IMAGE COMPRESSION CODER

摘要

PURPOSE: To improve the accuracy of code quantity control by calculating a code quantity when an output of a 2nd quantizer is subjected to variable coding and deciding a quantization table to adopt a prescribed code quantity for one image based on the result of calculation. CONSTITUTION: A DCT coefficient for each block outputted from a DC transformation circuit 3 is inputted to a quantizer 9, in which M-ways of scale factors α1-αM are assigned so that each block for one field has the same scale factor. Then each DCT coefficient is linearly quantized in a quantization table where the scale factors α1-αM and quantization steps are multiplied. A block code calculation circuit 12 calculates a code multi-value processing per block of variable length data when the quantized DCT coefficient is coded. A code quantity controller 80 predicts the code quantity based on M-ways of quantization table for one image according to the result of calculation and decides a quantization table to form a prescribed code quantity for one image pattern. The controller 80 outputs the decided code quantity to the quantizer 5.