• Annual Conference on Human and Language Technology
• /
• 2002.10e
• /
• pp.69-76
• /
• 2002

최근 음성 합성 과정에서 화자의 의도를 가장 많이 반영하는 언어 정보인 문맥 정보를 사용하려는 시도가 이루어지고 있으나 문맥 정보를 적은 비중으로 사용하기 때문에 자연성 향상에 큰 도움을 주지 못하고 있다. 본 연구에서는 구문 정보, 의미 정보를 억양 생성 과정에 이용함과 동시에 문맥 정보와 음성 정보와의 관계를 음성 데이터를 바탕으로 분석하여 다양한 문맥 정보를 음성 합성 과정에 반영하는 방법을 제안한다. 또한 한국어에서 나타나는 다양한 억양 곡선 유형을 형태소를 이용하여 의다 효율적으로 처리할 수 있는 방법을 제안하여 자연스러운 억양 생성 시스템을 구현하고 시스템의 결과를 음소 단위 억양 생성기와 VoiceXML을 이용하여 적용시켜보고 결과를 논의한다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2000.10b
• /
• pp.377-379
• /
• 2000

대화체 음성의 인식을 위해서는 음성 파형에 관한 음향학적인 연구뿐만 아니라 인식하려는 언어자체에 대한 언어학적인 연구를 필요로 한다. 본 논문에서는 숫자음의 언어학적인 요소를 고려하고, 포만트 주파수를 숫자음 검출과 숫자음 인식에 적용하는 방식을 제안한다. 시스템의 입력은 특정 질의에 대한 응답으로 대화체 문장이며, 끝점 추출 기술을 이용하여 고립단어로 분류한 후, 숫자음만을 검출해 내고, 검출된 숫자음을 인식하기 위해 포만트 주파수를 이용한다. 한국어 연결 숫자음 인식은 한국어 숫자음이 단음절로 구성된다는 점과 발음상의 조음효과 등으로 한계를 가지고 있다. 본 논문에서는 숫자음과 발성에 필요한 음소들을 추출하고, 숫자들을 모음에 따라 6개의 그룹으로 분류하여 인식의 범위를 좁히고, 포만트 주파수 정보와 음소 HMM 모델에 의한 두 단계에 걸친 인식을 수행함으로써 연결 숫자음 인식에 대한 성능을 향상시킨다.

• Proceedings of the Korean Information Science Society Conference
• /
• 2005.07b
• /
• pp.856-858
• /
• 2005

음성인식 시스템에서 입력된 음성 데이터에 대해 비인식 대상을 거부하는 기능은 신뢰도 보장 측면에 있어서 상당히 중요하며, 신뢰도를 높이기 위해서는 단순한 인식기능 외에 부적절한 입력 패턴의 거부 기능이 필요하다. 본 논문에서는 이러한 신뢰성 문제를 해결하기 위하여 음소기반 인식 네트워크에서 필러 모델 방법과 단어 검출률 방법을 사용하여 실험하였고, 문장의 단어 수에 따른 두 방법의 문장 거부 성능을 FAR과 FRR의 평균을 최소화 하는 값을 각각 구함으로써 비교${\cdot}$분석 하였다. 그 결과 필러모델 방법이 좀 더 나은 거부 성능을 보였고, 단어 검출률을 이용하는 방법이 인식 네트워크를 전부 거치지 않아도 되므로 실행속도와 메모리 절약에서 효과적이었다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2002.04b
• /
• pp.1073-1076
• /
• 2002

정보 검색시 질의어가 외래어일 경우에 검색의 재현율은 급격하게 감소된다. 이는 외래어에서 나오는 음역어의 불일치와 영어외래어, 한글음역어는 같은 색인으로 처리가 되지 않기 때문이다. 따라서 본 논문에서는 영어외래어를 한글음역어로 자동 변환시키고, 자동 변환시에는 영어음소에 해당하는 발음값을 한글음소로 모두 변환시킨 다음 조합하였다. 조합된 음역어들은 다시 동치부류 DB에 저장되어, 질의어 검색시 검색어가 동치부류 색인어로 확장되어 검색된다. 제안한 검색시스템의 성능을 평가하기 위해서 재현율을 측정하였다.

• Korean Journal of Cognitive Science
• /
• v.15 no.1
• /
• pp.1-11
• /
• 2004

The aim of this study is to reveal the order and the age of mastery of phonemic contrast in vowel sounds of Korean. For this purpose, we made an observation of the correspondences between the sounds produced by children of 12-35 months and the target sounds produced by adults. The provisional order and the age of contrast acquisition shown from the results of this study are as follows. First, the differential production of vowels by the feature relating to the body of the tongue precedes the differential production of vowels by the feature relating to the lip rounding. Second, as for the differential production of vowels by the feature relating to the body of the tongue, the contrast between the low vowels and the others is accomplished first, and the contrast between the high and low vowels and the contrast between the front and the back vowels are established around the age of 24 months. Third, as for the differential production of vowels by the feature relating to the lip rounding, the contrast between the rounded and the unrounded vowel is not accomplished until 36 months. Finally, we observed, prior to the completion of the differential production of phonemes, children use a specific phoneme excessively. This passing phrase could be interpreted as a result of over-application of a distinctive feature in the course of acquisition of it.

