By Topic

Acoustics, Speech and Signal Processing, IEEE Transactions on

Issue 8 • Date Aug. 1988

Filter Results

Displaying Results 1 - 22 of 22
  • Effect of course maneuvers on bearings-only range estimation

    Page(s): 1193 - 1199
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1524 KB)  

    The Cramer-Rao lower bound (CRLB) is used to examine the relative merits of tracking strategies in bearings-only target motion analysis. A formula is derived for the expected range error from a maximum-likelihood estimator (MLE) for the long-range situation. Monte Carlo simulations, with respect to the measurement noise, are used to show that for small amounts of noise the variance of the MLE range estimate follows the CRLB. It is also shown that there are often broad ranges of course changes for which acceptable ranging solutions can be found and large intervals of ownship course changes that should be avoided for target-motion analysis (TMA) range estimates.<> View full abstract»

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

    Page(s): 1223 - 1235
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1535 KB)  

    A speech model, referred to as the multiband excitation model, is presented. In this model the band around each harmonic of the fundamental frequency is declared voiced or unvoiced. Estimation methods for the parameters of the model are developed and methods to synthesize speech from the model parameters are described. To illustrate a potential application of the speech model, an 8 kb/s vocoder is developed and its performance is evaluated. Both informal listening and intelligibility tests show that the vocoder has very good performance both in speech quality and intelligibility, particularly for noisy speech.<> View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Robust image modeling techniques with an image restoration application

    Page(s): 1313 - 1325
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1424 KB)  

    A robust parameter-estimation algorithm for a nonsymmetric half-plane (NSHP) autoregressive model, where the driving noise is a mixture of a Gaussian and an outlier process, is presented. The convergence of the estimation algorithm is proved. An algorithm to estimate parameters and original image intensity simultaneously from the impulse-noise-corrupted image, where the model governing the image is not available, is also presented. The robustness of the parameter estimates is demonstrated by simulation. Finally, an algorithm to restore realistic images is presented. The entire image generally does not obey a simple image model, but a small portion (e.g. 8×8) of the image is assumed to obey an NSHP model. The original image is divided into windows and the robust estimation algorithm is applied for each window. The restoration algorithm is tested by comparing it to traditional methods on several different images View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On statistical identification of a class of linear space-invariant image blurs using non-minimum-phase ARMA models

    Page(s): 1360 - 1363
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (384 KB)  

    An optimal statistical parameter estimation technique was previously presented by the authors (ibid., vol.ASSP-34, p.963-72) for the identification of a class of unknown linear systems that can be modeled by a zero-phase finite impulse response filter. This technique is now extended to a more general case using a spectrally equivalent minimum-phase (SEMP) representation of the system. The original zero-phase system parameters are computed from the SEMP model parameters View full abstract»

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

    Page(s): 1236 - 1243
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (596 KB)  

    In an embedded delta modulation (EDM) encoder, a supplemental pulse-code-modulation (PCM) encoder processes the error signal of a delta modulator. The PCM and DM bit streams are multiplexed and then transmitted. The decoder adds and then filters the analog versions of the separately decoded PCM and DM signals. To enhance the efficiency of EDM, it is desirable to filter and decimate the DM error signal prior to PCM encoding. Corresponding filtering and decimation of the decoded DM signal is then required. The theory of both full-rate and reduced-rate EDM, which include filtering and decimation, is presented. The performance of these codes are those of compared to PCM, DM, and two embedded codes, embedded-differential PCM and multistage DM. It is found that reduced-rate EDM has a higher signal-to-noise ratio than DM and PCM. It is comparable in performance to multistage DM but simpler to implement. The signal-to-noise ratio of EDM is lower than that of embedded DPCM. However, EDM is simpler to implement, and the signal-to-noise ratio penalty is less than 2.6 dB over a range of transmission rates of practical interest View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Stack filters and the mean absolute error criterion

    Page(s): 1244 - 1254
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1044 KB)  

    A method to determine the stack filter which minimizes the mean absolute error between its output and a desired signal, given noisy observations of this desired signal, is presented. Specifically, an optimal window-width-b stack filter can be determined with a linear program with O(b2b) variables. This algorithm is efficient since the number of different inputs to a window-width-b filter is Mb if the filter has M-valued input and the number of stack filters grows faster than 2 raised to the 2b/2 power. It is shown that optimal stack filtering under the mean-absolute-error criterion is analogous to optimal linear filtering under the mean-squared-error criterion: both linear filters and stack filters are defined by superposition properties, both classes are implementable, and both have tractable procedures for finding the optimal filter under an appropriate error criterion View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Adaptive trimmed mean filters for image restoration

    Page(s): 1326 - 1337
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1216 KB)  

