Using Deep Neural Networks for Smoothing Pitch Profiles in Connected Speech

This paper presents a new pitch tracking smoother based on deep neural networks (DNN). It leverages Long Short-Term Memories, a particular kind of recurrent neural network, for correcting pitch detection errors produced by state-of-the-art Pitch Detection Algorithms. The proposed system has been ext...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Michele Ferro, Fabio Tamburini
Formato:	article
Lenguaje:	EN
Publicado:	Accademia University Press 2019
Materias:	Social Sciences H Computational linguistics. Natural language processing P98-98.5
Acceso en línea:	https://doaj.org/article/7241abc1b2644793b4cac8ebe884d74f
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:7241abc1b2644793b4cac8ebe884d74f
record_format	dspace
spelling	oai:doaj.org-article:7241abc1b2644793b4cac8ebe884d74f2021-12-02T09:52:34ZUsing Deep Neural Networks for Smoothing Pitch Profiles in Connected Speech2499-455310.4000/ijcol.476https://doaj.org/article/7241abc1b2644793b4cac8ebe884d74f2019-12-01T00:00:00Zhttp://journals.openedition.org/ijcol/476https://doaj.org/toc/2499-4553This paper presents a new pitch tracking smoother based on deep neural networks (DNN). It leverages Long Short-Term Memories, a particular kind of recurrent neural network, for correcting pitch detection errors produced by state-of-the-art Pitch Detection Algorithms. The proposed system has been extensively tested using two reference benchmarks for English and exhibited very good performances in correcting pitch detection algorithms outputs when compared with the gold standard obtained with laryngographs.Michele FerroFabio TamburiniAccademia University PressarticleSocial SciencesHComputational linguistics. Natural language processingP98-98.5ENIJCoL, Vol 5, Iss 2, Pp 33-48 (2019)
institution	DOAJ
collection	DOAJ
language	EN
topic	Social Sciences H Computational linguistics. Natural language processing P98-98.5
spellingShingle	Social Sciences H Computational linguistics. Natural language processing P98-98.5 Michele Ferro Fabio Tamburini Using Deep Neural Networks for Smoothing Pitch Profiles in Connected Speech
description	This paper presents a new pitch tracking smoother based on deep neural networks (DNN). It leverages Long Short-Term Memories, a particular kind of recurrent neural network, for correcting pitch detection errors produced by state-of-the-art Pitch Detection Algorithms. The proposed system has been extensively tested using two reference benchmarks for English and exhibited very good performances in correcting pitch detection algorithms outputs when compared with the gold standard obtained with laryngographs.
format	article
author	Michele Ferro Fabio Tamburini
author_facet	Michele Ferro Fabio Tamburini
author_sort	Michele Ferro
title	Using Deep Neural Networks for Smoothing Pitch Profiles in Connected Speech
title_short	Using Deep Neural Networks for Smoothing Pitch Profiles in Connected Speech
title_full	Using Deep Neural Networks for Smoothing Pitch Profiles in Connected Speech
title_fullStr	Using Deep Neural Networks for Smoothing Pitch Profiles in Connected Speech
title_full_unstemmed	Using Deep Neural Networks for Smoothing Pitch Profiles in Connected Speech
title_sort	using deep neural networks for smoothing pitch profiles in connected speech
publisher	Accademia University Press
publishDate	2019
url	https://doaj.org/article/7241abc1b2644793b4cac8ebe884d74f
work_keys_str_mv	AT micheleferro usingdeepneuralnetworksforsmoothingpitchprofilesinconnectedspeech AT fabiotamburini usingdeepneuralnetworksforsmoothingpitchprofilesinconnectedspeech
_version_	1718397952286785536

Using Deep Neural Networks for Smoothing Pitch Profiles in Connected Speech

Ejemplares similares