# Myovox: Reading Speech from the Muscles of the Face

arXiv: https://arxiv.org/abs/2609.17548 · HTML: https://mlsanity.com/paper/2609.17548/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 17 of its 40 references that are in the map.
- First posted on arXiv 15 Jul 2026.

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

- [Representation Learning with Contrastive Predictive Coding](https://mlsanity.com/paper/1807.03748/) (2018)
- [Qwen2.5 Technical Report](https://mlsanity.com/paper/2412.15115/) (2024)
- [QLoRA: Efficient Finetuning of Quantized LLMs](https://mlsanity.com/paper/2305.14314/) (2023)
- [HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units](https://mlsanity.com/paper/2106.07447/) (2021)
- [Conformer: Convolution-augmented Transformer for Speech Recognition](https://mlsanity.com/paper/2005.08100/) (2020)
- [WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://mlsanity.com/paper/2110.13900/) (2021)
- [Sequence-to-Sequence Speech Recognition with Time-Depth Separable Convolutions](https://mlsanity.com/paper/1904.02619/) (2019)
- [Digital Voicing of Silent Speech](https://mlsanity.com/paper/2010.02960/) (2020)
- [Brain-to-Text Decoding: A Non-invasive Approach via Typing](https://mlsanity.com/paper/2502.17480/) (2025)
- [An Improved Model for Voicing Silent Speech](https://mlsanity.com/paper/2106.01933/) (2021)
- [A Cross-Modal Approach to Silent Speech with LLM-Enhanced Recognition](https://mlsanity.com/paper/2403.05583/) (2024)
- [Brain-to-Text Decoding with Context-Aware Neural Representations and Large Language Models](https://mlsanity.com/paper/2411.10657/) (2024)
- [Can LLMs Understand Unvoiced Speech? Exploring EMG-to-Text Conversion with LLMs](https://mlsanity.com/paper/2506.00304/) (2025)
- [Accurate Decoding of Natural Sentences from Non-Invasive Brain Recordings](https://mlsanity.com/paper/2608.18114/) (2026)
- [emg2speech: Synthesizing speech from electromyography using self-supervised speech models](https://mlsanity.com/paper/2510.23969/) (2025)
- [Geometry of orofacial neuromuscular signals: speech articulation decoding using surface electromyography](https://mlsanity.com/paper/2411.02591/) (2024)
- [Non-invasive electromyographic speech neuroprosthesis: a geometric perspective](https://mlsanity.com/paper/2502.05762/) (2025)

## Abstract (arXiv)

Myovox, from myo (muscle) and vox (voice), decodes open-vocabulary English text from 31-channel surface electromyography (sEMG) recorded from the muscles of the face during vocalized speech. It takes the single-subject emg2speech General Corpus from a published 51.17% word error rate to 18.53%, in three separable moves, each measured in isolation. First, I recover the open-vocabulary decode settings missing from the public release and reach a faithful 40.63% WER / 39.02% PER baseline whose phone error rate matches the published one to within 0.8 points, so the acoustic model is reproduced faithfully. Second, I replace the causal encoder with a bidirectional Conformer trained by a four-term cross-modal distillation against the parallel audio's WavLM-Large layer-9 features, reaching 26.14% WER / 22.34% PER from the electromyography alone. Third, I ensemble two acoustic models, union their multi-scale n-best lists, and rerank with a QLoRA-fine-tuned 7B language model, reaching 18.53% WER, the best result reported on this corpus, though not the best reported for sEMG-to-text on other corpora (Section 2). I then report the negative result that bounds the whole approach: reranking is exhausted at 18.5% because the binding constraint is the electromyographic acoustic phone error rate (~20.9%), not the language model. The correct words are simply absent from the acoustic posteriors, so no reranker can reach the 9.30% n-best oracle. All test numbers are on the 400-sentence held-out test set under the authors' official 8,500 / 760 / 400 sequential split; every hyperparameter is tuned once on validation and applied once to test.