Browsing by Author "Hsieh, Shih-Fu"

Now showing 1 - 11 of 11

A temporal masking technique and its performace analysis for audio watermarking
(2007)
Chou, Shuang-An
;
Hsieh, Shih-Fu
;
Li, Ko-Chiang
;
電信工程研究所
;
Institute of Communications Engineering
Echo hiding embeds watermark information bit in the delay time. For imperceptibility, we propose a temporal masking technique to match the psycho acoustic model. Based on the spreading property of PN sequences and expansions of logarithm and binomial, two asymptotic detection gains are derived depending on the echo gain is small or large. This performance analysis of the detection gain explains the improvement with an increment of a small echo gain and the degradation when a large echo gain is raised even higher. Computer simulation confirms our analysis.
Common-Acoustic-Poles and Zeros Modeling Using Jointly Balanced Model Truncation
(2009)
Liu, C. J.
;
Hsieh, Shih-Fu
;
電子工程學系及電子研究所
;
Department of Electronics Engineering and Institute of Electronics
We propose a new technique for common-acoustic-poles and zeros (CAPZ) modeling. In 3-D sound processing, head-related transfer functions (HRTF's) record channels' characteristics from different locations to both ears. Compared to conventional pole/zero models, it is more efficient to model HRTF's by using CAPZ modeling. Based on balanced model truncation (BMT), jointly BMT is proposed to effectively model a group of HRTF's using common-poles IIR filters. The simulation result shows the proposed method is superior to the previous Prony and Shanks methods.
Fast Converging Nonlinear Echo Cancellation Based On Optimum Step Size
(2009)
Shih, Chia-Sheng
;
Hsieh, Shih-Fu
;
電子工程學系及電子研究所
;
Department of Electronics Engineering and Institute of Electronics
Adaptive Volterra filter is generally known to track nonlinear echo path at slow convergence rate. An optimum time- and tap- variant step-size for second-order Volterra filter is proposed to speed up its convergence rate. The step-size is derived based on the MMSE criterion. As the optimum step-size needs to know the real echo path coefficient, an exponential echo-path model is proposed for practical implementations. Computer simulations are provided to validate the proposed algorithms.
Nonlinear Acoustic Echo Cancellation Based on Higher Order Correlation
(2009)
Jiang, Guan-Yu
;
Hsieh, Shih-Fu
;
Jenq, Jang-Chyuan
;
電子工程學系及電子研究所
;
Department of Electronics Engineering and Institute of Electronics
In order to compensate the nonlinearly distorted echo in hands-free telephone, a memoryless power-series polynomial nonlinear acoustic echo canceller (AEC) can be used. Conventional nonlinear AEC employs a NLMS adaptive algorithm which can fail to converge in case of doubletalk or a very noisy environment. In this paper we propose a higher order correlation algorithm using a white Gaussian training signal. With comparable complexity as NLMS, it is most effective at small SNR. Computer simulations demonstrate that the proposed algorithm has a smaller steady-state echo and it is also very robust to background noise.
Nonlinear acoustic echo cancellation using orthogonal polynomial
(2006)
Jiang, Guan-Yu
;
Hsieh, Shih-Fu
;
電信工程研究所
;
Institute of Communications Engineering
In order to compensate the nonlinear distortion in the hands -free telephones or teleconferencing system, a memoryless power- series-based polynomial NLMS adaptive filter can be used to cancel nonlinear acoustic echo. Conventional polynomial model employs a power-series expansion. In this paper we propose an orthogonal polynomial adaptive filter to accelerate the convergence rate of the nonlinear adaptive filter. The convergence rates of residual echo power for both power-series and orthogonal polynomials are derived analytically. Computer simulations justify our analysis and show the improved performance of the proposed nonlinear acoustic echo canceller.
Phase-direct channel estimation for space-time OFDM
(2006)
Hsieh, Shih-Fu
;
Wu, T. Y.
