METHOD OF ZEALIZING APPROXIMATION AND QUANTIZATION OF DISCREET COSINUS TRANSFORMATION, METHOD OF ZEALIZING DEQUANTIZATION AND APPROXIMATION OF INVERZ DISCREET COSINUS TRANSFORMATION; FURTHER CODER AND DECODER

摘要

The present invention relates to the approximation and quantization of discrete cosine transformation, which steps are used sequentially on digital data to compress them. In order to improve the transformation, the invention proposes to simplify a particular transformation matrix, thereby requiring less operation when using the transformation matrix on digital data. In addition, the elements of the simplified transformation matrix formed by irrational numbers are maximized by rational numbers. These measures are compensated for by extending a predetermined quantization to compensate for the rational numerical approximation of the elements of the simplified transformation matrix. If the implementation is based on a simplified transformation matrix and extended quantization, good final quality fast transformation can be achieved. The approximation of the inverse cosine transformation used for decompressing the digested digital data can be similarly simplified. HE