• 제목/요약/키워드: Morpheme Boundaries

검색결과 4건 처리시간 0.017초

한국어 연속음성인식을 위한 형태소 경계에서의 발음 변화 현상 모델링 (Modeling Cross-morpheme Pronunciation Variation for Korean LVCSR)

  • 이경님;정민화
    • 대한음성학회:학술대회논문집
    • /
    • 대한음성학회 2003년도 5월 학술대회지
    • /
    • pp.75-78
    • /
    • 2003
  • In this paper, we describe a cross-morpheme pronunciation variation model which is especially useful for constructing morpheme-based pronunciation lexicon for Korean LVCSR. There are a lot of pronunciation variations occurring at morpheme boundaries in continuous speech. Since phonemic context together with morphological category and morpheme boundary information affect Korean pronunciation variations, we have distinguished pronunciation variation rules according to the locations such as within a morpheme, across a morpheme boundary in a compound noun, across a morpheme boundary in an eojeol, and across an eojeol boundary. In 33K-morpheme Korean CSR experiment, an absolute improvement of 1.16% in WER from the baseline performance of 23.17% WER is achieved by modeling cross-morpheme pronunciation variations with a context-dependent multiple pronunciation lexicon.

  • PDF

한국어 연속음성인식 시스템 구현을 위한 형태소 단위의 발음 변화 모델링 (Modeling Cross-morpheme Pronunciation Variations for Korean Large Vocabulary Continuous Speech Recognition)

  • 정민화;이경님
    • 대한음성학회지:말소리
    • /
    • 제49호
    • /
    • pp.107-121
    • /
    • 2004
  • In this paper, we describe a cross-morpheme pronunciation variation model which is especially useful for constructing morpheme-based pronunciation lexicon to improve the performance of a Korean LVCSR. There are a lot of pronunciation variations occurring at morpheme boundaries in continuous speech. Since phonemic context together with morphological category and morpheme boundary information affect Korean pronunciation variations, we have distinguished phonological rules that can be applied to phonemes in within-morpheme and cross-morpheme. The results of 33K-morpheme Korean CSR experiments show that an absolute reduction of 1.45% in WER from the baseline performance of 18.42% WER was achieved by modeling proposed pronunciation variations with a possible multiple context-dependent pronunciation lexicon.

  • PDF

일본어 회화문에 나타난 축약형의 음운론적 해석과 음향음성학적 분석 (Acoustical Analysis of Phonological Reduction in Conversational Japanese)

  • 최영숙
    • 음성과학
    • /
    • 제8권4호
    • /
    • pp.229-241
    • /
    • 2001
  • Using eighteen texts from various genera of present-day Japanese, I collected phonologically reduced forms frequently observed in conversational Japanese, and classified them in search of a unified. explanation of phonological phenomena. I found 7,516 cases of reduced forms which I divided into 43 categories according to the types of phonological changes they have undergone. The general tendencies are that deletion and fusion of a phoneme or an entire syllable takes place frequently, resulting in the decrease in the number of syllables. From a morphosyntactic point of view, phonological reduction often occurs at the NP and VP morpheme boundaries. The following findings are drawn from phonetical observations of reduction. (1) Vowels are more easily deleted than consonants. (2) Bilabials ([m], [b], and [w]) are the most likely candidates for deletion. (3) In a concatenation of vowels, closed vowels are absorbed into open vowels, or two adjacent vowels come to create another vowel, in which case reconstruction of the original sequence is not always predictable. (4) Alveolars are palatalized under the influence of front vowels. (5) Regressive assimilation takes place in a syllable starting with [r], changing the entire syllable into a phonological choked sound or a syllabic nasal, depending on the voicing of the following phoneme.

  • PDF

현대일본어의 회화문에 나타난 축약형의 음운론적 분석 (Analysis of Phonological Reduction in Conversational Japanese)

  • 최영숙;좌등자;박희태
    • 대한음성학회:학술대회논문집
    • /
    • 대한음성학회 1996년도 10월 학술대회지
    • /
    • pp.198-206
    • /
    • 1996
  • Using eighteen text materials from various goners of present-day Japanese, we collected phonologically reduced forms frequently observed in conversational Japanese, and classified them in search of unified explanation of phonological reduction phenomena. We found 7,516 cases of reduced forms which we divided into 43 categories according to the types of phonological changes they have undergone. The general tendencies ale that deletion and fusion of a phoneme or an entire syllable takes place frequently, resulting in the decrease in the number of syllable. Typical examples frequently observed throughout the materials are : $~/noda/{\rightarrow}~/nda/,{\;}-/teiru/{\rightarrow}~/teru/,{\;}~/dewa/{\rightarrow}~/zja/,{\;}~/tesimau/{\rightarrow}~/cjau/$. From morphosyntactic point of view phonological reduction often occurs at the NP and VP morpheme boundaries. The following findings are drawn from phonological observations of reduction. (1) Vowels are more easily deleted than consonants. (2) Bilabials(/m/, /b/, and /w/ are the most likely candidates for deletion. (3) In a concatenation of vowels, closed vowels are absorbed into open vowels, or two adjacent vowels come to create another vowel, in which case reconstruction of the original sequence is not always predictable. (4) Alveolars are palatalized under the influence of front vowels. (5) Regressive assimilation takes place in a syllable starting with ill, changing the entire syllable into phonological choked sound or a syllabic nasal, depending on the voicing of following phoneme.

  • PDF