freococo/myanmar_bible_audio_46books_jw_version

📖 JW Myanmar Bible Audio-Text Dataset (New World Translation) — 46 Books Edition

This dataset is a meticulously aligned, high-quality parallel corpus of Myanmar-language Bible audio recordings and faithful Burmese transcriptions, drawn from the New World Translation (NWT) published by Jehovah’s Witnesses (JW.org).

Spanning 46 books from both the Old and New Testaments, this release represents the largest open-source Burmese Bible audio-text dataset of its kind — crafted with care, curated with reverence, and presented with purpose.

---

✨ Dataset Highlights

• 🎧 622 chapter-level audio files in Unicode Myanmar (Zawgyi-free)
• 📝 Line-by-line aligned transcriptions with full-text audio for each chapter
• 🔗 Matched via consistent naming: file_name ↔️ audio_text
• 🕊️ Clear, natural narration for accessible Burmese speech modeling
• 📦 Ideal for training ASR, TTS, speech segmentation, or religious dialogue agents
• 🕓 Includes over 45 hours of Myanmar Bible speech and aligned text

“This is more than data — it’s scripture brought to life in the rhythms of our mother tongue.”

---

📌 Sample Row

| file_name               | audio_text             | book_number | book_code | chapter_num | book_name      | chapter_url |
|-------------------------|------------------------|--------------|------------|--------------|----------------|--------------|
| audio/nwt_20_Pr_BU_08.mp3 | ဉာဏ်ပညာက အော်ခေါ်နေတယ်... | 20           | Pr         | 8            | ပညာအလိမ္မာ     | https://...  |

⸻

🔍 Use Cases

•	🧠 Automatic Speech Recognition (ASR) fine-tuning for Burmese
•	🗣 Text-to-Speech (TTS) synthesis for religious voice agents
•	🕊 Speech-text alignment in low-resource language settings
•	📊 Linguistic and phonological research on spoken religious Burmese
•	🙏 Faith-based AI applications in scripture search, recitation, and study

⸻

📛 License

License Type: Custom Non-Commercial License

This dataset is offered solely for non-commercial, educational, and research purposes. It is a derived work based on JW.org content and shared under fair-use conditions.

⸻

🙌 Acknowledgments

This project honors and is made possible by:

    •	The JW.org Myanmar Bible team
    •	Community volunteers, narrators, and preservationists
    •	All who give voice to scripture and life to language

⸻

🙌 Community Inspiration

Bible school teachers and students are warmly encouraged to record their own voices reading these sacred chapters. It’s more than a language exercise — it’s a living expression of faith.

🎙 When students narrate scripture:

- They internalize truth
- Learn better pronunciation
- Hear the Bible in *many voices*, with *many accents*

🙏 Let each voice become a seed. One day, Burmese children may hear the Word of God in accents just like their own.

“May those who record His Word be blessed by the One who speaks it.”

⸻

💬 Notes from the Creator

This dataset was not funded or commissioned. It was born from quiet devotion, late nights, and a deep love for open-source language preservation. May it serve students, scientists, seekers, and saints alike.

— freococo

⸻

“In every syllable of scripture, there is soul. In every dataset like this, there is stewardship.”

⸻

📚 Citation

If you use this dataset in your research or project, please cite:

@misc{jw_myanmar_audio_2025,
  title        = {JW Myanmar Bible Audio-Text Dataset (46-Book Edition)},
  author       = {freococo},
  year         = {2025},
  url          = {https://huggingface.co/datasets/freococo/myanmar_bible_audio_46books_jw_version},
  note         = {Aligned Myanmar Bible audio and text dataset from JW.org, curated for ASR and NLP research}
}