Heart Sound Localization in Respiratory Sounds Based on Singular Spectrum Analysis and Frequency Features

Malihe Molaie,Mohammad Hassan Moradi 한국전자통신연구원 2015 ETRI Journal Vol.37 No.4

Heart sounds are the main obstacle in lung sound analysis. To tackle this obstacle, we propose a diagnosis algorithm that uses singular spectrum analysis (SSA) and frequency features of heart and lung sounds. In particular, we introduce a frequency coefficient that shows the frequency difference between heart and lung sounds. The proposed algorithm is applied to a synthetic mixture of heart and lung sounds. The results show that heart sounds can be extracted successfully and localizations for the first and second heart sounds are remarkably performed. An error analysis of the localization results shows that the proposed algorithm has fewer errors compared to the SSA method, which is one of the most powerful methods in the localization of heart sounds. The presented algorithm is also applied in the cases of recorded respiratory sounds from the chest walls of five healthy subjects. The efficiency of the algorithm in extracting heart sounds from the recorded breathing sounds is verified with power spectral density evaluations and listening. Most studies have used only normal respiratory sounds, whereas we additionally use abnormal breathing sounds to validate the strength of our achievements.

받침의 연음 방법 체계화와 교재 구성 방향 연구

신지현,최영환 한국초등국어교육학회 2016 한국초등국어교육 Vol.61 No.-

This study is aimed to set the stages of learnig prolonged sounds so that elementary school students can pronounce and read prolonged sounds correctly and to find a direction to the effective composition of a text book based on the systemization. With Para. 13 and 15 of ‘Korean Standard Pronunciation Rules’ examined, this study found out 2 reading elements of prolonged sounds: conditions for being a prolonged sound(kinds of final consonants of front syllables, kinds of vowels of back syllables), and pronunciation method of prolonged sounds. In this study, 4 stages of learning prolonged sounds were selected after the three reading elements of prolonged sounds were combined and changed palatalization by prolonged sounds was included in the patterns of reading prolonged sounds. Here, final consonants of ‘ㅇ’ and ‘ㅎ’ which are not associated with making a prolonged sound, were excluded in the elements. And then, a text book of reading prolonged sounds was designed into two directions: one is to reinforce the learning of prolonged sound principles and the other is to structure language material hierarchically to realize the learning of principle. In the composition of a text book, the conditions and reading method of prolonged sounds have to be proposed separately and a variety of repetitive activities has to be arranged so that learners can keep reminding and acquiring the laws of prolonged sounds. To attain the objective, the reading materials of prolonged sounds have to be arranged. In a unit of language material, words have to gradually expand to phrases; phrases to sentences; and sentences to paragraphs. Meanwhile, the basic practice of reading prolonged sounds start with words and a phrase of words. The levels of difficulty were set from ‘easy’ to ‘challenging’, considering such factors as learner’s familarity with language material, number of syllables, and the position and frequency of a prolonged sound in it. It will minimize the cogitative burden of a learner and help them concentrate on reading of prolonged sounds. This study expects this text book to be used in 3 ways for direct applicability in a short term and indirect applicability in a long term. First, it can be used as a supplementary material for a main text book when teaching students. Second, it can be utilized as a guideline to the order of a method of reading it to students. Last, it can be used a class maternal for a teacher who teaches to read prolonged sounds. 이 연구는 초등학생들이 받침의 연음을 정확하게 하고, 발음 오류를 교정할 수있도록 연음 학습 단계를 설정하고 이를 기반으로 효과적인 교재 구성 방안을 찾기 위한 것이다. 연음 교육을 위해 <표준 발음법> 13항과 15항을 중심으로 ‘연음 조건(앞 음절받침의 종류, 뒤 음절 초성의 종류), 연음 방법’이라는 두 가지 요소를 고려해야한다. 연음 조건 두 가지를 조합하고, 연음이 된 후에 발음이 변한 구개음화도 연음유형으로 포함하여 연음의 유형을 네 가지로 정리하였다. 용례의 수가 적고 사용빈도가 낮은 유형을 제외한 나머지 유형으로 연음 학습 3단계를 선정하였다. 효과적인 연음 학습 교재 구성을 위해서는 연음 원리 학습 강화와 언어 자료의위계적 배열을 고려해야 한다. 교재에는 연음의 조건과 연음 방법을 나누어 제시하는데, 이를 지속적으로 환기할 수 있는 다양한 반복 활동을 배열하여 연음 원리를 체득할 수 있도록 한다. 이를 위해서 언어 자료의 단위는 낱말과 한 어절에서어구나 문장, 문단으로 점차 확장하도록 하고, 언어 자료의 학습자 친숙성, 음절수, 연음의 위치, 연음의 횟수에 따라 발생하는 난도를 고려하여 쉬운 수준에서 어려운수준으로 배열한다. 이 연구는 학생들의 연음 지도 자료로서 교과서의 보조 자료로 활용할 수 있으며, 받아쓰기와 연계하여 언어 자료를 불러 주는 방법에 대한 지침으로 활용할수 있다. 마지막으로 교사들의 연음 교육용 자료로 활용 가능하다.

