By Topic

Acoustics, Speech and Signal Processing, IEEE Transactions on

Issue 6 • Date December 1977

Filter Results

Displaying Results 1 - 25 of 27
  • [Front cover and table of contents]

    Page(s): 0
    Save to Project icon | Request Permissions | PDF file iconPDF (174 KB)  
    Freely Available from IEEE
  • 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.
  • [Back cover]

    Page(s): c4
    Save to Project icon | Request Permissions | PDF file iconPDF (1053 KB)  
    Freely Available from IEEE
  • Line tracking using autoregressive spectral estimates

    Page(s): 510 - 519
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (768 KB)  

    This paper presents two new algorithms for tracking spectral lines of signals with time-varying spectra using a sequence of autoregressive spectral estimates. The algorithms are designed to discard false peaks due to noise and signal interference, and with the aid of high resolution autoregressive spectral estimates have the ability to track multiple lines. Simulation results indicate that line tracking using autoregressive spectral estimates can be accomplished for rapidly time-varying spectra with low signal-to-noise ratios (SNR) or closely spaced spectral lines. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Recursive least squares smoothing of noise in images

    Page(s): 520 - 524
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (704 KB)  

    In the recent past considerable attention has been devoted to the application of Kalman filtering to smoothing out observation noise in image data. A generalization of the one-dimensional Kalman filter to two dimensions was earlier suggested by Habibi, but it has since been shown that this generalization is invalid since it does not preserve the optimality of the Kalman filter. A new method is proposed here that enables well-established Kalman-filter theory to yield a simple two-dimensional filter for images that can be modeled by two-dimensional wide-sense Markov (WSM) random fields. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Deconvolution of cyclostationary signals

    Page(s): 466 - 476
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (832 KB)  

    Two algorithms are presented for the deconvolution of a cyclostationary (CS) signal which has been distorted by passage through an unknown linear, time-invariant system. It is necessary to know the statistics of the CS process, but no assumptions about the distorting system are required. Results of the application of the algorithms are presented as well as a discussion of the algorithms' efficiency in the presence of noise. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Interpolative models in restoration and enhancement of noisy images

    Page(s): 525 - 534
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1408 KB)  

    This paper considers the use of interpolative (noncausal) models in restoration and enhancement of noisy images. The first part of the paper presents a technique for restoring the noisy image. The second part involves enhancement of noisy images based on noisy data and uses a suboptimal estimator which is easily implementable. The identification algorithm presented has the capability of adapting the coefficients to the new incoming data. Several examples presented indicate that the algorithms work well. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Deconvolution when the convolution kernel has no inverse

    Page(s): 542 - 549
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (728 KB)  

    A short study on the general deconvolution problem when the kernel has no inverse proves that a priori information on the signal to be restored is a necessary condition for deconvolution. The proposed deconvolution method uses the following information: the signal to be restored has a bounded support; this support is known or is inside a known interval. This method concerns the convolution kernels whose Fourier transform has a cutoff frequency. This type of kernel has a wide practical field. The image restoration and the processing of "the principal value solution" of the deconvolution problem are the most characteristic elements. The method is derived from the general Liouville-Neuman theory of solving integral equations. This new method incorporates and extends Ville's analytic continuation and Van Cittert's successive convolution method. The iterative deconvolution algorithm is very simple. The advantages of this method are shown by numerical results and, in particular, by an experimental spectroscopic application. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Statistical tests and distance measures for LPC coefficients

    Page(s): 554 - 559
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (632 KB)  

    This paper considers the problem of comparing two sets of (LPC) coefficients or, more generally, that of comparing two short segments of speech via LPC techniques. It is shown that Itakura's prediction-residual ratio is intuitively unsatisfactory and theoretically misleading as a distance measure. Two slower, but more accurate statistical means of comparison are suggested, and these are supported by evidence from a simulation study. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Algorithm for pitch extraction using zero-crossing interval sequence

    Page(s): 559 - 564
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (600 KB)  

