• The Journal of the Acoustical Society of Korea
• /
• v.21 no.6
• /
• pp.543-551
• /
• 2002

In this paper, we present an efficient decoding method that performs in real time for 20k word continuous speech recognition task. Basic search method is a one-pass Viterbi decoder on the search space constructed from the novel language model network. With the consistent search space representation derived from various language models by the LM network, we incorporate basic pruning strategies, from which tokens alive constitute a dynamic search space. To facilitate post-processing, it produces a word graph and a N-best list subsequently. The decoder is tested on the database of 20k words and evaluated with respect to accuracy and RTF.

• The Journal of the Acoustical Society of Korea
• /
• v.33 no.4
• /
• pp.238-247
• /
• 2014

In this paper, rail vibration and its sound radiation are investigated, then the rail noise measurement by using microphone array is explored theoretically. A concrete slab track for domestic high speed trains is modeled as a Timoshenko beam on elastic support, regarding the moving of the excitation force on the rail. From the radiation characteristics of rail noise generated by a line source, the effect of moving load on sound radiation is obtained. Also it is found that the beam angle of the microphone array is a prominent factor to measure the rail noise level reliably because the rail noise propagates as a plane wave. In this investigation, a proper beam angle for the rail noise measurement by microphone array is suggested.

• The Journal of the Acoustical Society of Korea
• /
• v.36 no.2
• /
• pp.100-107
• /
• 2017

The focused MVDR (Minimum Variance Distortionless Response) can be applied for source localization in near field. However, if the number of sensors are increased, it requires a large amount of calculation to obtain the inverse of the covariance matrix. In this paper we propose a focused MVDR method using that beam space is formed from output of far field beamformer at the subarray. The performances of the proposed method was evaluated by simulation. As a result of simulation, the proposed method has the higher spatial resolution performance then the conventional delay-and-sum beamformer.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 2005.05a
• /
• pp.962-965
• /
• 2005

A high intensity progressive wave tube is installed at Korea Aerospace Research Institute (KARI) for acoustic environmental tests. The test facility has 700 mm x 800 mm cross-sectional area, and provides acoustic environment of 165 dB over the frequency range of $25Hz{\sim}10,000Hz$. The facility consists of a 6 m long acoustic wave tube, acoustic power generation systems, gases nitrogen supply systems, and acoustic control systems. This paper describes how the basic parameters of the facility and power generation systems are controlled to meet the requirement of the test. The shape and length of the tube has been designed by using the size of test objects and the wave propagation characteristics of the tube. The capacity of acoustic power generation systems is determined by the energy conversion of acoustic wave and the efficiency of acoustic modulators. Moreover, the paper introduces test run results of the tube. Overall of 163dB has been generated by using the test facility.

• Proceedings of the Korean Society for Noise and Vibration Engineering Conference
• /
• 1991.04a
• /
• pp.23-27
• /
• 1991

각종 홀 (음악홀, 극장, 사무실건물)의 공조 덕트계에는 미로형소음챔버가 설 치되는 경우가 많다. 이러한 소음장치를 건물내부에 설치하는 경우에는 건물 설계단계에서부터 소음챔버로 인한 감음양(투과손실 : Transmission Loss)의 예측계산이 중요하다. 그렇지만, 일반적인 소음장치는 그 형상이나 내표면의 흡음조건이 아주 복잡하기 때문에, 현단계에서는 간단한 이론만으로 투과손 실예측이 거의 불가능하다. 지금까지 이 문제에 대해서 유한요소법(Finite Element Method : FEM)을 이용해 검토한 예가 종종 소개되었으나, 대부분 소음챔버의 입구와 출구에서의 임의의 점에 대한 음압비를 투과손실로서 구 하고 있다. 그러나, 소음기자체의 실질적인 투과손실특성을 알기 위해서는 소음기의 입력 파워에 대한 출력파워의 비로서 구하지 않으면 안된다. 따라 서, 본 연구에서는 유한요소법에 의한 복소음향인텐시티(Complex sound intensity)의 수치계산법을 각종소음기 (팽창형, 미로형)의 투과손실해석에 적 용하기 위하여 이론적인 면에서 고찰했으며, 프로그램도 개발하여 모델해석 에 적용하였다. 또한, 위에서 언급된 수치해석법의 타당성의 검증을 위하여, 측정에 의한 투과손실예측방법으로서 크로스스펙트럼(Cross Spectrum)법에 의한 음향인텐시티계측법의 이용에 대해서 이론적으로 고찰했으며, 그 이론 을 기초로 한 축척 모형실험을 병행하였다.

