---
title: "Contributing to Speech-to-Speech — Ouedraogo 2026"
description: "Mooré is Burkina Faso's most widely used local language, the primary language of about 52.9% of the population."
authors: ["Fayçal S. A. Ouedraogo", "Maimouna Ouattara", "Rodrique Kafando", "Abdoul Kader Kabore", "Aminata Sabane", "Tegawendé F. Bissyandé"]
year: 2026
venue: "Proceedings of the Second Workshop on Language Models for Low-Resource Languages (LoResLM 2026)"
source: "https://aclanthology.org/2026.loreslm-1.54/"
doi: "10.18653/v1/2026.loreslm-1.54"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/ouedraogo-2026-contributing-speech-speech-translation-african/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/ouedraogo-2026-contributing-speech-speech-translation-african/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/3/) / Contributing to Speech-to-Speech Translation for African Low-Resource Languages : Study of French-Mooré Pair

# Contributing to Speech-to-Speech Translation for African Low-Resource Languages : Study of French-Mooré Pair

**Authors:** Fayçal S. A. Ouedraogo, Maimouna Ouattara, Rodrique Kafando, Abdoul Kader Kabore, Aminata Sabane, [Tegawendé F. Bissyandé](https://md.research.elamur.ai/free/en/authors/tegawende-f-bissyande/) · **Year:** 2026 · **Venue:** Proceedings of the Second Workshop on Language Models for Low-Resource Languages (LoResLM 2026) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.loreslm-1.54/) · [PDF](https://aclanthology.org/2026.loreslm-1.54.pdf) · [DOI 10.18653/v1/2026.loreslm-1.54](https://doi.org/10.18653/v1/2026.loreslm-1.54) · [OpenAlex](https://openalex.org/W7140136561)

**Citations:** 0 citations (openalex-snapshot · 2026-10-14) · 1 in our library, of the 140884 papers whose bibliographies we read

## Summary
Mooré is Burkina Faso's most widely used local language, the primary language of about 52.9% of the population. It is mostly spoken rather than written, so text-only translation tools leave many speakers cut off from services that run in French. The authors built Mooré speech recognition and speech synthesis models and combined them into a cascaded French–Mooré speech-to-speech translation system.

For data, they collected Bible text and audio from jw.org (using jw-soup) and bible.com. They cut chapter-long recordings into verse-level clips with fairseq, resampled them to 24 kHz, and added a small single-speaker legacy set from CITADEL. The resulting corpora total about 88,000 utterances and 150 hours.

They fine-tuned Orpheus-3B and XTTS-v2 for synthesis and Wav2Vec-BERT-2.0 for recognition. These were chained with a fine-tuned NLLB-200-distilled-600M translator, Whisper Turbo for French ASR and base XTTS-v2 for French TTS. On non-religious test texts, XTTS-v2 clearly beat Orpheus-3B, and the ASR model reached low error rates.

The results suggest that religious recordings can bootstrap usable speech technology for oral African languages.

## Key points
- **Data:** Mooré audio–text pairs come from jw.org and bible.com plus a small CITADEL dataset, about 88,000 utterances and 150 hours in total. Verse-level alignment was done with fairseq.
- **TTS evaluation:** 36 native speakers rated audio generated from news, sport, tales and proverbs.
  - MOS was 4.36/5 for XTTS-v2 and 3.47/5 for Orpheus-3B, which was trained on only 6.48 h.
  - UTMOS was 3.47 for XTTS-v2 and 2.80 for Orpheus-3B.
  - In A/B tests, listeners preferred XTTS-v2 77.8% of the time.
- **ASR:** Wav2Vec-BERT-2.0 was fine-tuned on dataset 2 with a 90/10 split and lowercased text reduced to Mooré letters and some punctuation. Its word error rate came to 4.24%, with a character error rate of 1.11%.
- **Model adaptation:** XTTS-v2 needed its vocabulary extended with Mooré tokens, Orpheus-3B was adapted directly, and Wav2Vec-BERT-2.0 needed a Mooré-alphabet vocabulary.
- **Design choice and limits:** A cascaded pipeline was used because aligned French–Mooré speech is scarce. Its drawbacks are error propagation between stages and high compute demands. Future work covers direct S2ST and synthetic data generation.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

Fields: [Linguistics & languages](https://md.research.elamur.ai/free/en/shelves/linguistics/) · [Machine translation](https://md.research.elamur.ai/free/en/shelves/machine-translation/) · [Natural language processing](https://md.research.elamur.ai/free/en/shelves/nlp/)

