• 디지털산업정보학회논문지
• /
• 제10권1호
• /
• pp.131-138
• /
• 2014

This paper is on a stereo acoustic echo canceller with the blind signal separation for post processing. The convergence speed of the stereo acoustic echo canceller is deteriorated due to mixing two residual signals which are update signals of each echo canceller. To solve this problem, we are to use the blind signal separation(BSS) method separating the mixed signals after the echo cancellers. The blind signal separation method can extracts the source signals by means of the iterative computations with two input signals. We had verified performances of the proposed acoustic echo canceller for stereo through simulations. The results of simulations show that the acoustic echo canceller for stereo using this algorithm operates stably without divergence in the normal state. And, when the speech signals were inputted, this echo canceller achieved about 2dB higher ERLE with the BSS post processing method than without this method. This stereo echo canceller showed the best performance in the case of inputting the real voice signal.

• 한국정보전자통신기술학회논문지
• /
• 제4권2호
• /
• pp.77-82
• /
• 2011

본 논문에서는 패닝 기법을 이용하여 믹싱된 스테레오 음원에서 음원을 분리하는 방법에 대하여 고찰한다. 음원 분리 알고리즘은 다채널 포맷 변환을 위한 업믹스나 음질 개선, 고품질 음원 분리 등 다양한 응용분야에 사용될 수 있다. 본 논문에서 사용하는 음원 분리 알고리즘은 믹싱된 스테레오 채널을 시간-주파수 별로 PCA(Principal Component Analysis) 분석 방법을 이용하여 각각의 음원들이 패닝된 방향을 추정하며, 추정된 방향의 성분만을 추출하는 방향 필터링 과정을 거쳐 음원들을 독립적으로 분리 해 낸다. 실험을 통해 각 음원 분리 알고리즘의 성능을 평가하였다.

• 디지털산업정보학회논문지
• /
• 제8권3호
• /
• pp.125-131
• /
• 2012

This paper is on a stereo acoustic echo canceller with the blind signal separation. The convergence speed of the stereo acoustic echo canceller is deteriorated due to mixing two residual signals in the update signal of each echo canceller. To solve this problem, we are to use the blind signal separation(BSS) method separating the mixed signals. The blind signal separation method can extracts the source signals by means of the iterative computations with two input signals. We had verified performances of the proposed acoustic echo canceller for stereo through simulations. The results of simulations show that the acoustic echo canceller for stereo using this algorithm operates stably without divergence in the normal state. And, when the speech signals were inputted, this echo canceller achieved about 3dB higher ERLE in the case of using the BSS algorithm than the case of not using the BSS algorithm. But this echo canceller didn't get good performances in the case of inputting the white noises as stereo signals.

• ETRI Journal
• /
• 제36권5호
• /
• pp.721-729
• /
• 2014

We propose a novel phase-based method for single-channel speech enhancement to extract and enhance the desired signals in noisy environments by utilizing the phase information. In the method, a phase-dependent a priori signal-to-noise ratio (SNR) is estimated in the log-mel spectral domain to utilize both the magnitude and phase information of input speech signals. The phase-dependent estimator is incorporated into the conventional magnitude-based decision-directed approach that recursively computes the a priori SNR from noisy speech. Additionally, we reduce the performance degradation owing to the one-frame delay of the estimated phase-dependent a priori SNR by using a minimum mean square error (MMSE)-based and maximum a posteriori (MAP)-based estimator. In our speech enhancement experiments, the proposed phase-dependent a priori SNR estimator is shown to improve the output SNR by 2.6 dB for both the MMSE-based and MAP-based estimator cases as compared to a conventional magnitude-based estimator.