차(茶)'의 독음을 통해 본 한자음의 특징 - 한국, 일본, 베트남어를 중심으로

나민구 ( Na Mingu ),이재준 ( Lee Jaejoon ) 한국외국어대학교 중국연구소 2017 中國硏究 Vol.72 No.-

The study of this manuscript has started, inspired by precedent studies about the correlations between the presence for friction of the first sounds of words which describe '茶' all over the world and the distribution of '茶' vocabulary. It contains consideration of spreading courses of '茶', according to the presence for friction in sounds, by comparing dialect sounds of various Chinese areas with sounds of related vocabulary of many other countries. It furthermore started with the tentative conclusion that differences of spreading courses of '茶' of Korea, Japan and Vietnam may have brought about those of reading sounds of '茶' of each country, because double reading sounds of each country, which had not been dealt with by previous studies, were handed down to different regions and different times in the above countries. Also, it has traced the origin of '차'·'다'·'チャ'·'タ'·'che'·'tra', reading sounds of '茶' of Korea, Japan and Vietnam, respectively. These three countries have common methods of sound reading in 2 ways contrary to other countries and they all belong to the cultural boundary of Chinese Character. In order to compare the similarities of these reading sounds, the study pursued investigation in the way of comparing reconstructed sounds of historical phonemes corresponding to the reading sounds of the three countries with those of China in the far-off times. It also shows that the sounds reconstructed by similar pronunciation significantly correlate to each other in the historical aspect. Lastly, the study investigated the features of reading sounds, among reading sounds of '茶', used like native tongue such as '차'·'チャ'·'che' in Korea, Japan and Vietnam and typical characteristics of the sounds. The three reading sounds, '차'·' チャ'·'che', got settled in the system of corresponding vocabulary after being introduced earlier than other sounds and used like native tongue. Afterward they formed a variety of vocabulary, combined with existing native morpheme. On the contrary, the three other words, '다'·'サ'·'tra' were introduced later than the former words, were not used like native tongue, and rather used in forming vocabulary pronouncing words as they are read. In the introduction we pursued the research based on the assumption that the sounds such as '다' or 'サ' may have originated from the reconstructed dialects, but against our expectations we have concluded that the differences of reading sounds of '茶' in Korea, Japan and Vietnam didn’t originate from the dialect differences of Chinese. Nevertheless, we are happy to get inspiration, thanks to the research, about relationships between the introduction of Chinese character sounds and the formation of the Chinese character cultural area in the East Asia.