• The Journal of the Acoustical Society of Korea
• /
• v.14 no.3
• /
• pp.49-60
• /
• 1995

This paper is a study on the design and implementation of the Korean Text-to-Speech system which is used for a small and simple system. In this paper, a parameter synthesis method is chosen for speech syntheiss method, we use PARCOR(PARtial autoCORrelation) coefficient which is one of the LPC analysis. And we use phoneme for synthesis unit which is the basic unit for speech synthesis. We use PARCOR, pitch, amplitude as synthesis parameter of voice, we use residual signal, PARCOR coefficients as synthesis parameter of unvoice. In this paper, we could obtain the 60% intelligibility by using the residual signal as excitation signal of unvoiced sound. The result of synthesis experiment, synthesis of a word unit is available. The controlling of phoneme duration is necessary for synthesizing of a sentence unit. For setting up the synthesis system, PC 486, a 70[Hz]-4.5[KHz] band pass filter for speech input/output, amplifier, and TMS320C30 DSP board was used.

• Proceedings of the Acoustical Society of Korea Conference
• /
• spring
• /
• pp.337-340
• /
• 2004

The exactly reproduced prosody of a TTS system is one of the key factors that affect the naturalness of synthesized speech. In general, rules about prosody had been gathered either from linguistic knowledge or by analyzing the prosodic information from natural speech. But these could not be perfect and some of them could be incorrect. So we proposed artificial neural network(ANN)s that can be trained to team the prosody of natural speech and generate it. In learning phase, let ANNs learn the pitch and energy contour of center phoneme by applying a string of phonemes in a sentence to ANNs and comparing the output pattern with target pattern and making adjustment in weighting values to get the least mean square error between them. In test phase, the estimation rates were computed. We saw that ANNs could generate the prosody of a sentence.

• Journal of Korea Multimedia Society
• /
• v.9 no.1
• /
• pp.23-32
• /
• 2006

This paper describes the optimization of a language model and an acoustic model to improve speech recognition using Korean unit digits. Since the model is composed of a finite state network (FSN) with a disyllable, recognition errors of the language model were reduced by analyzing the grammatical features of Korean unit digits. Acoustic models utilize a demisyllable pair to decrease recognition errors caused by inaccurate division of a phone or monosyllable due to short pronunciation time and articulation. We have used the K-means clustering algorithm with the transformed successive state splitting in the feature level for the efficient modelling of feature of the recognition unit. As a result of experiments, 10.5% recognition rate is raised in the case of the proposed language model. The demi-syllable fair with an acoustic model increased 12.5% recognition rate and 1.5% recognition rate is improved in transformed successive state splitting.

• The Journal of the Acoustical Society of Korea
• /
• v.22 no.6
• /
• pp.434-442
• /
• 2003

Software internationalization, the process of making software easier to localize for specific languages, has deep implications when applied to speech technology, where the goal of the task lies in the very essence of the particular language. A greatdeal of work and fine-tuning has gone into language processing software based on ASCII or a single language, say English, thus making a port to different languages difficult. The inherent identity of a language manifests itself in its lexicon, where its character set, phoneme set, pronunciation rules are revealed. We propose a decomposition of the lexicon building process, into four discrete and sequential steps. For preprocessing to build a lexical model, we translate from specific language code to unicode. (step 1) Transliterating code points from Unicode. (step 2) Phonetically standardizing rules. (step 3) Implementing grapheme to phoneme rules. (step 4) Implementing phonological processes.

• Journal of Digital Contents Society
• /
• v.4 no.2
• /
• pp.233-242
• /
• 2003

A TTS(Text-To-Speech) System is a computer-based system that should be able to read any text aloud. To output a natural voice, we need a general knowledge of language, a lot of time, and effort. Furthermore, the sound pattern of english has a variable pattern, which consists of phonemic and morphological analysis. It is very difficult to maintain consistency of pattern. To handle these problems, we present a system based on phonemic analysis for vowel and consonant. By analyzing phonological variations frequently found in spoken english, we have derived about phonemic contexts that would trigger the multilevel application of the corresponding phonological process, which consists of phonemic and allophonic rules. In conclusion, we have a rule data which consists of phoneme, and a engine which economize in system. The proposed system can use not only communication system, but also utilize office automation and so on.