By Topic

Acoustics, Speech and Signal Processing, IEEE Transactions on

Issue 2 • Date April 1985

Filter Results

Displaying Results 1 - 25 of 31
  • [Front cover and table of contents]

    Page(s): 0
    Save to Project icon | Request Permissions | PDF file iconPDF (217 KB)  
    Freely Available from IEEE
  • Comments on "Direct Fourier reconstruction in computer tomography"

    Page(s): 446 - 449
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (443 KB)  

    In this correspondence, several theoretical aspects of direct Fourier reconstruction in CAT as presented in [1]-[3] are revisited. Specifically, the following points are discussed: 1) a lack of appropriate references to analogous previous work in the theory of the proposed interpolation technique and its application in physics and engineering areas [1], 2) some extensions and remarks on well-known trigonometric interpolation formulas [1], [3], and 3) some ambiguities related to the optimality of the reconstruction technique and the exactness of the interpolation procedure [1]. 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 (638 KB)  
    Freely Available from IEEE
  • Detection of signals by information theoretic criteria

    Page(s): 387 - 392
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (728 KB)  

    A new approach is presented to the problem of detecting the number of signals in a multichannel time-series, based on the application of the information theoretic criteria for model selection introduced by Akaike (AIC) and by Schwartz and Rissanen (MDL). Unlike the conventional hypothesis testing based approach, the new approach does not requite any subjective threshold settings; the number of signals is obtained merely by minimizing the AIC or the MDL criteria. Simulation results that illustrate the performance of the new method for the detection of the number of signals received by a sensor array are presented. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Image design: Generation of a prescribed image through a diffraction-limited system with high-contrast recording

    Page(s): 460 - 465
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (584 KB)  

    Image restoration involves the recovery of an image which has been distorted by a given imaging system. "Image design," on the other hand, aims at determining the input image which when distorted by an imaging system (e.g., a display device) becomes a desired pattern. The image design problem is encountered in the design of masks for microphotography, microlithography, laser printing, and aids for the visually impaired. In this correspondence, we solve the "image design" problem using linear programming techniques for the case of an imaging system modeled by a band-limited linear system followed by a noninvertible point nonlinearity. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An RKHS analysis of sampling theorems for harmonic-limited signals

    Page(s): 437 - 440
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (432 KB)  

    A T-periodic signal x(t) is said to be K-harmonic limited if, for some integer K > 0, its complex Fourier coefficients{c_{n}}satisfyc_{n} = 0for |n| > K. Such signals may be completely reconstructed from a finite number of uniformly spaced signal samples taken over a period, a property which facilitates the representation of two-dimensional (polar form) signals used in computerized tomography. By employing a reproducing kernel Hilbert space (RKHS) setting, a generalized theorem for sampling harmonic-limited signals is derived. All the special representations which have appeared in the literature with separate proofs, including three recent versions, [1]-[3], are shown to be special cases of the generalized reconstruction formula. Connections with Fourier series theory and Kramer's generalized sampling theorem are also discussed. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Algorithms and structures for convolutions over Galois fields

    Page(s): 453 - 455
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (384 KB)  

    The computation of cyclic convolutions in Galois fields is an integral part of coding theory and formulation as well as of many signal processing applications. In this paper, we introduce a method for the computation of such convolutions that minimizes, in theory, the computational complexity of the algorithm. We also propose special-purpose computer architecture schemes for the efficient realization of the algorithm, and in general, for efficient calculation of convolutions in Galois fields. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Positive-definite Toeplitz embedding based on the cyclic extension of a data matrix

    Page(s): 393 - 400
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (872 KB)  

    Embedding in a block-Toeplitz matrix is one of the techniques for efficient inversion of arbitrary covariance matrices arising in least squares estimation. Delosme and Morf have shown how the concept of Toeplitz distance leads to a positive definite embedding via a mixed form representation of the given matrix. In this paper a new method is proposed for the positive definite block-Toeplitz embedding which has the particular feature of taking the signal itself as a starting point. It consists of a suitable cyclic extension and embedding of the given signal in a multichannel stationary signal. The main advantage of the method is that, given the signal, the embedding can be performed by inspection. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Speaker-dependent isolated word recognition using speaker-independent vector quantization codebooks augmented with speaker-specific data

    Page(s): 440 - 443
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (520 KB)  

    A hybrid approach to speaker-dependent isolated word recognition is discussed. The approach merges speaker-specific information obtained from a single training utterance with multisection vector quantization codebooks that were designed for speaker-independent recognition. The approach provides easily trained, computationally efficient, and accurate isolated word recognition. On the digits, the approach achieved an error rate less than 1 percent. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An FIR estimation filter based on the sampling theorem

    Page(s): 455 - 460
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (568 KB)  