東部慶南方言 漢字音 硏究

신기상 국어학회 2004 국어학 Vol.0 No.44

국어 한자음의 음운체계와 고유어의 음운체계는 일치하지 않는데, 한자음 음운체계는 고유어의 음운체계에 포함된다. 이점은 표준어에서도 그러하고, 동부경남방언에서도 그러하다. 동부경남방언의 음운체계는 표준어의 음운체계보다 단조로운데, 이 방언 한자음의 음운체계도 표준어의 것에 비하여 단조롭다. 그런 까닭으로 동부경남방언 한자음 수가 표준어 한자음 수에 비해 적다. 방언은 방언 나름의 음운변화의 특징적 경향이 강하여 그것이 쌓여 방언의 특징으로 나타난다. 방언 한자어의 음도 그에 준한다. 또, 한자어 개개의 변화는 대체로 한자음 음운체계와 유관하지만 한자어 중에는 그 한자어 독자적인 방향으로 형태변화를 겪기도 한다. 이 연구는 위와 같은 현상에 주목하여 이 방언의 한자음 음운체계를 초성, 중성, 종성, 고저장단별로 기술하고 이 방언의 한자음이 표준어의 어떠한 한자음과 변별되고 되지 않으며, 또 이 방언 한자음끼리 어떤 음이 변별되고 되지 않는지를 기술함으로써 이 방언 한자음의 수를 산출하였다. 그리고, 동부경남방언 한자어 중에서 개별적 방향으로 음이 변한 것들을 찾아내어 같거나 유사한 방향으로 변한 한자어를 묶어 변화의 경향을 기술하여 결과적으로 동부경남방언 한자음 전체를 개관하였다. This paper has three topics for Korean sounds of Chinese characters in Eastern Kyungsangnam province dialect. First, it describes the phonological structure of the Korean sounds of Chinese characters of this dialect. The structure has thirteen initial sounds, six simple vowels and nine diphthongs in the middle sounds, and six final sounds. The dialect has four suprasegmental phonemes : high-long, high-short, low-long, and low-short. Second, it shows the distinguishment of the Korean sounds of Chinese characters in the dialect. This dialect has less sounds of Chinese characters than the standard language does, so some of the sounds are not distinguished from each other in the dialect; about 90 out of 460 standard Korean sounds of Chinese characters are not distinguished. Third, it describes the change of the sounds of the words written in Chinese characters in the dialect. Generally, within this dialect, a Korean sound of a Chinese character has changed into another one. However, we can also find that some of the sounds have been turned into several different ones in some words written in Chinese characters.

현대 베트남어 한자음 종성 [k/c]와 [ŋ/□]에 관한 연구

이병운 ( Byung Woon Lee ) 부산대학교 과학교육연구소 2015 교사교육연구 Vol.54 No.4

이 논문의 목적은 베트남어의 한자음 종성(coda)에서 중국어의 중고음[Middle Chinese Sounds]과 한국어의 한자음과 다르게 연구개종성(Velar Consonants) [k/ŋ]이 경구개종성(Palatal Consonants) [c/μ]으로 분화되는 음운적 조건을 찾는 것이다. 현대 베트남 한자음의 연구개종성이 경구개종성으로 분화되는 음운적 조건을 찾기 위해서 중국어 중고음과 중세한국어 한자음을 비교·대조하였다. 그 결과는 다음과 같다. 첫째, 경구개종성(Palatal Consonants) [c/μ]은 현대 베트남 한자음에서 전설모음 [i/e] 뒤에서 나타난다. 둘째, 현대 베트남 한자의 경구개종성은 현대 베트남 한자음 [a] 모음 뒤에는 연구개종성 [k/ŋ]과 경구개종성 [c/□]이 함께 나타나지만, 베트남어의 경구개종성은 중국어 중고음의 음운적 조건이 [æ, □, i□, iε]일 때, 나타나고 연구개종성 [k/ŋ]은 중국어 중고음이 후설모음 [α,□]일 때, 나타난다. 셋째, 베트남어 한자음 경구개종성은 중세한국어 한자음에서 [j] 반모음이 핵모음에 선행하거나 후행할 때 나타난다. 그리하여 베트남어 한자음 경구개종성은 현대 한국어 한자음에서 애[ε] 또는 [j□]일 때 가장 빈도가 높게 나타난다. 그러므로 현대 한국어에서 [ε]는 중세한국어에서 하향이중모음 [aj], [Λj]에서 변천한 것이므로 베트남어의 경구개종성은 [j]가 잠재되어 있다고 할 수 있다. This study aimed to investigate the phonological conditions of the palatal consonants [c/η] which derive from the velar consonants [k/ŋ] in the coda position in Vietnamese-Chinese sounds. These conditions were expected to be different from those of both Middle Chinese and Korean-Chinese sounds. For this study, the Middle Chinese and Korean-Chinese sounds were compared and contrasted. The research results showed the following: First, in Vietnamese-Chinese sounds, the palatal consonants [c/μ] appeared in the back position of front vowels [i/e]. Second, the coda of Vietnamese-Chinese appeared with both [c/μ] (palatal) and [k/η] (velar) after the vowel [a], while the coda of Vietnamese appeared with [æ, □, i□, Iε] (Middle Chinese sounds) as well as Middle Chinese back vowels [α, □]. Third, the coda of Vietnamese-Chinese palatal sounds appeared when the semivowel [j] was pronounced either after or before the core vowels in Middle Korean-Chinese sounds. Therefore, the most frequent Vietnamese-Chinese sounds in the coda position were [ε] or [j□] similar to the modern Korean-Chinese sounds. Consequently the [ε] in modern Korean changed from the Middle Korean sounds [aj], [Λj], which may help to infer that the coda of Vietnamese contains the semivowel [j].

