This folder contains three scripts:
-
Goes through all the files in a folder or all the folders and files in a folder and writes in a txt information about formants, duration, intensity and F0. It extracts either the mid-point or the mean of the mid 50% of the vowel. The maximum formant is set automatically to 5000 or 5500 depending on the F0 of the file, it expects a file that contains a least a sentence.
-
Extracts the whole formant (formant-track) of labelled intervals that contain a vowel. From each time step it extracts the formants value, intensity, F0 and duration of the interval. It can be time normalized (30 points by segment) or with the same time-step for the whole folder. The time-normalized version is is useful to analyze formant tracks with GAMMs or similar.
-
Writes the F1 and F2 in a TextGrid to ease manual correction. You will need to use a script later on that extracts the labels.
A sound and a Textgrid with THE SAME filename and without spaces in the filename. For example this_is_my_sentence.wav and this_is_my_sentence.TextGrid The format of the TextGrid must be: tier1 interval for each sound, the script will analyse only intervals with labels that have one of these symbols: a, e, i ,o ,u, ɪ, ɛ, æ, ɑ, ɔ, ʊ, ʌ, ɝ, AEIOU@
- Open the script (Open/Read from file...), click Run in the upper menu and Run again.
- Set the parameters. a) Folder where the files you want to analyse are b) Name of the txt where the results will be saved c) Data to optimise the formantic analysis d) Data for optimizing the pitch (F0) detection
The output is a tab separated txt file (can be dragged to Excel, beware decimals are ".") with the following information in columns. Depending on the script that you are using you will find an interval per row or a sample point per row. a) file name b) number of the Interval c) label of the interval in the tier of the analysis: vowel or sentence d) F0 e) F1 f) F2 h) F3 i) F4 j) Duration of the vowel k) Intensity of the vowel at its mid point or mean intensity in the interval.
(c) Wendy Elvira García Contact Laboratori de Fonètica (Universitat de Barcelona)