ADPCM coding and decoding techniques for personal communication systems

摘要

The present invention provides a system and methods for coding and decoding digital data to improve the subjective quality of a voice signal. According to the invention, a sampled voice waveform is compressed by an ADPCM algorithm resulting in a sequence of samples. The sequence of samples are formed into an array and coded into parity bits. Row and column checksums are preferably performed on the parity bits to form row and column vectors. The samples and row and column vectors are transmitted to a receiving unit preferably having a diversity capability. The receiver processes each diversity block by generating row and column vectors from the received samples and comparing the generated row and column vectors to the row and column vectors received. Then based on the comparison, the samples of the best diversity block are used to reconstruct the voice waveform on a sample-by-sample basis.