Comparaison des propriétés acoustiques de la parole lue, préparée et conversationnelle en français
Abstract
In this paper, we investigate the acoustic properties of vocalic phonemes in three speaking styles: Read speech, broadcast news and casual spontaneous speech. Our aim is to understand better why speech recognition systems still fail to achieve good performances on spontaneous speech. Using Nakamura's method \cite{nakamura2008}, we use classical speech recognition features, MFCC, and try to represent the effects of the speaking styles on the spectral space. We happen to find some cues, and it also seems that phone duration also plays an important role regarding spectral reduction, especially for spontaneous speech.
Origin : Files produced by the author(s)
Loading...