Identification of Primitive Speech Signals using TMS320C54x DSP Processor

ct yet simple approach is presented to certain primitive speech signals using fixed-point DSP processor. The technique, based on the statistical derived from the captured voice adually saves around 90% of the pace as the statistical estimates are xtracted signal information rather than the original signal. The discussed does not increase the computational kes into account the dynamic nature of ch signals. The advantages of ing such a routine using a fixed-point essor (TMS320C54x) that supports le executable instructions unlike the pose ones are also discussed.