Adaptive time-frequency analysis for cognitive source separation [Elektronische Ressource] / submitted by Sylvia Kümmel

SSAATRIASVRSaarland UniversityFaculty of Natural Sciences and Technology IDepartment of Computer ScienceDissertationfor obtaining the title of Doctor of Engineering of the Facultiesof Natural Sciences and Technology of Saarland UniversityAdaptive Time-Frequency Analysis forCognitive Source Separationsubmitted bySylvia KümmelSupervisorProf. Dr.-Ing. Thorsten HerfetSaarbrücken, December 2009IEEVNISNIUSiiDate of Colloquium: 21.04.2010Dean of Faculty: Prof. Dr. Holger HermannsMembers of examination board:Prof. Dr.-Ing. Thorsten Herfet, Saarland UniversityProf. Udo Zölzer, Helmut Schmidt University HamburgProf. Dr. Antonio Krüger, Saarland UniversityDr. Mark Hillebrand, Saarland UniversityiiiStatutory declarationHereby I affirm in lieu of an oath, that I made the present thesis autonomously and withoutother than the indicated auxiliary means. The data used indirectly or from other sources andconcepts are characterized with lists of sources. The thesis has not been submitted for academicdegree consideration either nationally or internationally in identical or similar from to date.Saarbrücken, April 27, 2010Sylvia KümmelDeclaration of ConsentHerewith I agree that my thesis will be made available through the library of the ComputerScience Department.Saarbrücken, April 27, 2010Sylvia KümmelivvAbstractThis thesis introduces a framework for separating two speech sources in non-ideal, reverberantenvironments.