• The Journal of the Acoustical Society of Korea
• /
• v.31 no.3
• /
• pp.161-169
• /
• 2012

This paper presents a robust speech end-point detector using double talk detection in echoic conditioned speech recognition robot. The proposed method consists of combining conventional end-point detector result and double talk detector result. We have tested the proposed method in isolated word recognition system under echoic conditioned environment. As a result, the proposed algorithm shows superior performance of 30 % to the available techniques in the points of speech recognition rates.

• The Journal of the Acoustical Society of Korea
• /
• v.22 no.2
• /
• pp.121-128
• /
• 2003

Acoustic Echo Cancellation(m) is very active research topic having many applications like teleconference and hands-free communication and it employs Double Talk Detector(DTD) to indicate whether the near-end speaker is active or not. However. the DTD is very sensitive to the variation of acoustical environment and it sometimes provides wrong information about the near-end speaker. In this paper, we are focusing on the development of robust DTD algorithm which is a basic building block for reliable AEC system. The proposed AEC system consists of delayless subband AEC and narrow-band DTD. Delayless subband AEC has proven to have excellent performance of echo cancellation with a low complexity and high convergence speed. In addition, it solves the signal delay problem in the existing subband AEC. On the other hand, the proposed narrowband DTD is operating on low frequency subband. It can take most advantages from the narrow subband such as a low computational complexity due to the down-sampling and the reliable DTD decision making procedure because of the low-frequency nature of the subband signal. From the simulation results of the proposed narrowband DTD and wideband DTD, we confirm that the proposed DTD outperforms the wideband DTD in a sense of removing possible false decision making about the near-end speaker activity.

• The Journal of the Acoustical Society of Korea
• /
• v.16 no.3
• /
• pp.57-63
• /
• 1997

This paper proposes a new stereophonic acoustic echo canceller to preven the performance degradation at the instant of far-end talker's movement or change in transmission room environment in stereophonic teleconferencing system. In stereophonic acoustic echo canceller, the filter coefficients of the adaptive filters for echo cancellation do not have unique solutions and not converge to their optimum values. Therefore, the change of the far-end transmission room environment leads to degradation of ERLE(Echo Return Loss Enhanceement). Moreover, their computational complexity increases as the number of the filter coefficient increases, and their adaptive filters converge very slowly. So the real-time implementation is very difficult. To overcome these problems we propose a pre-processor consisting of an adaptive filter for making pseudo stereophonic signal, and it results in improved performance at the instant of environment change at far-end side and reduction of the total complexity.

• The Journal of the Acoustical Society of Korea
• /
• v.18 no.5
• /
• pp.3-11
• /
• 1999

The conventional stereophonic acoustic echo cancellers need two adaptive filters to estimate one channel echo signal. Since the two channel signals are strongly correlated, the ESR of the input signals is considerably increased whatever the input signals may be. This causes the slow convergence of the adaptive filter for echo cancellation. To speed up the convergence, the AP algorithm is frequently used for the stereophonic acoustic echo canceller although there isn't a fast version for 2-channel case. The AP algorithm can be approximated with the Gram-Schmidt orthogonalization and a TDL structure. We propose a two channel algorithm for stereophonic acoustic echo canceller with the approximated AP algorithm.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.15 no.8
• /
• pp.689-697
• /
• 1990

This paper proposes ADAPTIVE STOCHASTIC ITERATION ALGORITHM that can be used for the full-duplex digital data transmission over existing twisted-pair cables. The canceller step size changes according to the residual echo level. Thus convergence speed increases by an order of magnitude over the conventional stochastic interation alogorithm. And it can be combined with the sequential type adaptive digital filter that simplifies the conventional echo canceller circuit. Thus it is suitable for LSI implementation. Theoretical analysis on two algorithms is carried out and the algorithm simulation program is developed.

