PraatSauce

This is a suite of Praat scripts that facilitate the analysis of voice quality. They are used to extract measures such as f0, formants, spectral measures such as H1-H2, H1-A3 etc, cepstral peak prominence, harmonics-to-noise-ratios in select frequency bands, and amplitude from acoustic data.

Citation information

If you use PraatSauce in your research, you can cite the current version as follows:

Kirby, James & Rasmus Puggaard-Rode (2025) PraatSauce. Praat-based tools for spectral processing. Version 1.0.5. GitHub: kirbyj/praatsauce.

Getting started

For an overview of the workflow, parameters, and methods used by PraatSauce, see this introductory text.

Background

These scripts are inspired by the VoiceSauce suite for Matlab developed by Shue et al. (2011). PraatSauce extracts many of the same measures as VoiceSauce (although not all!), most of them by calling Praat's internal signal processing routines (Boersma & Weenink 2023). While VoiceSauce is in many ways more sophisticated than PraatSauce, PraatSauce arguably has advantages in terms of speed, stability, active development, and open-source software which allows anyone to edit the code base as they please. (While the VoiceSauce code is also open source, it relies on Matlab libraries that are not, and you cannot run a user-edited version of VoiceSauce without a Matlab license).

This version of PraatSauce is a fully rewritten version of the previous code base by James Kirby which is available as the praatsauce 0.2.7 release. Results should be very similar, but the workflow is quite different. An overview of changes can be found here.

Requirements

These scripts have been tested on Praat version 6.4.01. The scripts make extensive use of vectorized operations that are only available in newer versions of Praat, as well as the state-of-the-art filtered auto-correlation pitch tracking routine which is a relatively recent addition to Praat. So if your version of Praat is several years old, you should update it before running PraatSauce.

Basic workflow

All the PraatSauce files are in the src directory. There are bunch of parameters to toggle, and this is simply done using the params.csv file, also in the src directory, which is a comma-separated file that can be opened and edited in a simple text editor or spreadsheet editor (Excel or similar). You can run the PraatSauce scripts by running the code in the praatsauce.praat file, which will open a window that prompts the location of your parameters file.

If your workflow typically involves a combination of Praat and R, consider exploring the developmental R package sauceshelf which wraps the PraatSauce scripts from R and provides various methods for combining them with other sources of signal processing.

References

Boersma, Paul & David Weenink (2023) Praat. Doing phonetics by computer, version 6.4.01.

Shue, Yen-Liang, Patricia A. Keating, Chad Vicenik & Kristine Yu (2011) VoiceSauce. A program for voice analysis. Proceedings of the 17th International Congress of Phonetic Sciences, 1846--1849. Hong Kong.