## Cite

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@inproceedings{ouedraogo2026contributing,
  title = {{Contributing to Speech-to-Speech Translation for African Low-Resource Languages : Study of French-Mooré Pair}},
  author = {Ouedraogo, Fayçal S. A. and Ouattara, Maimouna and Kafando, Rodrique and Kabore, Abdoul Kader and Sabane, Aminata and Bissyandé, Tegawendé F.},
  booktitle = {Proceedings of the Second Workshop on Language Models for Low-Resource Languages (LoResLM 2026)},
  year = {2026},
  doi = {10.18653/v1/2026.loreslm-1.54},
  url = {https://doi.org/10.18653/v1/2026.loreslm-1.54}
}
```

- [← previous Contrastively Pre-trained Event Embeddings with Schema-free LLM Annotations](https://md.research.elamur.ai/free/en/papers/mtumbuka-2026-contrastively-pre-trained-event-embeddings/)
- [next → Controllable Sentence Simplification in Italian: Fine-Tuning Large Language Models on Automatically Generated Resources](https://md.research.elamur.ai/free/en/papers/papucci-2026-controllable-sentence-simplification-italian-fine/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/3/)

## References

27 references, 2 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Jesujoba O. Alabi, Xuechen Liu, Dietrich Klakow, and Junichi Yamagishi. 2024. Afrihubert: A self-supervised speech representation model for african languages. [↗](https://doi.org/10.48550/arxiv.2409.20201)
2. Alexei Baevski, Henry Zhou, Abdel-rahman Mohamed, and Michael Auli. 2020. wav2vec 2.0: A framework for self-supervised learning of speech representations. NeurIPS. [find](https://scholar.google.com/scholar?q=Alexei+Baevski%2C+Henry+Zhou%2C+Abdel%2Drahman+Mohamed%2C+and+Michael+Auli%2E+2020%2E+wav2vec+2%2E0%3A+A+framework+for+self%2Dsupervised+learning+of+speech+representations%2E+NeurIPS%2E) [our card](https://md.research.elamur.ai/free/en/papers/baevski-2020-wav2vec-2-0-framework-self/)
3. Mary E. Beckman and Jennifer J. Venditti. 2010. Tone and Intonation, chapter 16. John Wiley Sons, Ltd. [find](https://scholar.google.com/scholar?q=Mary+E%2E+Beckman+and+Jennifer+J%2E+Venditti%2E+2010%2E+Tone+and+Intonation%2C+chapter+16%2E+John+Wiley+Sons%2C+Ltd%2E)
4. Brian DeRenzi, Anna Dixon, Mohamed Aymane Farhi, and Christian Resch. 2025. Synthetic voice data for automatic speech recognition in african languages. [find](https://scholar.google.com/scholar?q=Brian+DeRenzi%2C+Anna+Dixon%2C+Mohamed+Aymane+Farhi%2C+and+Christian+Resch%2E+2025%2E+Synthetic+voice+data+for+automatic+speech+recognition+in+african+languages%2E)
5. Qianqian Dong, Fengpeng Yue, Tom Ko, Mingxuan Wang, Qibing Bai, and Yu Zhang. 2022. Leveraging pseudo-labeled data to improve direct speech-to-speech translation. [find](https://scholar.google.com/scholar?q=Qianqian+Dong%2C+Fengpeng+Yue%2C+Tom+Ko%2C+Mingxuan+Wang%2C+Qibing+Bai%2C+and+Yu+Zhang%2E+2022%2E+Leveraging+pseudo%2Dlabeled+data+to+improve+direct+speech%2Dto%2Dspeech+translation%2E)
6. Qingkai Fang, Shaolei Zhang, Zhengrui Ma, Min Zhang, and Yang Feng. 2024. Can we achieve high-quality direct speech-to-speech translation without parallel speech data? large, high-fidelity, multilingual, and uniquely african speech corpus. In Interspeech. ISCA. [find](https://scholar.google.com/scholar?q=Qingkai+Fang%2C+Shaolei+Zhang%2C+Zhengrui+Ma%2C+Min+Zhang%2C+and+Yang+Feng%2E+2024%2E+Can+we+achieve+high%2Dquality+direct+speech%2Dto%2Dspeech+translation+without+parallel+speech+data%3F+large%2C+high%2Dfidelity%2C+multilingu)
7. Haohan Guo, Fenglong Xie, Xixin Wu, Hui Lu, and Helen Meng. 2022. Towards high-quality neural tts for low-resource languages by learning compact speech representations. Preprint, arXiv:2210.15131. [↗](https://arxiv.org/abs/2210.15131)
