By Topic

Acoustics, Speech and Signal Processing, IEEE Transactions on

Issue 5 • Date October 1983

Filter Results

Displaying Results 1 - 25 of 42
  • [Front cover and table of contents]

    Page(s): 0
    Save to Project icon | Request Permissions | PDF file iconPDF (247 KB)  
    Freely Available from IEEE
  • Comments and corrections on the use of polar sampling theorems in CT

    Page(s): 1329 - 1331
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (222 KB)  

    In two recent papers, an exact polar interpolation formula was used to reconstruct computer tomography (CT) imagery by a procedure known as direct Fourier transform inversion. The resulting imagery compared favorably to CT imagery reconstructed by filtered convolution back projection. Strictly speaking, however, the interpolation formula was inappropriately used since it is valid only for an odd number of azimuthal samples while in CT one uses an even number of samples. When the number of samples N is large (say > 200 as in CT) the error is not noticeable and the "appropriateness" of the formula has no practical significance. However, when N is small, a large error can result. We derive, in this paper, exact azimuthal interpolation formulas for N even and arbitrary N. 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 (611 KB)  
    Freely Available from IEEE
  • Recursive FIR digital filter design using a z -transform on a finite ring

    Page(s): 1155 - 1164
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1008 KB)  

    Properties of a new complex number-theoretic z-transform (CNT z-transform) over a finite ring are presented here and related to the usual z-transform. Using the Chinese remainder theorem, it is convenient to use finite rings that are isomorphic to the direct sum of finite or Galois fields of the form GF(q2) where q is a Mersenne prime. Many properties of the usual z-transform are preserved in the CNT z-transform. This transform is used in the present paper to design both recursive and nonrecursive FIR filters on a finite ring. The advantages of the FIR filter on a finite ring are the following: 1) the absence of a roundoff error build up in the computation of either the recursive or nonrecursive realization of the filter; 2) when the FIR filter is recursive, the question of stability does not arise as long as the magnitudes of the impulse response and the input sequence do not exceed their design values; 3) for the frequency sample representation of the FIR filter an absolute error bound on the impulse response function can be obtained in terms of the power spectrum. The time required to compute a nonrecursive FIR filter on the Galois field GF(q2), where q is a Mersenne prime, is competitive with the similar nonrecursive realization on the usual complex number field, using the FFT algorithm. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A fast lattice algorithm for PARCOR parameter estimation based on an approximate likelihood

    Page(s): 1319 - 1323
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (432 KB)  

    A new fast algorithm for PARCOR parameter estimation based on an approximate likelihood is proposed. The exact likelihood of AR processes based on the Gaussian assumption is approximated by taking the exponential part only. The criterion is just a modification of that of the prewindowed method by using backward predictions. Then, a fast recursive lattice type algorithm is derived. Its performance is compared to those of the Lee-Morf ladder algorithm for the prewindowed method and the covariance ladder algorithm by simulations. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The fully recursive filter: A general 2-D recursive filter

    Page(s): 1327 - 1329
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (384 KB)  

    The conventional 2-D recursive filter can be thought of as performing a linear combination of 1-D filterings on past and present input and output data rows. All of these component 1-D filters are FIR except for that of the current output row. In a fully recursive filter all the component 1-D filters are IIR. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Performance of an FFT-based voice coding system in quiet and noisy environments

    Page(s): 1323 - 1327
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (528 KB)  

    This paper describes a 2400 bit/s vocoder based on spectral envelope estimation, spectral coding to 48 bits, pitch extraction, and decreasing-chirp excitation for voiced synthesis. Several spectral smoothing and coding schemes are described and intelligibility test results compared. This vocoder was implemented on the CSP-30 high speed digital processor at the RADC/EEV Speech Processing Research and Development Facility at Hanscom AFB, MA. This system yields high performance in a quiet environment and is robust in acoustic noise environments at a data rate of 2400 bits/s. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On the extrapolation of band-limited functions with energy constraints

    Page(s): 1222 - 1234
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (984 KB)  

    A modification of the algorithm proposed by Papoulis and Gerchberg for extrapolating band-limited functions is suggested to extend its applicability over data corrupted by noise. We assume that energy constraints are known either for the band-limited signal or for the noise. In addition, the discrete formulation of the iterative algorithm is derived, and the transition from the continuous algorithm to its digital implementation is presented. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The "Unity-lagged" short-time spectrum of a narrow-band Gaussian process

    Page(s): 1100 - 1109
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (872 KB)  

    A form of the short-time cross-spectrum of a signal with its delayed self is considered. Its statistical and signal-characterizing properties are derived for the case of a narrow-band random Gaussian signal in white Gaussian noise. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A fast convergence frequency domain adaptive filter

    Page(s): 1312 - 1314
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (384 KB)  

