Temporal and spectral characteristics of Korean phonation types

Temporal and spectral characteristics of Korean obstruents - i.e. stops, affricates, and sibilants - are the focus of this study. Comparing diverse obstruent manners necessitated discriminating subsegmental components, especially closure, frication, and aspiration (VOT). Detailed measurements were made for the obstruents in utterance-initial CV sequences (#Cada) and intervocalic VCV sequences (taCada#) for three speakers of Standard Korean (160 tokens each). A significant temporal compensation between the durations of aspiration and the following vowel was demonstrated. Also noteworthy is the patterning of the sole non-tense sibilant /s/ with Aspirated, not Lenis, in terms of all parameters and acoustic measures employed. To study the spectral characteristics of phonation types, I revised the formula for the combined source and radiation characteristics for Fant’s (1960) Source Filter Theory to permit spectral tilt variation. A full derivation of this extension is provided in Chapter 3. Once revised, I proposed a new phonation type index (k) to provide a single and simple measure of the overall spectral tilt and an acoustic measure of open quotient (k1). Both k and k1 were thus freed from the effects of fundamental frequency and vowel quality. Spectral characteristics of the vowel following Korean obstruents were measured in terms of these newly proposed measures of phonation type as well as previously established measures. The results show a significant difference in the overall level, the pattern of the time course, and the onset values, across speakers, between positions, and across phonation types for both the acoustic measures of open quotient (the new k1 as well as H1-H2, H1*-H2*, and Cor(H1-H2)) and the measures of spectral tilt (the new k as well as H1*-A1, H1-A2, and H1*-A3*) My finding supports the generally accepted view that the time courses and the onset values of F0 are a characteristic feature of Korean phonation types. New in this study is the finding that the time courses and onset values of B1 and the acoustic measures of open quotient and spectral tilt are a consequence of the transition of the prevocalic glottal state to modal vocal fold vibration.Linguistic