디지탈 신호처리를 위한 FFT 알고리즘에 관한 연구

金淳協,金長福 光云大學校 1981 論文集 Vol.10 No.-

The Fast Fourier Transform [FFT] is a computational tool which facilitates signal analysis such as power spectrum analysis and filter simulation by means of digital computers. It is a method for efficiently computing the Discrete Fourier Transform [DFT] of a series of data samples (referred to as a time series). In this paper, we organize a programming package for computing FFT. Once we have considered a program to calculate the FFT, it can efficiently be left to the user to process a digital signal.

韓國語 숫자音聲의 分析과 自動認識에 關한 硏究

金淳協 光云大學校 1983 論文集 Vol.12 No.-

This paper discusses a method of frequency analysis and theory used in automatic recognition of Korean isolated digits Sound. The analog speech signal is filtered by low pass filter of which the cut off frequency is 4.8[Kh??z,], and then sampled at 10 [??] of sampling rate. In speech signal precessing, Cadzow-Kay's closed form ARMA spectral estimation method is used. It has been demonstrated that the ARMA model provides better spectral estimation than the more specialized AR model and MA model. It has shown that a (10,10)th order ARMA spectral estimation is necessary in order to provide the desired frequency resolution It is shown that Dynamic Programming algorithm (DP) is of major importance for recognition of Korean isolated digits Sound. DP based on Itakura's method is used to achieve time alignment. As a result, the recognition rate of 97.3% for the three speakers is obtained.

TMS320C30을 이용한 음성 다이어링 시스템의 구현

김순협 光云大學校 1992 論文集 Vol.21 No.-

This paper presents an implementation of real time voice dialing system using the TMS320C30, digital signal processing chip. The DMS/SS recognition algorithm by DMS model is applied to speech recognition and 50 department names within university is utilized as a recognition vocabularies. This voice dialing system was implemented on personal computer(IBM-PC 386DX). As a result of experiment, recognition rate of 98% and recognition time like the existing dialing time are obtained. The validity of this system, therefore, is shown from experiment.

Acoustic modeling for Korean Language Vocabulary-Independent Speech recognition system

Tea-Woong choi,Soon-hyob Kim 한국소음진동공학회 2003 한국소음진동공학회 국제학술대회논문집 Vol.2003 No.8

Noise Estimation based on Standard Deviation and Sigmoid Function Using a Posteriori Signal to Noise Ratio in Nonstationary Noisy Environments

Soo-Jeong Lee,Soon-Hyob Kim 대한전기학회 2008 International Journal of Control, Automation, and Vol.6 No.6

In this paper, we propose a new noise estimation and reduction algorithm for stationary and nonstationary noisy environments. This approach uses an algorithm that classifies the speech and noise signal contributions in time-frequency bins. It relies on the ratio of the normalized standard deviation of the noisy power spectrum in 'time-frequency bins to its average. If the ratio is greater than an adaptive estimator, speech is considered to be present. The propose method uses an auto control parameter for an adaptive estimator to work well in highly nonstationary noisy environments. The auto control parameter is controlled by a linear function using a posteriori signal to noise ratio (SNR) according to the increase or the decrease of the noise level. The estimated clean speech power spectrum is obtained by a modified gain function and the updated noisy power spectrum of the time-frequency bin. This new algorithm has the advantages of much more simplicity and light computational load for estimating the stationary and nonstationary noise environments. The proposed algorithm is superior to conventional methods. To evaluate the algorithm's performance, we test it using the NOIZEUS database, and use the segment signal-to-noise ratio (SNR) and ITU-T P.835 as evaluation criteria.