    This correspondence presents a new fast convergence algorithm for frequency domain adaptive filter and its applicability to acoustic noise cancellation in speech signals. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Design and implementation of a single-chip 1-D median filter

    Page(s): 1164 - 1168
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (608 KB)  

    The design and implementation of a VLSI chip for the one-dimensional median filtering operation is presented. The device is designed to operate on 8-bit sample sequences with a window size of five samples. Extensive pipelining and employment of systolic data-flow concepts at the bit level enable the chip to filter at rates up to ten megasamples per second. A configuration for using the chip for approximate two-dimensional median filtering operation is also presented. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • C -matrix transform

    Page(s): 1304 - 1307
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (376 KB)  

    An approximation to the discrete cosine transform (DCT), called the C-matrix transform (CMT), has been developed by Jones et al. [3] for N = 8. This is extended to N = 16 and 32 and its performance is compared with the DCT based on some standard criteria. CMT is computationally simpler as it involves only integer arithmetic. It has potential in signal processing applications because of its closeness to the DCT. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A two-parameter class of Bessel weightings for spectral analysis or array processing--The ideal weighting-window pairs

    Page(s): 1309 - 1312
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (400 KB)  

    A unified theory for array processing in 1, 2, or 3 dimensions is pointed out and illustrated by a two-parameter class of Bessel weightings. This class subsumes theI_{0}-sinhweighting-window pair as well as the ideal space factor of van der Maas in one dimension. The weightings that realize the ideal space factor in 2 and 3 dimensions are generalized functions more singular than the delta function required in 1 dimension. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A hierarchical decision approach to large-vocabulary discrete utterance recognition

    Page(s): 1061 - 1066
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (728 KB)  

    Very short response time is a critical requirement for automatic discrete utterance recognition. The real-time vocabulary size of most of today's commercially available recognizers is limited to several hundreds of utterances, primarily due to the fact that detailed acoustic matching involves considerable computation. The method presented here offers an economical solution to the real-time large-vocabulary recognition problem by carrying out recognition in two stages. In the initial stage, the incoming utterance is linearly matched against the entire vocabulary using only two features-utterance duration and either two or three average spectra for each utterance. While the number of prototypes matched is large, the time required per match is substantially reduced. During this initial stage, a preset number of best-match prototypes is determined for each unknown input. In the second stage, matching is performed for the best-match list based upon more detailed features (e.g., 10-ms log-power spectra), using more elaborate matching methodology, e.g., dynamic programming. Evaluation experiments were conducted using the 2000 most frequent words in an office-correspondence corpus and three normal adult-male talkers. It was observed that first-stage best-match lists of 30-50 items included the "correct" words between 99.0 and 99.5 percent of the time. Using DP on 10-ms spectral samples for the second stage, recognition accuracy ranged from 86.5 to 94.5 percent. A match-limiter, when used with a 50-64-word, commercially available recognizer for the second stage, makes near-real-time large-vocabulary recognition feasible. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Optimality of high resolution array processing using the eigensystem approach

    Page(s): 1235 - 1248
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1216 KB)  

    In the classical approach to underwater passive listening, the medium is sampled in a convenient number of "look-directions" from which the signals are estimated in order to build an image of the noise field. In contrast, a modern trend is to consider the noise field as a global entity depending on few parameters to be estimated simultaneously. In a Gaussian context, it is worthwhile to consider the application of likelihood methods in order to derive a detection test for the number of sources and estimators for their locations and spectral levels. This paper aims to compute such estimators when the wavefront shapes are not assumed known a priori. This justifies results previously found using the asymptotical properties of the eigenvalue-eigenvector decomposition of the estimated spectral density matrix of the sensor signals: they have led to a variety of "high resolution" array processing methods. More specifically, a covariance matrix test for equality of the smallest eigenvalues is presented for source detection. For source localization, a "best fit" method and a test of orthogonality between the "smallest" eigenvectors and the "source" vectors are discussed. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Magnitude-only reconstruction of two-dimensional sequences with finite regions of support

    Page(s): 1256 - 1261
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (640 KB)  

    In this paper, a conceptual algorithm for reconstructing a two-dimensional (2-D) complex-valued finite sequence from an adequate set of samples of the magnitude of its Fourier transform is presented. In particular, one obtains, at least theoretically, all solutions of the 2-D magnitude-only reconstruction problem, provided that the modulus of the DFT is available in a sufficiently large set of points. However, the practicability of this algorithm is limited to sequences with relatively small regions of support. The key for developing the method is shown to be an appropriate mapping of 2-D finite sequences into 1-D ones, such that 2-D discrete correlation can be formulated in terms of ordinary 1-D discrete correlation. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Reconstruction of signals from phase: Efficient algorithms, segmentation, and generalizations

