• The Journal of the Acoustical Society of Korea
• /
• v.32 no.6
• /
• pp.518-524
• /
• 2013

In this paper, we propose a variable step size algorithm to enhance the normalized subband adaptive filter which has been proposed to improve the convergence characteristics of the conventional full band adaptive filter. The well-known Kwong's variable step size algorithm is simple, but shows better performance than that of the fixed step size algorithm. However, in case that large additive noise is present, the performance of Kwong's algorithm is getting deteriorated in proportion to the amount of the additive noise. We devised a variable step size algorithm which does not depend on the amount of additive noise by exploiting a normalized adaptation error which is the error subtracted and normalized by the estimated additive noise. We carried out a performance comparison of the proposed algorithm with other algorithms using a system identification model. It is shown that the proposed algorithm presents good convergence characteristics under both stationary and non-stationary environments.

• The Journal of the Acoustical Society of Korea
• /
• v.22 no.7
• /
• pp.528-538
• /
• 2003

A SCCRS (Speech and Character Combined Recognition System) is developed for working on mobile devices such as PDA (Personal Digital Assistants). In SCCRS, the feature extraction is separately carried out for speech and for hand-written character, but the recognition is performed in a common engine. The recognition engine employs essentially CHMM (Continuous Hidden Markov Model), which consists of variable parameter topology in order to minimize the number of model parameters and to reduce recognition time. For generating contort independent variable parameter model, we propose the SSMS(Successive State and Mixture Splitting), which gives appropriate numbers of mixture and of states through splitting in mixture domain and in time domain. The recognition results show that the proposed SSMS method can reduce the total number of GOPDD (Gaussian Output Probability Density Distribution) up to 40.0% compared to the conventional method with fixed parameter model, at the same recognition performance in speech recognition system.

• The Journal of the Acoustical Society of Korea
• /
• v.31 no.5
• /
• pp.288-297
• /
• 2012

This study proposes a design model of the variable type resonator which corrects the temperature variance according to the season, in order to maximize the resonance effect in the Sacred bell of the Great King Seongdeok. In the bell, the 1st natural frequency (64 Hz) and the 2nd natural frequency (168 Hz) are the most important partial tones. Resonance conditions of the two components are determined for the internal acoustic cavity system, which consists of bell body cavity, gap and the resonator. Acoustic frequency response characteristics of the internal cavity are determined by the boundary element analysis using SYSNOISE. As an external factor, temperature variance according to the season largely influences the resonance condition and the length of the resonator should be controlled to maximize the resonance effect. As a measure, this study proposes a design model of the variable type resonator for the Sacred Bell of the Great King Seongdeok, which can control the length at the belfry according to the season.

• The Journal of the Acoustical Society of Korea
• /
• v.28 no.8
• /
• pp.815-821
• /
• 2009

Information processing in variable and noisy environments is usually accomplished by means of adaptive filters. Among various adaptive algorithms, Least Mean Square (LMS) has become the most popular for its robustness, good tracking capabilities and simplicity, both in terms of computational load and easiness of implementation. In practical application of the LMS algorithm, the most important key parameter is the Step Size. As is well known, if the Step Size is large, the convergence rate of the algorithm will be rapid, but the steady state mean square error (MSE) will increase. On the other hand, if the Step Size is small, the steady state MSE will be small, but the convergence rate will be slow. Many researches have been proposed to alleviate this drawback by using a variable Step Size. In this paper, a new variable Step Size LMS(VSSLMS) called Categorized VSSLMS (CVSSLMS) is proposed. CVSSLMS updates the Step Size by categorizing the current status of the gradient, hence significantly improves the convergence rate. The performance of the proposed algorithm was verified from the view point of convergence rate, Excessive Mean Square Error(EMSE), and complexity through experiments.

• The Journal of the Acoustical Society of Korea
• /
• v.33 no.5
• /
• pp.326-334
• /
• 2014

This paper proposes a novel covariance matrix estimation scheme for minimum variance distortionless response (MVDR) beamforming. By accurately tracking direction-of-sound source arrival (DoA) information using audio-visual sensors, the covariance matrix is efficiently estimated by adopting a variable forgetting factor. The variable forgetting factor is determined by considering signal-to-interference ratio (SIR). Experimental results verify that the performance of the proposed method is superior to that of the conventional one in terms of interference/noise reduction and speech distortion.