A Study on the Department-Name Recognition For Voice Dialing System

Lee, Seong-Kwon,Kim, Soon-Hyob 광운대학교 신기술연구소 1997 신기술연구소논문집 Vol.26 No.-

본 논문은 OSDP(One Stage Dynamic Programming) 알고리즘과 DMS(Dynamic MultiSection)모델을 이용한 부서명 인식에 관한 연구이다. 이 논문은 대학에서 음성 다이얼링 시스템을 위해서 적용될 수 있다. 먼저, DMS 모델을 사용하여 부서명인 고립단어 모델을 만든다. 그리고 실험에서는 두 가지 형태의 모델이 사용됐다. 첫 번째 모델은 부서명과 교수명 뒤에 오는 직명과 존칭 즉 "교수님"이 포함된 모델이다. 다른 모델은 부서명과 교수님들 이름에 우리말 "교수님"이 제외되어 있다. 인식 실험에서는 OSDP 알고리즘을 사용하였다. 음성 데이터는 부서명과 교수님들의 이름들로 구성되어 있다. 실험에서는 우리말 "교수님"이 제외된 우리말 "교수님"이 DP를 수행하는 동안에 다른 템플리트 간에 오인식을 일으키는 것으로 생각되다. 레퍼런스 모델은 잡음환경하의 연구실에서 만들어 졌다. 인식 실험은 252개의 서로 다른 부서명과 교수님들 이름들에 대해서 이루어졌으며 화자독립에 대해서 모델 I을 사용하였을 경우 87.3%, 그리고 화자 독립에 대해서 모델 II를 사용하였을 경우 94.9%의 인식율을 얻었다. This paper proposes one way of recognition of department names by using DMS (Dynamic Multisection) Modeling with OSDP (One Stage Dynamic Programming). This proposed method can be applied to voice dialing system in university. First, we make a model of each department name by using DMS model. In the experiment, two types of model are used. One is the model which include department names and professor’s name included Korean word "gyo su nim". The other is the mode which does not include the Korean word "gyo su rum". And We perform the recognition experiment by using OSDP algorithm. Voice data consists of department names and professors’ names. In the experiment, the result of using second model which does not include Korean word "gyo su nim" is better than the first model. The Korean word "gyo su nim" which is included all professor’s name in model I cause the mis-recognition while the dp performs the distance between other templates. The references are made by a male speaker in the noisy laboratory environment. The recognition experiment is performed for 252 different department names and professors’ names. Recognition rate was 87.3% in speaker dependent using model I and 94.9% in speaker independent using model II.

A Study on the Speech Recognition for Department Names of University

Seong Kwon Lee,Soon Hyob Kim 대한전자공학회 1997 대한전자공학회 학술대회 Vol.1997 No.1

Noise Suppression Using Normalized Time-Frequency Bin Average and Modified Gain Function for Speech Enhancement in Nonstationary Noisy Environments

Lee, Soo-Jeong,Kim, Soon-Hyob The Acoustical Society of Korea 2008 韓國音響學會誌 Vol.27 No.e1

A noise suppression algorithm is proposed for nonstationary noisy environments. The proposed algorithm is different from the conventional approaches such as the spectral subtraction algorithm and the minimum statistics noise estimation algorithm in that it classifies speech and noise signals in time-frequency bins. It calculates the ratio of the variance of the noisy power spectrum in time-frequency bins to its normalized time-frequency average. If the ratio is greater than an adaptive threshold, speech is considered to be present. Our adaptive algorithm tracks the threshold and controls the trade-off between residual noise and distortion. The estimated clean speech power spectrum is obtained by a modified gain function and the updated noisy power spectrum of the time-frequency bin. This new algorithm has the advantages of simplicity and light computational load for estimating the noise. This algorithm reduces the residual noise significantly, and is superior to the conventional methods.

Noise Reduction Using the Standard Deviation of the Time-Frequency Bin and Modified Gain Function for Speech Enhancement in Stationary and Nonstationary Noisy Environments

Lee, Soo-Jeong,Kim, Soon-Hyob The Acoustical Society of Korea 2007 韓國音響學會誌 Vol.26 No.e3

In this paper we propose a new noise reduction algorithm for stationary and nonstationary noisy environments. Our algorithm classifies the speech and noise signal contributions in time-frequency bins, and is not based on a spectral algorithm or a minimum statistics approach. It relies on calculating the ratio of the standard deviation of the noisy power spectrum in time-frequency bins to its normalized time-frequency average. We show that good quality can be achieved for enhancement speech signal by choosing appropriate values for ${\delta}_t\;and\;{\delta}_f$. The proposed method greatly reduces the noise while providing enhanced speech with lower residual noise and somewhat higher mean opinion score (MOS), background intrusiveness (BAK) and signal distortion (SIG) scores than conventional methods.

적응적 자기 조직화 형상지도

이형준,김순협,Lee , Hyung-Jun,Kim, Soon-Hyob 한국음향학회 1994 韓國音響學會誌 Vol.13 No.6

본 논문에서는 코호넨(Kohonen)의 SOFM (Self-Organizing Feature Map) 알고리즘의 단점을 해결하기 위한 새로운 학습 알고리즘 ASOFM(Adaptive Self-Organized Feature Map)을 제안한다. 코호넨의 학습 알고리즘은 초기화된 연결 벡터에 대하여 극소점에 빠지는 경우도 있다. 그러나 제안된 알고리즘에서는 학습과정중에 네트워크의 상태를 평가할 수 있는 목적함수(object function)을 사용하였고, 이 함수의 출력에 따라 학습의 각 시점에서 적응적으로 학습률의 재조정이 가능하였다. 이 결과, 네트워크의 상태가 최소점에 수렴함이 보증 되고 학습률의 적응성에 의해 임의의 학습패턴에 대한 학습의 일반화 능력이 보장되었다. 또한 제안된 알고리즘은 코호넨의 알고리즘보다 약 $70\%$이상의 학습시간을 단축한다. 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.