There have been numerous studies on the enhancement of the noisy speech signal. In this paper, We propose a new speech enhancement method, that is, a DFF (Dissonant Frequency Filtering) scheme combined with NR (noise reduction) algorithm. The simulation results indicate that the proposed method provides a significant gain in perceptual quality compared with the conventional method. Therefore if the proposed enhancement scheme is used as a pre-filter, the output speech quality would be enhanced perceptually.

Superluminal group velocity in dispersive media has long been controversial. A partial source of confusion seems to be the absence of high precision numerical results concerning the waveform of the transmitted signal. This paper gives the precise waveforms of a causal half-sine-modulated pulse and a triangle-modulated pulse propagating in the Lorentz medium. Thus, the effects of analyticity of signal are clarified, which the analysis using Gaussian pulse cannot. Further, to deepen understanding of the mechanism of superluminal group velocity, we give a network theoretic consideration.

We propose a post-filter (digital filter applied after the correlator) to reduce multiple-access interference (MAI) in the correlator output in asynchronous communications. Optimum filter coefficients are derived for Markov and i.i.d. codes. It is shown that post-filter is not needed for Markov case. Variance of MAI is reduced in i.i.d. codes and it becomes equal to that of Markov codes; thus, both will have the same bit error rate (BER) performance. This post-filter reduces level of MAI in the correlator output for Gold codes as well.

In this paper, we examine the effect of random steering errors on the signal-to-interference-plus-noise-ratio (SINR) at the output of the recently addressed wavelet-based generalized sidelobe canceller (GSC). This new beamformer employs a set of P-regular M-band wavelet bases for the design of the blocking matrix of the GSC. We first carry out a general expression of the output SINR of the GSC with multiple interferers present. With this expression, we then examine the analysis of wavelet-based GSC by expressing the SINR in terms of parameters such as the regularity of wavelet filters, the number of bands of wavelet filters, the length of adaptive weights, and the input signal-to-noise ratio (SNR). Some simulation results verify the analytically predicted performance.

In this paper we consider an approximation method of a formal linearization which transform time-varying nonlinear systems into time-varying linear ones and its applications. This linearization is a kind of a coordinate transformation by introducing a linearizing function which consists of the Chebyshev polynomials. The nonlinear time-varying systems are approximately transformed into linear time-varying systems with respect to this linearizing functions using Chebyshev expansion to the state variable and Laguerre expansion to the time variable. As applications, nonlinear observer and filter are synthesized for time-varying nonlinear systems. Numerical experiments are included to demonstrate the validity of the linearization. The results show that the accuracy of the approximation by the linearization improves as the order of the Chebyshev and Laguerre polynomials increases.

Intensity-noise characteristics of stable multi-mode Fabry-Perot semiconductor lasers are analyzed experimentally and theoretically. Mode-partition noise caused by optical filtering and propagation through optical fibers is investigated by evaluating the relative intensity noise and signal-to-noise ratio. The experimental results indicate that the simplified two-mode analysis provides a good approximation. Suppression of the mode-partition noise by nonlinear gain is experimentally confirmed.

A high-attenuation waveguide filter using electromagnetic bandgap (EBG) substrates is introduced. With a simple design modification on the EBG covers, the waveguide filter produced an almost full Ku-band rejection bandwidth showing better than 20 dB input-to-output isolation from 12.3 to 17.2 GHz.

This paper describes the design of a programmable QAM transceiver for VDSL applications. A 12-b DAC with 64-dB spurious-free dynamic range (SFDR) at 75-MS/s and an 11-b ADC with 72.3-dB SFDR at 70-MS/s are integrated in this complete physical layer IC. A digital IIR notch filter is included in order to not interrupt existing amateur radio bands. The proposed dual loop AGC adjusts the gain of a variable gain amplifier (VGA) to obtain maximum SNR while avoiding saturation. Using several low power techniques, the total power consumption is reduced to 300-mW at 1.8-V core and 3.3-V I/O supplies. The transceiver is fabricated in a 0.18-µm CMOS process and the chip size is 5-mm 5-mm. This VDSL transceiver supports 13-Mbps data rate over a 9000-ft channel with a BER < 10-7.

