Application of speech recognition and synthesis on underwater acoustic speech transmission

This work employs state-of-the-art speech recognition and synthesis technologies for speech source coding and decoding. This approach decreases the coded speech data to a great extent, thus cuts down the bit-rate requirement greatly for underwater acoustic speech transmission. The experiments show that the reconstructed speech is clear, intelligible, and the codec performance is satisfactory. This work opens up a new road for underwater acoustic highly-intelligible speech transmission and has important application values.