Calculation of the Intervals

The algorithm for the arithmetic coding is not of complex nature. The most difficult part for realization is the fact, that the code word is a floating-point number with an extremely large amount of significant digits. All computer systems are restricted to a certain, comparable low range of values.

Beyond that, not all processors are able to process floating-point numbers directly. This requires a FPU (Floating Point Unit), which somtimes was placed in a separate coprocessor unit like for the traditional Intel x86 family. Most of the early computer systems had no internal floating-point arithmetic available and had to emulate those operations.

For the reasons mentioned usual coding procedures operate with pure integer arithmetics. These aspects will not be considered in the following, with exception of some notes.

 <   ^   >