Block floating point compression with exponent token codes

摘要

A method and apparatus for compressing signal samples uses block floating point representations where the number of bits per mantissa is determined by the maximum magnitude sample in the group. The compressor defines groups of signal samples having a fixed number of samples per group. The maximum magnitude sample in the group determines an exponent value corresponding to the number of bits for representing the maximum sample value. The exponent values are encoded to form exponent tokens. Exponent differences between consecutive exponent values may be encoded individually or jointly. The samples in the group are mapped to corresponding mantissas, each mantissa having a number of bits based on the exponent value. Removing LSBs depending on the exponent value produces mantissas having fewer bits. Feedback control monitors the compressed bit rate and/or a quality metric. This abstract does not limit the scope of the invention as described in the claims.

主权项

1. A method for data compression, comprising:
providing an exponent encoding table in a memory of a data processing system, the exponent encoding table including a first plurality of exponent token codes associated with a plurality of pairs of exponent difference values, wherein a particular exponent token code is associated with a particular pair of exponent difference values; determining block exponents for a first encoding group, a second encoding group and a third encoding group, respectively, wherein the encoding groups comprise samples processed by a processor of the data processing system; calculating a first exponent difference between the block exponents of the first encoding group and the second encoding group; calculating a second exponent difference between the block exponents of the second encoding group and the third encoding group; determining whether the first and second exponent differences match one of the pairs of exponent difference values in the exponent encoding table and if so, selecting the exponent token code associated with the matched one of the pairs to produce an exponent token that represents the first and second exponent differences; and providing the exponent token to a compressed data packet that stores compressed data for the encoding groups.