    An adaptive smoothing filter is proposed for reducing noise in digital signals of any dimensionality. The adaptive procedure is based on the selection of an appropriate inner or outer trimmed mean filter according to local measurements of the tail behavior (impulsivity) of the noise process. The set of trimmed means used provides robustness against a wide range of noise possibilities ranging from very shallow tailed to very heavy tailed. A Monte Carlo analysis using a family of generalized exponential distributions supports the choice of the trimmed mean selected for measured values of an impulsivity statistic. The assumption underlying the definition of the filter is that the signal to be filtered is locally smoothly varying, and that the noise process is uncorrelated and derives from an unknown, unimodal symmetric distribution. For image-processing applications, a second statistic is used to mark the location of abrupt intensity changes, or edges; in the vicinity of an edge, the trend-preserving median filter is used. Since the impulsivity and edge statistics used in defining the adaptive filter are both functions of order statistics, the extra computation required for their calculation is minimal. Examples are provided of the filter as applied to images corrupted by a variety of noises View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Image representation by one-bit Fourier phase: theory, sampling, and coherent image model

    Page(s): 1292 - 1304
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1596 KB)  

    The problem of recovering an image from its Fourier-transform phase quantized to one bit, or, equivalently, finding the locations of the zero crossings of the real part of the Fourier transform, is addressed. Theoretical results are presented that set an algebraic condition under which real zero crossings uniquely specify a band-limited image. It is then shown that sampling in the frequency domain presents a major obstacle to obtaining good reconstruction results. The one-bit Fourier phase reconstruction problem is then considered when the original image is coherent, i.e. the image phase is random and highly uncorrelated. Examples are given which demonstrate that the information loss produced by frequency sampling is not as severe as that in the nonclassical problem. Motivated by digital phase-only holograms, a known random diffuser is used as the image phase and a well-known iterative reconstruction procedure is extended to incorporate the knowledge of the image phase at each stage of the iteration. This reconstruction method produces good image quality by using a few iterations, unlike its noncoherent counterpart View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On resolving two-dimensional sinusoids in white noise using different spectral estimates

    Page(s): 1338 - 1350
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (888 KB)  

    Several well-known spectrum estimates are examined in terms of their performance in resolving two closely located sinusoids in white noise. Closed-form expressions for the estimates, using the different approaches and models, are derived and used to study the resolution question. The notions of 3-dB contour and single-peak area (SPA) are introduced and used to compare the different models. The harmonic mean of two single-quadrant linear prediction models is shown to be particularly promising as a high-resolution estimate View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Focussing matrices for coherent signal-subspace processing

    Page(s): 1272 - 1281
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (800 KB)  

    A class of focussing matrices is proposed for use in the coherent signal-subspace method (CSM) (H. Wang and M. Kaveh, ibid., vol.ASSP-33, Aug. 1985). When the directions-of-arrival of wideband sources fall into more than one group of one beamwidth each, this class leads to performance substantially better than those suggested in previous studies on direction-of-arrival estimation. New insight into the structures of various focussing matrices and their effect on the performance of CSM is presented. The performance of CSM is compared for several classes of these matrices on the simulations as well as relative sufficiency analyses View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Optimum FIR array filters

    Page(s): 1211 - 1222
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (776 KB)  

    An optimum finite impulse response (FIR) filter bank is derived that recovers a desired signal from an array of sensors in the presence of random sensor noise and one or more coherent interfering noise sources. The required filter order is determined by the relative delays between the signal and coherent noise at each sensor, and is independent of the time duration of either the signal or coherent noise. Usually, the presence of only three sensors is sufficient to be able to eliminate the effect of coherent noise sources provided the filter order M is sufficiently large View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Bias and resolution of the MUSIC and the modified FBLP algorithms in the presence of coherent plane waves

    Page(s): 1351 - 1352
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (164 KB)  

    A study of the characteristics of the MUSIC and the modified forward-backward linear prediction (MFBLP) methods of angle-of-arrival estimation, in terms of their sample bias and resolution in the fully correlated signal environment, is presented. Using Monte Carlo simulations, it is shown that the MFBLP algorithm has better capability in resolving two fully correlated signals than the spatially smoothed MUSIC algorithm View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Inhomogeneous Gaussian image models for estimation and restoration

    Page(s): 1305 - 1312
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1224 KB)  

    Two inhomogeneous Gaussian-image models are presented for estimation and restoration. By incorporating the local statistics of an image, a homogeneous autoregressive (AR) random field can be extended to an inhomogeneous AR field. This inhomogeneous random field can provide a better description of the image than the homogeneous one. As a consequence of this improved modeling, a minimum-mean-square-error estimator (MMSE), based on the inhomogeneous Gaussian model, can produce good results in both subjective and objective criteria. Two image models are proposed for use in image estimation and restoration: a residual image model (original image minus the space-variant mean) and a normalized image model (residual image divided by space-variant standard variation). The novel aspect of these models is the use of an autoregressive dynamical model for residual and normalized images. Some aspects of parameter estimation are discussed and simulation results are presented View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Adaptive algorithms for constrained ARMA signals in the presence of noise