    A novel digital computer algorithm for pitch period analysis of connected speech is presented. The principal idea employed is to find the periodic-like portions of the speech, from its zero-crossing interval (ZCI) sequence, by an adaptive search. The algorithm requires only additions and comparisons, no multiplication. An implementation of the algorithm on a 1.5 μs memory-cycle computer (PDP-8) performs the analysis in real time, as accurately as manually performed pitch measurements from a plot of the speech waveform low-pass filtered to a bandwidth of 900 Hz in such a way that each pitch boundary is marked on a zero-crossing. Although 12 threshold values are used for decision making, only two of them, the upper and lower limits of the speaker's one octave pitch range, are speaker dependent. However, these two threshold values of a speaker are easily extracted and set by the algorithm from a sample of only 4-5 s duration of his (or her) normal speech. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On number theoretic Fourier transforms in residue class rings

    Page(s): 585 - 586
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (272 KB)  

    The proof of the orthogonality conditions that must be fulfilled by the transform factor α of a NTFT of length N, is based upon the possibility of cancelling all nonzero factors of the form (αq- 1), q = 1, 2,..., N - 1. In a residue ring containing zero divisors, this is not allowed, unless all such factors can be shown not to be divisors of zero. It is shown that this is the case, when a is any primitive Nth root of unity, N being an allowed transform legnth. At the same time, a property is established that helps to reduce the amount of searching needed to find suitable transform factors. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A two-pass fixed point fast Fourier transform error analysis

    Page(s): 582 - 585
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (424 KB)  

    A statistical model is used to predict the output signal-to-noise ratio (SNR) when a two-pass fast Fourier transform (FFT) is computed using fixed-point arithmetic. The results show that the ratio varies essentially as the square root of the number of points in the transform. Also included are the results of the simulation of a fixed-point machine and the variation of the error as a function of the length of the coefficients. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An alternative formulation for the recursive solution of the covariance and autocorrelation equations

    Page(s): 574 - 577
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (384 KB)  

    The present paper gives an alternative formulation for solving autocorrelation and covariance equations, which is different from that based on a scalar product definition. A simple direct matrix formulation, which leads to recursive algorithms for both covariance and autocorrelation equations is given. For the covariance method k, α and β parameters are defined. Some useful definitions for ARMA models are given. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A comparison of algorithms for minimax design of two-dimensional linear phase FIR digital filters

    Page(s): 492 - 500
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (920 KB)  

    Linear programming algorithms for the design of finite impulse response (FIR), linear-phase digital filters can be extremely time-consuming. Two iterative design techniques using multiple-exchange ascent algorithms have been developed. These are the methods of Kamp and Thiran and Hersey and Mersereau. They are known to be much faster than the linear programming techniques. The mathematical basis for these algorithms is reviewed and differences in the algorithms are noted. Results of empirical efficiency comparisons are presented. A new algorithm for reducing the number of iterations for multiple-exchange ascent algorithms is also presented. View full abstract»

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

    Page(s): 579 - 581
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (216 KB)  

    Complete, orthonormal systems of piecewise polynomial functions are constructed from the Walshlike systems. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • The computation of two-dimensional cepstra

    Page(s): 476 - 484
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (920 KB)  

    In this paper we shall explore two methods of computing the complex cepstrum of a two-dimensional (2-D) signal. The two principal methods for computing 1-D cepstra, using discrete Fourier transforms (DFT's) and the complex logarithm function or using a recursion relation for minimum-phase signals, may be extended to two dimensions. These two algorithms are developed and simple examples of their use are given. As a matter of course, we shall also be drawn into considering the definitions of 2-D causality and 2-D minimum-phase signals. In addition, we shall explore the relationship among the nonzero regions of a signal, its inverse, and its cepstrum. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Another proof and a sharpening of Huang's theorem

    Page(s): 581 - 582
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (176 KB)  

    Two further proofs of Huang's theorem on the zeros of analytic functions of two variables are given: the first is similar to previous proofs, but is made shorter by the use of a known maximum-modulus principle; the second is completely different, using a theorem of Rudin which actually gives a sharper result than Huang's. Finally, it is indicated how a correspondingly sharper result may be obtained in higher dimensions. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Signal representation: An efficient procedure