8. Mahendra Gupta, Maitreyee Dutta, and Chandresh Kumar Maurya. 2025. Benchmarking hindi-to-english direct speech-to-speech translation with synthetic data. Language Resources and Evaluation, 59(3):2613–2651. [↗](https://doi.org/10.1007/s10579-025-09827-2)
9. Hackster.io. 2025. Canopy labs releases orpheus. Accessed: 2026-01-04. [find](https://scholar.google.com/scholar?q=Hackster%2Eio%2E+2025%2E+Canopy+labs+releases+orpheus%2E+Accessed%3A+2026%2D01%2D04%2E)
10. Rhonda L. Hartell and John T. Bendor-Samuel. 1989. The Niger-Congo languages: a classification and description of Africa’s largest language family. University Press of America. [find](https://scholar.google.com/scholar?q=Rhonda+L%2E+Hartell+and+John+T%2E+Bendor%2DSamuel%2E+1989%2E+The+Niger%2DCongo+languages%3A+a+classification+and+description+of+Africa%E2%80%99s+largest+language+family%2E+University+Press+of+America%2E)
11. Sukairaj Hafiz Imam, Babangida Sani, Dawit Ketema Gete, Bedru Yimam Ahamed, Ibrahim Said Ahmad, Idris Abdulmumin, Seid Muhie Yimam, Muhammad Yahuza Bello, and Shamsuddeen Hassan Muhammad. 2025. Automatic speech recognition for african low-resource languages: Challenges and future directions. [↗](https://doi.org/10.48550/arxiv.2505.11690)
12. Praveen Sai Krishna Paleti, Bhavesh Saluru, Sainadh Vatturi, Basaraboyina Yohoshiva, and Nagendra Panini Challa. 2024. Cascaded speech-to-speech translation system focusing on low-resource indic languages. In 2024 3rd Edition of IEEE Delhi Section Flagship Conference (DELCON), page 1–6. IEEE. [find](https://scholar.google.com/scholar?q=Praveen+Sai+Krishna+Paleti%2C+Bhavesh+Saluru%2C+Sainadh+Vatturi%2C+Basaraboyina+Yohoshiva%2C+and+Nagendra+Panini+Challa%2E+2024%2E+Cascaded+speech%2Dto%2Dspeech+translation+system+focusing+on+low%2Dresource+indic+langu)
13. Ki-Joong Kwon, Jun-Ho So, and Sang-Hoon Lee. 2025. Parameter-efficient fine-tuning for low-resource text-to-speech via cross-lingual continual learning. pages 1613–1617. [find](https://scholar.google.com/scholar?q=Ki%2DJoong+Kwon%2C+Jun%2DHo+So%2C+and+Sang%2DHoon+Lee%2E+2025%2E+Parameter%2Defficient+fine%2Dtuning+for+low%2Dresource+text%2Dto%2Dspeech+via+cross%2Dlingual+continual+learning%2E+pages+1613%E2%80%931617%2E)
14. M. Paul Lewis, Gary F. Simons, and Charles D. Fennig. 2016. Ethnologue: Languages of the world, mooré. SIL International. [find](https://scholar.google.com/scholar?q=M%2E+Paul+Lewis%2C+Gary+F%2E+Simons%2C+and+Charles+D%2E+Fennig%2E+2016%2E+Ethnologue%3A+Languages+of+the+world%2C+moor%C3%A9%2E+SIL+International%2E)
15. Xinjian Li, Ye Jia, and Chung-Cheng Chiu. 2022. Textless direct speech-to-speech translation with discrete speech representation. [find](https://scholar.google.com/scholar?q=Xinjian+Li%2C+Ye+Jia%2C+and+Chung%2DCheng+Chiu%2E+2022%2E+Textless+direct+speech%2Dto%2Dspeech+translation+with+discrete+speech+representation%2E)
16. Hsi-Che Lin, Yi-Cheng Lin, Huang-Cheng Chou, and Hung-yi Lee. 2024. Improving speech emotion recognition in under-resourced languages via speech-to-speech translation with bootstrapping data selection. [find](https://scholar.google.com/scholar?q=Hsi%2DChe+Lin%2C+Yi%2DCheng+Lin%2C+Huang%2DCheng+Chou%2C+and+Hung%2Dyi+Lee%2E+2024%2E+Improving+speech+emotion+recognition+in+under%2Dresourced+languages+via+speech%2Dto%2Dspeech+translation+with+bootstrapping+data+selection)
17. Josh Meyer, David Adelani, Edresson Casanova, Alp Öktem, Daniel Whitenack, Julian Weber, Salomon Kabongo Kabenamualu, Elizabeth Salesky, Iroro Orife, Colin Leong, Perez Ogayo, Chris Chinenye Emezue, Jonathan Mukiibi, Salomey Osei, Apelete Agbolo, Victor Akinode, Bernard Opoku, Olanrewaju Samuel, Jesujoba Alabi, and Shamsuddeen Hassan Muhammad. 2022. Bibletts: a Hamid Mojarad and Kevin Tang. 2025. Automatic speech recognition of african american english: Lexical and contextual effects. [find](https://scholar.google.com/scholar?q=Josh+Meyer%2C+David+Adelani%2C+Edresson+Casanova%2C+Alp+%C3%96ktem%2C+Daniel+Whitenack%2C+Julian+Weber%2C+Salomon+Kabongo+Kabenamualu%2C+Elizabeth+Salesky%2C+Iroro+Orife%2C+Colin+Leong%2C+Perez+Ogayo%2C+Chris+Chinenye+Emezue%2C+J)
