This paper presents a semi-automatic phonetic labeling method for processin
g in the MAT (Mandarin across Taiwan) speech database. MAT speech data are
collected through the telephone networks. Each utterance has been transcrib
ed into Chinese characters and Pinyin symbols. The proposed phonetic labeli
ng method will mark the syllable and sub-syllable boundaries in an utteranc
e. Phonetic symbols are assigned to each segmented syllable. The segmentati
on process is accomplished by using hidden Markov modeling (HMM) and Viterb
i decoding. The accuracy of syllable segmentation is detected by measuring
the syllable length and the distance of a syllable from its state models. T
he experimental results show that the proposed labeling method can achieve
segmentation accuracy around 90% for an allowed tolerance of 16 ms.