# Where Words Begin and End

arXiv: https://arxiv.org/abs/2610.17782 · HTML: https://mlsanity.com/paper/2610.17782/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 17 of its 38 references that are in the map.
- First posted on arXiv 13 Oct 2026.

## Builds on (most cited first, 17 of 17)

- [Robust Speech Recognition via Large-Scale Weak Supervision](https://mlsanity.com/paper/2212.04356/) (2022)
- [Scaling Speech Technology to 1,000+ Languages](https://mlsanity.com/paper/2305.13516/) (2023)
- [WhisperX: Time-Accurate Speech Transcription of Long-Form Audio](https://mlsanity.com/paper/2303.00747/) (2023)
- [CTC-Segmentation of Large Corpora for German End-to-end Speech Recognition](https://mlsanity.com/paper/2007.09127/) (2020)
- [Why does CTC result in peaky behavior?](https://mlsanity.com/paper/2105.14849/) (2021)
- [TorchAudio 2.1: Advancing speech recognition, self-supervised learning, and audio processing components for PyTorch](https://mlsanity.com/paper/2310.17864/) (2023)
- [Phone-to-audio alignment without text: A Semi-supervised Approach](https://mlsanity.com/paper/2110.03876/) (2021)
- [Tradition or Innovation: A Comparison of Modern ASR Methods for Forced Alignment](https://mlsanity.com/paper/2406.19363/) (2024)
- [Achieving Timestamp Prediction While Recognizing with Non-Autoregressive End-to-End ASR Model](https://mlsanity.com/paper/2301.12343/) (2023)
- [NeuFA: Neural Network Based End-to-End Forced Alignment with Bidirectional Attention Mechanism](https://mlsanity.com/paper/2203.16838/) (2022)
- [LLM-ForcedAligner: A Non-Autoregressive and Accurate LLM-Based Forced Aligner for Multilingual and Long-Form Speech](https://mlsanity.com/paper/2601.18220/) (2026)
- [Whisper Has an Internal Word Aligner](https://mlsanity.com/paper/2509.09987/) (2025)
- [The Mason-Alberta Phonetic Segmenter: A forced alignment system based on deep neural networks and interpolation](https://mlsanity.com/paper/2310.15425/) (2023)
- [Transcription Policy as a Latent Variable: Activating Controllable Verbatim ASR with Word-Level Timing](https://mlsanity.com/paper/2607.18934/) (2026)
- [Gradient-Based Speech-to-Text Alignment for Any ASR Model: From CTC to Speech LLMs](https://mlsanity.com/paper/2607.06831/) (2026)
- [Phoneme- and Word-Level Metrics Using Self-Supervised Speech Representations for Forced Alignment Evaluation](https://mlsanity.com/paper/2608.28508/) (2026)
- [Word Timestamps and Speaker Attribution with a Non-Autoregressive LLM](https://mlsanity.com/paper/2609.15218/) (2026)

## Abstract (arXiv)

Forced aligners locate each word of a known transcript in the audio, and many speech pipelines depend on them. In our own data, a widely used CTC aligner inserts a pause at nearly every word boundary, and we found almost no aligner evaluated on telephone audio. We compare nine aligner configurations against human word boundaries on read, conversational and telephone English. The Montreal Forced Aligner (MFA) is the most precise, while CTC aligners consistently start words late, because most frames inside a word go to the blank token. We then fine-tune Whisper with LoRA to predict each word's start and end as timestamp tokens, training only on boundaries placed by people: 47 hours in 87 languages. The resulting aligner is more precise than MFA on every English test condition, and than MFA's own Korean and Dutch models. With FA-Bench's protocol and code on TIMIT, its word-boundary error is 11.6 ms, against 18.2 ms for the best published system. Two choices matter most on telephone conversations: placing the timestamp slots after the word they time, and training on noisy transcripts, which also makes the aligner nearly insensitive to errors in the transcript. The released model reaches 16.0 ms on real landline calls, against 36.5 ms for MFA. The models are available at https://huggingface.co/collections/Scicom-intl/force-alignment