Automatic modeling of duration in a Spanish text-to-speech system using neural networks. de Córdoba, R.; Vallejo, J. Á.; Montero, J. M.; Gutiérrez Arriola, J. M.; López Carmona, M. Á.; and Pardo, J. M. In Eurospeech 1999. Proceedings of the 6th European Conference on Speech Communication and Technology, pages 1619-1622, Budapest, Hungary, September 5-9, 1999.
Automatic modeling of duration in a Spanish text-to-speech system using neural networks [pdf]Paper  abstract   bibtex   
Accurate prediction of segmental duration from text in a text-tospeech system is difficult for several reasons. One specially relevant is the great quantity of contextual factors that affect timing and how to model them. There are many parameters that affect duration, but not all of them are always relevant. We present a complete environment in which to decide which parameters are more relevant in different situations and the best way to code them. The system is based in a neural network absolutely configurable, and the main effort is made in the parameters to be used, including the contextual effects using windows of variable length.
@inproceedings{de_cordoba_automatic_1999,
	Address = {Budapest, Hungary, September 5-9, 1999},
	Author = {de Córdoba, Ricardo and Vallejo, José Ángel and Montero, Juan Manuel and Gutiérrez Arriola, Juana María and López Carmona, Miguel Ángel and Pardo, José Manuel},
	Booktitle = {Eurospeech 1999. Proceedings of the 6th European Conference on Speech Communication and Technology},
	Date = {1999},
	Date-Modified = {2016-09-24 18:56:01 +0000},
	Keywords = {duration, phonetics, prosody, Spanish, speech synthesis, speech technology, temporal factors, text-to-speech},
	Pages = {1619-1622},
	Title = {Automatic modeling of duration in a Spanish text-to-speech system using neural networks},
	Url = {http://www-gth.die.upm.es/research/documentation/AI-52Aut-99.pdf},
	Abstract = {Accurate prediction of segmental duration from text in a text-tospeech system is difficult for several reasons. One specially relevant is the great quantity of contextual factors that affect timing and how to model them. There are many parameters that affect duration, but not all of them are always relevant. We present a complete environment in which to decide which parameters are more relevant in different situations and the best way to code them. The system is based in a neural network absolutely configurable, and the main effort is made in the parameters to be used, including the contextual effects using windows of variable length.},
	Bdsk-File-1 = {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},
	Bdsk-Url-1 = {http://www-gth.die.upm.es/research/documentation/AI-52Aut-99.pdf}}
Downloads: 0