東國正韻 喉音의 ‘청⋅탁’ 연구 -초분절 자질의 관점에서-

김서영,문성호,김민경,안영실,이길경 중국학연구회 2021 중국학연구 Vol.- No.96

After Karlgren’s research was published, the phoneme quality of the qīngyīn(淸/clear) and zhuóyīn(濁/unclear) sounds of the Chinese ancient sounds began to be analyzed as a confrontation of voiceless sounds and voiced sounds. Having said that however, the qīngyīn(淸/clear) and zhuóyīn (濁/unclear) sounds of the Chinese ancient sounds before Karlgren’s research were not classified into conflicting sounds but described as qīng(輕) & zhòng(重) (heavy and light), gāo(高) & xià(下) (high and low) and finally as yīn(陰) & yáng(陽). The very first person to transcribe the qīngyīn(淸/clear) and zhuóyīn(濁/unclear) sounds of the Chinese ancient sounds into sound value characters was none other than King Sejong and if we were to look at Sejong’s Hunminjeongeum Haeryebon edition (訓民正音解例本) as well as his personal music scores, it is believed that King Sejong accepted the qīngyīn(淸/clear) and zhuóyīn(濁/unclear) sounds as suprasegmental qualities. In the Hunmongjahwae(訓蒙字會) written by Choi Sejin, unlike the Donggukjungwoon(東國正韻), the jeoncheong(全淸:voiceless consonant) ‘ㆆ’ and the jeontakja(全濁字:entering sounds of ‘ㄲ’, ‘ㄸ’, ‘ㅃ’, ‘ㅆ’, ‘ㅉ’, ‘ㆅ’ within the Hunminjeongeum) for the purpose of transcribing the gakjabyeongseo(各自並書:alignment of two consonants side by side) were not used so it is possible to carry out an objective comparison of the phoneme quality of the qīngyīn(淸/clear) and zhuóyīn(濁/unclear) sounds. In this paper, King Sejong, Choi Sejin and Kalgren’s transcription of the qīngyīn(淸/clear) and zhuóyīn(濁/unclear) sounds especially focusing on the hóuyīn(喉音/guttural sounds) sounds were carried out. As a result, an analysis was made where it was shown that individual values were clearly present in transcribing the qualities of the qīngyīn(淸/clear) and zhuóyīn(濁/unclear) sounds of the Chinese ancient sounds. It appears that when King Sejong was transcribing the qīngyīn(淸/clear) and zhuóyīn(濁/unclear) sounds, he was basing them upon the real Chinese sounds and it is actually possible to confirm that the qīngyīn(淸/clear) and zhuóyīn(濁/unclear) sounds were demonstrated as suprasegmental qualities as evidenced by the Chinese southern dialects which still exists today. The reason as to why King Sejong strictly adhered to the qīngyīn(淸/clear) and zhuóyīn(濁/unclear) sound transcription, which does not reveal themselves in the Korean sounds of Chinese characters, can be found in the musical perspective of Confucius (Rújiā:儒家) and Zhōuyì(周易) philosophy. As to why Choi Sejin didn’t transcribe the qīngyīn(淸/clear) and zhuóyīn(濁/unclear) sounds can be found in the fact that he had the sole purpose of phonetically notating the Chinese characters into Korean sounds. The reason why Kalgren viewed the qīngyīn(淸/clear) and zhuóyīn(濁/unclear) sounds of the Chinese ancient sounds as a confrontation between voiceless and voiced sounds appears to be an influence from the freshly established International Phonetic Alphabet system during that time. As it can be seen above until now, it is possible to surmise that there are various perspectives at work about the transcription of the qīngyīn(淸/clear) and zhuóyīn(濁/unclear) sounds. However, it was concluded that the qualities of the qīngyīn(淸/clear) and zhuóyīn(濁/unclear) sounds which King Sejong analyzed have been ultimately deemed to the most appropriate for the phonetic values of Chinese ancient sounds. 청⋅탁은 樂記에서부터 사용되었던 음악 용어였고, 전통적으로 음운학에서도 淸은 고음, 濁은 저음을 가리키는 것이 통설로 되어있었다고 한다. 淸代의 江永이 音學辨微(1759)에서 청⋅탁의 개념을 성모에 국한해서 쓰기 시작했고, 칼그렌(Klas Bernhard Johannes Karlgren, 1889∼1978)의연구 이후 청⋅탁은 무성음과 유성음의 분절 자질로 분석된다. 하지만 현재 한⋅중⋅일⋅베트남 어음을 살펴봤을 때, 후음에 해당하는 제로 음가와성문 마찰음에서 무성음과 유성음의 대립은 변별 자질로 활용되지 않는다. 이는 칼그렌이 주장의 근거로 삼았던 吳방언에서도 마찬가지이다. 趙元任 (1892∼1982)의 五度表記法1) 창안으로 초분절 자질에 대한 객관적 접근이가능해졌는데, 漢語 中古音의 청⋅탁과 중국 남방 방언의 음조와 양조가대체로 대응 관계를 이루고 있음을 알 수 있다. 청⋅탁과 음조와 양조가대응하는 것의 의미하는 바는 크지만 이와 관련한 연구는 많이 이뤄지지않았다. 東國正韻 후음 ‘ㆆ, ㅎ, ㆅ, ㅇ’은 조음 위치, 조음 방식, 청⋅탁표기를 위한 것이며, 여기서 청⋅탁은 초분절 자질의 관점에서 이해할 경우 납득할 만한 설명이 가능함을 訓蒙字會 및 칼그렌의 청⋅탁 재구음과 비교하여 고찰했다.

