Channel optimized switched split vector quantization for wideband speech LSF parameters

In this paper, we present an optimized switched split vector quantization (SSVQ) system developed for efficient and robust coding of the speech LPC parameters in terms of line spectral frequencies (LSF). Initially, the aim of this system (called LSF-SSVQ encoder) is to achieve a transparent quantization of wideband speech LSF parameters for ideal transmissions over noiseless channels. Subsequently, our interest was drawn to the improvement of the LSF-SSVQ encoder robustness for real transmissions over a noisy channel. To protect implicitly the transmission indices of our LSF encoders, we developed a non-redundant joint source-channel coding system based on a channel optimized switched split vector quantization technique.