Compression of multilevel signals

摘要

This invention relates to method and apparatus for compressing multilevel signals. The compression is based upon prediction errors and probability distributions. Compression is improved by conditioning the probability distributions using context of previous events. Storage required for storing the probability distributions is reduced by partitioning the prediction errors into predetermined ranges which become the coding events and contexts. Compression is made efficient with a partition which is a function of the sign and the number of significant bits in the prediction errors.