• 대한후두음성언어의학회지
• /
• 제9권1호
• /
• pp.22-26
• /
• 1998

Uvulopalatopharyngoplasty(UPPP) is one of the most popular surgical procedure for the treatment of obstructive sleep apnea syndrome(OSAS) occurring at the level of oropharynx. However, voice changes after UPPP have been a challenging issue for the professional voice users, because even minor changes in voice quality or articulation may be critical to professional singers, teachers, and so on. Several acoustic changes after UPPP have been proposed. However, based on the authors understanding, there is no report about voice changes after UPPP in Korean. We measured the first, second and third formant frequencies of /a/, /i/, /u/ phonations in 20 adult male patients who had undergone UPPP surgery, and the nasalances of Rabbit, Baby, and Mama passages. These parameters were measured preoperatively, at 1 month and 3 months after the operation. Any subjective voice changes were asked to be reported at the posto-perative visits. The third formant(F3) of /u/ phonation was significantly reduced at postoperative 1 month measurement. The nasalance of Mama passage was singnificantly increased at postoperative 3 months measurement. No one complained of subjective changes in voice quality, timbre, articulation or speech. Even though there are no complaints about postoperative voice changes subjectively, significant changes in the formant characteristics of certain vowel and changes in the nasality after UPPP require the clinicians to be mort cautious and careful in deciding UPPP for the professional voice users.

• 한국컴퓨터정보학회논문지
• /
• 제13권4호
• /
• pp.161-167
• /
• 2008

본 논문에서는 여러 대의 무전기간에 음성 및 데이터 정보를 원활하게 송수신 해 주는 전술 무선 이동 Ad-hoc 프로토콜인 MIL-STD-188-220C를 이용하여 VoIP통신을 하기 위한 패킷 크기 산출 및 전달 방법에 대해 연구하였다. 먼저 예상 데이터 전송시간을 산출한 후 사용자 입장에서의 VoIP 음성 품질과 무전기에서의 데이터 전송품질 요구수준을 동시에 고려하여 음성 패킷 길이 결정 및 패킷 전달 방법을 제시하였다. 전송 속도가 36Kbps인 무전기에서의 VoIP통신의 경우에는 90ms 재전송 패킷과 90ms 샘플링 패킷을 모아 짧은 프레임으로 전송하는 방법이 효율적이고, 36Kbps 이상의 경우에는 샘플링 패킷들을 1초 이상 모아서 전송 후 필요에 따라 재전송을 요청하는 방법을 고려할 수 있었다.

• 한국정보통신학회논문지
• /
• 제3권3호
• /
• pp.607-616
• /
• 1999

본 논문에서는 제한된 대역에서 낮은 전송률로 족은 품질을 유지하도록 하는 개선된 4.8kbps VSELP를 제안한다. 그러나 대부분의 경우에 있어서 낮은 전송률에는 좋은 품질을 유지하지 못하는 실정이다. 이러한 문제점을 해결하기 위해 많은 방법들이 제안되어 왔으나 대부분 외국어를 기준으로 맞추어져 우리 언어 구조에 적합하지 않다. 본 실험은 잡음이 없는 실험실에서 녹취한 데이터를 가지고 수행되었다. 본 논문은 저전송률을 가지며 한국어 음성에 적합한 코드북을 설계하고 8kbps의 VSELP와 4.8kbps의 VSELP를 SEGWSNR(Segmentally Weighted SNR) 평가와 MOS(Mean Opinion Score) 평가를 수행하고 주관적 평가에 있어서 4.8kbps의 VSELP의 우수함을 보인다.

