Kouichi Honma

Design of error correction methods using redundancy of speech coder data

Error correction methods which provide higher speech quality on channels of poor quality in digital mobile communications are presented. Conventionally, to mitigate the degradation of speech quality, the parameters which are more sensitive to channel errors are protected by more powerful channel codes. In that respect, source coding and channel coding are integrated. To expedite the integration of source coding and channel coding, the method to improve the error correction performance of convolutional code for its middle position in the trellis and the method to use redundancy of speech coder data such as correlation of frame energy or pitch predictor lag between consecutive frames are presented and evaluated. The results obtained indicate that the presented error correction methods are efficient for redundant source information such as voice signals.<<ETX>>