• 음성과학
• /
• 제11권2호
• /
• pp.151-164
• /
• 2004

This study examined the use of intonation in children with single word sentences and investigated the communicative functions of pitch range and pitch direction. Two children aged 13months were observed in interaction with their mothers for 10 months. The vocalizations were coded separately for communicative function and for prosodic feature. Results show that level tones are used most frequently, and pitch range is higher for request than declaration or answer and lower for answer than request or declaration. And trends in prosodic contours were observed in request, declaration, and answer respectively. For one child, rising tones were frequently associated with request whereas rising-falling tone with declaration. For the other child, rising-level tones were more frequently associated with request whereas falling-level appeared more often in conjunction with declaration. These trends appeared more distinct in proportion as they grow in months. This result indicate that the way to express communicative functions transfer gradually from differentiating pitch range to diversify pitch direction.

• 대한음성학회지:말소리
• /
• 제21_24호
• /
• pp.43-89
• /
• 1992

In this thesis, the prominent aspect of word accent in standard Korean is studied by auditory test and acoustic analysis experiment. The definition of 'accent' is, following Hoyoung Lee's discussion(1990), to be described as 'the means whereby a focused part of an utterance is made to stand out in order to concentrate the hearer's attention on it.' That is to say, the ten of 'accent' may be described in terms of phonological phenomenon and the accented syllable can be phonetically prominent as the result of those phonological process. Prosodic features may have different characteristics in different languages whether they contain linguistically important functions or not. Thus the characteristics of word accent in standard Korean will be determined as the content and trait of prosodic features. Following this viewpoint, present study looked over prosodic features which may effect the characteristics of word accent in standard Korean, through systematic experimental procedure. And the result of this experiment has been verified by statistical method, the T-test, for the purpose of identifying the relatedness among prosodic features(parameters). This thesis, therefore, aimed to investigate the intrinsic acoustic and physical qualities of the word accent in standard Korean. Nonsense words composed by 'mal' and 'ma' which can be divided into 'heavy syllable' and 'light syllable' quoted from Hyman(1975) have been classified into 28 types with respect to syllable numbers(2 syl., 3 sy1., 4 syl.) and these words have become the target of auditory test and acoustic experiment. As the result of those experimental Procedures, the word accent in standard Korean may be said that it has a tendency of fixing first two syllables regardless of syllable numbers. The syllable types of HH, HL, LL in the first two syllables may be prominent at first syllable and the type of H may be at second syllable. Various prosodic features(parameters) including duration, intensity, and Fo(purely phonetic terms) were also strengthened in those positions. The result of this experiment can be cleared up like these : 1. The most important feature is proved as 'duration', the feature of intensity resulted in more subsidiary one than the feature of duration. 2. Fo( fundamental frequency) could be observed as having some coherent contour through almost all syllable types(99 ％), that is, in 2 syllable types, it had rising contour, in 2 syllable types, rising-falling contour, and in 4 syllable types, it contained rising-falling-rising contour. The result of auditory test was different with those contour forms of all Fo surveyed. With respect to these results, the discuss for Fo is determined' to be excluded comparing other features. 3. Finally, this thesis resulted in a decision that the word accent in standard Korean may has fixed(somewhat weaker) accent, especially fixed at first two syllables in almost all words. 4. Various kinds of syllable types related with 2,3,4 syllables, therefore, can be reclassified into 4 types of HH, HL, LH, LL following the concept of accent fixing placement(i.e. first two syllables). In these 4 types, the types of HH, HL, LL were prominent at the position of the first syllable , and the type of LH was prominent at the second syllable otherwise.

