This Praat script allows to extract phonetic features of all occurrences of fricative consonants [f v s ʃ z ʒ] from sound recordings. It requires audio files (typically .wav) associated with their corresponding .TextGrid files, where word and phoneme tiers are well labeled and aligned, without any pause/silence inside each non-empty word interval.
The output will be a .tsv file made up of 11 columns listing the following information or features:
- The sound filename in which the fricative occurs
- The fricative represented by its SAMPA symbol
- The middle time point (in s)
- The duration (in ms)
- The mean intensity (in dB)
- The centre of gravity (CoG, in Hz)
- Number of phonemes between this and next fricatives
- Duration between this and next fricatives (in s)
- The word in which the fricative occurs
- Length of the word (in
$N$ of phonemes) - Position of the fricative in the word (
$i^{th}$ phoneme)
PS: This is my final assignment for the course LYSL042 - Informatique et phonétique (2024-2025) © Nicolas Audibert