Segmentation and labelling of Slovenian diphone inventories (CROSBI ID 515751)
Prilog sa skupa u zborniku | izvorni znanstveni rad | međunarodna recenzija
Podaci o odgovornosti
Gros, Jerneja ; Ipšič, Ivo ; Dobrišek, Simon ; Mihelič, France ; Pavešić, Nikola
engleski
Segmentation and labelling of Slovenian diphone inventories
Preparation, recording, segmentation and pitch labelling of Slovenian diphone inventories are described. A special user friendly interface package was developed in order to faciliate these operations. As acquisition of a labelled diphone inventory or adaptation of a speech synthesis system to synthesise further voices is manually intensive, an automatic procedure is required. A speech recognizer, based on Hidden Markov Models in forced segmentation mode is used to outline phone boundaries withhin spoken logatoms. A statistical evaluation of manual and automatic segmentation discrepancies is performed so as to estimate the reliability of automatically derived labels. Finally, diphone boundaries are determined and pitch markers are assigned to voiced sections of the speech signal.
tex-to-speech; diphone synthesis; HMM
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
Podaci o prilogu
298-303.-x.
1996.
objavljeno
Podaci o matičnoj publikaciji
COLING-96 / the 16th International Conference on Computational Linguistic, Copenhagen,
Kopenhagen: Center for Sprogteknologi,
Podaci o skupu
COLING-96 / the 16th International Conference on Computational Linguistic
predavanje
05.08.1996-09.08.1996
Kopenhagen, Danska