Phonetic decomposition for speech recognition of lesser-studied languages
Vijay George John · 2009
This paper deals with voice transcription systems for lesser-studied languages. In particular, it deals with creating phonetic decompositions for words in these languages, an important step in creating a voice transcription system. The two languages cited here as examples of lesser-studied languages are the San Juan Quiahije variety of Chatino and Vlax Romani. For these languages, an English-based phonemic decomposition is inadequate, because lesser-studied languages are not written with the orthographic rules used for English. The phonemic decomposition proposed here is composed of two stages: separation of words into sounds and expansion of sounds into phonemes.