• 한국음향학회:학술대회논문집
• /
• 한국음향학회 1999년도 학술발표대회 논문집 제18권 1호
• /
• pp.20-23
• /
• 1999

In this paper, The Split Model Analysis Algorithm, which can generate the wideband speech signal from the spectral information of narrowband signal, is developed. The Split Model Analysis Algorithm deals with the separation of the $10^{th}$ order LPC model into five cascade-connected $2^{nd}$ order model. The use of the less complex $2^{nd}$ order models allows for the exclusion of the complicated nonlinear relationships between model parameters and all the poles of the LPC model. The relationships between the model parameters and its corresponding analog poles is proved and applied to each $2^{nd}$ order model. The wideband speech signal is obtained by changing only the sampling rate

• 대한전자공학회:학술대회논문집
• /
• 대한전자공학회 1999년도 추계종합학술대회 논문집
• /
• pp.1135-1138
• /
• 1999

In this paper, The Split Model Analysis Algorithm, which can generate the wideband speech signal from the spectral information of narrowband signal, is developed. The Split Model Analysis Algorithm deals with the separation of the 10$\^$th/ order LPC model into five cascade-connected 2$\^$nd/ order model. The use of the less complex 2$\^$nd/ order models allows for the exclusion of the complicated nonlinear relationships between model parameters and all the poles of the LPC model. The relationships between the model parameters and its corresponding analog poles is proved and applied to each 2$\^$nd/ order model. The wideband speech signal is obtained by changing only the sampling rate.

• 한국컴퓨터산업학회논문지
• /
• 제5권2호
• /
• pp.269-274
• /
• 2004

기존 음성 인식의 실세계 적용에서 큰 문제점은 잡음이다. 본 논문에서는 잡음이 섞인 음성 신호로부터 잡음 성분을 분리해 내는 방법을 제안한다. 이 방법은 잡음이 섞인 음성 신호에 독립성분분석(ICA:Independent Component Analysis)을 사용한 암묵신호 분리(blind source separation)를 적용하여 잡음 성분을 제거하게 된다. 잡음이 혼합된 음성 신호에 독립성분분석을 전처리(preprocessing) 과정에 이용함으로써 인식성능을 향상시킬 수 있다. 깨끗한 음성 신호에 음악과 거리잡음을 섞었을 경우 인식률이 잡음 없는 음성의 인식률보다 각각 최대 14.98%, 13.78%까지 저하되었다. 그러나 독립성분분석으로 복원된 음성의 경우 잡음 없는 음성의 인식률 수준(각각 97.39%, 96.49%)으로 나타났으며, 독립성분분석을 이용한 음성의 잡음 제거가 인식률 향상에 좋은 결과를 가져옴을 확인 할 수 있다.

• 음성과학
• /
• 제12권2호
• /
• pp.29-41
• /
• 2005

This paper is to show the diphthong system of Jeju dialect speakers in their 70s or more on the basis of the acoustic analysis of their phonetic data. It is revealed through the analysis of their phonetic data that they clearly distinguish such diphthongs as [we], [w$\epsilon$], [yc] and [yo]. However, this paper shows that they are phonetically insensitive to the separation between [ye] and [y$\epsilon$] and they seldom make a precise pronunciation of diphthong [iy], which male speakers tend to pronounce to be [i] and female speakers to be [i].

• 한국소음진동공학회:학술대회논문집
• /
• 한국소음진동공학회 2002년도 추계학술대회논문초록집
• /
• pp.354.1-354
• /
• 2002

Pitch is a perception related to frequency, one of the psychological aspects or attributes of tones, and an important factor to determine sound quality of sound together with loudness and timber. while a study on pitch has been actively achieved In the part of speech recognition and speech separation, that for analysis and improvement of product sound quality is not yet enough. (omitted)

• 디지털융복합연구
• /
• 제11권10호
• /
• pp.453-458
• /
• 2013

HMM 음성 인식 시스템은 환경적 잡음과 여러 음성의 혼합으로 인하여 정확한 음성을 인지하지 못하는 단점이 있다. 따라서 본 논문은 잡음 음성으로 부터 원하는 음성만 선택하여 추출하기 위한 음성 특징 추출 기법을 CHMM을 이용하여 제안한다. 선택적 음성 추출을 위한 채널 유사성 상관 관계를 이용하여 음성 특징을 추출하는 방법을 사용하였다. 제안 기법의 실험 평가한 결과 평균 분리 왜곡도가 0.430dB 감소됨을 보임으로써 제안한 방법의 우수성을 확인하였다.