The Montreal Forced Aligner is a tool for time aligning orthographic and phonological forms from a pronunciation dictionary to orthographically transcribed audio files. The Montreal Forced Aligner will work on any dataset that is sufficiently large and can be used with any language for which you have a pronunciation dictionary. Single recordings and smaller datasets can be aligned using either pre-trained models included in the command line utility, or models that you have trained yourself from larger data.
The underlying technology is the Kaldi ASR toolkit (http://kaldi-asr.org/). Installing and using the Montreal Forced Aligner executables requires no prerequisites or compilation of third party toolkits (such as HTK). Further information is available on the project’s documentation (http://montreal-forced-aligner.readthedocs.io).
If you run into any issues, please check the mailing list for fixes/workarounds or to post a new issue.
Please visit the release page (https://github.com/MontrealCorpusTools/Montreal-Forced-Aligner/releases) and download the latest release.
McAuliffe, Michael, Michaela Socolof, Elias Stengel-Eskin, Sarah Mihuc, Michael Wagner, and Morgan Sonderegger (2017). Montreal Forced Aligner [Computer program]. Version 1.0.0, retrieved 05 May 2017 from http://montrealcorpustools.github.io/Montreal-Forced-Aligner/.
McAuliffe, Michael, Michaela Socolof, Sarah Mihuc, Michael Wagner, and Morgan Sonderegger (2017). Montreal Forced Aligner: trainable text-speech alignment using Kaldi. In Proceedings of the 18th Conference of the International Speech Communication Association. Paper PDF