论文部分内容阅读
在ICC’88会上有关信源压缩编码方面的文章,分散在各小组共有19篇论文。其中语声和图象编码各占7篇,在数量上它们占了优势。如果用当今感兴趣的新技术来划分,其分布是:有关矢量量化(VQ)的文章占4篇,分频带编码(subband coding)占3篇,自适应差分脉冲编码(ADPCM)占4篇,变速编码(主要是变速变换编码)占3篇。现分述如下。1.有关VQ的文章VQ也是一种块编码(block coding),它将输入样值分割成有限的子区间,对每一个子区间选择一个表示值输出,它不同于标量量化,这些分割和表示都在N维矢量空间进行。若矢量空间分割成2~(NR)个子集,它们中每一个有相应的表示值或码矢量,于是,含有N样值的
In the ICC’88 conference on source compression coding articles, scattered in each group a total of 19 papers. Among them, each of speech and image coding occupies 7 articles, and they occupy an advantage in quantity. If we divide it with new technologies that are of interest today, the distribution is as follows: four articles on vector quantization (VQ), three on subband coding, four on adaptive differential pulse code (ADPCM) Variable speed coding (mainly variable speed coding) accounted for 3 articles. Now is as follows. VQ article VQ is also a block coding, which divides the input sample into finite subintervals and selects a representation value output for each subinterval, which is different from scalar quantization. These partitions and representations All in N-dimensional vector space. If the vector space is divided into 2 ~ (NR) subsets, each of them has a corresponding representation value or code vector, then,