ARITHMETIC CODING FOR DATA-COMPRESSION

Arithmetic coding provides an effective mechanism for removing redunda ncy in the encoding of data. We show how arithmetic coding works and d escribe an efficient implementation that uses table lookup as a fast a lternative to arithmetic operations. The reduced-precision arithmetic has a provably negligible effect on the amount of compression achieved . We can speed up the implementation further by use of parallel proces sing. We discuss the role of probability models and how they provide p robability information to the arithmetic coder. We conclude with persp ectives on the comparative advantages and disadvantages of arithmetic coding.