Showing papers on "Cepstrum published in 1984"

Spectral analysis using generalized cepstrum

Abstract: In this paper, we present a generalization of the cepstral method from the viewpoint of spectral smoothing for speech. We use a natural generalization of the logarithmic function instead of the logarithmic function itself and we refer to the spectral representation parameter on the "generalized logarithmic" scale as the "generalized cepstrum," which corresponds to the cepstrum on the logarithmic scale. A number of properties of the generalized cepstrum are shown, and are compared to the cepstrum.

Measuring apparatus utilizing spectrum profile

Abstract: Measuring or imaging of a characteristic of the structure of a medium is carried out utilizing the power spectrum profile of a received signal wave, which is derived from an incident beam pulse applied to the medium after travelling through the medium by transmission and reflection. The received signal wave is processed to obtain its cepstrum, which is filtered to separate slowly varying cepstra (including a desired Gaussian shaped pulse signal) from fast varying cepstra. The filtered cepstrum is subject to a further Fourier transformation to eliminate the effects of line cepstra in a lower frequency region and to eliminate the effects of the frequency dependence of reflecting indices of the structure. The target power spectrum is recovered by a series of inverse transformations. Correction apparatus utilizing momenta of the distored power spectrum to obtain the target Gaussian shaped profile is disclosed.

Estimation of spectral slit width and blind deconvolution of spectroscopic data by homomorphic filtering

Abstract: A numerical method is developed to estimate the spectral slit width of a spectrophotometer which degrades the spectral resolution of the measured data. The slit width is estimated from the measured spectrum in certain conditions. This method is based on homomorphic filtering and utilizes the cepstrum of the average log Fourier spectrum of the section spectra. Computer simulations and experimental examples for infrared absorption spectra of ammonia gas demonstrate the potential of the method for practical applications to the preprocessing of deconvolution. A blind deconvolution method is also examined through the experimental result with an ammonia gas spectrum.

A performance comparison of pitch extraction algorithms for noisy speech

Abstract: Results of a performance comparison study of eight pitch extraction algorithms for noisy as well as clean speech are presented. These algorithms are the autocorrelation method with center clipping, the autocorrelation method with modified center clipping, the simplified inverse filter tracking (SIFT) method, the average magnitude difference function (AMDF) method, the pitch detection method based on LPC inverse filtering and AMDF, the data reduction method, the parallel processing method and the cepstrum method. It has been found that for pitch detection of noisy speech the algorithm that uses an AMDF or an autocorrelation function yields relatively good performance than others. A pitch detector that uses center clipped speech as an input signal is effective in pitch extraction of noisy speech. In general, preprocessing such as LPC inverse filtering or center clipping of input speech yields remarkable improvement in pitch detection.

Ultrasonic diagnosis apparatus for tissue characterization

Abstract: A diagnosis apparatus for discriminating a property of the tissue to be observed from the reflected ultrasonic wave uses the nature of the fine structure of the tissue, and more practically uses the interval of small reflecting bodies dispersely distributed in the tissue as the parameters. The intervals flucatuate. Therefore, an average value and/or a degree of fluctuation is calculated and is displayed. For obtaining the average interval, a method of using the cepstrum of the received signal or a method of using the self-correlation of the received signal can be used.

Orthogonal transformer and apparatus operational thereby

Abstract: Apparatus for obtaining an orthogonal transformer for use in signal analysis. This apparatus is especially useful in cepstral analysis where vocal tract characteristics of a signal are separated from a sound source. A buffer memory 10 temporarily stores input time series data and an orthogonal transformation is carried out by an orthogonal transform means 20 by multiplying each predetermined data read sequentially out of the buffer memory 10 by a predetermined twiddle factor. A controller 30 is used to specify the data read out of the buffer memory 10 and for giving the twiddle factor a value of 1 and r/2l corresponding to the magnitude of the real and imaginary part of the twiddle factor. Cepstrum information is obtainable using simple hardware to analyze the data generated by the orthogonal transformer.

On the use of discrete cosine transform in cepstral analysis

Abstract: If the original signal is defined to be symmetrical, the discrete Fourier transform (DFT) used in cepstral analysis can be replaced by a discrete cosine transform (DCT). This principle is applied to the evaluation of the real and complex pseudocepstrum of speech signals. In both the real and complex cepstrum cases, it is found that the use of the DCT does not degrade the information contained in the cepstrum while substantially reducing the computational complexity.

Spectral analysis using generalised cepstrum

Abstract: In this paper, we present a generalization of the cepstral method from the viewpoint of spectral smoothing for speech. We use a natural generalization of the logarithmic function instead of the logarithmic function and we refer to the spectral representation parameter on the "generalized logarithmic" scale as the "generalized cepstrum," which corresponds to the cepstrum on the logarithmic scale. A number of properties of the generalized cepstrum are shown, and are compared to the cepstrum.

