• The Journal of the Acoustical Society of Korea
• /
• v.19 no.1
• /
• pp.11-15
• /
• 2000

Speaker adaptive continuous speech recognition for the RM speech corpus is described in this paper. Learning of hidden markov models for the reference speaker is performed for the training data of RM corpus. For the evaluation, evaluation data of RM corpus are used. Parts of another training data of RM corpus are used for the speaker adaptation. After dynamic time warping of another speaker's data for the reference data is accomplished, error back propagation neural network is used to transform the spectrum between speakers to be recognized and reference speaker. Experimental results to get the best adaptation by tuning the neural network are described. The recognition ratio after adaptation is substantially increased 2.1 times for the word recognition and 4.7 times for the word accuracy for the best.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.11 no.3
• /
• pp.211-216
• /
• 1986

This paper presents theoretical analysis and experimental results on an improved interference rejection circuit in the presence of continuous wave interference at center frequency, particularly in a wide-band communication systems. This circuit is based upon a phase locked loop for estimating the CW interference phase and the MSE technique to minimize the estimation error. Therefore, the introduced rejection circuit that minimized the estimated phase error outperforms the rejection circuit with phase estimation only. As a result of experiment, we confirm that this proposed rejection circuit gives a high degree of interference rejection performances when the input J/S is large.

• Proceedings of the Korea Institute of Convergence Signal Processing
• /
• 2006.06a
• /
• pp.17-20
• /
• 2006

본 논문에서는 잡음이 첨가된 연속음성에서의 자동 음절분할을 위해 기존에 사용되고 있는 특징 파라미터인 단구간 에너지 이외에 잡음에 강인한 특성을 가지고 있는 새로운 특징인 스펙트럼 밀도비교척도와 의사역행렬을 이용한 선형판별함수를 제안한다. 기존에 사용되는 단구간 에너지는 잡음이 없는 환경에서는 좋은 성능을 나타내지만 잡음환경에서는 그렇지 못하다. 반면에 논문에서 제안한 척도들은 반대의 성능을 가지므로 주변잡음의 크기에 따라 각각의 파라미터를 적절한 가중치로 조합하는 음절구간 결정함수와 유한상태 머신을 추가로 사용면 무 잡음 환경뿐만 아니라, 잡음이 첨가된 연속음성에서도 일정수준 이상의 음절구간을 분리해 낼 수 있다.

• Economic and Environmental Geology
• /
• v.14 no.2
• /
• pp.93-102
• /
• 1981

Studies of model potential fields continued upward and downward show differences depending on the method of continuation. Beginning with a magnetic field computed over a buried vertical cylinder, the field was continued to various levels by a method introduced by Henderson (Lagrangian interpolation) and by a spectral method (frequency domain analysis). Resultant fields show (1) no significant differences in upward continued values, (2) in downward continuation, accurate values are obtained with the spectral method over the central part of the anomaly, and (3) accurate values are obtained with Henderson's method on the flanks of the anomaly, while oscillations usually characterize the spectral method in this region. Essentially the same observations are made for derivative calculations. Field oscillations are empirically predicted at levels continued to approximately two-thirds of the depth of the source. Our spectral computer program output yields marked oscillations at one-half of the depth of the source. Henderson's method shows no oscillations at this depth and only minor oscillations at the top of the body (some negative values appear on the flanks of the anomaly). The Henderson output is a smooth field even if continued below the top of the body. These results suggest that the presence of oscillations cannot be used to identify the top of a buried source without careful consideration of the method used to continue the field. Use of the derivative to outline and isolate anomalies must similarly include consideration of the method of calculation.

• Journal of Korea Multimedia Society
• /
• v.4 no.4
• /
• pp.306-315
• /
• 2001

In noise reduction method from noisy speech for speech recognition in noisy environments, conventional spectral subtraction method has a disadvantage which distinction of noise and speech is difficult, and characteristic of noise can't be estimated accurately. Also, noise reduction method in the wavelet transform domain has a disadvantage which loss of signal is generated in the high frequency domain. In order to compensate theme disadvantage, this paper propose spectral subtraction method in continuous wavelet transform domain which speech and non- speech intervals is distinguished by standard deviation of wavelet coefficient, and signal is divided three scales at different scale. The proposed method extract accurately characteristic of noise in order to apply spectral subtraction method by end detection and band division. The proposed method shows better performance than noise reduction method using conventional spectral subtraction and wavelet transform from viewpoint signal to noise ratio and Itakura-Saito distance by experimental.

