Image coding apparatus for correcting coarseness and encoding uncompressed data

An image conversion method, comprising:

• generating uncompressed image data including a plurality of pictures by decoding coded image data;
• determining, by circuitry, a coarseness set value indicating the coarseness of quantization of a current macroblock to be encoded when each picture of the plurality of pictures is to be encoded on a macroblock basis;
• correcting, by the circuitry, said coarseness set value by adding a correction value to said coarseness set value;
• and
• encoding said current macroblock by using said corrected coarseness set value, wherein
• said correcting by the circuitry includes
• calculating, when said current macroblock has been encoded, a first cumulative amount of generated codes by accumulating the respective amounts of generated codes of macroblocks in a picture being encoded ranging from a first macroblock which is first encoded up to and including said current macroblock;
• calculating, when said current macroblock has been encoded, a first cumulative target amount of codes by accumulating the target amounts of codes set for said macroblocks in said picture ranging from the first macroblock up to and including said current macroblock, respectively;
• calculating a first excessive amount of codes by subtracting the first cumulative target amount of codes from said first cumulative amount of generated codes;
• calculating a second excessive amount of codes by subtracting a sum total of the target amounts of codes set for one or more macroblocks, respectively, ranging from said first macroblock to a macroblock that has been encoded directly before said current macroblock, from a sum total of the respective amounts of generated codes of the one or more macroblocks;
• when the first excessive amount of codes is greater than zero:
• comparing the first excessive amount of codes with the second excessive amount of codes,
• increasing said correction value when the first excessive amount of codes is larger than the second excessive amount of codes, and
• maintaining said correction value when the first excessive amount of codes is smaller than the second excessive amount of codes;
• maintaining said correction value when the first excessive amount of codes is less than zero;
• and
• initializing the correction value, the first cumulative amount of generated codes and the first cumulative target amount of codes when the encoding of the picture is to begin.