CHAINS was a research project funded by Science Foundation Ireland from April 2005 to March 2009. Its goal was to advance the science of speaker identification by investigating those characteristics of a persons speech that make them unique. This goal was pursued in several ways:

  1. By recording and publishing a corpus of speech in which individual speakers willfully modify their normal manner of speaking,
  2. By developing methods for text-independent identification of individuals from their speech, using methods developed and tested on our corpus,
  3. By studying the form of speech movements underlying the sound patterns we observe, as movement patterns are highly individual, and
  4. By examining the organization of speech in time, in both rhythm and associated gesture, to better understand speech as a highly individualized form of movement.

The corpus was recorded in the early phase of the project. It was officially released at the 2006 SPECOM conference in St Petersburg, Russia. The corpus freely available to researchers, both through this website and through the Linguistic Data Consortium. Information about obtaining the corpus can be found by clicking on the 'corpus' link above.