SIMULATION AND EVALUATION OF PHONETIC SPEECH RECOGNITION TECHNIQUES. VOLUME II. SEGMENTATION OF CONTINUOUS SPEECH INTO PHONEMES.
NATIONAL CASH REGISTER CO DAYTON OHIO
Pagination or Media Count:
The objective of the program is the analysis of design parameters required for the simulation and evaluation of phonetic speech recognition methods. An approach to the segmentation of continuous speech into phonemes is described, and the results of segmentation experiments intended to evaluate the approach are presented in the report. Continuous speech is segmented by an analysis of the spectral power density function of the speech signal. Those different schemes for segmentation which differ in the form by which changes in the spectral power density function are detected and translated into segmentation criteria are described. The application of relatively simple segmentation rules allowed the direct segmentation into phonemes with an efficiency factor of more than 0.75. Author