• Proceedings of the Acoustical Society of Korea Conference
• /
• spring
• /
• pp.56-60
• /
• 1999

본 논문은, 자동차에서 발생할 수 있는 다양한 형태의 잡음이 섞인 음성을 대상으로, 잡음에 강인한 파라미터들을 사용하여 인식기들을 구축하였으며, 이들 파라미터를 비교 평가하였다. 실험에 사용된 음성 데이터는 차종, 속도, 도로 환경, 라디오 ON/OFF, 창문 개폐여부 등 다양한 잡음 환경에서 수집하였다. 실험에서 비교된 파라미터는 MFCC(Mel-Blrequency Cepstral Coefficient)와 PLP(Perceptually Linear Prediction) 이며, 각각의 파라미터에 대해서 MKM(Modified k-mean)을 이용하여 코드북을 작성하였고, DHMM(Discrete Hidden Markov Model)을 인식알고리즘으로 사용하였다. 실험 결과로서, 아스팔트 도로에서 창문을 닫고, 라디오를 켜지 않은 상태에서 60km/h로 주행시 $96.25\%$로 가장 높은 인식률을 얻었고, 고속도로에서 창문을 열고 100km/h로 주행시에는$60\%$로 가장 낮은 인식률을 얻었다.

• The Journal of the Acoustical Society of Korea
• /
• v.23 no.3E
• /
• pp.79-84
• /
• 2004

In this paper, the AEC(acoustic echo canceller) is designed and implemented using VHDL(VHSIC hardware description language). The designed Echo Canceller employs the pipeline and the master-slave structure, and is realized with FPGA. As an adaptive algorithm, the Normalized LMS algorithm is used. For the coefficient adjustment, the Stochastic Iteration Algorithm(SIA) which uses only current residual values is used and the number of registers are evidently reduced and convergence speed is also much improved comparing to existing methods by using EAB of FPGA for FIR filter structure of transceiver. The designed Echo Canceller is verified with the test board implemented for this paper. From the timing simulation echo signals at about 1500 sampling data are converged and ERLE is improved by about 42-dB.

• The Journal of the Acoustical Society of Korea
• /
• v.23 no.3E
• /
• pp.91-97
• /
• 2004

In this paper, we propose a classification system of TV program scenes based on audio information. The system classifies the video scene into six categories of commercials, basketball games, football games, news reports, weather forecasts and music videos. Two type of audio feature set are extracted from each audio frame-timbral features and coefficient domain features which result in 58-dimensional feature vector. In order to reduce the computational complexity of the system, 58-dimensional feature set is further optimized to yield l0-dimensional features through Sequential Forward Selection (SFS) method. This down-sized feature set is finally used to train and classify the given TV program scenes using κ -NN, Gaussian pattern matching algorithm. The classification result of 91.6% reported here shows the promising performance of the video scene classification based on the audio information. Finally, the system stability problem corresponding to different query length is investigated.

• The Journal of the Acoustical Society of Korea
• /
• v.16 no.6
• /
• pp.19-24
• /
• 1997

The acoustic characteristics of the sound-pipe of King Song-Dok bell, which is located on the top, are investigated by experiment and simulation. The experimental results;reflection coefficient, and radiation impedance, demonstrate that the sound-pipe is capable of radiating high frequency(above 300Hz) sound; it behave as damper. It is also found that a waveguide model well presens the acoustic characteristics of the sound-pipe.

• Proceedings of the Acoustical Society of Korea Conference
• /
• autumn
• /
• pp.223-226
• /
• 2001

Acoustic pressure transmission coefficient and phase velocity are measured as the functions of water porosity and air porosity in sand sediment slabs with water- and air-filled pores. Pores in the sand sediment slab we modeled as the structure of circular cylindrical tube shape filled with water and air. The first kind(fast) wave and second kind (slow) wave, identified by Biot, in the solid and fluid mixed medium are affected by the presence of water and air pores. Acoustic characteristics of such porous medium in water are also theoretically investigated in terms of the modified Biot-Attenborough (MBA) model, which uses the separate treatment of viscosity effect and thermal effect in non-rigid porous medium with water- and air-filed pores. The information on the fast waves introduces new concepts of the generalized tortuosity factor and dynamic shape factor.

