• 비교문화연구
• /
• 제37권
• /
• pp.287-307
• /
• 2014

The purpose of this paper is to introduce a new and effective classroom speaking activity helping student's communication in real situation. It will be one of useful teaching techniques for teachers because it can be used with other various types of speaking activities together. The activity is designed by the researcher, and named as the Partner's Speech Reconstitution(PSR) in this paper. In chapter 2, Noticing and Output hypothesis which is the theoretic basis of the PSR will be described and the chapter 3 will explain activity methods and examples of the PSR, and then describe its merits and demerits. The researcher applied and practiced the PSR in the speaking class for international students in the K university for three semesters. This paper systematically introduces its organized activity. It helped learners elicit speaking performance of students who avoided talking in the speaking class, made the students concentrate in speaking activity, and helped the learners to talk sufficiently by inducing each student to reconstitute partner's speech production.

• 한국음향학회지
• /
• 제18권2호
• /
• pp.11-17
• /
• 1999

본 논문에서는 DHMM(Discrete Hidden Markov Model) 기반의 음성 인식 시스템에서 소음에 강인한 인식 성능을 얻기 위하여, 관찰 확률 스무딩(observation probability smoothing) 방법을 제안하고, 자동차 소음하에서의 음성 인식에 적합한 소음처리 기법을 실험을 통해 제시한다. 제안된 관찰 확률 스무딩 방법은 입력되는 음성의 특징벡터가 소음에 오염되어 양자화(vector quantization) 과정에서 적절한 코드워드(codeword)가 아닌 다른 코드워드로 양자화됨으로써 발생하는 인식성능 저하를 막기 위하여, 각각의 코드워드와 거리가 가까운 코드워드들의 관찰 확률값을 높여주는 방법이다. 이 밖에 자동차 소음에 대한 대처 방안으로 특징 벡터의 거리 측정시의 리프터(lifter) 사용, 고역 통과 필터(high pass filter) 사용, 스펙트럴 차감법(spectral subtraction) 사용 등의 성능을 평가한다. 인식 실험은 자동차 정지 중과 주행 중의 두 가지 상황에서 녹음된 한국어 단독 숫자음과 명령어 14단어에 대해 수행하였으며, 정지 중 97.4%와 주행 중 59.1%의 인식률로부터, 제안된 관찰 확률 스무딩 방법과 리프터, 고역 통과 필터, 스팩트럴 차감법의 소음 처리 기법을 추가한 결과, 정지 중 98.3%와 주행 중 88.6%의 인식률을 얻을 수 있었다.

• 한국음향학회지
• /
• 제21권3호
• /
• pp.323-329
• /
• 2002

본 논문에서는 임의의 시점에서 발생하여 음성 신호의 일부분을 심하게 손상시키는 시간선택 잡음 (time-selective noise)을 보상하기 위한 프레임 신뢰도 가중 방법을 제안한다. 음성 프레임들은 서로 다른 정도의 신뢰도를 갖으며, 신뢰도는 프레임의 신호대잡음비 (signal-to-noise ratio)에 비례한다. 잡음이 일정한 경우에는 무음구간에서 획득한 잡음 정보를 이용하여 프레임의 신호대잡음비 추정이 용이하나, 시간선택 잡음은 잡음추정이 어렵다. 따라서, 본 연구에서는 프레임 신뢰도를 추정하기 위해 깨끗한 음성의 통계적 모델을 사용하였다. 제안한 MFR (model-based frame reliability) 방법은 탐조 모델의 평균 벡터열과 입력 MFCC (mel-frequency cepstral coefficient) 특징 벡터 열의 역변환에 의해 얻은 필터뱅크 에너지를 이용하여 프레임 신호대잡음비를 근사한다. 다양한 버스트 (burst) 잡음에 대한 인식 실험 결과, 제안한 방법은 프레임의 신뢰도를 효과적으로 나타낼 수 있었으며, 이 신뢰도를 우도 계산에서 가중치로 적용하여 인식 성능을 향상시킬 수 있었다.

