In this contribution a novel method is presented, allowing the detection of resonance frequencies of an externally excited vocal tract in a fast and robust non-invasive manner. The method can be applied alternatively to linear prediction coding (LPC) analysis for patients with voice and speech disorders. Sweep signals are emitted and recorded simultaneously from the small end of a tube that is ...

Segmentation of CSF and pulsative blood flow, based on a single phase contrast MRA (PC-MRA) image can lead to imperfect classifications. In this paper, we present a novel automated flow segmentation method by using PC-MRA image series. The intensity time series of each pixel is modeled as an autoregressive (AR) process and features including the Linear Prediction Coefficients (LPC), covariance ...

The objective and subjective classification of unvoiced stop consonants in varying vowel contexts were studied. The objective classification was based on auditory feature vectors obtained by warped linear prediction (WLP) and vector autoregressive (VAR) models for parameter trajectories. In the case of known vowel the unvoiced consonants were classified 98-100% correctly based on the auditory s...

It is known that voice quality plays an important role in expressive speech. In this paper, we present a methodology for modifying vocal effort level, which can be applied by text-to-speech (TTS) systems to provide the flexibility needed to improve the naturalness of synthesized speech. This extends previous work using low order Linear Prediction Coefficients (LPC) where the flexibility was con...

A non-linear predictive coding based algorithm is proposed in the paper for lossless image compression. The algorithm uses two neighbouring pixels, one left and the other top, as a pioneering block to search for the best matched blocks inside a pre-defined window. The corresponding pixels associated with the best matched blocks are then taken to produce the predictive value, together with the t...

The paper describes the use of two recognizers fed by different acoustic features. The first recognizer performs Multiple Resolution Analysis (MRA) while the other recognizer computes JRASTA Perceptual Linear Prediction Coefficients (JRASTAPLP). The two recognizers use the same denoising method but perform different partitions of their acoustic spaces. Experiments with the Italian and Spanish c...

The mel-cepstral coefficients are often calculated from the linear prediction coefficients by using recursion formulas. However, the obtained mel-cepstral coefficients have errors caused by truncation in the quefrency domain. The purpose of this report is to point out that the melcepstral coefficients can be calculated from the LP (Linear Prediction) coefficients using the recursion formulas wi...

A new method for speech parametric coding at medium outpul rate is presented. The availability of new el~ctronic components changes the hypotheses assumed in previous studies. A parametric coding scheme with medium outpul rate now seems appropriate for rule-based synthesis. The weak spot in most parametric synthesizers lies in the poor model used to represent the excitation signal. Our approach...

The performance of Support Vector Machines (SVMs) is highly dependent on the choice of a kernel function suited to the problem at hand. In particular, the kernel implicitly performs a feature selection which is the most important stage in any texture classification algorithm. In this work a new Gabor filter based kernel for texture classification with SVMs is proposed. The proposed kernel funct...

A new voice conversion algorithm that modifies a source speaker's speech to sound as if produced by a target speaker is presented. It is applied to a residualexcited LPC text-to-speech diphone synthesizer. Spectral parameters are mapped using a locally linear transformation based on Gaussian mixture models whose parameters are trained by joint density estimation. The LPC residuals are adjusted ...