By Topic

Acoustics, Speech and Signal Processing, IEEE Transactions on

Issue 3 • Date June 1983

Filter Results

Displaying Results 1 - 25 of 48
  • [Front cover and table of contents]

    Page(s): 0
    Save to Project icon | Request Permissions | PDF file iconPDF (244 KB)  
    Freely Available from IEEE
  • Rational modeling by pencil-of-functions method

    Page(s): 564 - 573
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (820 KB)  

    Pole-zero modeling of signals, such as an electromagnetic-scatterer response, is considered in this paper. It is shown by use of the pencil-of-functions theorem that a) the true parameters can be recovered in the ideal case [where the signal is the impulse reponse of a rational function H(z)], and b) the parameters are optimal in the functional dependence sense when the observed data are corrupted by additive noise or by systematic error. Although the computations are more involved than in all-pole modeling, they are considerably less than those required in iterative schemes of pole-zero modeling. The advantages of the method are demonstrated by a simulation example and through application to the electromagnetic response of a scatterer. The paper also includes very recent and promising results on a new approach to noise correction. In contradistinction with spectral subtraction techniques, where only amplitude information is emphasized (and phase is ignored), we propose a method that a) estimates the noise spectral density for the data frame, and then b) performs the subtraction of the noise correlation matrix from the Gram matrix of the signal. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Comments on "The reconstruction of a multidimensional sequence from the phase or magnitude of its Fourier transform"

    Page(s): 738 - 739
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (259 KB)  

    When one imposes a nonnegativity constraint, one usually can reconstruct a two-dimensional sequence of finite support from the modulus of its Fourier transform using an iterative algorithm, even when file initial estimate is an array of random numbers. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • [Back cover]

    Page(s): c4
    Save to Project icon | Request Permissions | PDF file iconPDF (863 KB)  
    Freely Available from IEEE
  • FIR filter design over a discrete powers-of-two coefficient space

    Page(s): 583 - 591
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (920 KB)  

    FIR digital filters with discrete coefficient values selected from the powers-of-two coefficient space are designed using the methods of integer programming. The frequency responses obtained are shown to be superior to those obtained by simply rounding the coefficients. Both the weighted minimax and the weighted least square error criteria are considered. Using a weighted least square error criterion, it is shown that it is possible to predict the improvement that can be expected when integer quadratic programming is used instead of simple coefficient rounding. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Improving resolution for autoregressive spectral estimation by decimation

    Page(s): 630 - 637
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (808 KB)  

    In this paper we present a method for efficiently improving the resolution of autoregressive spectral estimation algorithms. We derive the exact autoregressive spectrum for K complex sinusoids in additive white noise. From this equation resolution boundaries are constructed which give the resolution in terms of the model order and the signal-to-noise ratio. Simulation results are used to compare the resolution boundaries for decimated and undecimated spectra. Our results demonstrate that decimation by D with a model order M yields the same resolution as a model order MD used with the undecimated signal, and that decimation reduces the computation. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Minimum cross-entropy spectral analysis of multiple signals

    Page(s): 574 - 582
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (960 KB)  

    This paper presents a new information-theoretic method for simultaneously estimating a number of power spectra when a prior estimate of each is available and new information is obtained in the form of values of the autocorrelation function of their sum. One application of this method is the separate estimation of the spectra of a signal and additive noise, based on autocorrelations of the signal plus noise. A derivation of the method from the principle of minimum cross entropy is given, and the method is compared to minimum cross-entropy spectral analysis, of which it is a generalization. Some basic mathematical properties are discussed. Three numerical examples are included, two based oil synthetic spectra, and one based on actual speech data. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Comparison of the characteristics of linear least squares and orthonormal expansion in estimation

    Page(s): 755 - 759
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (616 KB)  

    It is frequently assumed in signal processing applications that expansion of a sequence by a weighted sum of mutually orthonormal sequences yields weighting coefficients that are identical to the estimates of the parameters which maximize the likelihood function if the linear sum of sequences is chosen as a model. Although this may be a valid approximation when signal-to-noise ratios are large, it is not generally the case and may lead to erroneous results when substantial noise exists. This paper explores the relationship between orthonormal expansion and linear least squares estimation. In doing so, the conditions under which orthonormal expansion coefficients are maximum likelihood estimates are identified. Several interesting properties related to both techniques are also revealed. The results are relevant to a wide range of signal processing applications such as the discrete Fourier transform and linear prediction theory and can be extended to non-linear least squares estimation. This should make the results of interest to those involved with the analysis of noisy data. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Desgin and comparative study of some sequential jump detection algorithms for digital signals

    Page(s): 521 - 535
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1312 KB)  