    The estimation of noise-perturbed bandlimited stochastic signal samples by FIR filtering is considered. The mean-square error of the estimate is used as the criterion of performance. We contrast three types of filters: all-pass, a sampling-theorem-based filter, and the minimum mean-square error (Wiener) filter. Although the Wiener filter is linearly optimal, its design requires detailed knowledge of the processes' second-order statistics. The sampling theorem filter does not. For large signal-to-noise ratios and large filter orders, the two filters perform nearly identically asymptotically. Furthermore, we demonstrate that for a fixed filter order, there exists an optimal sampling rate which decreases with increasing signal-to-noise ratio. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Improved dynamic time warping methods for discrete utterance recognition

    Page(s): 449 - 450
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (272 KB)  

    Two modifications of dynamic time warping methods for discrete utterance recognition are proposed. They compensate for inaccurate endpoint detection and emphasize the differentiating regions of similar sounding utterances. The methods proposed are shown to give an increased recognition accuracy on a difficult vocabulary containing many similar sounding words. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A simple noniterative speech excitation algorithm using the LPC residual

    Page(s): 432 - 434
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (368 KB)  

    This paper provides an analytical derivation of a simple noniterative technique for extracting a multiple impulse excitation model for synthesized speech directly from the LPC residual sequence. While suboptimal with respect to "multipulse" techniques, this method is very applicable for speech enhancement where processor capability is limited. The results suggest an additional "orthogonality" requirement between the excitation sequence and the resulting prediction error, which aids in the intuitive understanding of the method. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Moments of error-frequency response due to coefficient inaccuracy for sampled data filters

    Page(s): 436 - 437
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (256 KB)  

    McCrady and Spence recently developed a general coefficient fabrication error model for sampled data filters which makes use of higher order moments of the perturbed frequency response. By using previous results on cumulants and moments of a random variable, we derive an algorithm to compute recursively the coefficients of the Fourier series of these moments. The algorithm avoids the cumbersome combinatorial approach and permits a systematic determination of statistics of the perturbed frequency response. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Constrained signal restoration via iterated extended Kalman filtering

    Page(s): 472 - 475
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (488 KB)  

    The problem of estimating the input (signal restoration) to a known linear system (point spread function), given the noisy observations of the output, is considered. The input signal is assumed to satisfy certain known physical constraints such as positivity. It is proposed to incorporate the constraints by introducting a memoryless nonlinearity in the system. A statistical approach is taken leading to a closed-form type recursive solution in the form of iterated extended Kalman filtering with two local iterations at every new data point. A simulation example is presented which demonstrates the superiority of the proposed approach over the conventional Kalman-Wiener filtering. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Speech enhancement using a minimum mean-square error log-spectral amplitude estimator

    Page(s): 443 - 445
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (392 KB)  

    In this correspondence we derive a short-time spectral amplitude (STSA) estimator for speech signals which minimizes the mean-square error of the log-spectra (i.e., the original STSA and its estimator) and examine it in enhancing noisy speech. This estimator is also compared with the corresponding minimum mean-square error STSA estimator derived previously. It was found that the new estimator is very effective in enhancing the noisy speech, and it significantly improves its quality. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Fast recursive estimation of the parameters of a space-varying autoregressive image model

    Page(s): 469 - 472
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (584 KB)  

    The identification of two-dimensional (2-D) autoregressive (AR) image models has been previously shown to be an integral part of image estimation. Furthermore, because of the nonhomogeneous nature of images, much better results are obtained with space varying models. To this effect, the development of a fast recursive method is now proposed for estimating the parameters of a two-dimensional AR image model, at each pixel, based on a finite memory. This fast method can be coupled to a space-variant Kalman filter for on-line adaptive estimation or can be used to estimate 2-D spectra for space-variant fields. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Improvement of the excitation source in the narrow-band linear prediction vocoder

    Page(s): 377 - 386
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1272 KB)  

    The major weakness of the current narrow-band LPC synthesizer lies in the use of a "canned" invariant excitation signal, The use of such an excitation signal is based on three primary assumptions, namely, 1) that the amplitude spectrum of the excitation signal is flat and time invariant, 2) that the phase spectrum of the voiced excitation signal is a time-invariant function of frequency, and 3) that the probability density function of the phase spectrum of the unvoiced excitation signal is also time invariant. This paper critically examines these assumptions and presents modifications which improve the quality of the synthesized speech without requiring the transmission of additional data. Diagnostic acceptability measure (DAM) tests show an increase of up to five points in overall speech quality with the implementation of each of these improvements. These modifications can also improve the speech quality of LPC-based speech synthesizers. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Automatic glottal inverse filtering from speech and electroglottographic signals

    Page(s): 369 - 377
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1080 KB)  

