Typhoon Isan Corpus

This repository contains a collection of open linguistic resources for the Isan language (Northeastern Thai) released by the Typhoon team at SCB 10X. These datasets, standards, and research materials support dialect AI, ASR, speech technology, and linguistic research in Thailand.

📦 Contents

This repository includes:

• Isan Spelling Standard – Thai-script orthographic system for Isan (อักขรวิธีการสะกดคำภาษาอีสานด้วยอักษรไทย)

• Isan Speech Transcription Convention – Standardized transcription guidelines (แนวทางการถอดเสียงภาษาอีสานเป็นข้อความสำหรับการสร้างชุดข้อมูลเพื่องานด้านเทคโนโลยีทางเสียง)

• Isan Phonetic Dictionary – Word-to-phonetic mapping (พจนานุกรมคำอ่านที่เชื่อมโยงคำกับการออกเสียงในภาษาอีสาน)

• Technical Report – Protocol documentation and research methodology (รายงานอธิบายกระบวนการ)
  📄 Read the paper on arXiv

Each resource is designed to work together as a complete foundation for Isan speech and language technology.

Additionally, you can find the Isan Speech Corpus (audio + transcriptions + metadata) on our Hugging Face.

📄 Technical Report

Our methodology and findings are detailed in:

"Developing an Open Conversational Speech Corpus for the Isan Language"
Adisai Na-Thalang, Chanakan Wittayasakpan, Kritsadha Phatcharoen, Supakit Buakaw
arXiv:2511.21229 [cs.CL]
https://arxiv.org/abs/2511.21229

This paper introduces the first open conversational speech dataset for Isan, capturing natural speech phenomena including colloquialisms, spontaneous prosody, disfluencies, and code-switching with Central Thai. It addresses the challenges of establishing practical transcription protocols in the absence of standardized Isan orthography.

📜 License

Apache 2.0

📖 Citation

If you use these resources in your research, please cite:

@misc{nathalang2025developingopenisanspeechcorpus,
      title={Developing an Open Conversational Speech Corpus for the Isan Language}, 
      author={Adisai Na-Thalang and Chanakan Wittayasakpan and Kritsadha Phatcharoen and Supakit Buakaw},
      year={2025},
      eprint={2511.21229},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2511.21229}, 
}

🙏 Acknowledgements

Thanks to:

• Local speakers, teachers, and communities across Northeastern Thailand
• Linguists, engineers, and researchers contributing to Typhoon Isan
• Partner institutions supporting open-source dialect AI