---
title: "Same-Language Subtitles for Low-resource — Pradhan 2026"
description: "Regional-language YouTube videos in India rarely have same-language subtitles."
authors: ["Anirudh Pradhan", "Ayushi Pandey", "Divyansh Kushwaha", "Akshita Tiwary", "Vivek Seshadri"]
year: 2026
venue: "Proceedings of the Fifteenth Language Resources and Evaluation Conference"
source: "https://aclanthology.org/2026.lrec-1.246/"
doi: "10.63317/4938abev4keh"
licence: "CC BY-NC 4.0 (ELRA, LREC proceedings)"
licence_url: "https://creativecommons.org/licenses/by-nc/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/pradhan-2026-same-language-subtitles-low-resource/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/pradhan-2026-same-language-subtitles-low-resource/)

[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/8/) / Same-Language Subtitles for Low-resource Languages: A Case of Bundelkhandi

# Same-Language Subtitles for Low-resource Languages: A Case of Bundelkhandi

**Authors:** Anirudh Pradhan, Ayushi Pandey, Divyansh Kushwaha, Akshita Tiwary, Vivek Seshadri · **Year:** 2026 · **Venue:** Proceedings of the Fifteenth Language Resources and Evaluation Conference · **Licence:** [CC BY-NC 4.0 (ELRA, LREC proceedings)](https://creativecommons.org/licenses/by-nc/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.lrec-1.246/) · [PDF](https://aclanthology.org/2026.lrec-1.246.pdf) · [DOI 10.63317/4938abev4keh](https://doi.org/10.63317/4938abev4keh) · [OpenAlex](https://openalex.org/W7160053034)

**Citations:** 0 citations (openalex-snapshot · 2026-10-14)

## Summary
Regional-language YouTube videos in India rarely have same-language subtitles. This hurts accessibility for hearing-impaired viewers, limits engagement for independent creators, and leaves researchers without transcribed material. Because automatic captions still trail human ones in accuracy, the authors propose Project Saurakhi, a subtitling tool built into the Karya smartphone app. It pairs machine transcription with correction by native-speaker annotators. The tool is designed for rural conditions: it is mobile-first, works offline and syncs later, and has a localized interface. The pilot language is Bundelkhandi, an Indo-Aryan language of central India with over 5.6 million speakers but almost no digital resources. The authors collected 207 YouTube videos totalling 63 hours. A pipeline of download, speaker diarization and wav2vec-based ASR produces draft transcripts, which annotators then refine by tagging speakers and correcting the text. The paper describes the platform and the corpus but reports no evaluation results. The planned annotation phase will employ paid local women and produce a reusable resource for low-resource languages.

## Key points
- The Saurakhi Bundelkhandi Corpus contains 207 videos, 63:01 hours of audio (about 51 hours of actual speech) and an average of roughly 18 speakers per video. The paper gives the number of creators inconsistently: 19 in the abstract and 20 in the body and Table 1. The two largest channels supply 22.8% and 20.7% of the total duration.
- The processing pipeline has three steps:
  - videos are downloaded with yt-dlp;
  - audio is extracted with FFmpeg and converted to 16-bit, 16 kHz mono WAV;
  - automatic speaker diarization is applied, followed by AI4Bharat wav2vec-based ASR.
- The annotation workflow works on pre-cut segments. Annotators tag the speaker, correct the transcript, and then submit or skip the segment. Expert reviewers re-check a subset of submissions to measure Inter-Annotator Agreement (IAA).
- The roadmap plans 100 paid women annotators in the Damoh, Sagar, Jabalpur and Garhakota districts of Madhya Pradesh Bundelkhand, each guaranteed at least one hour of work.
- The data is currently used only internally. The corpus and subtitles will be released or uploaded to YouTube only after creators give written consent.

