Evidence of the significance of secondary excitations of the vocal tract for vocal intensity

In the production of voiced speech the main excitation of the vocal tract o ccurs during the glottal closing phase when the rate of change of the flow reaches its absolute maximum. The level of this maximum, the negative peak amplitude of the differentiated glottal flow, correlates strongly with soun d pressure level. This study presents a straightforward method that yields a numerical value to characterize the effect of the main excitation on voca l intensity. The method, energy ratio by modified excitation (ERME), utiliz es the glottal flow and the vocal tract transfer function estimated by inve rse filtering and it synthesizes two signals based on the linear source-fil ter theory. The first sound is produced using the glottal flow given by inv erse filtering per se. The second signal is synthesized by removing the mai n excitation from the differentiated flow. ERME is defined as the ratio bet ween the energy of the first synthesized signal and the energy of the secon d one. It is shown that when the loudness of speech increases, the value of ERME first rises but in the case of very loud voices it starts to decrease . Hence, ERME shows that effects of secondary excitations of the vocal trac t that occur during glottal opening become important in the production of v ery loud voices. Copyright (C) 2001 S. Karger AG. Basel.