没有合适的资源?快使用搜索试试~ 我知道了~
资源推荐
资源详情
资源评论
Lecture Notes in Artificial Intelligence 3445
Edited by J. G. Carbonell and J. Siekmann
Subseries of Lecture Notes in Computer Science
Gérard Chollet Anna Esposito
Marcos Faundez-Zanuy Maria Marinaro (Eds.)
Nonlinear
Speech Modeling
and Applications
Advanced Lectures and Revised Selected Papers
13
Series Editors
Jaime G. Carbonell, Carnegie Mellon University, Pittsburgh, PA, USA
Jörg Siekmann, University of Saarland, Saarbrücken, Germany
Volume Editors
Gérard Chollet
CNRS URA-820, ENST, Dept. TSI
46 rue Barrault, 75634 Paris cedex 13, France
E-mail: chollet@tsi.enst.fr
Anna Esposito
Second University of Naples, Department of Psychology
and
IIASS, International Institute for Advanced Scientific Studies
Via Pellegrino 19, Vietri sul Mare (SA), Italy
E-mail: iiass.annaesp@tin.it
Marcos Faundez-Zanuy
Escola Universitaria Politecnica de Mataro
Avda. Puig i Cadafalch 101-111, 08303 Mataro (Barcelona), Spain
E-mail: faundez@eupmt.es
Maria Marinaro
University of Salerno "E.R.Caianiello", Dept. of Physics
Via Salvatore Allende, Baronissi, 84081 Salerno, Italy
and
IIASS, International Institute for Advanced Scientific Studies
Via Pellegrino 19, Vietri sul Mare (SA), Italy
E-mail: iiass.vietri@tin.it
Library of Congress Control Number: 2005928448
CR Subject Classification (1998): I.2.7, J.5, C.3
ISSN 0302-9743
ISBN-10 3-540-27441-3 Springer Berlin Heidelberg New York
ISBN-13 978-3-540-27441-4 Springer Berlin Heidelberg New York
This work is subject to copyright. All rights are reserved, whether the whole or part of the material is
concerned, specifically the rights of translation, reprinting, re-use of illustrations, recitation, broadcasting,
reproduction on microfilms or in any other way, and storage in data banks. Duplication of this publication
or parts thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965,
in its current version, and permission for use must always be obtained from Springer. Violations are liable
to prosecution under the German Copyright Law.
Springer is a part of Springer Science+Business Media
springeronline.com
© Springer-Verlag Berlin Heidelberg 2005
Printed in Germany
Typesetting: Camera-ready by author, data conversion by Olgun Computergrafik
Printed on acid-free paper SPIN: 11520153 06/3142 543210
Preface
This volume contains invited and contributed papers presented at the 9th International
Summer School “Neural Nets E.R. Caianiello” on Nonlinear Speech Processing: Algo-
rithms and Analysis, held in Vietri sul Mare, Salerno, Italy, during September 13–18,
2004.
The aim of this book is to provide primarily high-level tutorial coverage of the
fields related to nonlinear methods for speech processing and analysis, including new
approaches aimed at improving speech applications.
Fourteen surveys are offered by specialists in the field. Consequently, the volume
may be used as a reference book on nonlinear methods for speech processing and anal-
ysis. Also included are fifteen papers that present original contributions in the field and
complete the tutorials.
The volume is divided into five sections: Dealing with Nonlinearities in Speech Sig-
nal, Acoustic-to-Articulatory Modeling of Speech Phenomena, Data Driven and Speech
Processing Algorithms, Algorithms and Models Based on Speech Perception Mecha-
nisms, and Task-Oriented Speech Applications.
Dealing with Nonlinearities in Speech Signals is an introductory section where non-
linear aspects of the speech signal are introduced from three different points of view.
The section includes three papers. The first paper, authored by Anna Esposito and Maria
Marinaro, is an attempt to introduce the concept of nonlinearity revising several nonlin-
ear phenomena observed in the acoustics, the production and the perception of speech.
Also discussed is the engineering endeavor to model these phenomena.
The second paper, by Marcos Faundez-Zanuy, gives an overview of nonlinear pre-
dictive models, with special emphasis on neural nets, and discusses several well-known
nonlinear strategies, such as multistart random weights initialization, regularization,
early stop with validation, committees of neural nets, and neural net architectures.
The third paper, by Simon Haykin, faces the problem of processing nonlinear, non-
Gaussian, and nonstationary signals describing the mathematical implications derived
by these assumptions. The topic has important practical implications of its own, not
only in speech but also in the field of signal processing.
Acoustic-to-Articulatory Modeling of Speech Phenomena deals with problems re-
lated to the acoustic-phonetic theory in which basic speech sounds are characterized
according to both their articulatory features and the associated acoustic measurements.
Fundamental and innovative ideas in speech production are covered. This section con-
tains three papers. The first paper, auth ored by Eric Keller, discusses voice quality
within a large predictive and methodological framework. Voice quality phenomena
are reviewed at two levels: (1) at the level of independent variables, topic-, affective-,
attitude-, emotion-, gender-, articulation-, language-, sociolect-, gender-, age- and
speaker-related predictors; and (2) at the level of dependent variables, where the em-
pirical identification of voice quality parameters in the speech signal are summarized.
Specifically, Fant’s original and revised source-filter models are reviewed.
剩余443页未读,继续阅读
资源评论
voipp2p
- 粉丝: 0
- 资源: 4
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功