Determining CPU and memory requirements for real-time speech recognition systems using the TMS320C3x/C4x

Developing a computer system using real-time speech recognition previously required a workstation using non-specialized CPUs. Limits to the system were imposed by the amount of memory and hardware required. The Texas Instruments (TI?) TMS320C3x/C4x (’C3x/’C4x) digital signal processor (DSP) is a high performance CMOS 32-bit floating point processor. This project uses the (’C3x/’C4x) DSP to count the cycles consumed by the algorithms involved and to know how many words can be recognized with a predetermined accuracy, allowing you to calculate the hardware required for your recognition system to perform. This document describes the CPU and memory requirements for DSP real-time speech recognition systems used in consumer applications. Included are sections on signal preprocessing, parameter measurement, vector quantization, veterbi algorithm, and results. This work was supported by the Spanish government under grant number C95-l022-C05-03. This document was part of the first European DSP Education and Research Conference that took place September 26 and 27, 1996 in Paris. For information on how TI encourages students from around the world to find innovative ways to use DSPs, see TI’s World Wide Web site at www.ti.com.