• The Journal of the Acoustical Society of Korea
• /
• v.28 no.5
• /
• pp.471-476
• /
• 2009

In this paper, we apply a discriminative weight training to a support vector machine (SVM) based speech/music classification for the selectable mode vocoder (SMV) of 3GPP2. In our approach, the speech/music decision rule is expressed as the SVM discriminant function by incorporating optimally weighted features of the SMV based on a minimum classification error (MCE) method which is different from the previous work in that different weights are assigned to each the feature of SMV. The performance of the proposed approach is evaluated under various conditions and yields better results compared with the conventional scheme in the SVM.

• Journal of The Korean Association of Information Education
• /
• v.8 no.3
• /
• pp.405-414
• /
• 2004

Even if Internet has been used very much, there is no method of music aptitude test for children. We develop the music aptitude test in the internet to make music aptitude test easily. We test children for music aptitude early. We provide children the learning and instruction environment. This research makes children improve the ability of music. We develop the music aptitude test system for children in the WWW for children to use it. This system is expected to be a great help for children to test music aptitude easily. This system establish a well-understood and maintenance through this system for replacing the data files of their children in a effective way.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.16 no.1
• /
• pp.66-71
• /
• 2006

Lately teaching-learning system on the web site offers broad educational environment without any restriction of time and place to the teachers and learners. Using a web site in music education turned out to be effective. It provokes learner's interest and guides self-directed learning. Therefore, on this research, the web site for self-directed learning in music subject was designed and embodied as an assistant material of learning which is suitable for learner's level and demand As a result of using the web site in music class, there are some responses of students. First, there were 27 students positive responses$(79.4\%)$ of 34 students in the learning effect of the web site, and 25 students$(73.5\%)$ of 34 showed their will to use the web site continually if it is modified and supplemented. Second, some students answered that using this program in Diagnosis Evaluation and Formative Test could have an instant result, which was used for the material of feedback. Third, there was a positive evaluation of helping self-directed learning.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2011.04a
• /
• pp.340-343
• /
• 2011

음악 감정 분류에 관한 기존의 연구들은 템포, 박자, 음정, 음표, 리듬 등과 같은 음악의 멜로디와 관련된 자질을 이용하여 음악 감정을 분류하였다. 그러나 노래(Song)와 같이 가사를 포함한 음악은 같은 스타일의 멜로디라도 가사의 내용에 따라 음악에 대하여 청자가 느끼는 감정이 크게 다르다. 본 논문에서는 감정 온톨로지를 활용하여 노래 가사를 감정에 따라 분류하는 방법에 대하여 제안한다. 기구축 된 감정 온톨로지를 바탕으로 네 가지 통사적 규칙을 적용하여 노래 가사로부터 감정 자질을 추출한다. 추출된 감정 자질을 이용하여 Naive Bayes, HMM, SVM과 같은 기계학습 기법을 이용하여 8개 감정 그룹에 대해 58.8%의 정확도를 보였다.

• The Journal of the Korea Contents Association
• /
• v.7 no.2
• /
• pp.132-141
• /
• 2007

Traditional music education for infants is the most important step of forming human knowledge. Since their rapid development of intelligence and emotion has an educational value not only it improves the human knowledge but also it makes infants understand the unique emotion of Korean people. Under the several circumstances, however, we have no enough contents for traditional music education. Thus, by researching and analyzing the existing educational contents and complementing them properly for multimedia environment as a form of animation, we perform a series of experiments to infants who are attending to the kindergarten with interactive animations which are familiar with infants. Infants become interested in the contents and after the lesson of the short bamboo flute content, they show the clear improvements for the playing on a musical instrument. This proves that the Korean music educational content realized with animation can be an alternative plan to improve the educational effect by causing enjoyment and interest of infants.

• 한국HCI학회:학술대회논문집
• /
• 2008.02a
• /
• pp.739-743
• /
• 2008

In this paper, we introduce a method for automatically generating accompaniment music, according to user's input melody. The initial accompaniment chord is generated by analyzing user's input melody. Then next chords are generated continuously based on markov chain probability table in which transition probabilities of each chord are defined. The probability table is learned according to reinforcement learning mechanism using sample data of existing music. Also during playing accompaniment, the probability table is learned and refined using reward values obtained in each status to improve the behavior of playing the chord in real-time. The similarity between user's input melody and each chord is calculated using pitch class histogram. Using our method, accompaniment chords harmonized with user's melody can be generated automatically in real-time.