    The purpose of this paper is to analyze the behavior of several jump detection algorithms when applied to the same real data (geophysical signals) and to compare these algorithms from different points of view: complexity, efficiency, robustness, and ability to characterize the detected jumps. Three types of algorithms are investigated: "filtered derivatives" detectors, cumulative sum (cusum) tests, and Willsky's generalized likelihood ratio (GLR) algorithm. A modified version of this last test is elaborated, and a new detector, mixing GLR and cusum tests, is presented. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Iterative time-limited signal restoration

    Page(s): 643 - 650
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (800 KB)  

    The purpose of this paper is to show that time-limited restoration of shift-invariant blurred signals can be done by means of fixed-point solutions of contraction mappings, under rather general conditions for the distortion operator. All our results are valid for multidimensional signals. An application of these results to the iterative extrapolation of band-limited discrete images is shown. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Some results in linear interpolation theory

    Page(s): 746 - 749
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (376 KB)  

    Using a well-known form for the inverse of a symmetric Toeplitz matrix, some results in linear interpolation theory are derived. For an autoregressive process it is shown that interpolation at the mid-point of a data record yields the minimum interpolation error. Also, some results for infinite length interpolators are simply derived. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Deconvolution of nonstationary seismic data using adaptive lattice filters

    Page(s): 591 - 598
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1416 KB)  

    This paper examines the results of the application of two lattice algorithm to the problem of adaptive deconvolution on non-stationary seismic data. A comparative study of the deconvolution performance of the recently proposed gradient lattice and least-squares lattice algorithms is made with the help of experiments on simulated and real seismic data. We show that the gradient lattice algorithm is computationally superior, but it suffers from a possible slow rate of convergence, while the least-squares lattice has better convergence properties and is more robust numerically. We also show that both algorithms can yield equally good deconvolution results with a moderate amount of computation. Finally we indicate that a modified deconvolved output, derived as a linear combination of the forward and backward residuals, improves the performance without involving any additional computational burden. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Performance model for square law detectors followed by accumulators and an ORing device

    Page(s): 759 - 761
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (312 KB)  

    A mathematical model is developed to characterize the performance of square law detectors followed by accumulators and an ORing device. The performance of a Gaussian signal in Gaussian noise is determined for a specified number of channels ORed. The ORing loss is computed as a function of the number of samples in the accumulator preceding the ORing device for various fixed input signal-to-noise ratios (SNR) to the square law detector. It is concluded that the ORing loss decreases with an increase in the number of samples in the accumulator preceding the ORing device. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The effect of correlation on accumulated detection probability

    Page(s): 536 - 540
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (416 KB)  

    A sonar model frequently used for accumulating probability of detection over a succession of independent opportunities is modified by the introduction of correlation between opportunities. Two Guassian models, Markov and unit-memory, are investigated. Monte Carlo calculations of accumulated detection probability and mean hold time are presented. The calculations of mean hold time can be used to estimate the correlation. The loss of accumulated detection probability compared with the independent model can then be estimated from the calculations of accumulated detection probability. Compared to the case of independent opportunities, nonzero correlation imposes a performance penalty which is modest when opportunities are weakly correlated but grows rapidly with increasing correlation. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A new approach to suppression of finite-length pulse interference using modified linear predictor

    Page(s): 622 - 629
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (880 KB)  

    In this paper the problem of pulse interference suppression using prediction technique is considered. It is shown that the conventional predictor cannot be used to suppress a finite-length pulse interference even when the data are ideal, for example, in the absence of a background random process. A new method is presented for the design of an adaptive tapped delay line filter used for finite-length pulse suppression. The method is based on a modified form of linear prediction procedure using a set of artificial state variables given by the product of input samples and their delayed versions. The advantages of this modified linear predictor are demonstrated by two examples, using computer simulation. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Space-time duality in digital filter structures

    Page(s): 550 - 556
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (736 KB)  

    An FIR filter with fixed-point coefficients is viewed as a two-dimensional binary pattern of ones and zeros, as seen by the digital machine. The operator z-1represents shift in one dimension, time, and the operator 2-1represents shift in the other dimension, space. This two-dimensional binary approach results in two digital filter structures, the split in space and merge in time structure and its dual the split in time and merge in space structure. The two structures are optimized to minimize a particular computational complexity measure. Extensions and potential applications of this work are discussed. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Integration and signal-to-noise ratio requirements for a signal processing system containing an ORing device

    Page(s): 650 - 656
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (584 KB)  

