Praat

About

Praat is a program for doing phonetics by computer, available for Macintosh, Windows, and Linux (including Raspberry Pi and Chromebook), with license and source code provided. Its features cover speech analysis, including spectrograms, pitch, formant, intensity, jitter, shimmer, voice breaks, cochleagram, and excitation pattern analysis. It also supports speech synthesis from pitch, formant, and intensity, articulatory synthesis, and Klatt acoustic synthesis, along with identification and discrimination listening experiments.

For labelling and segmentation, Praat lets users label intervals and time points on multiple tiers using the phonetic alphabet, and it handles sound files up to 2 gigabytes (3 hours). Speech manipulation tools allow changing pitch and duration contours and filtering. Praat also includes learning algorithms (feedforward neural networks, discrete and stochastic Optimality Theory), statistical tools (multidimensional scaling, principal component analysis, discriminant analysis), and graphics output as PDF, PNG, or EPS files with integrated mathematical and phonetic symbols.

Praat is programmable through a scripting language, can communicate with other programs, and can create hypertext manuals with sound I/O. Its binary files are machine-independent, it reads and writes many sound and other file types, and its menus and preferences (fonts, views, sound devices) are configurable. Help is available through an introductory tutorial, extensive manuals in the Help menus, and a user list.