Properties of Entropy Coding for 3D DCT Video Compression Method

Abstract

This contribution presents a basic structure of video compression coder based on Three Dimensional Discrete Cosine Transform (3D DCT). Mainly, this paper deals with a optimal method of forming the 3D DCT frequency coefficients into output bit stream. It is well known the number of non zero frequency coefficients depends on the video sequence character. The sequence with a slow motion evokes a low number of coefficients and vice versa the dynamic sequence produces a high number of frequency coefficients, thus the compress ratio of an encoded sequence tends to be low. It could be proved the compress properties of 3D DCT encoder do not depends only on scene character but can be improved by the entropy coder and by the way how the coefficients are scanned. The scanning order optimized for the 3D DCT and a content of the dictionary for Huffman coder is proposed in this paper.