• Proceedings of the Acoustical Society of Korea Conference
• /
• autumn
• /
• pp.65-68
• /
• 2001

현재 음성인식기는 다채널의 음성입력방식을 사용하고 있는 추세이다. 이런 방법으로 음성인식기를 사용할 때에 자동적으로 음성을 검출하는 음성입력방식은 발성자와 마이크간의 거리에 따라 발성음성, 반사음성과 잡음이 입력된 경우에 원 음성의 고차포만트 성분에 왜곡이 발생하게 된다 이러한 문제점을 개선하기 위하여 본 논문에서는 고주파영역의 특성을 강조시킬 수 있는 프리엠퍼시스를 이용하여 더욱 더 정확한 음성입력 방법을 제안한다. 제안한 방법은 원음과 유사한 합성입력음을 얻었고, 또한 기존의 시간영역 법 보다 더 높은 SNR을 얻을 수 있다는 것을 알 수가 있다.

• Proceedings of the Acoustical Society of Korea Conference
• /
• autumn
• /
• pp.57-60
• /
• 2001

This paper reports on speaker-Recognition of context independence-speaker recognition in the field of the speech recognition. It is important to select the parameter reflecting the characteristic of each single person because speaker-recognition is to identify who speaks in the database. We used Mel Frequency Cesptrum Coefficient and Vector Quantization to identify in this paper. Specially, it considered to find characteristic-vector of the speaker in different from known method; this paper used the characteristic-vector which is selected in MFCC Parameter Space. Also, this paper compared the recognition rate according to size of codebook from this database and the time needed for operation with the existing one. The results is more improved $3\sim4\%$ for recognition rate than established Vector Quantization Algorithm.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1996.10a
• /
• pp.14-18
• /
• 1996

쐐기각이 다른 4종류의 쐐기형 흡음타일을 설계,제작하고 각각의 흡음타일에 대한 음파 반사계수를 측정하였다. 흡음타일의 정면에서 측정한 반사계수는 실험한 4개의 타일중 쐐기의 꼭지각이 30$^{\circ}$일 때 최소값을 갖고 120$^{\circ}$ 일 때 최대값을 가졌다. 쐐기형 흡음타일의 꼭지각이 클수록 그 반사계수가 증가되는 것을 확인하였다. 또한 쐐기의 꼭지각이 같은 종류의 타일헤서는 본 실험에 사용한 10kHz~30kHz 입사음파의 주파수와 측정한 반사계수는 무관함을 보였다. 그리고 쐐기형 흡음타일 후면에서 측정한 반사계수는 타입에 관계없이 일정하게 나타났다.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1994.06a
• /
• pp.943-948
• /
• 1994

The vector sum excited linear prediction(VSELP) coding gives high quality of synthetic speech at bit rates as low as 4.8kbps, but its computational complexity is prohibitive for real time applications. In this paper, we propose a method to reduce the computations of the VSELP codebook search procedure. The proposed method reduces the search space efficiently, before applying every linear combination of the basis vectors to the codebook search procedure. It decides whether is can fix the combination coefficient of each basis vector using heuristics so that the number of combinations decreases. It has been shown that the proposed method retains good quality of synthetic speech and reduces the computations of codebook search procedure by more than 40% of the origin.

• Proceedings of the Acoustical Society of Korea Conference
• /
• 1994.06a
• /
• pp.734-739
• /
• 1994

Elastic wave propagation in discrete random medium is studied to evaluate the effects of particle resonance on dispersion and attenuation of composite materials containing spherical inclusions. The frequency-dependent wave speed and attenuation coefficient can be obtained from proposed self-consistent method. It can be observed that the abrupt increase of effective wave speed and the concurrent peak of attenuation at low frequency is due to the lowest resonance of particles, whereas those in high frequency region are due to higher ones. The lowest resonance is mainly caused by the density mismatch and higher resonances by the stiffness mismatch between matrix and particles. The dispersion and attenuation of elastic waves in particulate composites are affected by the lowest resonance much than by higher ones.