Some studies on text coding based very low bit rate speech coder

Abstract

There may be enormous advances in digital communication but newlinestill the voice is the superior one by which people can exchange more ideas newlineUnlike many other areas of technology speech technology is closely linked to newlinean understanding of basic human properties It comprises a variety of different newlinetechniques such as automatic speech recognition speaker recognition speech newlinecoding and speech synthesis This thesis mainly concentrates on speech newlinecoding Speech coding is the process of digitally representing a speech newlinesignal for the transmission of data over network with in the given limited newlinebandwidth while maintaining sufficient level of quality of the synthesized newlinespeech with reasonable computational complexity Especially a lower bit rate newlinecoding implies that a smaller bandwidth is required for transmission and less newlinememory is needed for storage Due to the large amounts of data that needs to newlinebe coded a high bit rate is required for a more natural representation of newlinehuman speech If memory or capacity of the channel is of concern then the newlinebit rate can be lowered at the cost of poorer quality of the reconstructed newlinespeech signal newline newline