• The Journal of the Acoustical Society of Korea
• /
• v.21 no.2
• /
• pp.167-172
• /
• 2002

The EVRC (Enhanced Variable Rate Codec) is currently in service as a speech cosec in digital cellular systems in North America and Korea. In the EVRC, the LSP (Line Spectral Pairs) related to energy distribution of speech signals in the frequency domain are coded by weighted split vector quantization. Considering that the LSP codebooks might be trained with the language of the develop country of the codebooks or English, it is expected that codebooks trained with Korean provide the performance improvements in the communication in Korean. In this paper, the EVRC LSP codebooks are designed with korean adopting the LBG algorithm based vector quantization, and the performance improvement of the vector quantization and the accompanying speech quality improvement are demonstrated by spectral distortion, SNR and SegSNR measurements, respectively.

• The Journal of the Acoustical Society of Korea
• /
• v.16 no.3
• /
• pp.81-85
• /
• 1997

In this paper, a recursive estimation algorithm for FIR systems is proposed using the 3rd and 4th order cumulants. To obtain the Overdetermined Recursive Instrumental Variable(ORIV) method type algorithm, we transform the 3'th and 4'th order cumulant relationship to a certain matrix form which is consist of only output data. From the matrix form, we induce the proposed algorithm procedure following the ORIV method. The proposed algorithm provides improved estimation accuracy with smaller data and can be applied to a time varying system as well. In addition, it reduces the estimation error due to the additive Gaussian noise compared to conventional 2'rd order based algorithms since it only uses higher than 2'rd order cumulant. Simulation results are presented to compare the performance with other HOS-based algorithms.

• The Journal of the Acoustical Society of Korea
• /
• v.18 no.8
• /
• pp.3-8
• /
• 1999

Previous variable vocabulary speech recognition systems with context-independent acoustic modeling, could not represent the effect of neighboring phonemes. To solve this problem, we use allophone-based context-dependent acoustic model. This paper describes the method to improve acoustic model of the system effectively. Acoustic model is improved by using allophone clustering technique that uses entropy as a similarity measure and the optimal allophone model is generated by changing the number of allophones. We evaluate performance of the improved system by using Phonetically Optimized Words(POW) DB and PC commands(PC) DB. As a result, the allophone model composed of six hundreds allophones improved the recognition rate by 13% from the original context independent model m POW test DB.

• The Journal of the Acoustical Society of Korea
• /
• v.20 no.1
• /
• pp.43-49
• /
• 2001

EVRC (Enhanced Variable Rate Codec) has been adopted as a standard coder for the CDMA digital cellular system in North America and Korea, and known to provide good call quality at 8kbps. In this paper, fast implementation algorithms for EVRC encoder are proposed. The proposed algorithms are based on both efficient pitch detection scheme and fast fixed codebook search algorithm. In the codebook search, computational complexity is reduced down to 70％ of the original EVRC by limiting the number of pulse position combination and by using a truncated impulse response. The proposed algorithms enable us to implement the EVRC with much smaller computational works. Also, informal subjective tests confirmed that the difference in the speech quality between the original EVRC and the proposed method was indistinguishable.

• The Journal of the Acoustical Society of Korea
• /
• v.29 no.6
• /
• pp.395-401
• /
• 2010

Since Philipp Reis succeeded in sending human voices through electric wires well ahead of Elisha Gray and A. G. Bell etc., he deserves to be acknowledged as the inventor of the telephone. Nevertheless, he did not enjoy any honor for his great invention while he was alive. Since he was working in a scientific community, his work was presented not as a patentable invention but as a scientific discovery. In addition, he used the intermittent electricity in accordance with the experimental tradition in European acoustics, occasioning the speech quality of his telephone to have a fatal shortcoming. On the contrary, Bell, who was a novice in electricity and acoustics, employed variable currents to transmit the sound signals, which guaranteed better speech qualities than Reis's.