Resolution and Visibility Enhancement of Ultrasonic Echoes Reflected from Targets Hidden by Highly Reverberant Thin Layers

Abstract: BO BI 83 In the ultrasonic imaging of many materials, the measured signal consists of multiple echoes (reverberation) which complicate direct characterization of the target. In an earlier report [7] it has been shown that ultrasonic echo classification can facilitate imaging of targets hidden by highly reverberant thin layers. Among all classes of echoes, two sequences of multiple echoes have been found to be essential in target characterization. In this report we present several processing techniques developed for separating these sequences and enhancing the visibility of those echoes backscattered from the hidden target. include substraction, correlation, spectrum and cepstrum analysis. niques enhances the signal-to-noise ratio and improves the resolution of the measurement. mathematical model of the reverberation and experimental results will be presented. Experimental results are in close agreement with the mathematical model. Techniques used

Pole-zero analysis of voiced speech using group delay characteristics

Abstract: The group delay characteristics of the vocal tract are derived from the approximate spectral envelope defined by peaks at the harmonics of the pitch frequency. Formant and anti-formant frequencies are extracted from the group delay by a simple peak-picking method.

On "Homomorphic deconvolution" by D.J. Jin and J.R. Rogers (Geophysics, 48, 1014-1016); discussion and reply

Abstract: The recent note by Jin and Rogers (1983) presented examples of the failure of the homomorphic transform to invert properly. Since this transform is not only of interest in geophysics, but has also found applications in other fields (Oppenheim and Schafer, 1975), these results are of concern. We consequently attempted to reproduce Jin and Rogers’ results. We failed to do so. In fact, in our experience, the transform has always inverted successfully. Our results using the first example of Jin and Rogers are shown in Figure 1. We used the algorithm of Tribolet (1977) with a modified Goertzel algorithm (Bonzanigo, 1978) for phase unwrapping. The figure is arranged as in Jin and Rogers’ paper. Figure 1a shows the input: impulses separated by 20 samples, of magnitude 2000 and 1999. Figure 1b shows its complex cepstrum. We have set the zero‐quefrency point to zero since this represents a scale factor and can dominate the plotting. Note the minimum delay cepstrum with a small amount of aliasing. The sequence retu...

The analysis of gear noise excitation

Abstract: This paper discusses the main forms of gear analysis which can be used to study transmission noise and ways in which they can be related to gear errors. Frequency analysis is a powerful tool in the study of gear noise problems. In order to understand fully the resulting spectra it is necessary to understand the nature of gear noise and its sideband structure. In many cases a simple Fourier analyser can give good results if used with care and with attention to speed variations. Signature analysis is a more complicated form of frequency analysis available on some sophisticated analysis systems. It gives a concise representation of the noise over any speed range of interest. It relies on speed variations during test runs. Because the speed is increasing (or decreasing) during runs, the data sample times are restricted. This is to ensure that speed changes by only a small amount during capture of each data block. The consequence of this is that frequency resolution is limited and precise identification of sidebands is sometimes impossible. Cepstrum analysis is a method of resolving particular gear meshes but may not always be related directly to subjective assessment of the noise problem. Probably the most useful form of analysis is the order locked spectrum. This can be performed using several of the relatively inexpensive portable analysers currently available with the provision of a suitable external sampling signal. This may cause difficulties depending on the frequency range and/or the resolution required. The sampling rate must be at least 2.5 times the maximum frequency of interest which, when higher harmonics are required, can be over 200 samples per revolution. (TRRL)

Method capable of extracting a value of a spectral envelope parameter with a reduced amount of operations and a device therefor

Abstract: A logarithmic frequency spectrum related to an input signal is converted by the use of an inverse Fourier transform into a cepstrum. The cepstrum has a first and a second frequency component which have a first peak and a second peak spaced apart from the first peak by a preselected period on an axis of frequency, respectively. The second frequency component is processed into a peak controlled frequency component having a controlled peak coincident with the first peak. The peak controlled frequency component and the first frequency component are summed up to produce an ultimate frequency component which corresponds to the value of the envelope parameter.

Talker identification by Cepstrum

Abstract: It is shown that the vocal tract impulse response magnitude should be less variable for a given speaker than other acoustic measures of his speech. Cepstrum analysis is used to deconvolve the vocal tract impulse response and the glottal pressure wave of each of 1850 speech segments taken from running English speech. Linear correlation coefficients derived from pairs of impulse responses are shown to differ, depending upon whether the two impulse responses were taken from the same speaker's utterances, from speakers of the same sex and/or vocal history, or from altogether different speakers. A feasible method of speaker identification which can, in principle, operate automatically is developed from this approach and tested.