Dynamic voice source parameters in a female voice

Dynamic variations of the voice source in speech have been studied for a female speaker. The speech material consisted of syllables and sentences with varying stress patterns. The speech signal was inverse filtered and then a parameter model of the voice source was fitted as closely as possible to the inverse filtered glottal pulses. The four parameter glottal source model, the LF-model, was used. This model is described by Fant, Liljencrants, & Lin, 1985, and is shown in Fig.