The phonetic characteristics of declarative sentence-ending 'ta' was examined based on the speech of one male Korean speaker in his 20s drawn from a large-scale speech corpus. An analysis using a Unicode-based phone alignment system was compared to an analysis based on manually corrected alignment and the two methods produced largely comparable results. The declarative-ending 'da,' which coincides with prosodic intonational boundaries in Korean, was marked by higher F1 as well as longer duration and lower F0 than medial 'ta', but the effect of utterance boundary on voice quality is inconclusive. The study implies that more accurate duration modeling to the baseline alignment system will enable us to investigate phonetic and prosodic substances of spoken Korean in the large-scale corpus.

Index Terms: Forced alignment, declarative-ending, F0, F1, F2, H1-H2, vowel duration, voice quality.