An effective speech compression based on syllable division
Yan Zhang, Xiao Dong, Qunyan Ren, Shengming Guo, Fuyuan Mo · Proceedings of meetings on acoustics · 2016
For medium and long distance underwater speech communication, transmission rate is relatively low due to various disadvantageous factors such as multipath propagation. Consequently, speech compression methods are necessary for effective transmission. Considering that the speech signal is usually redundant, a speech compression algorithm based on syllable division is proposed. It can extract the voiced speech, the unvoiced speech and the silence part, and subsequently delete the redundant part that does not affect the semantic comprehension. Through data processing, this compression algorithm is demonstrated that it can compress the speech signal to 75 % of the original one. Besides, listener recognition and semantic comprehension will not be affected.