Author(s)

Haibo Shi, Yonghong Li

Corresponding Author

Yonghong Li

ABSTRACT

This paper studies and realizes the problem of phoneme automatic segmentation of Tibetan Lhasa dialect.In this paper, the automatic segmentation of the related methods and details about based on hidden markov model (HMM) of the Tibetan capital Lhasa dialect phoneme principle and method for automatic segmentation of HTK as a set of specially set up and handle the HMM tools is the best tool to implement this method, the use of HTK toolkit experimental phonetics, combining with the characteristics of itself in the Tibetan language and language flow has been basically achieved the phoneme segmentation automatically, [1]based on the HMM model training, can undertake effective segmentation batch of corpora, greatly saves manpower and time cost.

KEYWORDS

Tibetanlhasadialect, HMM, HTK, Forcedalignment, Automatic segmentation