ISCA Archive SpeechProsody 2004
ISCA Archive SpeechProsody 2004

Analysis of autocorrelation-based parameters for creaky voice detection

Carlos Toshinori Ishi

Creaky voice carries important linguistic and paralinguistic information. Parameters based on autocorrelation of the glottal excitation waveform are proposed for automatic detection of creaky voice in spontaneous speech. Analysis results show the ratio of the first two peaks of the autocorrelation function as a primary parameter to detect creaky voice.