This paper proposes a new implementation of an adaptive noise canceller based upon a parallel block structure, which aims to raise the processing and convergence rates and to improve the steady-state performance. The procedure is as follows: First, an IIR bandpass filter with a variable center angular frequency using adaptive Q-factor control and two adaptive control signal generators are realized by the parallel block structure. Secondly, a new algorithm for adaptive Q-factor control with parallel block structure is proposed to improve the convergence characteristic. In addition, the steady-state performance of the filter is stabilized by using the variable step size parameter in adaptive control of the center frequency and the speed up of the convergence rate is achieved by adopting a normalized gradient algorithm for adaptive control. Finally, simulation results are given to demonstrate the convergence performance.

In this paper, we propose two adaptive filtering schemes for Stereophonic Acoustic Echo Cancellation (SAEC), which are based on the adaptive projected subgradient method (Yamada et al., 2003). To overcome the so-called non-uniqueness problem, the schemes utilize a certain preprocessing technique which generates two different states of input signals. The first one simultaneously uses, for fast convergence, data from two states of inputs, meanwhile the other selects, for stability, data based on a simple min-max criteria. In addition to the above difference, the proposed schemes commonly enjoy (i) robustness against noise by introducing the stochastic property sets, and (ii) only linear computational complexity, since it is free from solving systems of linear equations. Numerical examples demonstrate that the proposed schemes achieve, even in noisy situations, compared with the conventional technique, (i) much faster and more stable convergence in the learning process as well as (ii) lower level mis-identification of echo paths and higher level Echo Return Loss Enhancement (ERLE) around the steady state.

Classical designs of maximally flat finite impulse response digital filters need to perform inverse discrete Fourier transformation of the frequency responses, in order to calculate the impulse response coefficients. Several attempts have been made to simplify the designs by obtaining explicit formulas for the impulse response coefficients. Such formulas have been derived for digital differentiators having maximal linearity at zero frequency, using different techniques including interpolating polynomials and the Taylor series etc. We show that these formulas can be obtained directly by application of maximal linearity constraints on the frequency response. The design problem is formulated as a system of linear equations, which can be solved to achieve maximal linearity at an arbitrary frequency. Certain special characteristics of the determinant of the coefficients matrix of these equations are explored for designs centered at zero frequency, and are used in derivation of explicit formulas for the impulse response coefficients of digital differentiators of both odd and even lengths.

A new indirect approach for designing low-order linear-phase IIR filters is presented in this paper. Given an FIR filter, we utilize a new Krylov subspace projection method, called the rational Arnoldi method with adaptive orders, to synthesize an approximated IIR filter with small orders. The synthesized IIR filter can truly reflect essential dynamical features of the original FIR filter and indeed satisfies the design specifications. Also, from simulation results, it can be observed that the linear-phase property in the passband is stilled retained. This indirect approach is accomplished using the state-space realization of FIR filters, multi-point Pade approximations, the Arnoldi algorithm, and an intelligent scheme to select expansion points in the frequency domain. Such methods are quite efficient in terms of computational complexity. Fundamental developments of the proposed method will be discussed in details. Numerical results will demonstrate the accuracy and the efficiency of this two-step indirect method.

By modelling the quantization error as additive white noise in the transform domain, Wiener filter is used to reduce quantization noise for DCT coded images in DCT domain. Instead of deriving the spectrum of the transform coefficient, a DPCM loop is used to whiten the quantized DCT coefficients. The DPCM loop predicts the mean for each coefficient. By subtracting the mean, the quantized DCT coefficient is converted into the sum of prediction error and quantization noise. After the DPCM loop, the prediction error can be assumed uncorrelated to make the design of the subsequent Wiener filter easy. The Wiener filter is applied to remove the quantization noise to restore the prediction error. The original coefficient is reconstructed by adding the DPCM predicted mean with the restored prediction error. To increase the prediction accuracy, the decimated DCT coefficients in each subband are interpolated from the overlapped blocks.

High speed core networks with optical fibers have spread widely, but it is still difficult to access the core networks from many rural areas. Synchronous CDMA systems with GEO satellite links are attractive to solve this problem, since they have wide service areas and are suitable for packet-based networks due to their statistically multiplexing effects. Additionally, the synchronous CDMA systems have more effective frequency utilization and power efficiency than asynchronous ones. In the synchronous CDMA systems, transmitted signals from fixed earth stations are required to achieve synchronization with each other. The broadband systems require extremely precise timing control as their bit rates increase. In this paper, we propose a synchronization method for a synchronous CDMA communication system using a GEO satellite and verify the feasibility of Gigachip rate synchronous CDMA systems.