    Often it is not feasible to display all the data produced by many signal processing systems. It is desired to reduce the amount of data displayed and to reduce the amount of hardware needed. One technique that provides a reduction of data is ORing, a process where one picks that single channel with the most energy. This paper provides an analysis of a system composed of square law detectors followed by accumulators, an ORing device, and a display. The required input SNR for the ORing device is determined for a detection probability of 0.5 and for false alarm probabilities of 10-3, 10-4, and 10-6. The required input SNR is computed as a function of the number of channels ORed and the number of lines on the display. In addition, the number of samples in the accumulator required to achieve the desired input SNR for the ORing device is determined for various fixed input SNR's to the square law detector. Finally, the ORing loss is computed as a function of the number of channels ORed and the number of lines on the display. It is concluded that the ORing loss increases with an increase in the integration on the display. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Full text access may be available. Click article title to sign in or learn about subscription options.
  • Transform domain LMS algorithm

    Page(s): 609 - 615
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (688 KB)  

    The concept of transform domain adaptive filtering is introduced. In certain applications, filtering in the transform domain results in great improvements in convergence rate over the conventional time-domain adaptive filtering. The relationship between several existing frequency domain adaptive filtering algorithms is established. Applications of the discrete Fourier transform (DFT) and the discrete cosine transform (DCT) domain adaptive filtering algorithms in the areas of speech processing and adaptive line enhancers are discussed. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A stochastic approach to optimal linear digital equalizers

    Page(s): 775 - 780
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (592 KB)  

    This correspondence presents, from the stochastic point of view, a new method for the optimal design of linear digital equalizers using a quadratic performance index. It is well known that a Kalman filter can be used as an equalizer for a digital communication system. If the Kalman filter is applied, then it is useful that an optimal linear equalizer, which minimizes the mean-square error, can be obtained. However, the most disadvantageous point in the Kalman filter is that a great deal of computation is needed to obtain the estimates. Furthermore, the order of the Kalman filter is always equal to the sample number of the impulse response of a transmission channel. Therefore, a high-order equalizer becomes necessary if the sample number is large. Usually, however, it is not necessary to use such a high-order equalizer. Considering this fact, an optimal equalizer is presented under the condition that the order can be assigned arbitrarily by the designer. The proposed equalizer is optimal in the sense that it minimizes the mean-square error subject to this condition. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Minimum-variance and maximum-likelihood recursive waveshaping

    Page(s): 599 - 604
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (608 KB)  

    In this paper, we develop optimal recursive waveshaping filters in the framework of estimation theory and state-variable models. We develop a linear minimum-variance waveshaper and a nonlinear maximum-likelihood waveshaper. Both waveshapers consist of two components: 1) stochastic inversion and 2) waveshaping. The former is performed by means of minimum-variance deconvolution. Simulations are given which illustrate results that can be obtained by both waveshapers. In retrospect, we view the minimum-variance results of this paper as the recursive counterparts to those presented by Treitel and Robinson [14], which are for finite-impulse response waveshaping. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Representation of glottal shape data for signal processing

    Page(s): 766 - 769
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (720 KB)  

    A new representation of glottal-edge data derivable from high-speed laryngeal films is given. The resulting simple curves represent the glottal shape during phonation and are so arranged that various one-and two-dimensional discrete Fourier transforms can be used to test glottal symmetry and vocal fold vibration rates and phases. Possible discriminant features are suggested and illustrated. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Nonstationary spectral modeling of voiced speech

    Page(s): 664 - 678
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1600 KB)  

    The main purpose of this paper is to present a novel model for voiced speech. The classical model, which is being used in many applications, assumes local stationarity, and consequently imposes a simple and well known line structure to the short-time spectrum of voiced speech. The model derived in this paper allows for local non-stationarities not only in terms of pitch perturbations, but in terms of vocal tract variations as well. The resulting structure of the short-time spectrum becomes more complex, but can still be interpreted in terms of generalized lines. The proposed model supports new forms of spectral prediction, which can be put to advantage in speech coding applications. Experimental results are presented supporting the validity of both the model itself and the prediction relationships. Finally, a new class of speech coders, denoted harmonic coders, based on the presented model, is proposed, and a specific implementation is presented. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A note on stability and lattice filter relations

    Page(s): 770 - 772
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (328 KB)  

    In this note we draw attention to an earlier result by Marden [2] for establishing the stability of linear discrete-time systems. The stability algorithm then provides a simple constructive realization of stable filters in lattice form. The class of unstable filters which cannot be represented in lattice form is also partially characterized. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A particular example of phase unwrapping using noisy experimental data

    Page(s): 742 - 744
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (352 KB)  

    Obtaining a smooth continuous function of phase from a distorted, sampled set of phase measurements is addressed. Provided the smooth continuous function can be written as a power series, all the important coefficients can be derived from phase difference measurments which have proved to be less affected by phase noise than the phase measured directly. This simple technique is shown to have advantages in some experimental configurations. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.

Aims & Scope

This Transactions ceased production in 1990. The current retitled publication is IEEE Transactions on Signal Processing.

Full Aims & Scope