Prosody of read vs. spontaneous speech

This project compares prosodic structures and the acoustic features that cue prosody in read and spontaneous speech based on two corpora: the Boston University Radio News corpus of read speech and the Switchboard corpus of spontaneous telephone conversational speech.

Corpus annotation

This project employs linguists trained in the ToBI model of prosody transcription in the creation of an annotation for the Switchboard speech corpus that marks prosodic phrasing, accent, and disfluency structures.

Prosodic effects in Automatic Speech Recognition

Linguists in the lab collaborate with engineers in the use of speech recognition methods to investigate the role of prosodic context in conditioning acoustic variation in large databases of continuous, natural speech.

Disfluency and prosody

Prosody research that uses spontaneous speech data, such as the Switchboard corpus, must deal with the effect of disfluency on prosodic structure. This project explores the acoustic features that cue disfluency regions in running speech and the interaction of disfluency and prosodic structure on segmental and supra-segmental acoustic features. Current work on this project includes the analysis of acoustic cues to glottalization that often mark prosodic structure.