    Page(s): 461 - 465
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (568 KB)  

    The two concepts of orthogonal transformations and classical modeling are combined to form an effective procedure for efficiently representing (approximating) signals contained in a general Hilbert space. The signal to be represented is first operated upon by an appropriately chosen orthogonal transformation with the objective of generating a "simpler" transformed signal. This transformed signal is next approximated by the response of a fixed structured dynamic model to a given input. One selects the parameters which govern this model so that its response best matches the transformed signal. This general procedure is then illustrated by the specific task of representing an arbitrary sequence as the unit-impulse response of a linear recursive operator. The attractiveness of linear recursive modeling is demonstrated by the approximation of a sampled electrocardiogram signal. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Design of two-dimensional recursive digital filters

    Page(s): 577 - 579
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (328 KB)  

    This correspondence proposes a design method for two-dimensional recursive filters which permits one to supply the specifications in the frequency domain. The method involves an iterative optimization procedure and evaluates the z transfer function in a comparatively short time. It is worth pointing out that it is possible to orient the optimization procedure by introducing linear constraints between the coefficients of the transfer function. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Extensions of the Pade-approximant technique for the design of recursive digital filters

    Page(s): 501 - 509
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (744 KB)  

    A closed-form solution to the problem of synthesizing recursive filters with specified gain and phase characteristics is presented. The method is shown to be numerically well conditioned, and to result in BIBO stable realizations. By means of a free design parameter, the filter response can be modified according to given criteria in a particular application. Examples indicate that the method yields, with a standard setting of the free parameter, designs very close to optimal in terms of frequency-domain performance. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • BIBO stability of multidimensional filters with rational spectra

    Page(s): 549 - 553
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (544 KB)  

    Classical conditions relating bounded input-bounded output (BIBO) stability to the behavior of the impulse response for one-dimensional discrete filters with rational spectra are reviewed and generalized to the multidimensional case. It is shown, in this manner, that these conditions are fundamentally predicated by rationality and not by the existence of isolated filter natural frequencies. Furthermore, the conditions are seen to lead to novel stability criteria applicable in the frequency-domain and minimum mean-square design of multi-dimensional filters. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A pitch extraction algorithm based on LPC inverse filtering and AMDF

    Page(s): 565 - 572
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (752 KB)  

    We present an efficient algorithm for determining fundamental frequency and voiced/unvoiced (V/UV) decision of speech. The pitch extractor utilizes the cross-correlation average magnitude difference function (AMDF) waveform that is obtained from the linear prediction residual signal. The decision logic used in pitch extraction is simple and reliable. The periodicity and null depth of AMDF waveforms, together with the average residual energy and the past pitch information, are used in the decision logic for fundamental frequency and V/UV decision. Computer simulation of the algorithm yielded accurate results, even for difficult phonemes for pitch extraction. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A note on the eigenvectors of DFT matrices

    Page(s): 586 - 589
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (416 KB)  

    This note presents the derivation of orthogonal eigenvectors of DFT matrices using some of the interesting properties of centrosymmetric and involutory matrices. The computational requirements for computing DFT's using the eigenvectors of DFT matrices is also discussed. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Adaptive noise canceling applied to sinusoidal interferences

    Page(s): 484 - 491
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (696 KB)  

    This paper investigates a new method for eliminating sinusoidal or other periodic interference corrupting a signal. This task is typically accomplished by explicitly measuring the frequency of the interference and implementing a notch filter at that frequency. The method proposed herein uses an adaptive filter to eliminate the interference. The procedure is called adaptive noise canceling and is applicable when an auxiliary reference input is available containing the interference alone. The reference input is filtered in such a way that it closely matches the interfering sinusoid, and is then subtracted from the primary input leaving the signal alone. The results of this research show that when a sum of sinusoids is applied to an adaptive filter, the filter converges to a dynamic solution in which the weights of the filter are time varying. This time-varying solution implements a tunable notch filter, with a notch located at each of the reference frequencies. When used in noise-canceling applications, this adaptive notch filter provides a simple alternative to other methods of tracking and eliminating sinusoidal interferences. 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