By Topic

Acoustics, Speech and Signal Processing, IEEE Transactions on

Issue 2 • Date April 1984

Filter Results

Displaying Results 1 - 25 of 38
  • [Front cover and table of contents]

    Page(s): 0
    Save to Project icon | Request Permissions | PDF file iconPDF (236 KB)  
    Freely Available from IEEE
  • Comments and corrections on "On the eigenvectors of symmetric Toeplitz matrices"

    Page(s): 440 - 441
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (234 KB)  

    The necessary and sufficient condition for a matrix to be doubly symmetric, as stated in the above paper, is questioned, it is proved that the condition under consideration is necessary but not sufficient. Somecounterexamples are provided to substantiate the claim. In fact, what appears to be a new class of matrices possessing some interesting properties is discussed. 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 (708 KB)  
    Freely Available from IEEE
  • Maximum entropy power spectrum estimation with uncertainty in correlation measurements

    Page(s): 410 - 418
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (928 KB)  

    The purpose of this paper is to present a multidimensional MEM algorithm, valid for nonuniformly sampled arrays, which satisfies a "correlation-approximating" constraint. To this end, the correlation matching equality constraints of the usual MEM are replaced by a single inequality constraint whose form is based on a measure of the noise in the given autocovariance function (ACF). In this way, one can incorporate into the model knowledge of the noisy nature of the "given" ACF, since the "given" ACF is usually estimated from the samples of the wavefield. Specifically, the covariance matrix of the correlation estimates is used in a quadratic form that weights the difference between the "given" ACF and the one matched by the power spectrum. The maximization of entropy under this inequality constraint leads, ultimately, to a steepest-descent algorithm. The algorithm has been tested with 1-D synthetic data representing multiple sinusoids buried in additive white noise. The performance of this modified MEM algorithm is compared to a traditional MEM algorithm for extendible ACF's and for different SNR's. Examples of the MEM spectrum are given for the case of nonextendible ACF's. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Quadrature mirror filter design in the time domain

    Page(s): 353 - 361
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (672 KB)  

    A new technique for designing quadrature mirror filters is described. The formulation, carried out in the time domain, is shown to result in an optimization problem requiring minimization of a quartic multinomial. An iterative solution is suggested which involves (computation of) the eigenvector of a matrix with a dimensionality equal to one half the number of filter taps. Our experiments show that convergence to the optimum tap weights is stable, and the accuracy of the final solution is limited only by the accuracy of the eigenvalue-eigenvector routine. As in an earlier technique, the user can specify the stop: band frequency the relative weights of the passband ripple energy and the stopband residual energy, and, of course, the number of filter taps. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • System identification of the speech production process based on a state-space representation

    Page(s): 252 - 262
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1000 KB)  

    In this paper a new method for analysis of speech signals based on a representation in terms of an autoregressive moving-average (ARMA) process with variable order is presented. The process of speech production is approximated by an ARMA process described by a state equation and an observation equation. An algorithm for the minimal realization of the model of speech production process is proposed. The state transition matrix that characterizes the autoregressive (AR) scheme is determined to minimize the mean squared error in the estimation of the speech signal by Kalman filtering. The state transition matrix and the autocorrelation function of the signal are used to determine parameters of the moving-average (MA) scheme. Application to spectral density estimation of speech corrupted by additive noise is also discussed. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Comparison of the deskewed short-time correlator and the maximum likelihood correlator

    Page(s): 285 - 294
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (776 KB)  

    Cross correlation of broad-band Doppler-scaled signals in noise is degraded by the Doppler-induced loss of signal coherence. The maximum likelihood correlator (MLC) compensates for Doppler scaling by time companding one of the two input waveforms, providing optimal differential time delay and relative time companding estimates. The price paid in using the MLC is significant implementation complexity. This paper compares a more efficient suboptimal technique, the deskewed short-time correlator (DSTC), with the MLC. Expressions for the output SNR of the MLC and DSTC are derived, as are expressions for DSTC accuracy of estimating differential time delay and relative time companding. Numerical evaluation of these expressions for low-pass white signal and noise shows that DSTC output SNR is only fractional dB below optimum, and estimation accuracy approaches the appropriate Cramer-Rao lower bounds. Comparison of implementation complexity indicates that the DSTC requires similar storage while reducing the number of computations by more than an order of magnitude. The deskewed short-time correlator is thus shown to provide implementational simplicity and near-optimal performance. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On the computation of autocorrelation using polynomial transforms