• The Journal of the Acoustical Society of Korea
• /
• v.6 no.3
• /
• pp.43-51
• /
• 1987

In this work an implementation of SAW based serial type AWQPSK modulator is studied. The SAW AWQPSK modulator consisting of input apodized IDT and output uniform IDT with center frequency of 20 MHz and bit rate of 4MHz has been designed and fabricated on $YZ-LiNbO_3$ substrate. Measured center frequency and null-to-null bandwidth are 20MHz, respectively. The sidelobe suppression is achieved 60dB below the peak mainlobe level. Measured responses meet the theoretical values with tolerable amount of deviation. SAW-based modulator simplifies the implementation of AWQPSK which uses complex pulse shape as a baseband pulse.

• The Journal of the Acoustical Society of Korea
• /
• v.8 no.5
• /
• pp.33-50
• /
• 1989

In this paper, musical tone, especially instrument tones are synthesized using nonlinear distortion technique. Nonlinear distortion is very simple but versatile method when you synthesize musical instrument tones. It basically consists of one sine oscillator and amplifier which makes distortion to Input wave. Output wave has many harmonics that can be controlled by varying shaping function, which is the transfer function of nonlinear amplifier. Shaping function Is obtained from the analyzed harmonic amplitude data. Given harmonics amplitudes, Chebyshev polynomial is used to produce the shaping function that exactly makes the given harmonics at steady state. We contructed non -real time nonlinear distortion synthesizer program running at IBM-PC. To quantify the satis faction of synthesised tones, listening test is carried out, and the result is presented.

• The Journal of the Acoustical Society of Korea
• /
• v.13 no.6
• /
• pp.83-90
• /
• 1994

In this paper, we propose a new learning algorithm, ASOFM(Adaptive Self Organizing Feature Map), to solve the defects of Kohonen's Self Organiaing Feature Map. Kohonen's algorithm is sometimes stranded on local minima for the initial weights. The proposed algorithm uses an object function which can evaluate the state of network in learning and adjusts the learning rate adaptively according to the evaluation of the object function. As a result, it is always guaranteed that the state of network is converged to the global minimum value and it has a capacity of generalized learning by adaptively. It is reduce that the learning time of our algorithm is about $30\%$ of Kohonen's.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1998.06e
• /
• pp.381-384
• /
• 1998

본 논문은 주행중인 자동차 환경에서 운전자의 안전성 및 편의성의 동시 확보를 위하여, 보조적인 스위치 조작없이 상시 음성의 입, 출력이 가능하도록 한다. 이때 잡음에 강인한 threshold 값을 구하기 위하여, 일정한 시간마다 기준 에너지와 영교차율(Zero Crossing Rate)을 변경하며, 밴드패스 필터(bandpass filter)를 이용하여 1차, 2차로 나누어 실시간 상태에서 자동으로, 정확하게 끝점검출(End Point Detection)을 처리한다. 기준패턴(reference pattern)은 DMS(Dynamic Multi-Section)을 사용하며, 화자의 변별력을 높이기 위하여 2개의 모델사용을 제안한다. 또한 주행중인 차량의 잡음환경에 강인하기 위하여 일반주행(80km/h 이내), 고속주행(80km/h 이상)등으로 나누며 차량의 가변잡음 크기에 따라 자동으로 선택하도록 한다. 음성의 특징 벡터와 인식 알고리즘은 PLP 13차와 One-Stage Dynamic Programming (OSDP)를 이용한다. 실험결과, 자주 사용되는 차량 편의장치 제어명령 33개에 대하여 중부, 영동 고속도로(시속 80Km/h 이상)에서 화자독립 89.75%, 화자종속 90.08%의 인식율을 구하였으며, 경부 고속도로에서는 화자독립 92.29%, 화자종속 92.42%의 인식율을 구하였다. 그리고 저속 주행중인 자동차 환경(80km/h 이내, 시멘트, 아스팔트 등의 서울시내 및 시외독립)에서는 화자독립 92.89%, 화자종속 94.44% 인식율을 구하였다.

• The Journal of the Acoustical Society of Korea
• /
• v.10 no.3
• /
• pp.5-12
• /
• 1991

In this paper, 10 city names of Korean were recognized. The name are articulated each 5 times by 10 male speakers. Filter bank output on total 500 words were extracted and they were used as feature parameters. Filter bank was constructed of 15 channels with 1/3 octave spacing from 200[Hz], using RC active circuit. Reference templates were created by clustering algorithm. DTW algorithm was used to compare similarity between reference templates and input words. Euclidean distance equation and Chebyshev distance equation were used to know the distinction between the recognition results obtained by the method of distance caculation, error rates are 16.4[%], 15.0[%], respectively.