An Acoustic-Phonetic Analysis of a Ventriloquist's Speech

I will present research that I have carried out concerning the acoustics of ventriloquism. This research investigates the techniques employed by ventriloquists to alter their acoustic output to auditorily mimic labials that they cannot produce due to articulatory constraints. I look at the articulatory compensations and whether the acoustic output betrays the place of articulation or if it is more similar to the target sound. To do this, I conducted an interview with a ventriloquist producing both natural speech and altered ventriloquist-style speech (i.e. without moving his lips). I then carried out an acoustic analysis of the recordings and compared formant transitions, VOT, and burst intensities in the two speech styles. I made fascinating discoveries not only in the effects of constraints on bilabials, but also on associated non-labial consonants. Prior to carrying out the research, I looked into two competing theories of how the compensation would be achieved. Miller and Nicely (1955)'s confusion matrices would give more support to a prediction that labials would be compensated for by dentals, whereas Jakobson, Halle, and Fant's Grave feature theory would predict compensation by velar or uvular sounds. I also consulted numerous ventriloquism textbooks, dating back as far as 1801, and discovered a shift in mainstream techniques, as well as several surprising suggestions.