    The flow of air through the glottis, the glottal volume-velocity, reflects the action of the vocal folds and is thus an important indicator of laryngeal function. However, it cannot be measured directly because of vocal tract filtering. We have developed an automated on-line method to determine the glottal volume-velocity waveform from normal and pathological speech based on digital inverse filtering. The method developed addresses the problems of accurate identification of vocal tract parameters and reduction of low-frequency noise. The vocal tract filter is estimated by analysis of the undriven vocal tract response during closed glottis, as identified from an electroglottographic signal. Low-frequency noise is attenuated by a high-pass filtering operation followed by a low-pass compensation. The complete inverse filtering method provides reliable glottal volume-velocity waveforms for both normal and pathological speech. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A zero-crossing consistency method for formant tracking of voiced speech in high noise levels

    Page(s): 349 - 355
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (824 KB)  

    A technique is described for the determination of the first three formant frequencies of a voiced speech signal corrupted by high levels of noise. The technique involves a rough frequency-domain analysis followed by an accurate time-domain statistical analysis. The frequency-domain analysis is performed with a bank of overlapped, shifted, bandpass digital filters chosen to cover the frequency range of the first three formant frequencies. The time-domain analysis involves a calculation of the consistency of the intervals between successive zero-crossings at the output of each filter. As the algorithm proceeds, those filters which contain peaks in the energy spectrum are chosen as a first estimate of the location of the first three formants. The final choice of filter containing each formant is made, in the neighborhood of the first estimate, by choosing that filter with a signal output having the maximum zero-crossing interval consistency. A decision procedure is applied to several calculated parameters to result in a measurement of each of the first three formant frequencies. The results of extensive testing of the method are reported for speech contaminated by high levels of additive white Gaussian noise. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Properties of the overdetermined normal equation method for spectral estimation when applied to sinusoids in noise

    Page(s): 406 - 413
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (688 KB)  

    In this paper we study the spectral properties of an auto-regressive (AR) spectral estimator proposed by Cadzow, which uses a large number of correlations to set up the normal equations. This Set of overdetermined equations is then solved in a least squares sense. The main contribution of this paper is the derivation of asymptotic relationships, i.e., as the number of samples goes to infinity, between the number of correlations used, the model order and the signal-to-noise ratio of the signal, and the characteristics of the resulting spectral estimate when the signal under study is composed of sinusoids in noise. The characteristics studied are the height, bandwidth, and area of the peaks in the estimated spectrum. The method is shown to be a spectral density estimator like the ME method, where spectral areas rather than spectral values should be interpreted as estimates of power. The role of the number of correlations as a signal-to-noise ratio enhancer is discussed. Computer simulations are presented which verify the theoretical results. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Recursive quadrature mirror filters--Criteria specification and design method

    Page(s): 414 - 420
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (816 KB)  

    Criteria are found for a pair of recursive infinite impluse response (IIR) filters which have mirror image amplitude responses and whose outputs are in phase quadrature at all frequencies. A method is suggested for designing filters with these characteristics, and their performance and processing requirements are compared with a typical nonrecursive finite impluse response (FIR) realization. A particular filter is described in more detail, and it is shown how it could be incorporated in a multiband processing scheme. It is concluded that recursive quadrature mirror filters could offer considerable savings in terms of signal processing compared with nonrecursive filters with similar performances. Additionally, it is suggested that, because they can be designed to have relatively short absolute delays, the performance of the IIR filters may prove superior for certain applications. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Bias-free estimates of the variance of time of arrival differences

    Page(s): 483 - 486
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (352 KB)  

    In the frequency domain estimation of time of arrival differences, phase data of the cross spectrum may show correlation between frequency bins due to either insufficient time-bandwidth product or the use of a nonrectangular window on the input signal samples. This correlation severely biases variance estimates if standard results of generalized least square theory are applied indiscriminately. A technique has been developed to eliminate the bias of variance estimates. Simulation results are provided. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Effects of input quantization in floating-point digital pulse compression

    Page(s): 434 - 435
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (232 KB)  

    A computer simulation of an FFT-based digital LFM pulse compressor using vector floating point arithmetic is presented, showing the effects of retaining various number of mantissa bits at the input quantizer. Plots of the compressed pulse waveforms and rms and peak sidelobe levels as a function of the number of mantissa bits are shown. 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.
  • Estimation of local statistics for digital processing of nonstationary images

    Page(s): 465 - 469
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (1080 KB)  

    In this correspondence we report results from our experiments to find useful measures of local image autocovariance parameters from small subblocks of data. Our criterion for the reliability of parameter estimates is that they should correlate with observed signal activity and yield high quality results when used in adaptive processing. We describe a method for estimating the correlation parameters of first-order Markov (nonseparable exponential) autocovariance models, The method assumes that image data are stationary within N × N pixel sub-blocks. Values of the autocovariance parameters may be calculated at every pixel location. A value of N = 16 yields results which fit our criterion, even when the original data are degraded by blur and noise. An application to data compression is suggested. 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