• 한국음향학회지
• /
• 제20권3호
• /
• pp.56-62
• /
• 2001

본 논문에서는 손실이 발생하는 상황에서 높은 인식률을 유지하기 위해서 손실 데이터 이론을 음성 인식기에 적용하였다 손실 데이터 이론은 일반적으로 이용되는 통계적 정합 방법인 은닉 마코프 모델 (HMM: hidden Markov model) 중 연속 Gaussian확률 밀도 함수를 이용하여 음성 특징들의 출력 확률을 나타내는 경우에 쉽게 적용할 수 있다는 장점을 갖고 있다. 손실 데이터 이론의 방법 중 계산량이 적고 인식기에 적용이 쉬운 주변화(marginalization)방법을 사용하였으며 특징 벡터의 특정 차수나 시간열의 손실 검출 방법은 음성 신호의 에너지와 주위 배경 잡음의 에너지의 차이가 임계치보다 작게 되는 부분을 찾는 주파수 차감 방법을 이용하였다. 본 논문에서 제안한 손실 영역의 신뢰도 평가는 분석 구간이 모음일 확률을 계산해서 비교적 잉여 정보가 많이 포함된 모음화된 구간의 손실만을 처리하도록 하였다. 제안한 방법을 사용하여 여러 잡음 환경에 대해서 기존의 손실 데이터 처리 방법만을 사용한 경우보다 452 단어의 화자독립 단어 인식 실험을 수행한 결과 오류율측면에서 평균적으로 약 12％의 성능 향상을 얻을 수 있었다.

• 음성과학
• /
• 제9권2호
• /
• pp.109-132
• /
• 2002

The purpose of this study is to examine the focus realization of [Adjective+Noun] phrases which are used in English classroom situations. In order to examine this, two production and one perception experiments were designed. The noun phrases in the first two production experiments are divided into three patterns according to the location of focus. The difference between the two production experiments is that in the first experiment the focused words are contextually given in the classroom situation, but in the second experiment they are presented in written form. We compare the native English teachers' focus realization of noun phrases with that of Korean teachers from the point of view of intonational phonology. In the perception test, we examine how the uttered sentences are perceived by English native speakers and Korean native speakers. The results from the three experiments show that native English teachers' focus realization is quite consistent with informational structure. Also, there is a significant difference in pitch range of adjectives and nouns when the native speakers give pitch accents on the two content words, and the uttered sentences are mostly perceived as well as the speakers' intentions. As for Korean speakers, however, they usually focus only on the adjective or they focus on both the adjective and the noun, regardless of the relative informativeness of these words. From these findings, we can conclude that focus realization of Korean teachers is rather inconsistent with respect to informational structure when compared to that of native English teachers.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• 제15권1호
• /
• pp.1-11
• /
• 2015

A concept of Deep Level Situation Understanding is proposed to realize human-like natural communication (called casual communication) among multi-agent (e.g., humans and robots/machines), where the deep level situation understanding consists of surface level understanding (such as gesture/posture understanding, facial expression understanding, speech/voice understanding), emotion understanding, intention understanding, and atmosphere understanding by applying customized knowledge of each agent and by taking considerations of thoughtfulness. The proposal aims to reduce burden of humans in humans-robots interaction, so as to realize harmonious communication by excluding unnecessary troubles or misunderstandings among agents, and finally helps to create a peaceful, happy, and prosperous humans-robots society. A simulated experiment is carried out to validate the deep level situation understanding system on a scenario where meeting-room reservation is done between a human employee and a secretary-robot. The proposed deep level situation understanding system aims to be applied in service robot systems for smoothing the communication and avoiding misunderstanding among agents.

• 대한임베디드공학회논문지
• /
• 제8권4호
• /
• pp.185-194
• /
• 2013

Abstract- In distant-talking environments, speech recognition performance degrades significantly due to noise and reverberation. Recent work of Michael L. Selzer shows that in microphone array speech recognition, the word error rate can be significantly reduced by adapting the beamformer weights to generate a sequence of features which maximizes the likelihood of the correct hypothesis. In this approach, called Likelihood Maximizing Beamforming algorithm (Limabeam), one of the method to implement this Limabeam is an UnSupervised Limabeam(USL) that can improve recognition performance in any situation of environment. From our investigation for this USL, we could see that because the performance of optimization depends strongly on the transcription output of the first recognition step, the output become unstable and this may lead lower performance. In order to improve recognition performance of USL, some post-filter techniques can be employed to obtain more correct transcription output of the first step. In this work, as a post-filtering technique for first recognition step of USL, we propose to add a Wiener-Filter combined with Feature Weighted Malahanobis Distance to improve recognition performance. We also suggest an alternative way to implement Limabeam algorithm for Hidden Markov Network (HM-Net) speech recognizer for efficient implementation. Speech recognition experiments performed in real distant-talking environment confirm the efficacy of Limabeam algorithm in HM-Net speech recognition system and also confirm the improved performance by the proposed method.

