• 한국컴퓨터정보학회논문지
• /
• 제26권10호
• /
• pp.1-7
• /
• 2021

현재 대부분의 음성인식 오류 교정에 관한 연구는 영어를 기준으로 연구되어 한국어 음성인식에 대한 연구는 미비한 실정이다. 하지만 영어 음성인식에 비해 한국어 음성인식은 한국어의 언어적인 특성으로 인해 된소리, 연음 등의 발음이 있어, 비교적 많은 오류를 보이므로 한국어 음성인식에 대한 연구가 필요하다. 또한, 기존의 한국어 음성인식 연구는 주로 편집 거리 알고리즘과 음절 복원 규칙을 사용하기 때문에, 된소리와 연음의 오류 유형을 교정하기 어렵다. 본 연구에서는 된소리, 연음 등 발음으로 인한 한국어 음성인식 오류를 교정하기 위하여 LSTM을 기반으로 한 인공 신경망 모델 Sequence-to-Sequence와 Bahdanau Attention을 결합하는 문맥 기반 음성인식 후처리 모델을 제안한다. 실험 결과, 해당 모델을 사용함으로써 음성인식 성능은 된소리의 경우 64%에서 77%, 연음의 경우 74%에서 90%, 평균 69%에서 84%로 인식률이 향상되었다. 이를 바탕으로 음성인식을 기반으로 한 실제 응용 프로그램에도 본 연구에서 제안한 모델을 적용할 수 있다고 사료된다.

• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 2005년도 학술발표회(2)
• /
• pp.1064-1066
• /
• 2005

• 음성과학
• /
• 제7권1호
• /
• pp.109-141
• /
• 2000

Acoustic and aerodynamic characteristics of Cheju consonants were examined with the focus on the well-known three-way distinction among stops (i.e., lenis, fortis, aspirated) and the two-way distinction between sand s*. Acoustic parameters examined for the stops included VOT, relative stop burst energy, Fo at the vowel onset, H1-H2, and H1-F2 at the vowel onset. For the fricatives s and s*, acoustic parameters were fricative duration, Fo, centroid of the fricative noise, RMS energy of the frication, H1-H2 and Hl-F2 at the onset of the following vowel. In investigating aerodynamics, intraoral pressure and oral flow were included for the bilabial stops. Results indicate that, although Cheju and Korean are not mutually intelligible, acoustic and aerodynamic properties of Cheju consonants are very similar in every respect to those of the standard Korean. Among other findings there are three crucial points worth recapitulating. First, stops are systematically differentiated by the voice quality of the following vowel. Second, stops are also differentiated by aerodynamic mechanisms. The aspirated and fortis stops are similar in supralaryngeal articulation, but employ a different relation between intraoral pressure and flow. Finally, our study suggests that the fricative s is better categorized as 'lenis' than as 'aspirated' in terms of its phonetic realization.

• 대한음성학회:학술대회논문집
• /
• 대한음성학회 2003년도 10월 학술대회지
• /
• pp.163-167
• /
• 2003

The present thesis examines the correlation of VOT and F0 in the three-way distinction of Korean obstruents, conducting production and perception tests. In the production test, one female native speaker of Korean with a Seoul dialect (the author) recorded 15 repetitions of a monosyllabic word list including /ka, kha, k*a, pa, pha, p*a, ta, tha, t*a, ca, cha, c*a/ in random order, VOT and F0 of the following vowels were measured, and the result was significant for the three-way distinction with a strong correlation between VOT and F0, and also in the VOT-F0 plot, no overlapping among the domains was observed. As for the perception test, I manipulated the data recorded in the production test, heightening or lowering their F0 values. In all, 14 subjects (seven males and seven females) participated in the identification test. The result was as follows: the fortis stimuli were not influenced by F0 changes, and the VOT and F0 values at the lenis-aspirated boundary were negatively correlated. From these results I concluded the following: 1) VOT and F0 can distinguish the three domains of Korean obstruents without overlapping; 2) the fortis perception does not need F0 as its acoustic cue; and 3) VOT and F0 in the distinction between the lenis and aspirated are in the phonetic trading relation[2].