;
電信工程研究所
;
Institute of Communications Engineering
Space time (ST) OFDM can achieve high data rate transmissions with diversity gains. However, channel estimation is necessary at the receiver end to recover the signal. To further improve the subspace-based estimated channel, we introduce phase direct (PD) in the ST OFDM system. PD is originally used in OFDM system, and not suited in ST OFDM. Hence, we exploit the transmitted data's spatial and temporal correlation to develop a sum-difference square algorithm to solve this problem for ST OFDM.
Phase-direct channel estimation for space-time OFDM
(2006)
Hsieh, Shih-Fu
;
Wu, T. Y.
;
電信工程研究所
;
Institute of Communications Engineering
Space time (ST) OFDM can achieve high data rate transmissions with diversity gains. However, channel estimation is necessary at the receiver end to recover the signal. To further improve the subspace-based estimated channel, we introduce phase direct (PD) in the ST OFDM system. PD is originally used in OFDM system, and not suited in ST OFDM. Hence, we exploit the transmitted data's spatial and temporal correlation to develop a sum-difference square algorithm to solve this problem for ST OFDM.
Phase-direct channel estimation for space-time OFDM
(2006-01-01)
Hsieh, Shih-Fu
;
Wu, T. Y.
;
電信工程研究所
;
Institute of Communications Engineering
Space time (ST) OFDM can achieve high data rate transmissions with diversity gains. However, channel estimation is necessary at the receiver end to recover the signal. To further improve the subspace-based estimated channel, we introduce phase direct (PD) in the ST OFDM system. PD is originally used in OFDM system, and not suited in ST OFDM. Hence, we exploit the transmitted data's spatial and temporal correlation to develop a sum-difference square algorithm to solve this problem for ST OFDM.
Robust 3-D crosstalk canceller design
(2007)
Huang, Chun-Rong
;
Hsieh, Shih-Fu
;
電信工程研究所
;
Institute of Communications Engineering
It is well known that the effectiveness of loudspeaker-based 3D sound systems is critically dependent on the listener known position. Traditional approaches try to find the optimum loudspeaker location with least sensitivity to the listener movement, or to use multi-loudspeaker to reduce the effect of head movement. In this paper, we propose a region-based crosstalk canceller with only a pair of loudspeakers to provide increased robustness to perturbations such as head movement and head shapes. In the design of the crosstalk canceller, we propose a new common-pole model structure with reduced complexity.
Robust crosstalk cancellation for 3D sound using multiple loudspeakers
(2007)
Su, Da-Jhuang
;
Hsieh, Shih-Fu
;
電信工程研究所
;
Institute of Communications Engineering
Crosstalk cancellation is needed for 3D sound playback over two loudspeakers. A critical problem is the degraded spatial hearing due to head movements. Previous research has shown that multiple loudspeakers arrangement is more robust to such perturbation than two loudspeakers setup. In this paper, we will propose novel simplified shuffler structures of crosstalk cancellers for multiple loudspeakers. Their computational costs are lower than conventional setup. Perturbation analysis and computer simulation will demonstrate the advantages of using multiple loudspeakers.
Two-Staged Nonlinear Acoustic Echo Cancellation and Its Convergence Analysis
(2009)
Li, Ko-Chiang
;
Hsieh, Shih-Fu
;
Jenq, Jang-Chyuan
;
電子工程學系及電子研究所
;
Department of Electronics Engineering and Institute of Electronics
Hand-free telephone system commonly drives the power-amplification and loudspeaker into saturated nonlinear region such that performance of conventional acoustic echo cancellation (AEC) is degraded. A computationally efficient cascade AEC model is used to include a memoryless piecewise linear (PWL) processor and a linear filter. In order to overcome the start-up divergence problem encountered in nonlinear AEC adaptation, a two-stage update scheme is adopted that starts with the linear filter update, and then joint adaptation of PWL and linear coefficients follows. Theoretical convergence analysis and stability criterion will be derived and validated by the computer simulation.