• Journal of Internet Computing and Services
• /
• v.18 no.5
• /
• pp.17-22
• /
• 2017

In a MPC coding using excitation source of voiced and unvoiced, it would be a distortion of speech waveform. This is caused by normalization of synthesis speech waveform of voiced in the process of restoration the multi-pulses of representation section. This paper present PCFBD-MPC( Position Compensation Frequency Band Division-Multi Pulse Coding ) used V/UV/S( Voiced / Unvoiced / Silence ) switching, position compensation in a multi-pulses each pitch interval and Unvoiced approximate-synthesis by using specific frequency in order to reduce distortion of synthesis waveform. Also, I was implemented that the PCFBD-MPC( Position Compensation Frequency Band Division-Multi Pulse Coding ) system and evaluate the SNRseg of PCFBD-MPC in coding condition of 8kbps. As a result, SNRseg of PCFBD-MPC was 13.4dB for female voice and 13.8dB for male voice respectively. In the future, I will study the evaluation of the sound quality of 8kbps speech coding method that simultaneously compensation the amplitude and position of multi-pulse source. These methods are expected to be applied to a method of speech coding using sound source in a low bit rate such as a cellular phone or a smart phone.

• Journal of Welding and Joining
• /
• v.23 no.5
• /
• pp.14-19
• /
• 2005

In high speed $CO_2$ welding, it has been blown to produce no sound bead with undercut or humping bead. In this study welding phenomena through synchronized high speed camera and output welding voltage and current waveform has been analyzed. For the purpose to improve a welding bead, effect of short circuit current rise slope has been examined. With commercial power source it was produced no sound bead by instantaneous short circuit, long arc period and stubbing at welding speed 2.5mm/min Humping bead or undercut were showed by instantaneous short circuit and long arc period. Also, the weld bead was not formed during the long short circuit period after stubbing start and long arc extinguishment period after wire sticking by failure of arc regeneration, because the droplet was not transferred to weld pool. With increasing short circuit current rise slope the frequency of stubbing was decreased and the normal short circuit rate was increased. A control of short circuit current rise slope was effective factor in high speed welding.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2005.11a
• /
• pp.800-803
• /
• 2005

Impact sounds, such as those created by footsteps, the dropping of an object or the moving of furniture, can be a source of great annoyance in residential buildings. Running and jumping impact sound by child and walking by adult are one of the most irritating noises in an apartment buildings. It's necessary to know that the impact power characteristics of real impact source in an apartment buildings. This study aims to investigate the impact power and impact time of normal walking and fast walking for 62 adults. It is shown that when the weight of the person increase, the maximum impact power increases. The impact power waveform for the adults walking varies for subjects walking types. The normal walking impact power lower than that of fast walking and impact time is higher than that of fast walking. The range of the impact power generated by adults walking is less than 1000 N.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.10 no.6
• /
• pp.1229-1236
• /
• 2009

Puretone audiometer, which is a machine used for measuring the minimum hearing threshold, can be cost-effectively implemented using computer with sound card and software. In this paper, we describe a puretone audiometer which has been designed and implemented based on a general PC with sound card. It supports air conduction and bone conduction test taking with automatic masking. It also provides multiple modes consisted of self-test, auto-test and manual test mode. Such multiple modes makes it possible to use in various environments like as home and/or hospital. Through measure of waveform of output voltage and sound pressure, we verified that puretone audiometer of this paper properly operates.

• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.26 no.1
• /
• pp.89-96
• /
• 2016

TDOA(time delay of arrival) position estimate from acoustic measurement of artillery shell impact is studied in order to develop a targeting evaluation system. Impact position is calculated from the intersections of hyperbolic estimates based on the least square Taylor series method. The mathematical process of Taylor series estimation is known to be robust. However, the concern lays with the accuracy because it is sensitive to the bias caused by the randomness of measurement situation. The measurement error typically occurs from the distortion of waveform, change of travelling path, and sensor position error. For outdoor measurement, a consideration should be made on the atmospheric condition such as temperature and wind which can possibly change the trajectories of rays of sound. It produces wrong propagation time events accordingly. Ray tracing and optimization techniques are introduced in this study to minimize the bias induced by the ray of sound. The numerical simulation shows that the atmospheric correction improves the estimation accuracy.

• The Journal of the Acoustical Society of Korea
• /
• v.37 no.5
• /
• pp.338-342
• /
• 2018

In order to solve the floor impact sound problem in the upper and lower floors, the Ministry of Land, Transport and Maritime Affairs also notifies the physical properties of the resilient material affecting the floor impact sound level. The dynamic modulus of elasticity and the loss factor before and after heating are most related to the floor impact noise, especially for the cushioning material. Therefore, in this study, the rate of change with respect to the dynamic modulus and loss factor with temperature change was examined by increasing $10^{\circ}C$ by $10^{\circ}C$ from the temperature condition of $70^{\circ}C$ specified in the standard. The dynamic modulus of elasticity and the loss modulus were measured by using the pulse excitation method for eight kinds of samples. The calculation method was calculated by the time series analysis method using the damped vibration waveform.