18. Alvin Nahabwe, Sulaiman Kagumire, Denis Musinguzi, Bruno Beijuka, Jonah Mubuuke Kyagaba, Peter Nabende, Andrew Katumba, and Joyce Nakatumba-Nabende. 2025. Benchmarking automatic speech recognition models for african languages. [↗](https://doi.org/10.48550/arxiv.2512.10968)
19. Uchenna Nzenwata and Daniel Ogbuigwe. 2024. Automatic speech recognition for the ika language. [↗](https://doi.org/10.48550/arxiv.2410.00940)
20. Peter Ochieng and Dennis Kaburu. 2024. Phonology-guided speech-to-speech translation for african languages. [↗](https://doi.org/10.1016/j.specom.2025.103287)
21. Myle Ott, Sergey Edunov, Alexei Baevski, Angela Fan, Sam Gross, Nathan Ng, David Grangier, and Michael Auli. 2019. fairseq: A fast, extensible toolkit for sequence modeling. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics (Demonstrations), pages 48–53, Minneapolis, Minnesota. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Myle+Ott%2C+Sergey+Edunov%2C+Alexei+Baevski%2C+Angela+Fan%2C+Sam+Gross%2C+Nathan+Ng%2C+David+Grangier%2C+and+Michael+Auli%2E+2019%2E+fairseq%3A+A+fast%2C+extensible+toolkit+for+sequence+modeling%2E+In+Proceedings+of+the+2019) [our card](https://md.research.elamur.ai/free/en/papers/ott-2019-fairseq-fast-extensible-toolkit-sequence/)
22. Maimouna Ouattara, Abdoul Kader Kaboré, Jacques Klein, and Tegawendé F. Bissyandé. 2025. Bridging literacy gaps in African informal business management with low-resource conversational agents. In Proceedings of the First Workshop on Language Models for Low-Resource Languages, pages 193–203, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Maimouna+Ouattara%2C+Abdoul+Kader+Kabor%C3%A9%2C+Jacques+Klein%2C+and+Tegawend%C3%A9+F%2E+Bissyand%C3%A9%2E+2025%2E+Bridging+literacy+gaps+in+African+informal+business+management+with+low%2Dresource+conversational+agents%2E+In+Proc)
23. Andrei Popescu-Belis, Alexis Allemann, Teo Ferrari, and Gopal Krishnamani. 2025. Speech-to-speech translation pipelines for conversations in low-resource languages. [↗](https://doi.org/10.48550/arxiv.2506.01406)
24. B P Pradeep Kumar, S Monishaa, Shreehari Menon, and Syeda Aayesha Aiman H. 2025. Real-time speech-to-speech translator: Analysis and implementation. In 2025 4th International Conference on Distributed Computing and Electrical Circuits and Electronics (ICDCECE), page 1–6. IEEE. [find](https://scholar.google.com/scholar?q=B+P+Pradeep+Kumar%2C+S+Monishaa%2C+Shreehari+Menon%2C+and+Syeda+Aayesha+Aiman+H%2E+2025%2E+Real%2Dtime+speech%2Dto%2Dspeech+translator%3A+Analysis+and+implementation%2E+In+2025+4th+International+Conference+on+Distributed)
25. Alice Tan. 2023. Deep learning in speech recognition for low resource languages. International Journal for Research Publication and Seminar, 14(5):530–537. [↗](https://doi.org/10.36676/jrps.v14.i5.1577)
26. Harish Thangaraj, Ananya Chenat, Jaskaran Singh Walia, and Vukosi Marivate. 2024. Cross-lingual transfer of multilingual models on low resource african languages. [↗](https://doi.org/10.48550/arxiv.2409.10965)
27. Haitong Zhang and Yue Lin. 2020. Unsupervised learning for sequence-to-sequence text-to-speech for low-resource languages. In Interspeech 2020, pages 3161– 3165. [find](https://scholar.google.com/scholar?q=Haitong+Zhang+and+Yue+Lin%2E+2020%2E+Unsupervised+learning+for+sequence%2Dto%2Dsequence+text%2Dto%2Dspeech+for+low%2Dresource+languages%2E+In+Interspeech+2020%2C+pages+3161%E2%80%93+3165%2E)