• 대한음성학회지:말소리
• /
• 제35_36호
• /
• pp.1-12
• /
• 1998

This study aims to investigate the prosodic characteristics of Korean discourse sentences, especially focusing the initial and final part of a sentence. 50 disourse sentences were read in two different styles; one, sentence by sentence, the other, continuous of all 50's. First, we tried to get two kinds of ratios from the acoustic results: first, ratio of the final syllable to the initial syllable in first word in a sentence; second, ratio of the final syllable to the initial syllable in last word in a sentence. We, then, calculated statistical values of the ratios including mean, standard deviation, minimum, maximum, and p-values in t-test. With respect to duration, there were little difference between two different styles. If any, we could see tiny unharmonious durational aspect in the initial of continuous reading. More concisely, there could be observed some deviation from standard. In case of F0, there was prominent statistical difference between ratios of last words in two styles. This difference might play a role as a prosodic feature. Energy seems to show similar pattern with that of F0. The results showed that final syllable in last word was pronounced with about 85 % of initial syllable in the same context and the last words in continuous speech were strongly articulated compared with those of sentence by sentence reading.

• 음성과학
• /
• 제10권2호
• /
• pp.7-25
• /
• 2003

Predicting the right prosodic elements is a key factor in improving the quality of synthesized speech. Prosodic elements include break, pitch, duration and loudness. Pitch, which is realized by Fundamental Frequency (F0), is the most important element relating to the quality of the synthesized speech. However, the previous method for predicting the F0 appears to reveal some problems. If voiced and unvoiced sounds are not correctly classified, it results in wrong prediction of pitch, wrong unit of triphone in synthesizing the voiced and unvoiced sounds, and the sound of click or vibration. This kind of feature is usual in the case of the transformation from the voiced sound to the unvoiced sound or from the unvoiced sound to the voiced sound. Such problem is not resolved by the method of grammar, and it much influences the synthesized sound. Therefore, to steadily acquire the correct value of pitch, in this paper we propose a new model for predicting and classifying the voiced and unvoiced sounds using the CART tool.

• 한국지능시스템학회:학술대회논문집
• /
• 한국퍼지및지능시스템학회 2004년도 춘계학술대회 학술발표 논문집 제14권 제1호
• /
• pp.470-473
• /
• 2004

In this paper, the comparison of speech feature parameters for emotion recognition is studied for emotion recognition using speech signal. For this purpose, a corpus of emotional speech data recorded and classified according to the emotion using the subjective evaluation were used to make statical feature vectors such as average, standard deviation and maximum value of pitch and energy. MFCC parameters and their derivatives with or without cepstral mean subfraction are also used to evaluate the performance of the conventional pattern matching algorithms. Pitch and energy Parameters were used as a Prosodic information and MFCC Parameters were used as phonetic information. In this paper, In the Experiments, the vector quantization based emotion recognition system is used for speaker and context independent emotion recognition. Experimental results showed that vector quantization based emotion recognizer using MFCC parameters showed better performance than that using the Pitch and energy parameters. The vector quantization based emotion recognizer achieved recognition rates of 73.3％ for the speaker and context independent classification.

• Journal of Information Processing Systems
• /
• 제15권1호
• /
• pp.91-99
• /
• 2019

An innovative tone modeling framework based on deep neural networks in tone recognition was proposed in this paper. In the framework, both the prosodic features and the articulatory features were firstly extracted as the raw input data. Then, a 5-layer-deep deep belief network was presented to obtain high-level tone features. Finally, support vector machine was trained to recognize tones. The 863-data corpus had been applied in experiments, and the results show that the proposed method helped improve the recognition accuracy significantly for all tone patterns. Meanwhile, the average tone recognition rate reached 83.03%, which is 8.61% higher than that of the original method.