    Page(s): 1282 - 1291
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (668 KB)  

    A family of algorithms is developed for adaptive parameter estimation of constrained autoregressive moving-average (ARMA) signals in the presence of noise. These algorithms utilize a priori information about the signal's properties, such as its spectral type (for example, low-pass, bandpass, etc.) or a spatial-domain characteristic. Special applications include modeling of autoregressions (AR) and signals of known spectral type in the presence of noise, signal deconvolution, image deblurring and multipath parameter estimation. Selected results of simulations are included to demonstrate the performance of the algorithms View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On determining optimum multirate structures for narrow-band FIR filters

    Page(s): 1255 - 1271
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1360 KB)  

    Several lower bounds on the storage and multiplication rate required to implement a bandpass filter using multirate structures are developed. These lower bounds suggest hypothetical optimum structures that actual optimum structures are found to resemble. The hypothetical structures are useful for predicting the number of stages, decimation and interpolation factors, and cost of actual structures from knowledge of only the passband width, transition bandwidths, and tolerance specifications of the desired filter. The determination of the actual optimum structure for a given bandpass filter specification requires a search over the space of all possible solutions. The size of this space grows very rapidly as filter bandwidth decreases, and excessive computer time is required to search it exhaustively. A fast branch-and-bound algorithm to search this space is also presented. This algorithm, which is applicable for a large class of cost criteria, improves search time by one to two orders of magnitude for filters with passband width in the range 10-3--5 cycles/sample View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A note on line spectral frequencies [speech coding]

    Page(s): 1355 - 1357
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (232 KB)  

    The main properties of line spectral frequencies (LSFs) are analyzed by means of the root-locus technique. This leads to a better insight into the nature of this efficient predictor filter parametrization. The effects of quantization of the LSFs are illustrated by a numerical example View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A robust approach to sequential detection

    Page(s): 1200 - 1210
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (828 KB)  

    The problem of signal detection in severe and/or changing noise environments, often encountered in underwater acoustics, radar, and some communications applications, is considered. The detector operates at near optimum levels for a particular noise environment, and is robust by virtue of maintaining high efficiency in other than nominal noise environments by adapting its optimum nonlinearity using an m-interval polynomial approximation (MIPA). It is shown that the MIPA detectors have the same basic structure for fixed sample size and variable sample size, i.e. sequential operation. The sequential MIPA detectors are asymptotically optimum, increase their transmission rate up to four times as compared to their fixed-sample size counterparts, and are highly insensitive to variations of noise compared to detectors based on min-max theory. The estimation and updating of the detector parameters can be accomplished using parallel processors operating in a recursive mode without disturbing the decision process View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Target parameter estimation in the near field with two sensors

    Page(s): 1357 - 1360
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (260 KB)  

    A performance-prediction procedure is presented for the evaluation of a passive-tracking technique for the localization of targets in the near field of two sensors. Parameter identifiability is established with the aid of the Fisher information matrix, which is evaluated for various combinations of measurements. This is used to determine bounds on localization performance. The corresponding uncertainty ellipses associated with the target position are evaluated for various tracking scenarios View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Computationally efficient stepup and stepdown procedures for real predictor coefficients

    Page(s): 1353 - 1355
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (256 KB)  

    Algorithms are presented for the stepup and stepdown problem for the real case. It is shown that the split Levinson algorithm and similar algorithms can be used to reduce the multiplicative complexity of the problem considerably View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Minimum worst case processing loss

    Page(s): 1369 - 1371
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (152 KB)  

    A figure of merit-the worst-case processing loss-is defined to characterize a data window's influence on the detection of a tone in broadband noise using the discrete Fourier transform. The minimum loss is close to -3 dB and occurs when the window's normalized second moment is about 1/5 View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Maximum likelihood estimation of the autoregressive model by relaxation on the reflection coefficients

    Page(s): 1363 - 1367
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (452 KB)  

    A method for autoregressive parameter estimation, which successively maximizes the likelihood with respect to each reflection coefficient while keeping the others fixed, is presented. The algorithm generalizes the recursive-maximum-likelihood technique of S.M. Kay (1983), which corresponds to performing only one iteration cycle. An interesting application is the estimation of a Toeplitz covariance matrix. Simulations show that the algorithm converges quite fast and provides much better estimates than current procedures for short record length View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A channel efficient method for high-resolution active direction finding of multiple targets

    Page(s): 1367 - 1369
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (276 KB)  

    A high-resolution direction-finding method for active systems is presented. Unlike the available high-resolution method, which requires the number of coherent receiver channels to be larger than the number of targets, this method requires only two (three) channels for one (two)-dimensional direction-finding, regardless of the number of targets to be resolved, thus reducing the implementation cost. The idea of this method is to incorporate the signaling waveform information, available in the active systems, in using the signal-subspace processing approach 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