• Journal of Biomedical Engineering Research
• /
• v.43 no.2
• /
• pp.124-130
• /
• 2022

In this paper, we proposed a portable electronic cardiogram and stethoscope (ECGS) that can simultaneously perform the electrocardiogram (ECG) and auscultation tests to increase the reliability of diagnosis of heart disease. To measure the ECG and heart sound (HS) at the same time, three ECG electrodes and a microphone sensor were combined into a triangular shape with a width of 90 mm and a height of 97 mm that can be held in one hand. In order to prevent skin problems when they contact the patient's skin, a capacitive coupled electrode was selected as the ECG electrode and a silicone material was used in a chest piece with the microphone sensor. For the signals measured from the electrodes and the chest piece, filters were respectively configured to pass only the signals of 0.01-100 Hz and 20-250 Hz, which are frequency bands for ECG and HS. The filtered ECG and HS analog signals were converted into digital signals and transmitted to a PC using wireless communication for monitoring them. The HS could be auscultated simultaneously using an earphone. The monitored ECG had an SNR of about 34 dB and a P-QRS-T waveform is clearly visible. In addition, the HS had an SNR of about 28 dB and both S₁ and S₂ are clearly visible. It is expected that it can aid doctors' inexperience in analyzing the ECG and HS.

• Speech Sciences
• /
• v.6
• /
• pp.103-117
• /
• 1999

This paper suggests that the onset-coda asymmetry found in languages like Korean and others should be dealt with in terms of one acoustic feature rather than other articulatory features, claiming that the acoustic feature involved here is [noise], i.e., 'aperiodic waveform energy'. It determines the structural well-formedness of the languages in question whether a coda ends in [noise] or not, regardless of the intensity, the frequency, and the time duration of the [noise]. Fricatives, affricates, aspirated stops, tense stops, and released stops are all disallowed in the coda position due to the acoustic feature [noise] they, commonly end with if they were, posited in the coda. The proposal implies that the three seemingly separate prohibitions of consonants in the coda position -- i) no fricatives/affricates, ii) no aspirated/tense stops, and iii) no released stops -- are directly correlated with each other. Incorporation of the one acoustic feature [noise] in the feature theory enables us to see that the aspects of onset-coda asymmetry are derived from one single source: ban, of [noise] in the coda.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2017.11a
• /
• pp.82-83
• /
• 2017

파노라마 영상은 최근 몇 년 사이 스마트 폰을 사용하는 일반 사용자에게 손쉽게 제작할 수 있게 되었다. 하지만 파노라마 동영상은 기술상의 어려움이 있다. 파노라마 동영상을 제작하기 위해서는 두 대 이상의 스마트 폰으로 동시에 촬영을 하여야 하는데, 하드웨어의 차이로 인해 시간축을 동기화하는데 어려움이 발생한다. 이러한 어려움을 극복하기 위해 영상 정보를 이용한 시간축 동기화도 시도되었다. 하지만 영상 정보를 이용한 시간축 동기화에도 어려움이 있다. 영상 정보를 이용하기 때문에 계산복잡도가 높고 영상 내 움직임으로 인해 시간축 동기화 정확도도 떨어진다. 본 논문에서는 영상 정보를 이용하는 대신 소리 파형을 이용한 시간축동기화 기법을 제안한다. 소리 파형 이미지를 이용하여 시간축을 동기화하기 때문에 정확도가 향상되고 시간축 동기화 프로세싱 시간이 단축되는 것을 확인할 수 있다.

• Journal of Broadcast Engineering
• /
• v.24 no.3
• /
• pp.472-484
• /
• 2019

This paper proposes a time-domain sound event detection algorithm using DNN (Deep Neural Network). In this system, time domain sound waveform data which is not converted into the frequency domain is used as input to the DNN. The overall structure uses CRNN structure, and GLU, ResNet, and Squeeze-and-excitation blocks are applied. And proposed structure uses structure that considers features extracted from several layers together. In addition, under the assumption that it is practically difficult to obtain training data with strong labels, this study conducted training using a small number of weakly labeled training data and a large number of unlabeled training data. To efficiently use a small number of training data, the training data applied data augmentation methods such as time stretching, pitch change, DRC (dynamic range compression), and block mixing. Unlabeled data was supplemented with insufficient training data by attaching a pseudo-label. In the case of using the neural network and the data augmentation method proposed in this paper, the sound event detection performance is improved by about 6 %(based on the f-score), compared with the case where the neural network of the CRNN structure is used by training in the conventional method.