Praat

Praat is a computer program for phonetic analysis. It analyzes, synthesizes, and manipulates speech, and generates high-quality graphics for scientific publications. Its functions encompass spectral analysis, pitch and formant measurement, audio segmentation and annotation, and the creation of auditory perception experiments.

★★★★★

5.0(1 ratings)

• Speech Analysis. Praat examines speech signals in detail. It performs spectral analysis, measures fundamental frequency (pitch), formants, intensity, jitter, and shimmer. It provides quantitative data on the acoustic properties of sound for phonetic and acoustic research.
• Speech Synthesis. The program generates artificial speech sounds. It creates speech from formulas, through articulatory synthesis, or by manipulating parameters like pitch and formants. It is used in speech production research and creating stimuli for experiments.
• Speech Manipulation. It modifies existing voice recordings. It allows altering the duration, pitch contour, and vocal quality of a signal. These operations are applied in studies of perception and digital signal processing of voice.
• Labeling and Segmentation. It facilitates the annotation of audio files. With the TextGrid object, intervals and points are marked on multiple levels or "tiers." This process is fundamental for phonetic analysis, transcription, and forced text-audio alignment.
• Auditory Experimentation. It includes tools for designing and running speech perception experiments. It supports identification and discrimination paradigms. It allows controlled presentation of stimuli and collection of listener responses for psycholinguistic research.
• Graphics Generation. It creates visual representations of acoustic analyses. It generates publication-quality spectrograms, waveform graphs, pitch curves, and formants. Graphics are exported in EPS, PDF, PNG, and other formats for inclusion in documents.
• Statistical Analysis. It offers functionalities for statistical data analysis. It performs statistical tests and creates models. These analyses are applied to measurements extracted from Praat objects for scientific hypothesis validation.
• Scripting Language. It has an integrated scripting language. Users automate complex or repetitive tasks. Scripts allow batch processing and the creation of customized analysis procedures, extending standard capabilities.
• Cross-platform Portability. It works on multiple operating systems. Versions exist for Windows, macOS, Linux, and Raspberry Pi. This portability ensures its use in different research and teaching environments.
• Learning Algorithms. It implements methods for model development. It includes algorithms for data classification and neural network training. They are used in research on automatic speech recognition and computational phonetic modeling.
• Sung Voice Analysis. It contains specific functions for the analysis of sung voice. It measures vibrato and tone quality in musical performances. This capability is used in musicology, musical acoustics, and vocal performance studies.
• Customizable Configuration. Users adapt aspects of the work environment. Visualization preferences, keyboard shortcuts, and audio settings are personalized. The interface supports adjustments for different methodologies and specific research needs.

The Praat program was created by Paul Boersma and David Weenink from the Department of Phonetic Sciences at the University of Amsterdam. Its development began in 1992. The software is primarily written in the C++ programming language. Praat is distributed with public access to its source code, facilitating its review and continuous development by the scientific community.

Alternatives to Praat: