To precisely decode a code, which is obtained by decoding such as expressing digital data by a smaller number of bits, to original digital data without deteriorating precision.

The continuing number of 0 is investigated from the lowest order bit of digital data 103, a value obtained by subtracting a number larger than this continuing number of 0 by one from the bit number of data 103 is set to be the number of assigned bits 105 and the bit is assigned from the highest order bit of data 103 according to this number 105. The number of bits reduced by using the number of assigned bits 114 (105) and a low-order bit which is provided with the size and where a highest-order is 1 and the other bits are 0 is generated. This low-order bit is connected to the low-order side of a code 115 (106) to decode to original digital data 117.