复用时序与逻辑,码流合成以及缓存控制等.该系统改进了时钟匹配与音视频同步方案;基于复用器产生音视频时间标签,并同步送给音频编码器与视频编码器;复用器采用新型的级联结构来完成多节目的复用与再复用,无需对时钟参考值的抖动进行额外的校正;采用DSP与FPGA联合构建节目特定信息和包结构,节约了存储空间;改进了PES打包单元的缓存逻辑与码流的输出缓存控制.测度显示该系统结构灵活,性能稳定,码流的各项指标均达到国家广播级要求.

Design and Implementation for MPEG2 Stream Synthesis System

According to the requirement of the DTV development, the scheme is designed for a novel MPEG2 stream construction system integrating a video-audio source encoder with the multiplexer. Major issues for implementation of the video, the audio packetizer, and the TS multiplexer are discussed, such as time matching, multiplexer scheduling, stream synthesis, and memory control. Schemes for timing matching and the synchronization are modified. DTS is generated by the multiplexer and sent to video/audio encoders. The multi-program multiplexing and the re-multiplexing are implemented with the novel cascade mechanism, which needs no additional program clock reference (PCR) corrections and maintains synchronization constraints of the standard. The novel architecture of DSP and FPGA cooperation is employed for constructing program specific information and packet structure resulted in the high operational efficiency and the memory space saving. The scheduling for PES packetization and the data flow is modified. Experiments show that the system is flexible and stable, and the perfect performance meets with national requirements for the broadcasting level.