Proposal of a new method for learning of diesel generator sounds and detecting abnormal sounds using an unsupervised deep learning algorithm

조훤기,김송현,김창락 한국원자력학회 2023 Nuclear Engineering and Technology Vol.55 No.2

This study is to find a method to learn engine sound after the start-up of a diesel generator installed in nuclear power plant with an unsupervised deep learning algorithm (CNN autoencoder) and a new method to predict the failure of a diesel generator using it. In order to learn the sound of a diesel generator with a deep learning algorithm, sound data recorded before and after the start-up of two diesel generators was used. The sound data of 20 min and 2 h were cut into 7 s, and the split sound was converted into a spectrogram image. 1200 and 7200 spectrogram images were created from sound data of 20 min and 2 h, respectively. Using two different deep learning algorithms (CNN autoencoder and binary classification), it was investigated whether the diesel generator post-start sounds were learned as normal. It was possible to accurately determine the post-start sounds as normal and the pre-start sounds as abnormal. It was also confirmed that the deep learning algorithm could detect the virtual abnormal sounds created by mixing the unusual sounds with the post-start sounds. This study showed that the unsupervised anomaly detection algorithm has a good accuracy increased about 3% with comparing to the binary classification algorithm.

도(鼗)ㆍ고(鼓)의 덕음적(德音的) 성격의 고찰

이창신 청주대학교 학술연구소 2010 淸大學術論集 Vol.15 No.-

The Book of Rites indicates the six musical instruments - Do(鼗), Go(鼓,) Gang(椌), Ga(楬)l, Hun(壎), Ji(篪) - as those of virtuous sounds, and says that social order can be established through these instruments. Such utility of these instruments comes from their virtuous sounds. Yao(堯) King's and Hoahgje's(黃帝) music, Daejang(大章) and Hamji(咸池), mean virtue that shines benevolently worldwide. Likewise, in ancient China, kings strove to shine their virtue through music. Virtue comes from personality, music shines virtue, and iron(金), stone(石), thread(絲) and bamboo(竹) are the tools of music. Thus, the purpose of making instruments of virtuous sounds is to start the beginning of promoting virtue. Do, Go, Gang, and Gal sounds genuine sounds, and were named as instruments of virtuous sounds(德音). This is the expression of the thought of ceremony and music(禮樂思想) that regards genuine sounds as valuable, and that performs these instruments of such sounds to lead the people into the right way. With the performance of ceremony and music(禮樂), the heaven and the earth are brightened, promoting the spirit of reconciliation(和) and the harmony of Yim and Yang, and creating all things in the world. In this way, saints, through ceremony and music(禮樂), participate in the heaven and the earth harmoniously fostering all things in the world. As such, humanity being together with the heaven and the earth is the thought of the three things, namely, the heaven, the earth, and man(三才思想). The characteristics of virtuous sounds of Do, Go, Gang, and Gal are outlined below. Do and Go are composed of the three kinds of instruments, namely, Noego(雷鼓) and Noedo(雷鼗), Yeonggo(靈鼓) and Yeongdo(靈鼗), and Nogo(路鼓) and Nodo(路鼗), each of which symbolizes heavenly sounds(天聲), the earth's virtue(地德), humans' way(人道), expressing the thought of three things(三才思想). They are also related to the Book of Changes'(『주역』) six stroke bars(6획괘), become Yim and Yang, and correspond to changes of things(易道) and the principle of lives(性命).