*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/)

## Cite

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

```bibtex
@inproceedings{pradhan2026same,
  title = {{Same-Language Subtitles for Low-resource Languages: A Case of Bundelkhandi}},
  author = {Pradhan, Anirudh and Pandey, Ayushi and Kushwaha, Divyansh and Tiwary, Akshita and Seshadri, Vivek},
  booktitle = {Proceedings of the Fifteenth Language Resources and Evaluation Conference},
  year = {2026},
  doi = {10.63317/4938abev4keh},
  url = {https://doi.org/10.63317/4938abev4keh}
}
```

- [← previous SALT-31: A Machine Translation Benchmark Dataset for 31 Ugandan Languages](https://md.research.elamur.ai/free/en/papers/nsumba-2026-salt-31-machine-translation-benchmark/)
- [next → Sanskrit Travelogue: A Large-Scale Unified and Annotated Corpus of Sanskrit Texts](https://md.research.elamur.ai/free/en/papers/luca-2026-sanskrit-travelogue-large-scale-unified/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/8/)

## References

28 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. 2023. Revitalising languages and safeguarding cultural diversity (rise up) project fact sheet. CORDIS, European Commission. Grant agreement ID: 101095048. Accessed: 15 October 2025. [find](https://scholar.google.com/scholar?q=2023%2E+Revitalising+languages+and+safeguarding+cultural+diversity+%28rise+up%29+project+fact+sheet%2E+CORDIS%2C+European+Commission%2E+Grant+agreement+ID%3A+101095048%2E+Accessed%3A+15+October+2025%2E)
2. Niyati Bafna, Josef van Genabith, Cristina España-Bonet, and Zdeněk Žabokrtskỳ. 2022. Combining noisy semantic signals with orthographic cues: Cognate induction for the indic dialect continuum. In Proceedings of the 26th Conference on Computational Natural Language Learning (CoNLL), pages 110–131. [find](https://scholar.google.com/scholar?q=Niyati+Bafna%2C+Josef+van+Genabith%2C+Cristina+Espa%C3%B1a%2DBonet%2C+and+Zden%C4%9Bk+%C5%BDabokrtsky%CC%80%2E+2022%2E+Combining+noisy+semantic+signals+with+orthographic+cues%3A+Cognate+induction+for+the+indic+dialect+continuum%2E+In+Pr)
3. Billion Readers (BIRD). 2025. A roadmap for accessible entertainment in india: Recommendations for access, literacy and language enhancement at population scale. Technical report, Billion Readers. [find](https://scholar.google.com/scholar?q=Billion+Readers+%28BIRD%29%2E+2025%2E+A+roadmap+for+accessible+entertainment+in+india%3A+Recommendations+for+access%2C+literacy+and+language+enhancement+at+population+scale%2E+Technical+report%2C+Billion+Readers%2E)
4. Edmund Burke Boatner. 1981. Captioned films for the deaf. American annals of the deaf, 156(5):520–525. [find](https://scholar.google.com/scholar?q=Edmund+Burke+Boatner%2E+1981%2E+Captioned+films+for+the+deaf%2E+American+annals+of+the+deaf%2C+156%285%29%3A520%E2%80%93525%2E)
5. Steven Coats. 2023. Dialect corpora from youtube. Language and linguistics in a complex world, pages 79–102. [find](https://scholar.google.com/scholar?q=Steven+Coats%2E+2023%2E+Dialect+corpora+from+youtube%2E+Language+and+linguistics+in+a+complex+world%2C+pages+79%E2%80%93102%2E)