• 음성과학
• /
• 제13권4호
• /
• pp.51-67
• /
• 2006

This study examined the acoustic properties of initial stops in Korean, produced by Korean native speakers and English Korean learners. The productions of Korean native speakers were compared with those of beginners and advanced learners of Korean. Fundamental frequency(F0) and Voice Onset Time(VOT) were measured in condition of one or two syllable words, containing word-initial lenis, fortis, and aspirated stops. English Korean Learners showed that they produced stops with relatively shorter VOT and lower F0, compared with those of Korean native speakers. In case of the manner of articulation, English Korean learners have production difficulties in order of lenis stops, aspirated stops, and fortis stops. In regard to the place of articulation, English Korean learners showed production troubles in order of labial stops, velar stops, and alveolar stops. In the experiment of perception, it is hard for English Korean learners to distinguish stops of lenis and aspirated. Therefore, the results of production experiment were almost consistent with those of the perception experiment. Finally, according to both groups of proficiency, the results demonstrated that the advanced learners produce or perceive Korean stops easier than the beginners.

• 말소리와 음성과학
• /
• 제16권1호
• /
• pp.17-24
• /
• 2024

The Korean three-way stop contrast (lenis, aspirated, fortis) is currently undergoing a sound change, such that the primary cue distinguishing lenis and aspirated stops is shifting from voice onset time (VOT) to F0. Despite recent discussions of this shift, research on voice quality, traditionally considered an additional cue signaling the contrast, remains sparse. This study investigated the extent to which the associated voice quality [as reflected in the acoustic measurements of H1^*-H2^*, H1^*- A1^*, and cepstral peak prominence (CPP)] contributes to the three-way stop contrast, and how the realization is conditioned by prominence- vs. boundary-induced prosodic strengthening amid the ongoing sound change. Results for 12 native Korean speakers indicate that there was a substantial distinction in voice quality among the three stop categories with the breathiness of the vowel being the greatest after the lenis, intermediate after the aspirated, and least after the fortis stops, indicating the role of voice quality in the maintenance of the three-way stop contrast. Furthermore, prosodic strengthening has different effects on the contrast and contributes to the enhancement of the phonological contrast contingent on whether it is induced by prominence or boundary.

• 한국동물학회지
• /
• 제19권3호
• /
• pp.101-112
• /
• 1976

등줄쥐(Striped field mouse, Apodemus agrarius coreae Thoma)의 염색체수는 2n=48이다. 핵형은 상염색체가 형태적으로 명백히 3군으로 나누어 지는데, 1쌍의 가장 큰 제 1번차단부가 염색체군, 4쌍의 작은 중부 내지 차중부 염색체군과 18쌍의 아크로 센트릭 염색체군이다. 제 1번 염색체군을 제외한 다른 염색체군들에서의 상동염색체 구별은 크기가 점차적인 차이를 보이기 때문에 쉽지 않다. X염색체는 제 3번 염색체와 크기의 것이며 Y염색체는 작은 아크로센트릭 염색체와 차이를 나타내어서 새로운 형임이 밝혀졌다. 본 종의 집단내 또는 집단간 염색체 多形現象은 없었다. 갈밭쥐(Manchurian red vole, Microtus fortis pelliceus Thomas)의 염색체수는 2n=52이다. 핵형은 상염색체가 역시 3군으로 나누어 지는데, 5쌍의 중부내지 차중부 염색체군, 2쌍의 차단부 염색체군과 18쌍의 아크로센트릭 염색군이다. 아크로센트릭 염색체군에서의 상동염색체 구별은 점차적인 차이를 보이기 때문에 역시 어렵다. X염색체는 가장 큰 중부 염색체임으로 쉽게 구별이 되나 Y는 작은 아크로센트릭 염색체중 하나가 아닌가 생각된다. 또한 본 종은 아크로센트릭 염색체군중에 서로 다른 크기의 2차 응축현상을 보이는 상동염색체들이 2쌍 있음이 밝혀졌다. 본 연구를 통해서 본 종의 핵형은 쏘련산 M. fortis 아종의 핵형과 같지 않음이 밝혀졌는데, 쏘련산 M. fortis의 핵형은 상염색체가 5쌍의 중부 염색체군, 1쌍의 차중부염색체군과 19쌍의 아크로센트릭 염색체군으로 되어있으며, X는 비교적 큰 아크로센트릭이고, Y는 보다 작은 아크로센트릭 염색체이다. 본 종의 한국내에서의 염색체 다형현상 및 성염색체의 대형화현상은 없었다. 비단털쥐(Korean giant hamster, Cricetulus trition nestor Thomas)의 염색체수는 2n=28이며, 염색체의 크기는 위의 2종 보다 커서 $7.5\\mu - 1.5\\mu$이다. 핵형은 상염색체가 아주 뚜렷한 2군으로 나뉘어 지는데, 11쌍의 큰 아크로센트릭 염색체군과 2쌍의 아주 작은 중부 염색체군이다. X염색체는 비교적 큰 차단부 염색체로서 쉽게 구별이 된다. 본 실험에서 숫컷을 재료로 상요하지 못했기 때문에 Y염색체는 판정할 수가 없었다. 또한 본 종의 핵형은 쏘련산 Tscherskia triton 의 핵형과 동일하다. 따라서 Cricetulus 속과 Tscherskia속과의 분류에 있어서 동일함이 핵형으로도 증명된 셈이다. 본 종의 염색체수는 다른 햄스터류보다 많으나, 아주 작은 2쌍의 중부 염색체군이 있음이 특이하며, 한국산 햄스터로서의 세포유전학적 실험동물로서 이용가치가 있다고 사료된다.