• 한국차세대컴퓨팅학회논문지
• /
• 제13권6호
• /
• pp.96-105
• /
• 2017

본 논문에서는 실제 응급상황센터에 접수된 신고전화의 음성분석을 통하여 발화자의 연령을 분류하고자 한다. 2가지 발화행태적 특징요소인 무성휴지(Silent Pause), 대화반응시간(Turn-taking latency)를 활용하여 성인과 노인을 분류할 수 있는 특징에 대한 분류기준을 선정하고, 이를 기계학습 분류기인 SVM(Support Vector Machine)을 활용하여 분류정확도를 확인하였다. 먼저, 응급상황센터의 실제 신고전화에 대하여 발화행태적 특징 요소를 기반으로 청취분석을 통하여 발생길이에 대하여 성인과 노인사이에 통계적으로 유의하다는 것을 확인하였다(p<0.05). 또한, 성인과 노인 각 100개, 총 200개의 음성데이터를 5차 교차검증방법을 사용하여 기계학습을 실행한 결과, 2가지의 발화행태를 모두 사용한 복합기준(무성휴지+대화반응시간)일 경우, 70%의 가장 높은 분류정확도를 확인할 수 있었다. 본 연구의 결과는 음성에 기반한 연령을 분류하는 연구에 있어서, 기존의 음성정보와 더불어, 새로운 발화행태적 특징요소와의 결합을 통하여 연령구분을 가능하게 하는 새로운 방법으로 제안할 수 있을 것이다. 또한, 향후 음성기반 상황판단 시스템 기술 개발에 있어서 기초자료로 적용이 가능하며, 이를 통하여 신속한 연령분류를 판단을 통한 상황대처가 가능하도록 하는 데에 기여할 수 있을 것이다.

• 영어어문교육
• /
• 제15권2호
• /
• pp.225-243
• /
• 2009

This study investigates the use of different speech act of apology strategies between male and female EFL college students by comparing the components of intensity, stylistic competence, and semantic formulas. The data was collected from 37 participants who were studying freshmen English reading course at the Department of English Education of C University in Seoul. Most students were English majors taking pre-teacher course of teaching English for secondary school students. The participants were divided into two gender groups of male and female. The discourse completion test (DCT) which was revised from the speech act of apology by Olshtain and Cohen (1990) was provided with the participants after the researcher explained the speech act of apology in ten situations. The speech act of apology depends on situation variables: social solidarity, severity of offense, and social status. The results show that in the preference of intensity, male and female have almost the similar ratio in high (female: 24.7%, male 24%) and low intensity (female: 75.3%, male: 76%). In the use of stylistic competence, male group (21%) expresses more diversely formal features than female group (12%), while female (87%) use more informal features than male (66%). Most of participants show a limitation in the use of speaking four types of semantic formulas: expression of apology (APOL), acknowledgment of responsibility (RESP), offer of repair (REPR), and promise of forbearance (FORB). As nonnative speakers, the participants cannot conduct the semantic formula in some situations regardless of the tasks provided. The results suggest that English teachers should recognize pragmatic variations in which students feel difficulty in appropriate speaking strategies on apology. This study also contributes to teaching learners the strategies and speaking patterns in the course of various apology situations.

• 교육시설
• /
• 제11권1호
• /
• pp.5-14
• /
• 2004

Since 1990s in Korea, elementary school classrooms have been designed toward open education system in pursuit of variety of educational purpose. Also, the architectural designs of schools have been acomplished for individual school not based on the standard design code. The present paper aims to investigate the acoustic environment of existing classrooms and to compare the sound insulation capacity between the ordinary classrooms and the newly built classrooms for open education. The current acoustical situation of elementary classrooms was analyzed using field measurements and questionnaire survey. In order to this, Three elementary schools were selected which were built in 1978, 1996 and 2000 respectively. Room acoustical parameters including Reverberation time(RT), Definition(D50), Speech Intelligibility(RASTI), Transmission loss(TL) and STC were measured in a classroom in each elementary school classroom. Each measurement was undertaken with the windows and doors being open or closed. As the result, it was found that the transmission loss between rooms in open classrooms is, $5{\sim}6dB$ in average, inferior than the ordinary classrooms. The RASTI of 0.70 was measured in newly built classrooms which is better than old classrooms(0.70) and open classrooms(0.73). This was shown as same in the speech definition measurements. This results from the condition of sealing and airtightness of classrooms and floor materials. The results denote that open classrooms have poor acoustic condition in sound insulation and speech intelligibility.