Speech segmentation

Research output: Chapter in Book/Report/Conference proceedingChapter

Author(s)

Department/unit(s)

Publication details

Title of host publicationSpeech Perception and Spoken Word Recognition
DateAccepted/In press - 2015
PublisherTaylor and Francis
EditorsGareth Gaskell, Jelena Mirkovic
Original languageEnglish

Abstract

Speech segmentation refers to the challenge for language users and learners to isolate words from fluent speech. This chapter compares the effectiveness of lexically-driven segmentation with segmental, sub-segmental, and suprasegmental heuristic cues. The relative contribution of the two types of information is examined in the mature system, the developing system, in common adverse conditions, and in special populations. We propose that speech segmentation is intrinsically lexically-driven regardless of language, age, or listening condition. However, non-lexical cues are consulted in a predictable fashion whenever access to lexical information is limited.

Discover related content

Find related publications, people, projects, datasets and more using interactive charts.

View graph of relations