6. Christopher R Cooper. 2023. The identification of youtube videos that feature the linguistic features of english informal speech. Applied Corpus Linguistics, 3(3):100068. [find](https://scholar.google.com/scholar?q=Christopher+R+Cooper%2E+2023%2E+The+identification+of+youtube+videos+that+feature+the+linguistic+features+of+english+informal+speech%2E+Applied+Corpus+Linguistics%2C+3%283%29%3A100068%2E)
7. Ministry of Information & Broadcasting Government of India. 2019. Office memorandum: Accessibility standards for persons with disabilities in television programmes. Technical Report No. 9/10/2012-BP&L (Vol-II), Ministry of Information & Broadcasting, Government of India. Issued under Sections 29(h)… [find](https://scholar.google.com/scholar?q=Ministry+of+Information+%26+Broadcasting+Government+of+India%2E+2019%2E+Office+memorandum%3A+Accessibility+standards+for+persons+with+disabilities+in+television+programmes%2E+Technical+Report+No%2E+9%2F10%2F2012%2DBP%26L)
8. Ali Gürkan. 2021. Comparison of subtitling for the deaf and hard-of-hearing guidelines implemented accross countries. Karamanoğlu Mehmetbey Üniversitesi Uluslararası Filoloji ve Çeviribilim Dergisi, 3(2):249–278. [find](https://scholar.google.com/scholar?q=Ali+G%C3%BCrkan%2E+2021%2E+Comparison+of+subtitling+for+the+deaf+and+hard%2Dof%2Dhearing+guidelines+implemented+accross+countries%2E+Karamano%C4%9Flu+Mehmetbey+%C3%9Cniversitesi+Uluslararas%C4%B1+Filoloji+ve+%C3%87eviribilim+Dergisi%2C+3)
9. Ideal Insight. 2023. Netflix and captions?: A new study shows that 85% of netflix customers in the uk use subtitles while watching. Accessed: 21 November 2023. [find](https://scholar.google.com/scholar?q=Ideal+Insight%2E+2023%2E+Netflix+and+captions%3F%3A+A+new+study+shows+that+85%25+of+netflix+customers+in+the+uk+use+subtitles+while+watching%2E+Accessed%3A+21+November+2023%2E)
10. Paweł Kamocki, Vladimir Popescu, Isabelle Gavanon, Camille Gaffiot, Khalid Choukri, and Valérie Mapelli. 2021. Elrc report on legal issues in web crawling (version 1.1). Revision date: 2021-02-09; originally published 2018; Accessed: 15 October 2025. [find](https://scholar.google.com/scholar?q=Pawe%C5%82+Kamocki%2C+Vladimir+Popescu%2C+Isabelle+Gavanon%2C+Camille+Gaffiot%2C+Khalid+Choukri%2C+and+Val%C3%A9rie+Mapelli%2E+2021%2E+Elrc+report+on+legal+issues+in+web+crawling+%28version+1%2E1%29%2E+Revision+date%3A+2021%2D02%2D09%3B+ori)
11. Hyunju Kim, Yan Tao, Chuanrui Liu, Yuzhuo Zhang, and Yuxin Li. 2023. Comparing the impact of professional and automatic closed captions on video-watching experience. In Extended Abstracts of the 2023 CHI Conference on Human Factors in Computing Systems, pages 1–6. [find](https://scholar.google.com/scholar?q=Hyunju+Kim%2C+Yan+Tao%2C+Chuanrui+Liu%2C+Yuzhuo+Zhang%2C+and+Yuxin+Li%2E+2023%2E+Comparing+the+impact+of+professional+and+automatic+closed+captions+on+video%2Dwatching+experience%2E+In+Extended+Abstracts+of+the+2023+)
12. Ritesh Kumar, Meiraba Takhellambam, Bornini Lahiri, Amalesh Gope, Shyam Ratan, Neerav Mathur, and Siddharth Singh. 2023. Collecting speech data for endangered and under-resourced indian languages. In Proc. 2nd Annual Meeting of the ELRA/ISCA SIG on Under-resourced Languages (SIGUL 2023), pages 14– 1… [find](https://scholar.google.com/scholar?q=Ritesh+Kumar%2C+Meiraba+Takhellambam%2C+Bornini+Lahiri%2C+Amalesh+Gope%2C+Shyam+Ratan%2C+Neerav+Mathur%2C+and+Siddharth+Singh%2E+2023%2E+Collecting+speech+data+for+endangered+and+under%2Dresourced+indian+languages%2E+In+)
