Informal (spontaneous) speech and matched read sentences

Informal (spontaneous) speech was a recorded story about a holiday trip, told face-to-face to an inteviewer. The speech was transliterated and randomized. Sentences were read aloud in a second recording. Speech was collected, labeled and segmented according to [Van Son et al., 2001]. 13 Pairs of matched informal and read speech were selected from a single speaker.

The label files have 9 tiers. From top to bottom these are:

1. PHONEMES: Phoneme labels
2. SYLLPARTS: Syllable onsets, kernels, and offsets
3. LEXSYLL: Lexical syllable transcriptions
4. SYLLABLES: Syllable labels, transcribed as spoken
5. LEXCGN: Word transcriptions from the lexicon (CGN format)
6. CGN: Word transcriptions as spoken (CGN format)
7. WORDS: Orthographic word labels
8. TRANSLIT: Orthographic sentence transliteration as spoken
9. SENTENCE: Orthographic sentence text as presented to the speaker
   (for Informal speech this is the global transliteration used for text generation)

Speech from:

Below the table, there is a link to a compressed archive with all the files (Intas10files_archive.tgz). There is also a link to a Praat script that will display all the files and aligned label files (the Inspect button).
Recordings of full paragraph "context" are available in a compressed audio format ($DHTMLServer/OggVorbis.html">Ogg Vorbis). Click on the entries of the Full paragraph speech column. You might need a $DHTMLServer/OggVorbis.html">browser/audio player plugin to listen to the sounds.

---