Skip to Main Content
An instructional device capable of displaying a continuous speech passage with its corresponding printed text has been developed for use at The National Technical Institute for the Deaf. A series of computer routines are employed to segment prerecorded speech into syllable-like segments, and to separate the corresponding text into syllables. A speech/text synchronization routine then combines the two inputs on a syllable-by-syllable basis in such a way as to maximize the likelihood of their correct synchronization. This paper describes the synchronization routine's performance and discusses its ability to detect and correct a variety of speech segmentation errors. A demonstration of the speech/text synchronization system's output will also be provided.