Factors in perceptual processing of interrupted speech and text

Speech perception is notoriously robust to signal distortions and can succeed despite substantial reduction in acoustic cues. Systematic removal of portions of acoustic information from the speech signal by replacing it with silence or masking noise has been frequently used to investigate contributions of low-level peripheral and high-level central factors to linguistic closure, observed as accurate perception of fragmented speech. A similar approach has been also used to investigate perception of interrupted text. Findings from experiments with interrupted speech and text provide a comparison of domain-general factors in linguistic processing to those related to signal quality and modality-specific peripheral encoding. Moreover, when interrupted speech and text are presented together, both facilitory and inhibitory effects on speech perception accuracy can be observed, reflecting complex cross-modal interactions. These findings will be discussed in the context of prior research on linguistic closure in listeners with normal and impaired hearing. Overall, perception of parametrically interrupted speech and text has served as a useful model for predicting degraded speech perception and elucidating its underlying mechanisms.