13. Shantanu Kumar and Narayan Choudhary. 2025. Maithili language technology: A survey. Strength for Today and Bright Hope for Tomorrow Volume 25: 6 June 2025 ISSN 1930-2940, page 160. [find](https://scholar.google.com/scholar?q=Shantanu+Kumar+and+Narayan+Choudhary%2E+2025%2E+Maithili+language+technology%3A+A+survey%2E+Strength+for+Today+and+Bright+Hope+for+Tomorrow+Volume+25%3A+6+June+2025+ISSN+1930%2D2940%2C+page+160%2E)
14. Margaret Root Kustritz, Ryan Rupprecht, and Perle Zhitnitskiy. 2023. Comparison of accuracy of machine-generated or human-generated captions of zoom live lectures in a comparative theriogenology course. Clinical Theriogenology, 15:52–56. [find](https://scholar.google.com/scholar?q=Margaret+Root+Kustritz%2C+Ryan+Rupprecht%2C+and+Perle+Zhitnitskiy%2E+2023%2E+Comparison+of+accuracy+of+machine%2Dgenerated+or+human%2Dgenerated+captions+of+zoom+live+lectures+in+a+comparative+theriogenology+cours)
15. Institut Ramon Llull. 2025. Grants for the subtitling of audiovisual works produced and/or co-produced in catalonia (2025 call). Accessed: 15 October 2025. [find](https://scholar.google.com/scholar?q=Institut+Ramon+Llull%2E+2025%2E+Grants+for+the+subtitling+of+audiovisual+works+produced+and%2For+co%2Dproduced+in+catalonia+%282025+call%29%2E+Accessed%3A+15+October+2025%2E)
16. Nathalie Mälzer and Maria Wünsche. 2024. Subtitling for people who are d/deaf or hard-of-hearing (sdh). In Handbook of Accessible Communication, pages 313–328. Frank & Timme, Berlin. [find](https://scholar.google.com/scholar?q=Nathalie+M%C3%A4lzer+and+Maria+W%C3%BCnsche%2E+2024%2E+Subtitling+for+people+who+are+d%2Fdeaf+or+hard%2Dof%2Dhearing+%28sdh%29%2E+In+Handbook+of+Accessible+Communication%2C+pages+313%E2%80%93328%2E+Frank+%26+Timme%2C+Berlin%2E)
17. Helard Becerra Martinez, Alessandro Ragano, Diptasree Debnath, Asad Ullah, Crisron Rudolf Lucas, Martin Walsh, and Andrew Hines. 2024. Dialogue understandability: Why are we streaming movies with subtitles? ArXiv, abs/2403.15336. — [arXiv:2403.15336](https://arxiv.org/abs/2403.15336) [find](https://scholar.google.com/scholar?q=Helard+Becerra+Martinez%2C+Alessandro+Ragano%2C+Diptasree+Debnath%2C+Asad+Ullah%2C+Crisron+Rudolf+Lucas%2C+Martin+Walsh%2C+and+Andrew+Hines%2E+2024%2E+Dialogue+understandability%3A+Why+are+we+streaming+movies+with+subt)
18. Smith Mehta. 2020. Localization, diversification and heterogeneity: Understanding the linguistic and cultural logics of indian new media. International Journal of Cultural Studies, 23(1):102–120. [find](https://scholar.google.com/scholar?q=Smith+Mehta%2E+2020%2E+Localization%2C+diversification+and+heterogeneity%3A+Understanding+the+linguistic+and+cultural+logics+of+indian+new+media%2E+International+Journal+of+Cultural+Studies%2C+23%281%29%3A102%E2%80%93120%2E)
19. Sriram Mohan and Aswin Punathambekar. 2019. Localizing youtube: Language, cultural regions, and digital platforms. International journal of cultural studies, 22(3):317–333. [find](https://scholar.google.com/scholar?q=Sriram+Mohan+and+Aswin+Punathambekar%2E+2019%2E+Localizing+youtube%3A+Language%2C+cultural+regions%2C+and+digital+platforms%2E+International+journal+of+cultural+studies%2C+22%283%29%3A317%E2%80%93333%2E)
