In speech signal processing, the amount of data analyzing requires a long time process. One of the pre-processing techniques to make the speech processing faster is the voiced and unvoiced (V/UV) classification. This article presents an improvement V/UV classification technique based on real time processing using Field Programmable Gate Array (FPGA) board. The Virtex-II Pro board which consists of XC2VP30 chip as central processor unit is used in this experiment. The XC2VP30 chip consists of 30,816 logic cells and it can operate with external memory. The experiment results show that this system can be function on real time system. It used only 4.17-50 ms for time processing which does not effect to delay time process especially in real time system. Moreover, the output speech signal quality is still similar to the original speech signal. This is the major point that the XC2VP30 chip can be develop to use in speech compression and speech recognition.
Keywords: Voiced and unvoiced classification, Xilinx XC2VP30, Real time processing system
Email: Jakkree.s@en.rmutt.ac.th
Sutacha, C. ., & Srinonchat*, J. . (2018). Improvement Voiced and Unvoiced Classification Technique Based on Real Time Processing Using FPGA Board. CURRENT APPLIED SCIENCE AND TECHNOLOGY, 28-32.
https://cast.kmitl.ac.th/articles/135267