• 한국음향학회지
• /
• 제20권1호
• /
• pp.43-49
• /
• 2001

EVRC (Enhanced Variable Rate Codec)는 북미 및 우리 나라 CDMA 디지털 셀룰러 시스템에 채택되었으며 8kbps의 전송률에서 우수한 성능을 갖는 부호화기이다. 본 논문에서는 복잡한 알고리듬으로 인해 많은 계산량을 갖는 EVRC 부호화기를 성능 저하 없이 고속으로 구현할 수 있는 알고리듬을 제시한다. 제안된 고속 알고리듬에서는 효율적인 피치 검색과 고정 코드북 탐색 과정이 구현되는데, 고정 코드북 탐색 과정에서는 펄스 위치 조합의 수를 제한하는 방법과 줄여진 임펄스 응답을 사용하여 연산량을 기존의 방법의 70％ 정도로 감소시킨다. 주관적인 음질 평가를 통해 제안된 고속 EVRC 알고리듬이 기존의 방법에 비해 적은 계산량에 구현되지만 음질의 저하는 초래하지 않는다는 것을 확인하였다.

• 한국음향학회지
• /
• 제21권4호
• /
• pp.349-360
• /
• 2002

인터넷 폰 시스템은 네트워크 트래픽 문제로 인한 지연, 지터 그리고 패킷 손실을 경험하고 이로 인한 통화품질의 저하가 문제가 되어 통화품질 (QoS) 향상 기술이 필요하게 되었다. 본 논문에서는 인터넷상에서 통화품질을 저해하는 요소들을 분석하고 실시간 전송 프로토콜/실시간 전송제어 프로토콜 (RTP/RTCP)을 이용하여 네트워크 상태를 진단하여 송, 수신 단말기간 네트워크 트래픽에 알맞은 방식으로 인코딩된 패킷을 송,수신하는 동적인 손실 복구 알고리즘을 제안한다. 실험결과 제안한 부가정보를 이용한 동적인 손실 복구 알고리즘은 연속 패킷손실인 경우 63%의 손실패킷 복원률을 보여주며, 비연속 패킷손실인 경우 42%의 패킷손실 복원률을 보여준다.

• 한국산업정보학회논문지
• /
• 제5권2호
• /
• pp.16-21
• /
• 2000

백터 양자화기는 음성 부호화, 오디오 부호화, 그리고 비디오 부호화와 같은 많은 고품질 고전송률 데이터 압축응용에서 널리 사용되고 있다. 백터 양자화기의 코더북의 크기가 매우 클 때, 코더북 전체를 찾는 방식은 많은 응용의 경우에서 계산량 때문에 상당한 문제점이 된다. 계산량을 낮추기 위하여 삼각형의 변 길이에 대한 부등식과 같은 코더북의 특성을 활용하는 많은 알고리즘들이 제안되고 연구되어 왔다. 본 논문에서는 최적의 코더백터를 찾기 위하여 다단구조에 기반한 신속 코더백터 찾기 알고리즘을 제안하고자 한다. 간단한 2 단계 구조의 이 알고리즘을 사용하여도 상당한 계산 복잡성을 압축대상의 품질을 손상시키지 않고 줄일 수 있다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제15권5호
• /
• pp.1590-1609
• /
• 2021

Laughter is one of the most important nonverbal sound that human generates. It is a means for expressing his emotions. The acoustic and contextual features of this specific sound are different from those of speech and many difficulties arise during their modeling process. During this work, we propose an audio laughter generation system based on unsupervised generative models: the autoencoder (AE) and its variants. This procedure is the association of three main sub-process, (1) the analysis which consist of extracting the log magnitude spectrogram from the laughter database, (2) the generative models training, (3) the synthesis stage which incorporate the involvement of an intermediate mechanism: the vocoder. To improve the synthesis quality, we suggest two hybrid models (LSTM-VAE, GRU-VAE and CNN-VAE) that combine the representation learning capacity of variational autoencoder (VAE) with the temporal modelling ability of a long short-term memory RNN (LSTM) and the CNN ability to learn invariant features. To figure out the performance of our proposed audio laughter generation process, objective evaluation (RMSE) and a perceptual audio quality test (listening test) were conducted. According to these evaluation metrics, we can show that the GRU-VAE outperforms the other VAE models.

