The acoustics of a spoken message depends on the speaker,
their dialect, the rate of speech, background noise among other things. Despite this variability, human listeners are able to reliably perceive speech seemingly effortlessly especially compared to contemporary speech recognition systems. How is this possible?
In the Speech, Language and Cognition Lab we investigate such questions related to Language use, in particular, and Cognition, in general.
The Pattern Playback machine used to synthesize speech (courtesy: Haskins Laboratories)