Review of hardware implementation of Speech-To-Text Engine for Jawi Character

Jawi is a Malay writing script based on the Arabic script which adds six extra characters to accommodate Malay vocal sounds. Real-Time Speech-To-Text Engine for Jawi Character aims to help students to read Jawi more effectively. In this paper, we investigate the suitable method and device that best for speech-to-text hardware implementation. A complete set of Jawi Speech-To-Text tools will be implements by integrating the hardware with a microphone and Jawi word processor. The speech recognition will use Linear Predictive Coding (LPC) as feature extraction and Neural Network (NN) with Backpropagation will be use as a pattern classification. The implementation of this system will be tested using Xilinx Field Programmable Gate Array (FPGA) board.