A method is presented for selecting items asked for new users to input their preference rates on those items in recommendation systems based on the collaborative filtering. Optimal item selection is formulated by an integer programming problem and we solve it by using a kind of the Hopfield-network-like scheme for interior point methods.

In this paper we propose a parallel composition based adaptive notch filter for eliminating sinusoidal signals whose frequencies are unknown. The proposed filter which is implemented using second order all-pass filter and a band-pass filter can achieve high convergence speed by using the output of an additional band-pass filter to update the coefficients of the notch filter. The high convergence speed of the proposed notch filter is obtained by reducing an effect that an updating term of coefficient for adaptation of a notch filter significantly increases when the notch frequency approaches the sinusoidal frequency. In this paper, we analyze such effect obtained by the additional band-pass filter. We also present an analysis of a convergence performance of cascaded system of the proposed notch filter for eliminating multiple sinusoids. Simulation results have shown the effectiveness of the proposed adaptive notch filter.

In this paper, we present a design and implementation of the M-channel linear-phase filter banks with unequal-length and same center of symmetry. The filter banks are separated into paraunitary and biorthogonal case. We discuss both cases. A novel filter bank can be regarded as a special class of generalized lapped transform with arbitrary number of channels M. In image coding applications, long basis functions should be used to avoid the blocking artifacts in low-frequency bands, while short basis functions should be used to reduce the ringing artifacts in high-frequency bands. Having the same center of symmetry is suitable for progressive image coder [SPIHT]. Filter banks with such characteristics can be achieved structurally by taking acount of the lattice structure. Finally, several design and image coding examples are shown.

We have proposed a new low-IF transceiver architecture to simultaneously achieve both a small chip area and good minimum input sensitivity. The distinctive point of the receiver architecture is that we replace the complicated high-order analog filter for channel selection with the combination of a simple low-order analog filter and a sharp digital band-pass filter. We also proposed a high-speed convergence AGC (automatic gain controller) and a demodulation block to realize the proposed digital architecture. For the transceiver, we further reduce the chip area by applying a new form of direct modulation for the VCO. Since conventional VCO direct modulation tends to suffer from variation of the modulation index with frequency, we have developed a new compensation technique that minimizes this variation, and designed the low-phase noise VCO with a new biasing method to achieve large PSRR (power-supply rejection ratio) for oscillation frequency. The test chip was fabricated in 0.35-µm BiCMOS. The chip size was 3 3 mm2; this very small area was realized by the advantages of the proposed transceiver architecture. The transceiver also achieved good minimum input sensitivity of -85 dBm and showed interference performance that satisfied the requirements of the Bluetooth standard.

We compared two edge-blending methods for multi-projection displays, elliptic and rectangular blending, by simulating three common situations: (1) an inaccurately estimated calibration parameter, (2) a worn projector lamp, and (3) a shifted viewpoint. We used a two-level-of-detail display including a high-gain rear-projection screen in the simulation to demonstrate an extreme case. The comparisons showed how strongly inaccurate elements affect a composite besides affecting the appearance itself. A subjective assessment was also carried out to obtain the evaluations of actual users. The simulation results showed that in many cases elliptic blending is more effective than rectangular blending.

The accuracy of channel estimation significantly affects the performance of coherent rake receiver in DS-CDMA systems. It is desirable for improved channel estimation to employ a channel estimation filter (CEF) whose bandwidth is adjustable to the channel condition. In this paper, we consider the use of moving average (MA) FIR filters as the CEF since it is simple to implement and can provide relatively good receiver performance. First, we optimize the tap size of the MA FIR CEF so as to minimize the mean squared error of the estimated channel impulse response. For practical applications, we propose a low-complexity adaptive channel estimator (ACE), where the tap size of the MA FIR CEF is adjusted based on the estimated channel condition by exploiting the correlation characteristics of the received pilot signal. Numerical results show that the use of the proposed ACE can provide the receiver performance comparable to that of Wiener CEF without exact a priori information on the operating condition.