Title: On building phonetically and prosodically rich speech corpus for text-to-speech synthesis
Other Titles: VytvářenĂ­ foneticky a prozodicky bohatĂ˝ch Ĺ™eÄŤovĂ˝ch korpusĹŻ v Ăşloze syntĂ©zy Ĺ™eÄŤi z textu
Authors: Matoušek, Jindřich
Romportl, Jan
Citation: MATOUĹ EK, JindĹ™ich; ROMPORTL, Jan. On building phonetically and prosodically rich speech corpus for text-to-speech synthesis. In: Proceedings of the second IASTED international conference on Computational intelligence: 20-22 November 2006, San Francisco, USA. San Francisco: ACTA Press, 2006, p. 442-447. ISBN 0-88986-602-3.
Issue Date: 2006
Publisher: ACTA Press
Document type: článek
article
URI: http://www.kky.zcu.cz/cs/publications/MatousekJ_2006_Onbuilding
http://hdl.handle.net/11025/17022
ISBN: 0-88986-602-3
Keywords: zpracovánĂ­ pĹ™irozenĂ©ho jazyka;syntĂ©za Ĺ™eÄŤi;vĂ˝bÄ›r vÄ›t;Ĺ™eÄŤovĂ˝ korpus;prozodie
Keywords in different language: natural language processing;speech synthesis;sentence selection;speech corpus;prosody
Abstract: ÄŚlánek navrhuje metodu přípravy a pořízenĂ­ Ĺ™eÄŤovĂ©ho korpusu pro Ăşlohu syntĂ©zy Ĺ™eÄŤi z textu s dynamickĂ˝m vĂ˝bÄ›rem jednotek řízenou pomocĂ­ symbolickĂ© prozodie. SoustĹ™edĂ­ se na algoritmus vĂ˝bÄ›ru foneticky a prozodicky bohatĂ˝ch vÄ›t. Foneticky pĹ™epsanĂ© vÄ›ty jsou obohaceny o symbolickĂ˝ popis na hrubĂ© prozodickĂ© Ăşrovni s respektovánĂ­m typu prozodĂ©mu, ve kterĂ©m se fony objevujĂ­. VĂ˝slednĂ˝ algoritmus pak vybĂ­rá vÄ›ty s ohledem na fonetická i prozodická kritĂ©ria. Abychom tĂ©Ĺľ pokryli i supravÄ›tnĂ© prozodickĂ© jevy, náhodnÄ› jsme vybrali odstavce a nahráli je. NovĂ˝ Ĺ™eÄŤovĂ˝ korpus se mĹŻĹľe využít k syntĂ©ze Ĺ™eÄŤi s dynamickĂ˝m vĂ˝bÄ›rem jednotek a takĂ© k trĂ©novánĂ­ datovÄ› orientovanĂ©ho prozodickĂ©ho parseru.
Abstract in different language: This paper proposes a way of preparing and recording a speech corpus for unit selection text-to-speech speech synthesis driven by symbolic prosody. The research is focused on a phonetically and prosodically rich sentence selection algorithm. Symbolic description on a deep prosody level is used to enrich the phonetic representation of sentences (by respecting the prosodeme types phones appear in). The resulting algorithm then selects sentences with respect to both phonetic and prosodic criteria. To cover supra-sentential prosody phenomena, paragraphs were selected at random and recorded as well. The new speech corpus can be utilised in unit selection speech synthesis and also for training a data-driven prosodic parser.
Rights: Â© JindĹ™ich Matoušek - Jan Romportl
Appears in Collections:Články / Articles (MMI)
Články / Articles (KKY)
Články / Articles (KIV)

Files in This Item:
File Description SizeFormat 
MatousekJ_2006_Onbuilding.pdfPlný text138,96 kBAdobe PDFView/Open


Please use this identifier to cite or link to this item: http://hdl.handle.net/11025/17022

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