YOU ARE HERE: HOMESPEECH DATA COMPRESSION BASED ON A SPEEDUP ALGORITHM USING VECTOR RUNLENGTH CODING

SPEECH DATA COMPRESSION BASED ON A SPEEDUP ALGORITHM USING VECTOR RUNLENGTH CODING

Pages: 19 to 34

Ahmed A. Abdelwahab

Abstract:

This paper proposes a new speech data compression technique based on a new speedup algorithm using Vector Runlength Coding (VRLC). VRLC is proposed and software implemented for speech speedup based on some speech specific properties both in the temporal and transform domains.  The successful implementation of VRLC for real-time secure speech transmission system for multimedia applications is described.  In this paper, there are two main objectives which are speech data compression and secure transmission for the compressed (speedup) speech by making the transmitted signal completely imperceptible using a lower complexity algorithm compared to other similar methods in the speech coding literature to compare with. Therefore, the reconstructed received speech signal would be as much as possible of a good perceptual quality. The algorithm is applied in both temporal domain and one dimensional discrete cosine transform (1D_DCT) domain. The simulation results show the superiority of applying the new algorithm in transform domain over temporal domain, in terms of the achieved total imperceptible compressed speech transmitting signal and the intelligibility of the reconstructed speech at the receiver.

PDF