    Page(s): 1135 - 1147
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (992 KB)  

    In this paper, we develop several new formulations of the phase-only reconstruction problem which lead to new, efficient algorithms for signal reconstruction. We describe a procedure for approximately reconstructing a long duration signal from short segments. Phase-only blind deconvolution for linear phase filters relies on mirror image symmetry in the filter transfer function H(z). We generalize this deconvolution result to include several other types of symmetries for H(z). View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Application of two-channel prediction filtering to the recursive filter design problem

    Page(s): 1169 - 1177
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (752 KB)  

    We consider the recursive filter design problem for the case where the numerator and denominator polynomials have the same order and represent the problem in the context of two-channel prediction filtering. A new method of solution is then obtained by applying Strand's two-channel extension of the Burg scheme. The new method is shown to yield a stable recursive filter while providing an improvement in performance over previous linear methods. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Convolutions of long integer sequences by means of number theoretic transforms over residue class polynomial rings

    Page(s): 1125 - 1134
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (936 KB)  

    In a recent paper, Dubois and Venetsanopoulos [6] have derived methods for convolving sequences of numbers belonging to a ring S, using number theoretic transforms (NTT's) over an extension ring R of S. In this paper we obtain more explicit expressions for some of their results and, more important, improve the efficiency of their methods. Attention is focused on the case R = S[z]/(P(z)), that is, R is the quotient ring of S[z], modulo the principal ideal generated by a polynomial P(z) of degree n. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Finite data lattice algorithms for instrumental variable recursions

    Page(s): 1202 - 1210
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (856 KB)  

    The last few years have seen a rapid development of the so-called lattice algorithms for the fast solution of finite data least-squares problems. While a fast algorithm has been given for finite data instrumental variable recursions, as yet no finite data lattice schemes have been given. In this work, a lattice algorithm is derived for a finite data instrumental variable recursion, and its use in both ARMA and ARX time series models is indicated. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Separation of superimposed signals by a cross-correlation method

    Page(s): 1084 - 1089
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (640 KB)  

    An approach is presented for separating a superposition of delayed finite-duration signals, with unknown gains and delays, which overlap both in the time and frequency domains. The method is partly based on the cross-correlations among the signals and some of their derivatives. Given the data, an initial detection scheme formulates hypotheses on signals which may be present in the mixture, and provides estimates of the signal positions for each hypothesis. In the second stage, the most likely hypotheses are selected and accurate values of the signal amplitudes and their time positions are determined by a nonlinear least squares approach. The method is illustrated by a simulation with myoelectric (EMG) signals. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Convolution using a conjugate symmetry property for number theoretic transforms over rings of regular integers

    Page(s): 1121 - 1125
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (624 KB)  

    In a recent paper, Dubois and Venetsanopoulos [7] have derived a method for convolving sequences over a ring S, using number theoretic transforms (NTT's) over an extension ring R of S. Until now, their assumptions have only been verified for the special case that R is an extension field of S. This paper examines their assumptions for the very general class of rings of regular integers, a class of rings that was introduced by the authors in [5]. In this paper, we also present an algorithm for the synthesis of rings suitable for number theoretic transforms and with low computational complexity. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Restoring causal signals by analytic continuation: A generalized sampling theorem for causal signals

    Page(s): 1294 - 1298
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (480 KB)  

    A band-limited part x0(t) of a non-band-limited causal signal x(t) is used to completely restore x(t) by using analytic continuation. As a consequence, a generalized sampling theorem for causal signals is derived; the total causal signal x(t) can be completely restored by sampling the band-limited part x0(t) with finite sampling frequency. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Multirate recursive digital filters--A general approach and block structures

    Page(s): 1148 - 1154
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (696 KB)  

    Multirate filters, in which the input and output sample rates differ, are examined from a general point-of-view with the objective of providing an approach that will permit the filter designer to use input and output sample rates as design variables to be used to optimize system performance. The emphasis is on multirate recursive (IIR) filters that can be computationally efficient when implemented with block structures, with input and output blocks of different lengths. Classes of multirate filters that possess a generalized shift-invariance property, and are susceptible to transfer function analysis, are identified. A general approach to deriving matrix transfer functions for block implementations of multirate filters is derived, and a general method for obtaining state-space realizations of the matrix transfer function is presented. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The aliasing problem in discrete-time Wigner distributions

    Page(s): 1067 - 1072
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (664 KB)  

    There is no straightforward way to proceed from the continuous-time Wigner distribution to a discrete-time version of this time-frequency signal representation. A previously given definition of such a function turned out to yield a distribution that was periodic with period π instead of 2π and this caused aliasing to occur. Various alternative definitions are considered and compared with respect to aliasing and computational complexity. From this comparison it appears that no definition leads to a function that is optimum in all respects. This is illustrated by an example. 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