| 摘要 |
A plurality l of data blocks, each comprised of at least two data words and a parity word associated bit for bit with the data words, together with an error correcting code word associated with each data and parity word, are transmitted in interleaved relation. L data words and associated error correcting code words followed by l parity words and associated error correcting code words and followed by l data words and associated error correcting code words, all transmitted in seriatim, are received and stored, and each error correcting word is decoded to ascertain the presence of an error in each received data or parity word. All of the data words and the parity word in a block are read out simultaneously, together with an indication of the presence of errors in any of the read out data and parity words; and in the event of an error, an erroneous data word is corrected in accordance with the parity word in that block; and if parity correction cannot be achieved, an erroneous data word is replaced by the average of the data word in the preceding and next following data block. In a preferred application, each data word is formed of a plurality of multi-bit samples of, for example, interleaved left channel and right channel audio signals. The overall length of l data words and associated error correcting words plus l parity words and associated error correcting words is greater than the longest expected error length, such as a burst error. |