• 대한전자공학회논문지SP
• /
• 제40권5호
• /
• pp.371-377
• /
• 2003

본 논문에서 음성신호를 사용하여 인간의 감정를 인식하기 위한 특징 파라메터 비교에 관하여 연구하였다. 이를 위하여 여러 가지 감정 상태에 따라 분류된 한국어 음성 데이터 베이스를 이용하여 얻어진 음성 신호의 피치와 에너지의 평균, 표준편차와 최대 값 등 통계적인 정보 나타내는 파라메터와 음소의 특성을 나타내는 MFCC 파라메터가 사용되었다. 파라메터들의 성능을 평가하기 위하여 문장 및 화자 독립 감정 인식 시스템을 구현하여 인식 실험을 수행하였다. 성능 평가를 위한 실험에서는 운율적 특징으로 피치와 에너지와 각각의 미분 값을 사용하였고, 음소의 특성을 나타내는 특징으로 MFCC와 그 미분 값을 사용하였다. 벡터 양자화 방법을 사용한 화자 및 문장 독립 인식 시스템을 사용한 실험 결과에서 MFCC와 델타 MFCC를 사용한 경우가 피치와 에너지를 사용한 방법보다 우수한 성능을 나타내었다.

• 대한음성학회지:말소리
• /
• 제40호
• /
• pp.51-63
• /
• 2000

There are sentences where sentence stress is imposed on a specific word. These sentences are called 'focused sentences'. The purpose of this paper is to investigate the variation of pitch, duration, amplitude in focused words. It is noted that pitch of a focused word is higher than that of unfocused words irrespective of the accentual pattern, and that contour tones such as HL or LH are realized longer when these tones appear in focused words. Not only the noun but also the following particle like '-boda' is higher when these words are in focus. Hence pitch is proved to be the most salient prosodic feature of the focused sentence.

• 한국컴퓨터정보학회논문지
• /
• 제17권9호
• /
• pp.75-84
• /
• 2012

HMM 기반 음성합성시스템은 성능향상을 위해 일반적으로 대용량 음성 DB로부터 생성된 문맥의존 tri-phone을 이용한다. 그리고 대용량 DB의 경량화를 위해서 문맥의존정보를 이용하여 결정트리 방식으로 발화특성이 유사한 문맥의존음소들을 군집화한다. 군집화에 사용하는 문맥의존정보는 음소열 뿐만 아니라 운율정보도 포함하는데 이는 합성음의 자연성이 끊어 읽기, 억양패턴, 음의 장단과 같은 운율에 의해 크게 좌우되기 때문이다. 그러나 복잡한 운율정보를 사용할 경우 훈련과정에 포함되지 않은 문맥의존음소는 하나의 대표값으로 평활화되며 이로 인해 합성음의 자연성이 크게 저하된다. 본 논문에서는 합성음의 자연성을 향상시키기 위해 복잡한 운율정보 대신 억양 변화를 상승, 평탄, 하강으로 구분함으로써 운율정보표현을 간소화시킨 운율경계정보를 포함하는 문맥의존정보에 대한 문맥질의, 그리고 해당 질의의 패턴을 정의하는 방법을 제안하였다. 본 논문에서 제안하는 세 가지 운율경계정보를 포함한 문맥의존정보를 이용하여 합성음을 생성하고 MOS평가를 수행한 결과 운율경계정보를 이용한 HMM기반 한국어 TTS 합성음의 자연성이 향상됨을 확인하였다.

• 한국음향학회지
• /
• 제20권3호
• /
• pp.15-23
• /
• 2001

본 논문에서는 임의의 사람이 발성한 음성을 마치 다른 사람이 발성한 것처럼 들리도록 하는 음성변환 기술에 대하여 설명하고, 화자간의 성도 특성과 여기신호 특성 파라미터 변환을 독립적으로 수행하기 위한 변환방법을 실험한다. 성도 특성 파라미터 변환은 입력되는 음성신호에서 LPC (Linear Predictive Cofficient)켑스트럼을 추출하여 선형다변회귀모델에 적용하여 수행하고, 여기신호 특성 파라미터 변환은 잔차신호를 추출하여 LP-PSOLA (Linear Predictive-Pitch Synchronous Overlap and Add) 합성방식을 이용한 화자간의 평균 피치주기 변환으로 수행된다. 실험결과는 선형다변회귀모델과 LP-PSOLA 합성방식을 이용하여 변환된 음성이 대상화자의 음성에 유사함을 보여준다