Procedure: Utterance identification labels

Each utterance in each sound file is identified by a label in the corresponding label file, either for test syllable labels (.lab) or additional (unstressed) syllable labels (.uns). These utterance labels are placed in the label file at the position corresponding to the end of the utterance in the speech file.

Utterance labels are formed by separating the words of the whole sentence with underscores (see SENTENCES page for a complete list of sentences) and adding a further underscore and the number 1 or 2 to indicate the whether this is the first or second repetition of that sentence by that speaker, for example:

• RECOMMEND_it_again_for_me_please_1
  
  
• JOHN_threw_the_captaincy_BADGE_again_2

As these examples illustrate, block capitals are used to indicate which words are intended to be accented in the utterance.