• 말소리와 음성과학
• /
• 제1권3호
• /
• pp.39-48
• /
• 2009

Previous studies on the effects of speaking rate on voice onset time (VOT) of stops in English, French, Icelandic, and Thai indicate that speaking rate asymmetrically affects VOT values. That is, pre-voiced and long-lag stops vary due to the rate factor more than short-lag stops do. One suggested explanation for this asymmetry is that it is due to the necessity of maintaining phonetic contrasts among the stop categories. Since pre-voiced and long-lag stops represent the ends of the VOT scale, they encompass broad swathes of that range and consequently allow for large variations. On the other hand, the VOT variations of short-lag stops may result in overlap with the VOTs of long-lag stops. This study aimed to explore the effects of speaking rate on the VOTs of Korean stops and see whether Korean fortis and lenis stops are limited in the degrees of variation as a function of rates due to the existence of stops with larger VOT values, lenis and aspirated stops respectively. Conversely, aspirated stops were expected to show more variation since there are no other categories with longer VOTs. Fortis, lenis, and aspirated stops in /CVn/ words (C = bilabial or velar stop, V = /i/ or /a/) were examined in isolation, and at normal and fast rates in a carrier sentence. Speaking rates were controlled by alternating words or sentences on a computer screen at intervals of two seconds for the isolation- and normal-rate conditions and one second for the fast-rate condition. This study found that while the VOTs of fortis stops did not change significantly, those of lenis and aspirated stops showed considerable changes as a function of speaking rates. Also, overlap between lenis and aspirated stops occurred considerably at all speaking rates. These phenomena were interpreted to relate to the fact that VOT contrasts between lenis and aspirated stops in Korean are currently being collapsed. Large variations of lenis stops as a function of rates seem to occur due to a weak motivation to limit the degree of variations for the purpose of maintaining phonetic contrasts. The significant overlap between lenis and aspirated stops at all rates was interpreted to occur because the VOT merger between the two categories became considerably fixed. Also the percentage of correctly-classified VOTs by optimal-boundary values between lenis and aspirated stops turned out to be lower than in previously-studied languages. This was interpreted to be further evidence that VOTs are losing their role in contrasting the two stop categories in Korean.

• 대한음성학회지:말소리
• /
• 제46호
• /
• pp.1-11
• /
• 2003

This study examines the acoustic characteristics of Korean stops of two dialects, Seoul and Kyungsang, focusing on VOT(Voice Onset Time). 8 speakers of these two dialects were asked to read 590 words which contain the stops of different places of articulation and phonation types. The results showed that overall the VOTs of Kyungsang dialect were shorter than those of Seoul dialect. This was more prominent in lenis stops than in fortis or aspirated stops. It was also shown that there were significant VOT overlapping differences between the two dialects.

• 대한음성학회:학술대회논문집
• /
• 대한음성학회 2004년도 춘계 학술대회 발표논문집
• /
• pp.139-142
• /
• 2004

This study examines the acoustic characteristics of Korean stops of two dialect, Seoul and Daegu, 20 speakers of these two dialects were asked to read 15 words containing the stops of different places of articulation and phonation types at initial. The stops in the two dialects show mainly two acoustic differences. Firstly, There was a difference in distinctive features for phonetic types in the two dialects. Secondly, lenis revel fortis`s characters in Daegu dialect.