One of the most difficult problems in speech analysis is reliable discrimination among silence, unvoiced speech, and voiced speech which has been transmitted over a telephone line. Although several methods have been proposed for making this 3-level decision, these schemes have met with only modest success. In this paper a novel approach to the voiced-unvoiced-silence detection problem is proposed in which a spectral characterization of each of the 3 classes of signal is obtained during a training session, and an LPC distance metric and an energy distance are nonlinearly combined to make the final discrimination. This algorithm has been tested over conventional switched telephone lines, across a variety of speakers, and has been found to have an error rate of about 5%, with the majority of the errors (about 2/3) occurring at the boundaries between signal classes. The algorithm is currently being used in a speaker independent word recognition system.

Voiced-unvoiced-silence detection using the Itakura LPC distance measure

One of the most difficult problems in speech analysis is reliable discrimination among silence, unvoiced speech, and voiced speech which has been transmitted over a telephone line. Although several methods have been proposed for making this three-level decision, these schemes have met with only modest success. In this paper, a novel approach to the voiced-unvoiced-silence detection problem is proposed in which a spectral characterization of each of the three classes of signal is obtained during a training session, and an LPC distance measure and an energy distance are nonlinearly combined to make the final discrimination. This algorithm has been tested over conventional switched telephone lines, across a variety of speakers, and has been found to have an error rate of about 5 percent, with the majority of the errors (about \frac{2}{3} ) occurring at the boundaries between signal classes. The algorithm is currently being used in a speaker-independent word recognition system.

Application of an LPC distance measure to the voiced-unvoiced-silence detection problem

This paper presents a study of speech scramblers based on permutations of samples within an N-block. It has been found that a new family of “uniform” (U) permutations (defined by the address mapping i → ki modulo N; k prime to N) is as effective as pseudorandom (pr) permutations in destroying speech intelligibility. Analytical results show the relation between input and scrambled-signal spectra, while computer simulations compare the effects of scrambling on pam samples and on codes based on adm, apcm, and adpcm. Scrambling is increasingly effective in that order, and encoding delays in adpcm can be as low as 1 to 2 ms. Finally, scrambling has been compared with frequency inversion, which corresponds to sign-inversion in every other Nyquist-rate waveform sample.

On speech encryption using waveform scrambling

The paper is a review of the subject of analogue encryption. The rank correlation approach to the study of permutations is also reviewed. Some families of permutations that have been used in scrambling systems are described. The problems of analogue encryption, as well as its scope, are analysed.

Overview of analogue signal encryption

The performances of five discrete orthogonal transforms in speech encryption systems are compared. The transforms considered are the discrete Fourier transform, discrete cosine transform, Walsh-Hadamard transform, Karhunen-Loeve transform, and discrete prolate spheroidal transform. Four objective measures are used to grade the encryption systems with respect to residual intelligibility and recovered voice quality. A figure of merit based on all the four objective measures is formed. It gives good correlation to the subjective results of residual intelligibility and recovered speech quality. >

Speech encryption using discrete orthogonal transforms

This paper discusses several manipulations of lpc (linear predictive coding) parameters for providing speech encryption. Specifically, the paper considers temporal rearrangement or scrambling of the lpc code sequence, as well as the alternative of perturbing individual samples in the sequence by means of pseudo-random additive or multiplicative noise. The latter approach is believed to have greater encryption potential than the temporal scrambling technique, in terms of the time needed to “break the secrecy code.” The encryption techniques are assessed on the basis of perceptual experiments, as well as by means of a quantitative assessment of speech-spectrum distortion, as given by an appropriate “distance” measure.

http://ieeexplore.ieee.org/iel7/6731005/6768888/06768892.pdf

Speech encryption by manipulations of LPC parameters

This paper discusses several manipulations of LPC (linear predictive coding) parameters for providing speech encryption Specifically, the paper considers temporal rearrangement or scrambling of the LPC code sequence, as well as the alternative of perturbing individual samples in the sequence by means of pseudorandom additive or multiplicative noise The latter approach is believed to have greater encryption potential than the temporal scrarabling technique, in terms of the time needed to “break the secrecy code” The encryption techniques were assessed on the basis of perceptual experiments, as well as by means of a quantitative assessment of speech‐spectrum distortion, as given by an appropriate “distance” measure

M. R. Sambur

Papers

Speech encryption by manipulations of LPC parameters

Speech encryption by manipulations of LPC parameters