An open-source software platform for the annotation and analysis of speech corpora.
Phonometrica provides an integrated environment for managing, annotating, querying, and statistically analyzing spoken language data. It is designed for phoneticians, phonologists, and anyone working with time-aligned speech corpora.
- Project management: organize files into projects with extensible metadata (typed properties).
- Sound visualization: waveforms, spectrograms, pitch tracks, formant tracks, intensity curves, and spectral slices.
- Annotation: create and edit multi-layer annotations based on annotation graphs; seamless import/export of Praat TextGrids.
- Queries: search for text patterns across annotation layers using simple or complex multi-constraint queries (alignment, dominance, precedence, and more); extract formant, pitch, and intensity measurements.
- Concordance and dataset views: browse, filter, recode, transform, and merge query results; toggle wide/long formats; perform set operations.
- Statistical modeling (preview): fit linear, logistic, Poisson, and negative binomial regression models with fixed and random effects (LMM/GLMM); post-hoc testing (estimated marginal means); model comparison with LRT and information criteria; diagnostic plots and exploratory visualizations. Validated against reference implementations in R, but offered as a preview: check results you intend to publish against R, and please report any discrepancy.
- Scripting: extend Phonometrica with a built-in scripting language and JSON-based plugins.
- Cross-platform: runs on Windows, macOS, and Linux.
Pre-built binaries are available from the releases page and from the project website.
Phonometrica is written in C++ (C++20) and uses Qt 6 for its graphical interface. You need:
- A C++20-compliant compiler (GCC 10+, Clang 13+, or MSVC 2019+)
- CMake 3.21 or later
- Qt 6.2 or later (Widgets, Svg, PrintSupport modules)
- QScintilla for Qt 6
- Boost (for Boost.Math)
- libsndfile
All other dependencies (Eigen, PCRE2, pocketfft, CppAD, etc.) are bundled in the source tree. See BUILD.md for detailed instructions.
git clone https://github.com/jeychenne/phonometrica.git
cd phonometrica
mkdir build && cd build
cmake ..
cmake --build .
The documentation is available online at phonometrica-ling.org and is also accessible from within the application via the Help buttons.
If you use Phonometrica in your research, please cite:
Eychenne, Julien & Léa Courdès-Murphy (2025). Annotation et analyse de données sociophonologiques sur grands corpus : présentation de la plateforme Phonometrica. In Wim Remysen & Hélène Blondeau (eds.) (Re)donner la parole aux corpus montréalais : Regards rétrospectifs et prospectifs. Montreal: Presses Universitaires de Montréal, pp. 255–270.
Phonometrica is free software distributed under the terms of the GNU General Public License (version 3).
For questions, bug reports, or feature requests: jeychenne@gmail.com