• The Journal of the Acoustical Society of Korea
• /
• v.19 no.7
• /
• pp.34-41
• /
• 2000

This paper proposes a new double-talk detection algorithm which is based on the fuzzy rules, in the adaptive echo canceller of telecommunication system. In this method, the two inputs of the fuzzy inference for detecting double-talk condition are used. One is the cross-correlation coefficient between the error signal and the primary signal which is the summation of the real echo signal and the near-end signal. The other one is the cross-correlation coefficient between the estimation error signal and the primary signal. The fuzzy controller makes a fuzzification for two inputs by the membership functions of trapezoid does the max-min composition using if-then rules. The composed result is defuzzificated by the center gravity method. And by defuzzificated values, the double-talt the echo path variance, and the echo path variance during the double-talk are detected. It is confirmed by computer simulation that this fuzzy double-talk detector is able to estimate the double talk and the echo path variation condition, and even track echo path variation more accurately than the conventional algorithm during the double-talk period.

• Proceedings of the IEEK Conference
• /
• 2001.09a
• /
• pp.927-930
• /
• 2001

본 논문에서는 웨이블렛 변환 알고리듬을 이용 한 음향반향 제거 기법을 제안하였다. 일반적으로 LMS 알고리듬은 간단하지만 그 수렴특성이 고유치 분포율과 탭 수에 의해 결정되는 단점이 있고 MDF (multidelay filter)를 이용한 적응 알고리듬은 블록단위로 수행되는 LMS를 수렴속도 측면에서 보완해주지만 긴 차수일때는 FFT에 비례하여 연산량이 늘어나는 단점이 있다. 본 논문은 웨이블렛의 직교성을 이용한 변환 적응 알고리듬을 음향반향 제거기에 도입하여 그 성능을 비교·평가하였다. 모의실험을 통해 시간영역 LMS, MDF 알고리듬과 비교하여 제안한 알고리듬의 수렴성능이 우수함을 보였다.

• The Journal of the Acoustical Society of Korea
• /
• v.28 no.5
• /
• pp.454-460
• /
• 2009

In this paper, we propose a residual echo and noise reduction scheme by using linear predictor for hands-free telephony applications. The proposed scheme whitens residual echo by the linear prediction during the non double-talk. But whitened residual echo signal still has speech characteristics. In this scheme, the whitened residual echo signal is more whitened by using the power of the linear prediction error signal and the linear predicted signal. After whitening process, near-end speech and ambient noise is present during double-talk but white noise will appear during non double-talk situation. By linearly predicting again the combined signal of the near-end speech and the whitened signal, the ambient noise is removed. Through computer simulation, it is shown that the proposed method performs well at the side of AIC (acoustic interference cancellation).

• Proceedings of the Korea Multimedia Society Conference
• /
• 2001.06a
• /
• pp.124-127
• /
• 2001

핸즈프리 음성 통신에 있어 음향 반향 제거 시스템은 필수적인 요소이다. 본 논문에서는 음향 반향을 효과적으로 제거하기 위한 새로운 시변 적응 상수를 갖는 적응 알고리즘을 제안한다. 제안하는 알고리즘은 최적 필터의 직교 원리를 이용하여 적응 알고리즘의 적응 상수를 시변한다. 직교 원리에 따라 적응 상수를 시변하면 적응 초기에는 적응 필터가 반향 경로를 추정하지 못해서 큰 값을 가지지만 수렴 후에는 적응 필터의 입력 신호와 추정 오차 신호간의 상호 상관도가 작아져서 정상 상태 오차가 작으며 근단화자 등 의 주변 잡음 신호에 적응필터 계수가 오조정되는 것을 줄일 수 있다. 본 논문에서 음향반향 시뮬레이션을 통해서 제안하는 적응 알고리즘이 기존의 방법에 비해서 우수함을 보인다.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.16 no.11
• /
• pp.2388-2396
• /
• 2012

This paper is on an simplified orthogonal projection method which cancel the acoustic echo signals in the stereo acoustic echo canceller. Comparing with the NLMS algorithm which is widely used for simplicity and stability, it shows that this method has the improvement of the convergence performances for signals with the high auto-correlation, and has small computational quantities. To verify the convergence characteristics of the proposed algorithm, we simulated about various input signals. And we compared the results of simulation for this algorithm with the ones for the NLMS algorithm. By these works, it was proved that the stereo acoustic echo canceller adopting the proposed algorithm shows about 3dB more high ERLE than the NLMS algorithm for the white noise signals, and 5dB for the colored voice signals.