    Page(s): 448 - 450
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (328 KB)  

    Polynomial transforms (PT's) are known to be efficient for computing convolutions. In this correspondence, the use of polynomial transforms for computing autocorrelation has been considered. A method has been described for using polynomial transforms instead of the FFT in the Rader's algorithm for autocorrelation, and the number of arithmetic operations compared; it is found that the use of PT's does not result in the expected computational advantage over the FFT method. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Discrete Fourier transform algorithms for real valued sequences

    Page(s): 390 - 396
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (608 KB)  

    It is well known that the discrete Fourier transformation (DFT) of a real-valued sequence contains some redundancies. More precisely, approximately half of the DFT coefficients suffice to completely determine the DFT. In this paper, it is shown that the choice of the set of (nonredundant) DFT coefficients to be calculated affects the efficiency of the resulting algorithm. One especially interesting choice is discussed in detail for the case of mixed radix-(2, 3) DFT algorithms. Algorithms for the DFT calculation of both one and more dimensional real-valued arrays are discussed. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A two-step bilinear filtering approximation

    Page(s): 344 - 352
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (760 KB)  

    A new approximation technique to a certain class of non-linear filtering (signal processing) problems is considered here. The method is based on an approximation of a nonlinear, partially observable system by a bilinear model with fully observable states. The filter development proceeds from the assumption that the unobservable states are conditionally Gaussian with respect to the observation initially. The method is shown to be promising for real-time communication and sonar applications as demonstrated by computer simulations. Moreover, some of the traditional techniques evolve as special cases of this methodology. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Another simplification in multidimensional stability tests

    Page(s): 453 - 455
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (224 KB)  

    The main computational burden of checking stability of a multidimensional system is to check whether a multivariable polynomial has zeros on the distinguished boundary of a certain region of analyticity. A transformation is performed yielding an expression which has zeros on the distinguished boundary if and only if the original polynomial has such zeros. In some cases, especially where a linear dependence of certain auxiliary variables defined in the text can be presented in terms of nonnegative integers, the transformed expression is considerably simpler, and its test is easier to perform than the original one. Examples are provided. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Criteria for the absence of limit cycles in two-dimensional discrete systems

    Page(s): 432 - 434
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (344 KB)  

    Sufficient conditions for the nonexistence of zero input limit cycles in two-dimensional systems which contain a finite number of memoryless nonlinearities are presented. Two different conditions are given. The first one is based on the frequency domain representation of the linear part of the discrete system, and it is shown to be more general than the one already presented in [1]. The second condition is formulated using the state space representation and is based on the properties of quasidominant matrices and on partial results on the 2-D Lyapunov equation. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A unified approach to noniterative linear signal restoration

    Page(s): 403 - 409
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (776 KB)  

    The main goal of this paper is to describe a unified framework for several noniterative algorithms for signal extrapolation reported in the literature. This unification is achieved through integral equation and Hilbert space theories. The importance of this unification is that we can bring to bear the vast body of techniques in these theories to the solution of the extrapolation problem. We will show that the so-called two-step procedures for extrapolation with different underlying models can be unified by means of noniterative algorithms for solving optimization problems in Hilbert spaces. In particular, we show that two-step procedures under a discrete-continuous model [1], [2] belong to a general class of well-known algorithms for solving linear integral equations of the first kind: giveng(x), x in Afindf(t), t in Omegasuch thatg(x) = intmin{Omega} K(x,t)f(t)dt, x in A(1) In addition, we will show that the prolate spheroidal expansion technique is also a special case of the well-known Picard's eigenfunction procedure for the general integral problem (1). This theoretical unification, together with that presented in [3] for iterative least-squares algorithms, demonstrates that most of the well-known procedures for band-limited extrapolation can be considered as special cases of standard techniques in integral equations and operator theory. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Signal estimation from modified short-time Fourier transform

    Page(s): 236 - 243
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1656 KB)  

    In this paper, we present an algorithm to estimate a signal from its modified short-time Fourier transform (STFT). This algorithm is computationally simple and is obtained by minimizing the mean squared error between the STFT of the estimated signal and the modified STFT. Using this algorithm, we also develop an iterative algorithm to estimate a signal from its modified STFT magnitude. The iterative algorithm is shown to decrease, in each iteration, the mean squared error between the STFT magnitude of the estimated signal and the modified STFT magnitude. The major computation involved in the iterative algorithm is the discrete Fourier transform (DFT) computation, and the algorithm appears to be real-time implementable with current hardware technology. The algorithm developed in this paper has been applied to the time-scale modification of speech. The resulting system generates very high-quality speech, and appears to be better in performance than any existing method. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Estimation of coherence and time delay with ARMA models

