On automatic recognition of spectrally reduced speech synthesized from amplitude and frequency modulations - Archive ouverte HAL Accéder directement au contenu
Rapport (Rapport De Recherche) Année : 2009

On automatic recognition of spectrally reduced speech synthesized from amplitude and frequency modulations

Résumé

This report investigates the behavior of automatic speech recognition (ASR) system with spectrally reduced speech (SRS) synthesized from subband amplitude modulations (AMs) and frequency modulations (FMs). Acoustic analysis shows that the resynthesis of SRS from only AM components helps alleviate certain non-linguistic iabilities in the original speech signal. When the SRS spectral resolution is suciently good, this alleviation not only has no consequence but also yields comparable or even better ASR word accuracy compared to that attained with original clean speech signal. In contrast, FM components support human speech recognition but yield no significant improvement in terms of ASR word accuracy when the SRS spectral resolution is sufficiently good.
Fichier principal
Vignette du fichier
DO2.pdf (317.42 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)
Loading...

Dates et versions

hal-02316498 , version 1 (15-10-2019)

Identifiants

  • HAL Id : hal-02316498 , version 1

Citer

Cong Thanh Do, Dominique Pastor, André Goalic. On automatic recognition of spectrally reduced speech synthesized from amplitude and frequency modulations. [Research Report] Laboratoire en sciences et technologies de l'information, de la communication et de la connaissance (UMR CNRS 6285 - Télécom Bretagne - Université de Bretagne Occidentale - Université de Bretagne Sud - ENSTA Bretagne - Ecole Nationale d'ingénieurs de Brest); Dépt. Signal et Communications (Institut Mines-Télécom-Télécom Bretagne-UEB). 2009, pp.17. ⟨hal-02316498⟩
11 Consultations
10 Téléchargements

Partager

Gmail Facebook X LinkedIn More