Abstract
The past decade has witnessed substantial progress towards the application of low bit rate speech coders to civilian and military communications as well as computer related voice applications. Central to this progress has been the de- velopment of new speech coders capable of producing high quality speech at low data rates. Most of these coders incorporate mechanisms to: synthesize speech using Linear Predictive Coding, represent the spectral properties of speech, pro- vide speech waveform matching, and "optimize" the coder's performance for the human ear. A number of these coders have already been adopted in national and international cellular telephony standards. The objective of speech is communication, whether face to face or cell phone to cell phone. Though intelligibility is the key factor in Speech Coding, naturalness of speech is desirable in most cases. There are vocoders like LD-CELP & MPELP which provide high quality and naturalness at medium bitrates. Although CELP has good quality at a low bitrate of 4.8 kbps, it is undesirable for real-time com- munication purposes due to it's highly complex algorithm. Therefore the search for a low complexity vocoder that can achieve reasonably good intelligibility and quality with low bit rate is still on. This thesis suggests a new vocoder to achieve low bit rate, acceptable quality and low complexity. The main factor a®ecting natural sounding speech synthesis is the excitation signal used. This thesis work presents a new excitation signal and designs a vocoder based on the new excitation signal and a new parametric representation for the excitation signal. The new excitation signal is obtained by placing pulses at peaks and valleys of the residue signal. The new vocoder provides good quality at low codec delay and low bitrate
[attachment=311]