• The Journal of the Acoustical Society of Korea
• /
• v.28 no.4
• /
• pp.383-391
• /
• 2009

In speech signal processing, speech signal corrupted by noise should be enhanced to improve quality. Usually noise estimation methods need flexibility for variable environment. Noise profile is renewed on silence region to avoid effects of speech properties. So we have to preprocess finding voice region before noise estimation. However, if received signal does not have silence region, we cannot apply that method. In this paper, we proposed SNR estimation method for continuous speech signal. The waveform which is stationary region of voiced speech is very correlated by pitch period. So we can estimate the SNR by correlation of near waveform after dividing a frame for each pitch. For unvoiced speech signal, vocal track characteristic is reflected by noise, so we can estimate SNR by using spectral distance between spectrum of received signal and estimated vocal track. Lastly, energy of speech signal is mostly distributed on voiced region, so we can estimate SNR by the ratio of voiced region energy to unvoiced.

• Journal of the Korea institute for structural maintenance and inspection
• /
• v.15 no.5
• /
• pp.124-135
• /
• 2011

The capacity of bearings installed at abutments and piers for continuous bridges is usually determined by the magnitude of the maximum vertical reaction at each support and the capacity of bearings placed at piers is higher than that at abutments. In this study, the possibility of the improved seismic performance of base-isolated continuous skew bridges was investigated by analysing the variation of the seismic behavior of them according to three arrangements of bearings. Based on the conventional arrangement of bearings(Case A), three arrangements of bearings such as Case A, Case B and Case C were selected considering the variation of the horizontal stiffness of the lead rubber bearing(LRB) installed at the pier. The seismic behavior of the total 36 skew bridges was investigated by conducting the response spectrum analysis using the hybrid response spectrum considered the effect of LRB's damping. Results of analyses show that a more desirable seismic behavior of base-isolated continuous skew bridges can be obtained by reducing the magnitude of the horizontal stiffness of LRB placed at the pier to similar to or less than that of LRB installed at abutments. The variation of LRB's stiffness at the pier brings about period elongation and the change of mode shapes of base-isolated skew bridges and results in the reduction of the total base shear, the maximum base shear at the pier and the girder stresses. Although positive effects on the seismic behavior of base-isolated skew bridges caused by the change of arrangement of bearings decreased slighty with an increase in the flexibility of the substructure, the proposed arrangements of bearings bring about the improved seismic performance of base-isolated continuous skew plate girder bridges with less than 10m height of piers.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.1 no.1
• /
• pp.85-93
• /
• 1997

In this paper, a novel method for obtaining spectra of short pulses is proposed. This method will be well-applied to perform spectral analyses of ultrashort laser pulses which are known to be difficult to evaluate their exact forms in frequency domain because of their narrow-width characteristics in time domain. It must be noted that the method can be represented by a recursive formula derived from the PRS(partial-response signaling) model, and thus more easily available for numerical solutions than the conventional methods such as consecutive differentiations and convolutions. The pulses whose fundamental frames can be exactly determined or approximately represented by truncated Fourier series have well-behaving conditions for applications of this new method. Moreover, the transversal 9-type PRS model can offer various approximations for spectral analyses of ultrashort pulses currently used in transmission systems.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.52 no.1
• /
• pp.162-172
• /
• 2015

In this paper, we propose an adaptive detection algorithm of footstep and a classification algorithm for activities of the detected footstep. The proposed algorithm can detect and classify whole movement as well as individual and irregular activities, since it does not use continuous footstep signals which are used by most previous research. For classifying movement, we use feature vectors obtained from frequency spectrum from FFT, CWT, AR model and image of AR spectrogram. With SVM classifier, we obtain classification accuracy of single footstep activities over 90% when feature vectors using AR spectrogram image are used.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.10 no.3
• /
• pp.1-6
• /
• 1990

The dynamic behaviour of a four-span continuous girder railway bridge subjected to multiple support excitations is investigated using the response spectrum method. Small-amplitude oscillations and linear-elastic material behaviour are assumed. Soil-structure interaction effects are disregarded and only the out-of-plane response of the bridge is considered. The results of the response spectrum analysis are compared with those from a time history analysis. Different combination rules for the superposition of modal maxima as well as supports are employed, such as square-root-of-sum-squares, double sum and p-norm methods.