20. Ayushi Pandey and Indranil Dutta. 2014. Bundeli folk-song genre classification with knn and svm. In Proceedings of the 11th International Conference on Natural Language Processing (ICON 2014), pages 133–138, Goa, India. NLP Association of India. Accessed: 11 October 2025. [find](https://scholar.google.com/scholar?q=Ayushi+Pandey+and+Indranil+Dutta%2E+2014%2E+Bundeli+folk%2Dsong+genre+classification+with+knn+and+svm%2E+In+Proceedings+of+the+11th+International+Conference+on+Natural+Language+Processing+%28ICON+2014%29%2C+pages+1)
21. Pablo Romero-Fresco and Nazaret Fresno. 2023. The accuracy of automatic and human live captions in english. Linguistics and the Human Sciences (LANS–TTS), 22(1). [find](https://scholar.google.com/scholar?q=Pablo+Romero%2DFresco+and+Nazaret+Fresno%2E+2023%2E+The+accuracy+of+automatic+and+human+live+captions+in+english%2E+Linguistics+and+the+Human+Sciences+%28LANS%E2%80%93TTS%29%2C+22%281%29%2E)
22. P Roy. 2019. Why india has world’s cheapest mobile data. BBC News. [find](https://scholar.google.com/scholar?q=P+Roy%2E+2019%2E+Why+india+has+world%E2%80%99s+cheapest+mobile+data%2E+BBC+News%2E)
23. Hui Yang, Weiju Shen, Xinyan Ma, Bo Liu, and Yi Wang. 2025. Research on availability and reception of intralingual subtitles by the deaf and hard of hearing in china. SAGE Open, 15(3):21582440251365754. [find](https://scholar.google.com/scholar?q=Hui+Yang%2C+Weiju+Shen%2C+Xinyan+Ma%2C+Bo+Liu%2C+and+Yi+Wang%2E+2025%2E+Research+on+availability+and+reception+of+intralingual+subtitles+by+the+deaf+and+hard+of+hearing+in+china%2E+SAGE+Open%2C+15%283%29%3A2158244025136575)
24. Soledad Zárate. 2021. Captioning and subtitling for d/deaf and hard of hearing audiences. UCL Press. [find](https://scholar.google.com/scholar?q=Soledad+Z%C3%A1rate%2E+2021%2E+Captioning+and+subtitling+for+d%2Fdeaf+and+hard+of+hearing+audiences%2E+UCL+Press%2E)
25. Nagrani, Arsha and Chung, Joon Son and Zisserman, Andrew. 2017. Voxceleb: a large-scale speaker identification dataset. arXiv. [find](https://scholar.google.com/scholar?q=Nagrani%2C+Arsha+and+Chung%2C+Joon+Son+and+Zisserman%2C+Andrew%2E+2017%2E+Voxceleb%3A+a+large%2Dscale+speaker+identification+dataset%2E+arXiv%2E)
26. Siegert, Ingo and Marquenie, Jan and Grawunder, Sven. 2025. Queer Waves: A German Speech Dataset Capturing Gender and Sexual Diversity from Podcasts and YouTube. ISCA. [find](https://scholar.google.com/scholar?q=Siegert%2C+Ingo+and+Marquenie%2C+Jan+and+Grawunder%2C+Sven%2E+2025%2E+Queer+Waves%3A+A+German+Speech+Dataset+Capturing+Gender+and+Sexual+Diversity+from+Podcasts+and+YouTube%2E+ISCA%2E)
27. VAANI Team. 2025. VAANI: Capturing the Language Landscape for an Inclusive Digital India (Phase 1). Indian Institute of Science (IISc) and ARTPARK. [find](https://scholar.google.com/scholar?q=VAANI+Team%2E+2025%2E+VAANI%3A+Capturing+the+Language+Landscape+for+an+Inclusive+Digital+India+%28Phase+1%29%2E+Indian+Institute+of+Science+%28IISc%29+and+ARTPARK%2E)
28. Wichmann, S. and Brown, C. H. and Holman, E. W. and Dryer, M. S. and Ran, Q. 2018. ASJP Database. Cross-Linguistic Linked Data (CLLD). Accessed: 11 October 2025. [find](https://scholar.google.com/scholar?q=Wichmann%2C+S%2E+and+Brown%2C+C%2E+H%2E+and+Holman%2C+E%2E+W%2E+and+Dryer%2C+M%2E+S%2E+and+Ran%2C+Q%2E+2018%2E+ASJP+Database%2E+Cross%2DLinguistic+Linked+Data+%28CLLD%29%2E+Accessed%3A+11+October+2025%2E)