A Pipelined Adaptive Differential Vector Quantizer for Low-power Speech Coding Applications

A fine-grain pipelined adaptive differential vector quantizer architecture i s proposed f o r low-power speech coding applications. T h e pipelined architecture is developed by employing the relaxed look-ahead technique. The hardware overhead due t o pipelining i s only the pipelining latches. Simulat ions with speech sampled at 8 h'hz show that , f o r a vector dimension o 8, the degradation in the signal-to-noise ratio (,!"/due t o pipe lin ing is negligible. Furthermore, t his degradation is independent of the level of pipelining. Thus , the proposed architecture is attractive from an integrated circuit implementat ion point of view.