. . . "Generally, if the encoding is performed in the unit of (l???n) bits wherein l is an integer, the number of pattern is expressed (n+1)l which can be substituted with X bit(s) having a relation of 2X-1 <(n+1)l ???2X." .