Title: | Reducing footprint of unit selection TTS system by excluding utterances from source speech corpus |
Other Titles: | Snižování paměťových nároků systému TTS pracujícího na principu výběru jednotek vyhozením promluv ze zdrojového řečového korpusu |
Authors: | Matoušek, Jindřich Tihelka, Daniel Hanzlíček, Zdeněk |
Citation: | MATOUŠEK, Jindřich; TIHELKA, Daniel; HANZLÍČEK, Zdeněk. Reducing footprint of unit selection TTS system by excluding utterances from source speech corpus. In: Speech processing 19th czech – german workshop 29th September – 1st October 2009. Prague: Institute of Photonics and Electronics Academy of Sciences of the Czech Republic, 2009, p. 92-98. ISBN 978-80-86269-18-4. |
Issue Date: | 2009 |
Publisher: | Institute of Photonics and Electronics Academy of Sciences of the Czech Republic |
Document type: | článek article |
URI: | http://www.kky.zcu.cz/cs/publications/MatousekJ_2009_ReducingFootprintof http://hdl.handle.net/11025/17017 |
ISBN: | 978-80-86269-18-4 |
Keywords: | syntéza řeči;výběr jednotky;korpus řeči |
Keywords in different language: | speech synthesis;unit selection;speech corpus |
Abstract in different language: | Current unit selection speech synthesis systems are capable of producing speech of a high quality at the expense of enormous computational and storage requirements. In this paper, the analysis of an existing large speech corpus employed for unit-selection-based synthesis of Czech speech is performed. Subsequently, a procedure for the exclusion of some amount of utterances from the source speech corpus is proposed. The procedure is based on the statistics of the utilisation of all utterances during text-to-speech synthesis of a large portion of texts. The exclusion of whole utterances was preferred over the exclusion of the particular instances of speech units in order to preserve the main feature of unit selection framework - to select as longest sequence of contiguous speech units as possible. After the exclusion, the footprint of the system was reduced approximately by 42 %. The resulting synthetic speech was then judged by means of 5-scale CCR listening tests and evaluated in average as only "slightly worse" than speech generated by the baseline (i.e. not reduced) system. |
Rights: | © Jindřich Matoušek - Daniel Tihelka - Zdeněk Hanzlíček |
Appears in Collections: | Články / Articles (NTIS) Články / Articles (KIV) |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
MatousekJ_2009_ReducingFootprintof.pdf | Plný text | 742,59 kB | Adobe PDF | View/Open |
Please use this identifier to cite or link to this item:
http://hdl.handle.net/11025/17017
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.