미얀마인 한국어 학습자들의 종성 발음 오류 분석 - 비음과 유음을 중심으로

퓨퓨아웅 ( Phyu Phyu Aung ),노채환 ( Roh Chaehwan ) 한국외국어대학교 동남아연구소 2021 東南亞硏究 Vol.31 No.1

This paper aims to analyze the errors made by Myanmar learners in producing Korean final consonants and to reveal the cause of the errors. There are seven sounds /k, n, t, l, m, p, ŋ/ that can be pronounced in the Korean final consonant position. However the only sound that can be pronounced in the Myanmar final consonant position is the glottal stop /ʔ/. As a result Myanmar learners have a hard time pronouncing the Korean final consonants. Especially Myanmar learners make many errors in the pronunciation of final nasal and liquid sounds. Myanmar learners make errors by replacing them with other sounds because they do not clearly recognize the final nasal sounds /m, n, ŋ/ in Korean. Because the exact articulation position of the final Korean nasal sounds is not well known, it is difficult to pronounce it, so the nasal vowel in the native language is applied to the pronunciation of the final nasal sounds in Korean. In addition, the final nasals /m, n, ŋ/ are confused with consonants in different positions to make pronunciation errors. These errors can be interpreted as being because learners are not familiar with the final nasal sound and their discrimination power for the three phonemes is poor. And, Myanmar learners make an error in pronunciation by dropping out when pronouncing the final sound /l/ and replacing it with a different sound[ɬ]. This can be interpreted as committing this error because Myanmar learners do not recognize the final liquid sound /l/ well and feel it difficult to pronounce it because there is no liquid sound /l/ in the final consonant sound of Myanmar.

Onomatopoeia and Universality of Sound Symbolism:A Phonological Comparison between Animal Sounds in Korean and English

김우정 대한영어영문학회 2024 영어영문학연구 Vol.50 No.2

In contrast to the mainstream arbitrariness of sound-meaning relationships in modern linguistics, sound symbolism manifests a direct resemblance between sign and referent. Cross-linguistic observations have indicated that onomatopoeia, the direct representations of natural sounds in an iconic way, are spread worldwide and may provide a scaffolding mechanism for children in the early stages of language acquisition. Focusing on the universality and language-specific idiosyncrasy of sound symbolism in onomatopoeic forms, this study compares and contrasts the phonological features of animal sounds in Korean and English. A comparative analysis based on articulatory criteria is performed to describe the features of consonants and vowels in onomatopoeic words. The results suggest that even though the onomatopoeic animal sounds of Korean and English manifest themselves differently according to their phonotactics and phonological systems, these highly dissimilar languages with distinct phoneme inventories employ conventionalized forms that bear certain phonetic similarities. Both demonstrate similar patterns in the distributions of segments and share analogous representations for animal sounds to some extent. Along with cross-lingual phonological alikeness, sound symbolic correspondences between sound and size are observed in the animal sounds of both languages. Such findings support the universality of iconic links and sound symbolism in language. (Korea University)