    Page(s): 295 - 303
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (808 KB)  

    Magnitude squared coherence (MSC) and time delay are two important quantities needed for passive detection and localization of a radiating source using several sensors. This paper presents a novel approach to estimating the MSC and time delay from the outputs of two sensors. It first models the MSC as the product of the transfer functions of two autoregressive moving-average (ARMA) filters. The ARMA coefficients are then determined by a least squares, matrix pseudoinverse solution to the estimation equation. Time delay estimation follows as a byproduct, since both phase and MSC estimates are now available. Simulation results obtained from this modeling approach are compared against those estimated from the traditional periodograms. In the majority of cases studied, the ARMA approach appears to be superior, giving smaller bias and variances in both MSC and time delay estimations. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Comparison of least-squares and stochastic gradient lattice predictor algorithms using two performance criteria

    Page(s): 441 - 445
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (528 KB)  

    The least-squares (LS) and stochastic gradient (SG) lattice prediction algorithms are compared using two different performance criteria. These are a) output mean squared error and b) the accuracy of the autoregressive, spectral estimate obtained from the mean values of the lattice coefficients, assuming a stationayinput. It is found that the second performance criterion is more sensitive than the first. This "spectral" performance criterion is a measure of the accuracy of the estimatcd autoregressive model coefficients. Bias in the LS and SG coefficient estimates can cause significant deviation of the asymptotic spectral estimates from the actual input spectrum: The similarly between the LS and SG lattice algorithms enables comparative simulations with analogous initial conditions and exponential weighting constants. For both types of comparisons, the LS algorithm offers a modest performance improvement over the SG algorithms simulated. This improvement is more noticeable when the input is highly correlated. It is also found that slight changes in the SG lattice algorithm may significantly affect its performance. 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.
  • Full text access may be available. Click article title to sign in or learn about subscription options.
  • Digital SAR processing using a fast polynomial transform

    Page(s): 419 - 425
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1008 KB)  

    In this paper, a new digital processing algorithm is developed for producing images from SEASAT synthetic aperture radar (SAR) data. For reasons of ecomony, SEASAT presently operates in the stripmapping mode rather than the potentially higher resolution spotlight mode [1]. This algorithm applies the fast polynomial transform (FPT) for computing the two-dimensional cyclic correlation of the raw echo data with the impulse response of a point target. This two-dimensional correlation algorithm solves the range migration problem. It is demonstrated that this SAR processing technique is readily implemented on a general purpose computer. Actual results of SEASAT SAR imagery are given in this paper. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • State-space realizations of fractional-step delay digital filters with applications to array beamforming

    Page(s): 371 - 380
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (840 KB)  

    A general approach to the design of single-rate fractional-step delay (FSD) filters for array beamforming, based on a state-space formulation, is presented. This approach is applicable to both FIR and IIR implementations. In this approach, state-space realizations of FSD filters can be derived from state-space realizations of parent interpolating filters. The realizations so produced have the property that the fractional-step delay is determined solely by the B (input coupling) matrix. The A (system) and C (output coupling) matrix operations are independent of the fractional-step delay, and thus can be shared by all array element channels. All FSD beamformers are subject to spurious spatial response lobes generated by aliasing in the interpolating filter. However, these spurious lobes can be effectively suppressed by appropriate design of the magnitude response of the interpolating filter. If the magnitude response of the interpolating filter is chosen so that the spurious sidelobes are effectively suppressed, then although the phase response of the interpolating filter will affect the temporal response of the system, it will have no effect on the spatial response of the array; i.e., it is not necessary to have linear phase response in the interpolating filter to obtain ideal beam patterns. Comparisons of computation rates for FIR and IIR implementation of FSD beamformers are presented. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The use of a one-stage dynamic programming algorithm for connected word recognition

    Page(s): 263 - 271
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (992 KB)  

