|Research Area:||Speech Analysis||Year:||2009|
|Type of Publication:||In Proceedings|
|Authors:||Sri Harish Reddy M., Guruprasad S., B. Yegnanarayana|
This paper presents an analysis of speech produced at different speaking rates using the features of speech signal at subsegmental level. These features are mostly related to the source of excitation in speech production mechanism. The effect of speaking rate is studied using three parameters of excitation source, namely, the instantaneous fundamental frequency, the strength of excitation of epoch, and a loudness measure derived from the Hilbert envelope of the linear prediction residual. Of these, the instantaneous fundamental frequency and the strength of excitation show a systematic variation with speaking rate, as illustrated for the speech data collected from six speakers. The effect of speaking rate on the perception of loudness is also studied.