• 한국HCI학회:학술대회논문집
• /
• 2007.02b
• /
• pp.439-444
• /
• 2007

정서 전달에는 언어적 의사소통뿐만 아니라 비언어적 의사소통이 거론되곤 한다. 하지만 지금까지의 비언어적 의사소통 연구는 대부분 언어의 음향학적 연구나 얼굴표정 연구에 국한되었다. 또한 음악의 정서에 대한 연구는 음악적 구조 혹은 스타일과 여러 성격적 특질 및 행동 간의 관계, 혹은 생리적 효과 등 어떤 음악(what music)이 특정한 효과를 일으키느냐에 중점을 두고 있었다. 따라서 본 연구에서는 Gabrielsson & Juslin(1996)의 연구에 의거, 음악을 통하여 작곡가가 아닌 연주자의 의도된 정서가 청자에게 얼마나 잘 지각될 수 있는지 알아보고자 하였다. 봉 녕구의 가설은 다음과 같다; 첫째, 연주자의 정서적 의도는 음악의 모든 물리적 변인에 영향을 준다; 둘째, 청자는 의도된 정서를 지각하는 데 일반적으로 성공적이다; 셋째, 특정 정서는 다른 정서보다 더 잘 구분된다. 본 연구에 사용된 곡은 학습 효과를 배제하기 위하여 새롭게 전문 작곡가에 의해 작곡되었으며, 전문 연주자는 동일한 곡을 7종류의 정서(행복한, 슬픈, 화난, 두려운, 다정한, 엄숙한, 정서 표현 없음)를 표현하도록 연주하였다. 하나의 완전한 음악을 표현하기 위하여 각 곡은 멜로디(악기구성: 일렉트릭 기타, 베이스, 그랜드 피아노)와 리듬(드럼)을 포함하였다. 실험참가자는 각 곡을 듣고 7개의 정서 종류 각각에 점수를 평정하였다. 그 결과 이전 연구와 마찬가지로 청자는 연주자의 의도된 정서를 일반적으로 지각하는 데 성공하였으며 7개의 정서 중 특징적인 정서("행복한", "슬픈", "화난", "다정한") 는 다른 정서 보다 더 잘 구별되었다. 본 연구에 사용된 "두려운" 정서 곡의 음향분석 결과 소리 강도의 큰 변산이 특징이었다. 이는 이전 연구에서 "두려운" 정서의 특징인 타이밍의 큰 변산과 함께, "두려운" 정서가 '불규칙적이거나 변화가 커서 예측하기 어려운' 속성을 지니고 있음을 시사한다. 또한 "다정한" 정서에 대해서 본 연구에서는 다른 모든 정서와 유의미한 차이를 보인 반면, 이전 연구에서는 "슬픈"과 유의미한 차이가 없었다. 이는 본 연구에 쓰인 "다정한" 정서의 곡은 리듬 패턴을 다른 정서 버전과는 다르게 보사노바 리듬을 사용하였다. 이전 연구와는 다르게 빠른 템포였음에도 불구하고 구별이 잘 된 이유는 이와 같이 장르 특징적인 영향이 컸기 때문이라고 할 수 있다. 이는 연주자나 음악 스타일 자체의 성격이 정서 판단에 있어 큰 영향을 준다는 사실을 시사한다. 종합적으로 음악을 통한 정서 전달에 있어 연주자, 청자, 악기, 음악 스타일의 차이가 영향을 줄 수 있다는 사실을 확인하였으며 이러한 결과는 Scherer & Oshinsky(1977)가 언급하였듯 음악 연주가 다른 비언어적 의사소통 방법과 공유하는 특징이 있음을 시사한다.

• Journal of Gifted/Talented Education
• /
• v.24 no.6
• /
• pp.1073-1097
• /
• 2014

The working concept of the disadvantaged gifted is meant by those who hardly can have a chance to develop one's own gifts because of the environmental conditions and factors. The main purpose of this study is to explore the inherent trait of the music giftedness, which is composed of musical ability, creativity, and task commitment. Grounded upon the phenomenal research, this study purposefully samples ten disadvantaged students who participated in the arts gifted camp hosted by Korean National Research Institute for the Gifted in Arts. Thus, it explores the inherent traits of the disadvantaged gifted by coding, segmenting, and analyzing the interview with ten disadvantaged, and the evaluation reports of the expert group. The inherent trait of music giftedness of the disadvantaged gifted is shown as this: the inherent trait of the disadvantaged gifted in music is closely interrelated with one's own concept of giftedness. Whereas they show the high motivation and task commitment, they are hardly observed to have the creativity in their musical activities and learning. Finally, this study give a proposition for the strategy to stimulate and improve the inherent trait of the disadvantaged gifted in music: the importance of parents education and parent screening, meta-cognitive skills, the leaning-based creativity education, and the significant role of intrapersonal catalyst.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.20 no.6
• /
• pp.884-889
• /
• 2010

Recently, a music genre classification has been studied. However, experts use different criteria to classify each of these classifications is difficult to derive accurate results. In addition, when the emergence of a new genre of music genre is a newly re-defined. Music as a genre rather than to separate search should be classified as emotional words. In this paper, the feelings of people on the basis of brightness and darkness tries to categorize music. The proposed classification system by applying VCM(Variance Considered Machines) is the contrast of the music. In this paper, we are using three kinds of musical characteristics. Based on surveys made throughout the learning, based on musical attributes(beat, timbre, note) was used to study in the VCM. VCM is classified by the trained compared with the results of the survey were analyzed. Note extraction using the MATLAB, sampled at regular intervals to share music via the FFT frequency analysis by the sector average is defined as representing the element extracted note by quantifying the height of the entire distribution was identified. Cumulative frequency distribution in the entire frequency rage, using the difference in Timbre and were quantified. VCM applied to these three characteristics with the experimental results by comparing the survey results to see the contrast of the music with a probability of 95.4% confirmed that the two separate.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2014.06a
• /
• pp.3-4
• /
• 2014

음악 장르 분류 분야에서는 다양한 특징을 모아서 특징 벡터를 만들고 이를 support vector machine (SVM)와 같은 분류기에 입력하는 시스템이 주로 사용되고 있다. 이 논문에서는 거리 함수 학습를 음악 장르 분류를 위한 특징 벡터의 간소화에 적용하였다. 여러 거리 함수 학습 방법 중 하나의 방법을 선택하고, 기존의 논문들에서 사용되었던 특징 셋을 활용하여 기존 특징 셋에 대해서 성능을 떨어뜨리지 않으면서 특징 셋의 길이를 줄일 수 있는지 살펴본다. 우리의 실험에서는 168차원의 특징 셋을 10차원까지 줄였는데, 이 경우 분류 정확도가 2% 이내로 저하되었다.