Turkish pairwise variability index

Three scripts are provided for

  • Extracting target words from a single wav file, creating one wav file for each target
  • Hand-correcting vowel boundaries in each target word
  • Estimating vowel duration, intensity, and F0 in each target word and calculating pairwise variability index

    • Input data

    Sample audio data are not provided to protect participants' privacy. The wav file is expected to contain all repetitions of all target words. The accompanying textgrid was generated in Phon, with expected and actual productions with broad transcription.

    Scripts

    Script 1: 1_extractWord.Praat

    This script extracts and saves the pre-defined target words listed in the text files swList.txt and wsList.txt as separate wav files.

    Script 2: 2_correctBoundary.Praat

    This script provides a Praat interface for hand-correcting vowel boundaries in the individual target words.

    Script 3: 3_correlatesOfStress.Praat

    This script extracts F0, intensity, and duration measures form the vowels based on the manually corrected segmentation and saves it together with speaker, word, repetition, and stress (strong-weak or weak-strong) information in a comma-separated file. The scripts expects words to be transcribed using IPA symbols for Turkish, as specified in vowelList.txt

    All scripts, together with list of target words and vowels: Scripts 1-3