Acoustic Analysis of Actresses' Voices Long-term average spectrum analysis of the voices of 18 professional Dutch actresses, comparing speaking with emotion to speaking in a neutral voice
Summary
The actor's formant, a clustering of the 4th and the 5th formant around the area of 3500 Hz, has been widely described when it concerns male actors. Earlier research mentioned that this clustering of formants could be obtained by specific voice training and that there is a high correlation with the actor's formant and 'good voice quality'. Much less research has been conducted with regards to actresses’ voices. Therefore, the current research was set up in order to see what happens when 18 Dutch professional actresses do a dramatic reading of the text compared to when using their neutral voice. Both of their recordings were researched acoustically by making Long-term Average Spectra (LTAS) in the program Praat, and were perceptually evaluated. The evaluators could trace - practically without error - which recording was the acted one. However, this study did not reveal what the evaluators listened for in order to trace which one was the acted recording due to a lot of variation between actresses. The reinforcement in the 3-4 kHz area found in this study was not as strong as with male actors, but, on average, the actresses did exhibit significantly more spectral energy in the 3-4 kHz part of the LTAS when they acted compared to when they used their neutral voice. Moreover, we found something new when examining the 5-10 kHz part of the LTAS. The actresses showed significantly more reinforcement between 5-8 kHz in the acting condition. While the reinforcement in the 3-4 kHz area is probably linked to resonances in the vocal tract, the reinforcement between 5-8 kHz is likely related to vocal effort and thus to a glottal mechanism. Further research should examine the exact causes of these findings and how audible these differences actually are.