Name:Proy, Matthew

School/Affiliation:University at Buffalo, SUNY

Co-Authors:Peter Q. Pfordresher

Virtual or In-person:Virtual

Short Bio:

Matthew Proy is a graduate student at the University at Buffalo researching music cognition and psycholinguistics. His research interests concern how different levels of linguistic information—syntactic, lexical, etc.—affect how we parse prosodic information in speech. Likewise, he is interested in differences regarding the cognitive representations underpinning speech and song and how these bear on vocal motor control abilities.

Abstract:

Current work suggests pitch stability differentiates speech and song such that song tends to have greater pitch stability than speech (Medeiros et al., 2021; Ozaki et al., 2024). It is not clear, however, whether manipulations of pitch stability in auditory sequences lead to transformations from speech to song, or vice versa, when other acoustic variables are held constant. To investigate this, recordings of spontaneously produced speech and song recordings were morphed into song- and speech-like stimuli, respectively. We started with a set of spontaneously produced short sentences and melodies that were matched with respect to text and global pitch contour. We manipulated the pitch stability of each item by creating morphed versions of stimuli that approximated the pitch patten of speech versus song in varying proportions. These manipulations led to a set of stimuli that varied along a speech-to-song continuum in different steps. Participants vocally imitated stimuli and rated how speech- or song-like they perceived them to be. Preliminary analyses showed that participants more accurately matched pitch for song-like items than speech-like items; ratings likewise indicated that stimuli on extreme ends of the continuum were heard as speech or song. Moreover, imitation accuracy and perceptual ratings varied continuously across morphed items—rather than shifting categorically—as a function of pitch stability. These results demonstrate a potentially robust manipulation for gradually shifting spoken, non-sung stimuli into song-like items, and vice versa, through pitch manipulations alone.