北京邮电大学学报

  • EI核心期刊

北京邮电大学学报 ›› 2008, Vol. 31 ›› Issue (1): 116-119.doi: 10.13190/jbupt.200801.116.073

• 研究报告 • 上一篇    下一篇

MPEG-2/MPEG-4实时转换I帧的快速补偿算法

覃团发, 唐振华   



  1. (广西大学 计算机与电子信息学院, 南宁 530004)
  • 收稿日期:2007-03-24 修回日期:1900-01-01 出版日期:2008-02-28 发布日期:2008-02-28
  • 通讯作者: 覃团发

A Quick Compensation Method for I Frame in MPEG-2 to MPEG-4 Real-time Transcoding

QIN Tuan-fa, TANG Zhen-hua   



  1. (School of Computer and Electronic Information, Guangxi University, Nanning 530004, China)
  • Received:2007-03-24 Revised:1900-01-01 Online:2008-02-28 Published:2008-02-28
  • Contact: QIN Tuan-fa

PDF

405

摘要：

基于离散余弦变换(DCT)域的MPEG-2到MPEG-4实时转码中，MPEG-2 的帧内编码帧(I帧)进行部分解码后所得到的DCT数据不能直接在 MPEG-4 重量化中使用，提出了一种基于DCT域的I帧快速补偿算法.该算法避免了重新进行DCT变换和逆变换，每个宏块节省6次DCT变换和6次逆DCT变换的时间.实验结果表明，与采用像素域补偿算法进行I帧转码相比，采用快速补偿算法节省了65.05％的时间，并且转码后I帧的峰值信噪比(PSNR)值偏差不超过0.001dB.

关键词: MPEG-2, MPEG-4, 码流转换, 离散余弦变换域, 重量化

Abstract:

The real-time transcoding from MPEG-2 to MPEG-4, the discrete cosine transform (DCT) domain data of intra frame(I-frame) decoded in part by MPEG-2 decoder could not be directly used in re-quantization of MPEG-4. A quick compensation method based on DCT domain is proposed, in which the calculation time of six DCTs and six inverse DCTs are saved in each macroblock for implementing DCT and inverse DCT are not needed. Experimental results show that the transcoding time for I frame is saved by 65.05% and the peak signal-to-noise ratio (PSNR) error is in 0.001dB compared with the pixel domain compensation method.

Key words: MPEG-2, MPEG-4, transcoding, discrete cosine transform domain, requantization

中图分类号: 

版权所有 © 2018 《北京邮电大学学报》编辑部 京ICP备14033833号     在线人数：今日访问 当前在线 总访问

通信地址：北京市海淀区西土城路10号北京邮电大学195信箱    邮编：100876    电话：(010)62281995    QQ：307514248    微博    留言    电子邮箱：byxb@bupt.edu.cn