3 - Modèle ARMA à coefficient dépendant du temps: estimateurs et applications

Show full item record

Files in this item

PDF 003.PDF TEXTE.pdf 1.355Mb

Pour citer ce document :
URI: http://hdl.handle.net/2042/1609
Title: 3 - Modèle ARMA à coefficient dépendant du temps: estimateurs et applications
Author: GRENIER (Y.)
Abstract: On présente une méthodologie de modélisation des signaux non stationnaires au moyen d'une classe de modèles autorégressifs à moyenne ajustée. Application à la synthèse de la parole
Description: This paper describes a framework for time-dependent modelling of nonstationary signais, based upon autoregressive or ARMA models with time-varying coefficients . Previous works by the author are summarized here : starting front ideas proposed by Rao, Mendel and then Liporace, it is assumed that the coefficients of the model are expressed as linear combinations of known time functions . A class of estimators is described in this text, comprising autoregressive models, moving average models, and also lattice filters, parametrized through reflection coefficients or through Log Area Ratios . Speech synthesis, which is one of the applications for which these models are efficients, concludes the paper.
Subject: Traitement parole; Parole; Synthèse parole; Codage; Modèle autorégressif; Modélisation; Filtre treillis; Signal non stationnaire
Publisher: GRETSI, Saint Martin d'Hères, France
Date: 1986

This item appears in the following Collection(s)

Show full item record





Advanced Search