An image-inspired audio sharpness index

Abstract : We propose a new non-intrusive (reference-free) objective measure of speech intelligibility that is inspired from previous works on image sharpness. We define the audio Sharp- ness Index (aSI) as the sensitivity of the spectrogram sparsity to the convolution of the signal with a white noise, and we calculate a closed-form formula of the aSI. Experiments with various speakers, noise and reverberation conditions show a high correlation between the aSI and the well-established Speech Transmission Index (STI), which is intrusive (full-reference). Additionally, the aSI can be used as an intelligibility or clarity criterion to drive sound enhancement algorithms. Experimental results on stereo mixtures of two sounds show that blind source separation based on aSI maximization performs well for speech and for music
Type de document :
Communication dans un congrès
EUSIPCO 2017 : 25th European Signal Processing Conference, Aug 2017, Kos Island, Greece. IEEE Computer Society, Proceedings EUSIPCO 2017 : 25th European Signal Processing Conference, pp.683 - 687, 2017, 〈10.23919/EUSIPCO.2017.8081294〉
Liste complète des métadonnées

https://hal.archives-ouvertes.fr/hal-01682856
Contributeur : Médiathèque Télécom Sudparis & Institut Mines-Télécom Business School <>
Soumis le : vendredi 12 janvier 2018 - 15:23:34
Dernière modification le : mardi 11 septembre 2018 - 15:20:04

Lien texte intégral

Identifiants

Citation

Gael Mahé, Lionel Moisan, Mihai Mitrea. An image-inspired audio sharpness index. EUSIPCO 2017 : 25th European Signal Processing Conference, Aug 2017, Kos Island, Greece. IEEE Computer Society, Proceedings EUSIPCO 2017 : 25th European Signal Processing Conference, pp.683 - 687, 2017, 〈10.23919/EUSIPCO.2017.8081294〉. 〈hal-01682856〉

Partager

Métriques

Consultations de la notice

139