算术编码是将累积分布函数的区间（0，1）分成许多互不重叠的小区间，每个信源符号对应于各个小区间，每个小区间的长度等于这个信源符号的概率分布值，在此小区间内取一点，取该点二进制小数点后l位作为这个信源符号的码字。把这基本思想运用到信源符号序列中来，能计算出信源符号序列的累积分布函数，使每个符号序列对应于累积分布函数上的不同区间，在此区间上取一点，将其二进制小数点后l位作为这符号序列的码字，只要这些区间不重叠，编的的码即为算术编码。-Arithmetic coding is the cumulative distribution function of the interval (0,1) is divided into many non-overlapping between the community, each letter of the source symbols corresponding to the various inter-community, each community equal length between the source symbol probability distribution of values, In this inter-community within the check point, take the point after the l-bit binary decimal point as the source of the codeword symbols. The basic idea of this letter applied to the source symbol sequence in years, can calculate the source symbol sequence cumulative distribution function, so that each symbol sequence corresponding to the cumulative distribution function on different intervals, in this interval on admission that its binary l-bit after the decimal point, as this symbol of the codeword sequence, as long as these do not overlap interval, a provision of the Code is arithmetic coding.