Please use this identifier to cite or link to this item: http://hdl.handle.net/10603/183498
Full metadata record
DC FieldValueLanguage
dc.coverage.spatial
dc.date.accessioned2017-11-30T09:12:51Z-
dc.date.available2017-11-30T09:12:51Z-
dc.identifier.urihttp://hdl.handle.net/10603/183498-
dc.description.abstractSpeech is the most natural, convenient, and useful means of communication. Moreover, speech can convey other information such as emotion, attitude, and speaker individuality. Therefore, it is important to realize a man-machine interface to facilitate communication between people and computers. newlineThe objective of the research is to develop high quality speech synthesizer, primarily in Hindi language based on corpus speech synthesis using concatenation technique. This technique is used as the naturalness obtained is highest as compared to other techniques. A major problem associated with implementation of speech synthesizers for Hindi is lack of availability of database. For development of efficient speech synthesizer, it is mandatory to have phonetically rich sentences, consistent recording conditions and availability of properly annotated database tagged with prosodic features. Literature survey has revealed availability of such database for English and other European languages but they are not available in Indian context. Hence, it poses a major challenge to develop synthesizer for Hindi language. A sincere effort has been made to address this crucial and time consuming issue i.e. Development of an algorithm for automatically creating appropriate synthesis units. newlineSince larger domain synthesizer cannot be created using syllable as the basic unit the phoneme has been chosen as the basic unit. In the proposed technique, the base line Hidden Markov Model (HMM) has been used for segmentation of speech signals. It is applied on single speaker segmentation task, using Hindi speech database. The automatic phoneme segmentation framework evolved imitates the human phoneme segmentation process. A set of 42 Hindi phonemes were chosen for the segmentation experiment, wherein, Continuous Density Hidden Markov Model (CDHMM) with a mixture of Gaussian distribution were used...
dc.format.extent
dc.languageEnglish
dc.relation
dc.rightsuniversity
dc.titleCreating an Unlimited Voice Response in Hindi
dc.title.alternative
dc.creator.researcherArchana Balyan
dc.description.note
dc.contributor.guideAmita Dev and S.S. Agrawal
dc.publisher.placeDelhi
dc.publisher.universityGuru Gobind Singh Indraprastha University
dc.publisher.institutionUniversity School of Engineering and Technology
dc.date.registered2007
dc.date.completed2016
dc.date.awarded18/03/2016
dc.format.dimensions
dc.format.accompanyingmaterialCD
dc.source.universityUniversity
dc.type.degreePh.D.
Appears in Departments:University School of Engineering and Technology

Files in This Item:
File Description SizeFormat 
01_front pages .pdfAttached File930.53 kBAdobe PDFView/Open
02_declaration.pdf222.53 kBAdobe PDFView/Open
03_thesis.pdf4.57 MBAdobe PDFView/Open


Items in Shodhganga are protected by copyright, with all rights reserved, unless otherwise indicated.