Correction of random and burst errors.

摘要

A method and means is described for correcting multiple error bursts in data recorded on a storage medium in blocks, comprising a plurality of sub-blocks. After reading the data, decoded block check syndromes are algebraically summed with estimated block check syndromes to provide a set of syndromes for a code for locating sub-blocks having an error burst. This set of syndromes is decoded to identify each sub-block having an error burst. Concurrently block level syndromes are computed to identify the locations and values of errors within the sub-blocks having error bursts. During writing, the data in all sub-blocks of a block is encoded and block level syndromes are generated for these sub-blocks. These block level syndromes are multiplied by a series of preselected weighting factors ( alpha <l> ... alpha <l(B-1)>) according to the location index l of the sub-block within the block and as multiplied, each is stored in a different one of B buffers. These are cumulatively summed to produce block check syndromes, which are encoded after the last sub-block of the block is written to provide check bytes for their protection. These check bytes and the weighted cumulative sums are stored on the medium at the end of the block as block check syndromes. , and more specifically relates to such a method and means which includes error correction code (ECC) for which decoded and estimated block check syndromes are generated