Praat is a program language often used in the speech/acoustics research community to extract complex acoustic features for (often) University Research. Although a fairly old language and the software looks somewhat unintuitive and dated, it still masters some of the most complex functions compared to similar programs. The challenge with PRAAT programming language is that it is very different from popular other languages and that proper examples are difficult to find online (it is an old language :-)). Therefore I descided to share some of the scripts I used during the earlier stages of my PhD to transform .Eaf files into .TextGrid files and extract acoustic features from all the .wav files in a directory based on annotations that we extracted. I used it for some work in the fields of Affective Computing and Social Signal Processing. I hope that other researchers can quickly adopt the code to their needs and use my scripts to get a basic understanding of how to do these tasks in PRAAT.
- Extracting all .eaf in a (sub)directory and converting them in .TextGrid files using Pympi in Python.
- Loop through all .TextGrid files and .Wav files and extract all relevant information using Praat.
This forum (https://www.fon.hum.uva.nl/praat/) really helped me understand PRAAT language. It shows far more complex codes for PRAAT sorted by topic. You can also download PRAAT on the forum. I recommend first taking a short peek at my code and then checking out the forum.
If you need more help or have questions, please contact me on michel[dot]jansen[at]live[dot]nl.