• Asian-Australasian Journal of Animal Sciences
• /
• 제12권5호
• /
• pp.815-836
• /
• 1999

The role of animal agriculture for the quality of human life has always been emphasized during 20th century and it is expected to be even more important in terms of food supplies and in providing additional functions in the future. The world human population has almost tripled during a period of half century. The world population of animals has increased 2~3 times (6 times for chicken) during the last 60 years, and the total amount of livestock products has increased 5~6 times (more than 10 times in pork) with higher annual growth rate (9%) in developing countries. Increased personal income certainly encouraged demand for animal products over grains and lower animal production costs resulted from scientific and technological advances. Similarly the production of total grains has more than doubled owing to the advances in agricultural science during the later part of the 20th century. The average life span of world people in 1950s was only 46 years, which will be increased to almost 66 years in the year 2000. Present date clearly indicate that the life span of people is proportional to their income (GNP) and/or animal protein intake. Animals can provide other resources than foods. The increase of human population indicates that the number of animals as well as per capita consumption of animal products will be increased in the 21st century. The other resources we get from animals are drafts, packing, riding, hunting and herding. Guiding the blind, protection and companionship are also examples of what we can expect from animals. In the very near future, animals will become major donors of organs, skin and producers of drugs or special functional foods. It may be concluded that animals are very closely associated and related to the quality of human life, and they are expected to remain the same way in the 21st century.

• 대한음성학회:학술대회논문집
• /
• 대한음성학회 1996년도 10월 학술대회지
• /
• pp.362-366
• /
• 1996

This paper is generally divided in 2 parts. One is the study on vowels about korean singer's lyric song in view of Daniel Jones' Cardinal Vowel. The other is acoustic study on vowels in my singing about korean lyric song. Analysis data are KBS concert video tape and CSL's. NSP file on my singing and Informants are famous singers i.e. 3 sopranos, 1 mezzo, 2 tenors, 1baritone, and me. Analysis aim is to find out Korean 8 vowels([equation omitted]) quality in singing. The methods of descrition are used in closed vowels, half closed vowels, half open vowels, open vowels and rounded vowels, unroundes vowels and formants. The study of the former is while watching the monitor screen to stop the scene that is to be analysixed. The study of the latter is to analysis the spectrogram converted by CSL's. SP file. Analysis results are an follows: Visual and auditory korean vowels quality in singing have the 3 tendency. One is the tendency of more rounded than is usual Korean vowels. Another is the tendency of centralized to center point in Cardinal Vowel and the other is the tendency of diversity in vowel quality. Acoustic analysis is studied by means of 4 formants. Fl and F2 show similiar step in spoken. In Fl there is the same formant values. This seems to vocal organization be perceived the singign situation. The width of F3 is the widest of all, so F3 may be the characteristics in singing. In conclude, the characteristics of vowels in Korean lyric songs are seems to have the tendencies of rounding, centralizing to center point in Cardinal Vowel, diversity in vowel quality and, F3'widest width in compared with usual Korean vowels.

• Korean Journal of Otorhinolaryngology-Head and Neck Surgery
• /
• 제61권12호
• /
• pp.663-668
• /
• 2018

Background and Objectives The main objectives of this study were to determine the clinical usefulness of the program-assisted and real ear measurement (REM)-assisted fitting of hearing aids. Subjects and Method Fifteen participants with moderate to moderately severe hearing loss were enrolled in this study. Objective and subjective fitting results were assessed to compare the benefits between the program-assisted fitting (using a software fitting program) and the REM-assisted fitting. Real ear insertion gain (REIG), sound-field audiometry using warble tone, and Korean Hearing in Noise Test (K-HINT) were performed as objective tests. Sound quality rating was performed as a subjective test. Results In the program fitting, 48.89% of fitting points failed to come within ${\pm}10dB$ of the REIG target. In the REM fitting, however, the percentage of failure significantly decreased to 23.33% (p=0.013). In K-HINT test, the reception threshold for speech in quiet situation significantly decreased from 50.1 dB HL with the program fitting to 44.7 dB HL after the REM fitting (p<0.001). In front noise condition, signal-to-noise ratio improved from 4.53 dB to 3.50 dB with the REM fitting without statistical significance (p=0.099). In the sound quality rating, the REM fitting ($4.27{\pm}0.56$) showed a significantly better sound quality ratings than the program fitting ($3.69{\pm}0.74$) (p=0.017). Conclusion The REM fitting showed better results in both subjective and objective measurements than the program fitting.