Perceptually Motivated Non-Intrusive Assessment of Speech Quality

Christophe Veaux, Vincent Barriac

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In this contribution, we present a non-intrusive assessment technique that aims to reproduce some human perceptual mechanisms. We model a set of perceptually relevant features extracted after an auditory transform of clean speech. This internal source-model is used to characterize any occurring distortions. Preliminary results are given to illustrate the characterization of distortions by some perceptual features. It seems that the temporal resolution of our auditory transform provides interesting information to characterize non-linear distortions induced by low bit-rate coders
Original languageEnglish
Title of host publicationProceedings: Measurement of Speech and Audio Quality in Networks 2002
Publication statusPublished - 2002
EventMeasurement of Speech and Audio Quality in Networks - On-Line, Prague, Czech Republic
Duration: 25 Jan 20021 Mar 2002

Workshop

WorkshopMeasurement of Speech and Audio Quality in Networks
CountryCzech Republic
CityPrague
Period25/01/021/03/02

Fingerprint Dive into the research topics of 'Perceptually Motivated Non-Intrusive Assessment of Speech Quality'. Together they form a unique fingerprint.

Cite this