Some acoustic characteristics of syllable nuclei in conversional speech
M. A. Earle, Larry L. Pfeifer · The Journal of the Acoustical Society of America · 1975
Approximately 5 min of conversional speech from an interview of one male speaker of Midwestern American was transcribed and digitized. The digitized speech is bandlimited at 5 kHz, sampled at 10 kHz, and quantized to 12 bits per sample. Steady-state portions of the vowels and initial and final portions of the diphthongs are located by visual inspection of a time-synchronized display of the waveform, spectral peaks, and the rms of the signal. Formant parameters derived from the inverse filter method of linear prediction are then extracted from four consecutive spectral samples representing 29.6 msec of the steady state of each vowel or of the most stable region in the initial and final portions of each diphthong. The effect of various segmental environments on some acoustic parameters, such as formant frequencies, is discussed. A comparison is made of the vowel and diphthong samples in terms of formant frequency means and standard deviations. [This research is supported by the Air Force Office of Scientific Research under Contract F44620-74-C-0034.]