Codage conjoint source-canal des paramètres lsf du codeur FS1016: application de la quantification vectorielle par treillis optimisée pour un canal bruité

In this paper, we present an optimized trellis coded vector quantization (tcvq) coding system designed for the effective and robust coding of lsf spectral parameters at low bit rate. The aim of this system, called at the beginning « lsf-otcvq Encoder », is to achieve a low bit rate transparent quantization of lsf parameters of the us Federal Standard (fs1016) speech coder. Once the effectiveness of our lsf-otcvq encoders (with weighted distance) was proven in the case of ideal transmissions over noiseless channel, we were interested after in the improvement of their robustnesses for real transmissions over noisy channel. To implicitly protect the transmission indices of our lsf-otcvq coders incorporated in the Fsl016, we used a joint source-channel coding carried out by the channel optimized vector quantization (covq) method. In the case of transmissions over noisy channel, we will show that our new encoding system, called “covq-lsf-otcvq Encoder”, would be able to contribute significantly in the improvement of the fs1016 performances by ensuring a good coding robustness of its lsf spectral parameters.     

Mesure automatique des paramétres de bruit des MESFET hyperfréquences

An automatic noise parameter measurement test set is described in two sections : (i) A novel noise parameter extraction procedure from a set of noise figure measurements is developped. The experimental minimum noise figure is no longer required and the sensitivity to measurement incertainties is found to be very low. (ii) An automatic input tuner with a large number of precalibrated and reproducible positions is designed and realized. The last part give examples of broadband (8 to 17 GHz) noise parameters characterization ofmmic GaAsmesfets when both the automatic tuner and the extraction procedure are used.     

La quantification vectorielle des signaux: approche algébrique

Spherical vector quantization (svq)consists of quantizing separately on the one hand the norm of a vector and, on the other hand, its orientation.Implementing such quantizer requires a set of normalized vectors (points on the unit sphere) which are the possible rounded-off values for the vector’s orientation. A clever way to design such a quantizer consists of taking a spherical subset from a regular point lattice. In particular a fast algorithm is given for the Gosset lattice in 8 dimensions. A second algorithm, equally optimal, is given for the Leech lattice in 24 dimensions. The performances of such quantizers for white Gaussian noise is compared to information theory limits. Finally, the application of these techniques for encoding speech residual is discussed.     

Influence des propriétés de surface sur les paramètres électriques des TEC GaAs

The study of gate leakage current, low frequency current noise, transconductance frequency dispersion and lagging on 660 low noise GaAs fet,led us to analyse the effect of surface properties on these parameters. It appears two groups of uncorrelated parameters which are related to surface properties.     

Caractérisation des images stationnaires par des modèles non gaussiens bidimensionnels à moyenne ajustée

In this paper, four batch least squares linear approaches are presented for identification of non minimum phase bidimensional non Gaussian moving average (MA) models, and a relationship between autocorrelation and cumulant sequences is given. One of the proposed methods is cumulant-based only but the others use both autocorrelations and m-th order cumulants (m. > 2). Three of them are derived from the Brillinger-Rosenblatt’s non linear relation by using the Tugnait’s closed-form solution. Also, we generalize to m-th order cumulants the 2-D version of Giannakis-Mende’s approach. By simulations, we test and compare the Tugnait’s closed-form solution and the proposed methods, and we evaluate the performance of our relationship in noisy environment. Finally, we propose to characterize textured images by a 2-D ma model witch will be identified using our approaches in noisy and free noise cases.     

Étude de la précision de l’estimation de la densité spectrale de puissance croisée,obtenue par transformation de fourier de la fonction de corrélation

The authors compare two different estimators of the crosscorrelation function and show that they give equivalent estimators in the frequency domain. They calculate the first and second order moments of the estimator of the cross-spectral-density (DSPC)due to the statistical errors and the quantification of inputs. From these general formulas, they perform numerical calculations for a particular case and they compare them to practical measurements. They show the interesting fact, that the variance of the statistical errors on the dspc is inferior by estimating its phase than by estimating its module. Furthermore, the quantification of the inputs may influence the variance of the phase of the DSPC estimated.     

Définition théorique des paramètres en fort niveau d’un multipôle actif non linéaire

In this paper, large signal non linear P parameters are rigorously defined, allowing the characterization of non linear active n-port circuits in the frequency domain with the aim of computer aided design of power microwave devices. No a priori conditions bind the parameters to the variables describing the input signal. However the following rules apply to the parameters: they may be function of any input variables of the n-port circuit; the expression of the large-signal parameters should become identical to the expression of the linear parameters characterizing the n-port circuit under linear operating conditions when the values of the input variables tend to zero. An application of the definition to afet is demonstrated.     

Du droit des télécommunications au droit des communications électroniques: quel changement de modèle?

The sector-based model stemming from the regulatory framework for telecommunications, was extended to electronic communications, a term which takes into account the convergence of the telecommunications, broadcasting andIt sectors. At the same time, this model tends to dissolve in the general competition rules which are supposed to replace the sectoral regulation. Already, this regulation applies exclusively to operators having a significant market power on the relevant markets. But this competition model itself is subordinated to general interest objectives, as universal service or cultural diversity. Thus, a new evolution is initiated, where the objectives of contents regulation could help to redefine some aspects of network regulation.     

Amélioration des performances des systèmes à haut débit par une méthode de réduction de la diaphonie dans les câbles multiconducteurs de télécommunications

dsl (Digital Subscriber Line) technology provides low cost access to the consumers and sources of information for any one subscribing to classical analogical telephony. Indeed xdsl services enable high-speed digital transmission on conventional telephone lines. For sixteen years that the internet has risen, a great number of offers proposing increasing bit rate have been developed. Considering this increasing demand of high bit rate access, telecommunication operators perpetually work on finding solutions to optimize the copper lines channel. In that way, we present here a new method of crosstalk reduction in a multiconductor telecommunication cable. Interests and limitations of this method are presented and some simulations on a frequency band extending to 30 MHz will picture those theoretical results.     

Antennes en réseau à bas niveau de lobes secondaires par commutation de la loi d’alimentation. Optimisation par l’algorithme génétique

The paper presents an approach of synthesis from a new thinned microstrip antennas array with binary feeding law. The operating principle of this type of antennas array consists in establishing a number of elements supplied at the center of the array as well as the feeding of the extreme elements. The problem of synthesis with constraints is dealt with only by the research of the binary feeding law of the array eccentric elements, thus allowing to reduce to the maximum the sidelobes level. To solve this type of problem, the genetic algorithm is proposed.