Description

This document presents an implementation of a hidden Markov model (HMM) speech recognition system using the 16-bit fixed-point TMS320C2xx or TMS320C5x digital signal processors (DSPs). It discusses system resource requirements vocal flexibility and future enhancements to this system. The application is verified using a TMS320C53 DSP platform.