MontrealCorpusTools/Montreal-Forced-Aligner

Command line utility for forced alignment using Kaldi

What it solves

Montreal Forced Aligner (MFA) provides a way to automatically align speech datasets, matching the spoken words in an audio file to their corresponding text transcripts. This process, known as forced alignment, forced alignment of speech datasets using Kaldi.

How it works

MFA is a command-line utility that leverages the Kaldi speech recognition toolkit to perform the alignment of text and speech. It provides a set of tools to synchronize audio recordings with their written transcripts.

Who it’s for

Researchers and developers working with speech datasets, linguistics researchers, and anyone needing to precise timing information for when specific words or sounds are spoken in an audio recording.

Highlights

  • own a command-line interface for processing speech datasets.
  • Built on top of the Kaldi toolkit for high-quality alignment.
  • Supports installation via conda-forge for simplified dependency management.
  • Provides pretrained models for various languages to accelerate the alignment process.

Related

  • Project
    strob/gentle

    A robust forced-aligner built on Kaldi that aligns speech audio with text transcripts to provide precise word-level timing.

  • Project
    MahmoudAshraf97/ctc-forced-aligner

    A Python package for efficient forced alignment of audio and text using Hugging Face CTC models, supporting over 1,100 languages.

  • Project
    readbeyond/aeneas

    A Python/C library and toolset for forced alignment, automatically synchronizing audio narrations with their corresponding text fragments.

  • Project
    pytorch/audio

    A PyTorch-based audio library that provides tools for processing audio data for machine learning, featuring GPU acceleration and specialized audio transforms.

  • Project
    daanzu/kaldi-active-grammar

    Kaldi‑Active‑Grammar is a Python wrapper for the Kaldi speech‑recognition engine that lets you compile many small grammars and activate only the needed ones per utterance. It ships binary wheels for all major OSes, includes a pre‑trained English model, and provides a backend for the Dragonfly and Caster voice‑control frameworks. Install via pip, download the model, define rules, and dynamically enable them for context‑aware command‑and‑control.