Relation between character of real video sequences and 3D-DCT compression

This paper is about video compression based on the 3D-DCT. This algorithm is similar used in the specification MPEG (Motion Picture Experts Group). In spite of width application of the standard MPEG, better results in compression can be obtained with method of the 3D-DCT. In the next text we firstly describe the architecture of encoder and decoder including the definitions of mathematical operations like the forward and inverse three-dimensional discrete cosine transform 3D-FDCT and 3D-IDCT, quantisation and thresholding and finally the calculation of the NRMSE (normalized root mean square error). The calculation of the 3D-DCT with two real video sequences is also realized.