    This paper is of tutorial nature and describes a one-stage dynamic programming algorithm for file problem of connected word recognition. The algorithm to be developed is essentially identical to one presented by Vintsyuk [1] and later by Bridle and Brown [2] ; but the notation and the presentation have been clarified. The derivation used for optimally time synchronizing a test pattern, consisting of a sequence of connected words, is straightforward and simple in comparison with other approaches decomposing the pattern matching problem into several levels. The approach presented relies basically on parameterizing the time warping path by a single index and on exploiting certain path constraints both in the word interior and at the word boundaries. The resulting algorithm turns out to be significantly more efficient than those proposed by Sakoe [3] as well as Myers and Rabiner [4], while providing the same accuracy in estimating the best possible matching string. Its most important feature is that the computational expenditure per word is independent of the number of words in the input string. Thus, it is well suited for recognizing comparatively long word sequences and for real-time operation. Furthermore, there is no need to specify the maximum number of words in the input string. The practical implementation of the algorithm is discussed; it requires no heuristic rules and no overhead. The algorithm can be modified to deal with syntactic constraints in terms of a finite state syntax. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A spectral matching technique for ARMA parameter estimation

    Page(s): 338 - 343
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (552 KB)  

    An iterative frequency domain technique is presented for estimating AR-plus-noise and autoregressive moving-average (ARMA) parameters. The technique is based on minimizing the error between the sample power spectrum and a spectral model. The variance of the estimation error is shown to be close to the Cramer-Rao bound for some examples. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Maximum likelihood spectral estimation and its application to narrow-band speech coding

    Page(s): 243 - 251
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (952 KB)  

    Itakura and Saito [1] used the maximum likelihood (ML) method to derive a spectral matching criterion for autoregressive (i.e., all-pole) random processes. In this paper, their results are generalized to periodic processes having arbitrary model spectra. For the all-pole model, Kay's [2] covariance domain solution to the recursive ML (RML) problem is cast into the spectral domain and used to obtain the RML solution for periodic processes. When applied to speech, this leads to a method for solving the joint pitch and spectrum envelope estimation problems. It is shown that if the number of frequency power measurements greatly exceeds the model order, then the RML algorithm reduces to a pitch-directed, frequency domain version of linear predictive (LP) spectral analysis. Experiments on a real-time vocoder reveals that the RML synthetic speech has the quality of being heavily smoothed. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On the application of embedded digit training to speaker independent connected digit recognition

    Page(s): 272 - 280
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1048 KB)  

    In recent years, several algorithms have been proposed for recognizing a string of connected words (typically digits) by optimally piecing together reference patterns corresponding to the words in the string. Although the algorithms differ greatly in details of implementation, storage requirements, etc., they all have essentially the same performance in that their ability to match the unknown string is related to how well words spoken in isolation can match their counterparts in connected speech. For low rates of articulation (i.e., about 100-130 words per minute) the performance of such connected word recognition systems is excellent. However, as the articulation rate approaches that of continuous discourse (180-300 words per minute) the performance of such connected word recognizers falls dramatically. To partially alleviate these problems a modified training procedure was devised in which multiple versions of each reference word were used. The multiple versions included an isolated form for each word, and 2 versions of the word extracted from the middle of 3 word sequences. One of these embedded reference patterns represented a noncontextual token of the word (i.e., spoken in a format where the words on either side had minimal effect on the acoustic properties at the boundaries), and the second represented a highly contextual token of the word. It was shown that a training algorithm could be devised to obtain these embedded reference tokens, and that when using the multiple reference patterns, the performance in a speaker trained system was greatly improved at faster talking rates. In this paper we show how the embedded training technique can be extended to the case of speaker, independent connected word recognizers. In particular, we show that improved recognition performance on connected digit strings is obtained by using standard clustering procedures on the embedded tokens to give a speaker-independent embedded reference set. We also show that the use of the K-nearest neighbor (KNN) rule leads to additional real improvements in performance for recognizing strings of connected digits. A discussion of the types of problems that remain is given. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • 2-D spectral estimation combining parametric estimation of background and maximum entropy estimation

    Page(s): 220 - 228
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1064 KB)  

    A new algorithm for the estimation of the two-dimensional (2-D) power spectrum from finite correlation data is proposed. This method is based on the combination of parametric estimation of a background power spectral component and maximum entropy estimation for point-like spectral components. For the discrimination of the background power spectrum, we adopt a new measure which reflects the information of the point-like spectral components. A practical iterative algorithm is presented and the effectiveness of the method is demonstrated through several numerical examples. 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