Bengali-Loop: Community Benchmarks for Long-Form Bangla ASR and Speaker Diarization
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917275641053184 |
|---|---|
| author | Tabib, H. M. Shadman Rifti, Istiak Ahmmed Ehsan, Abdullah Muhammed Amimul Dasgupta, Somik Sowdha, Md Zim Mim Siddiqee Sarker, Abrar Jahin Nijamy, Md. Rafiul Islam Hossain, Tanvir Khatun, Mst. Metaly Mahmood, Munzer Debnath, Rakesh Biswas, Gourab Karim, Asif Navid, Wahid Al Azad Muztahid, Masnoon Udoy, Fuad Ahmed Rahman, Shahad Shahriar Shifat, Md. Tashdiqur Rahman Khatun, Most. Sonia Rahman, Mushfiqur Hasan, Md. Miraj Saha, Anik Nobo, Mohammad Ninad Mahmud Bhattacharjee, Soumik Bhomik, Tusher Swapnil, Ahmmad Nur Kabir, Shahriar |
| author_facet | Tabib, H. M. Shadman Rifti, Istiak Ahmmed Ehsan, Abdullah Muhammed Amimul Dasgupta, Somik Sowdha, Md Zim Mim Siddiqee Sarker, Abrar Jahin Nijamy, Md. Rafiul Islam Hossain, Tanvir Khatun, Mst. Metaly Mahmood, Munzer Debnath, Rakesh Biswas, Gourab Karim, Asif Navid, Wahid Al Azad Muztahid, Masnoon Udoy, Fuad Ahmed Rahman, Shahad Shahriar Shifat, Md. Tashdiqur Rahman Khatun, Most. Sonia Rahman, Mushfiqur Hasan, Md. Miraj Saha, Anik Nobo, Mohammad Ninad Mahmud Bhattacharjee, Soumik Bhomik, Tusher Swapnil, Ahmmad Nur Kabir, Shahriar |
| contents | Bengali (Bangla) remains under-resourced in long-form speech technology despite its wide use. We present Bengali-Loop, two community benchmarks to address this gap: (1) a long-form ASR corpus of 191 recordings (158.6 hours, 792k words) from 11 YouTube channels, collected via a reproducible subtitle-extraction pipeline and human-in-the-loop transcript verification; and (2) a speaker diarization corpus of 24 recordings (22 hours, 5,744 annotated segments) with fully manual speaker-turn labels in CSV format. Both benchmarks target realistic multi-speaker, long-duration content (e.g., Bangla drama/natok). We establish baselines (Tugstugi: 34.07% WER; pyannote.audio: 40.08% DER) and provide standardized evaluation protocols (WER/CER, DER), annotation rules, and data formats to support reproducible benchmarking and future model development for Bangla long-form ASR and diarization. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_14291 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Bengali-Loop: Community Benchmarks for Long-Form Bangla ASR and Speaker Diarization Tabib, H. M. Shadman Rifti, Istiak Ahmmed Ehsan, Abdullah Muhammed Amimul Dasgupta, Somik Sowdha, Md Zim Mim Siddiqee Sarker, Abrar Jahin Nijamy, Md. Rafiul Islam Hossain, Tanvir Khatun, Mst. Metaly Mahmood, Munzer Debnath, Rakesh Biswas, Gourab Karim, Asif Navid, Wahid Al Azad Muztahid, Masnoon Udoy, Fuad Ahmed Rahman, Shahad Shahriar Shifat, Md. Tashdiqur Rahman Khatun, Most. Sonia Rahman, Mushfiqur Hasan, Md. Miraj Saha, Anik Nobo, Mohammad Ninad Mahmud Bhattacharjee, Soumik Bhomik, Tusher Swapnil, Ahmmad Nur Kabir, Shahriar Sound Audio and Speech Processing Bengali (Bangla) remains under-resourced in long-form speech technology despite its wide use. We present Bengali-Loop, two community benchmarks to address this gap: (1) a long-form ASR corpus of 191 recordings (158.6 hours, 792k words) from 11 YouTube channels, collected via a reproducible subtitle-extraction pipeline and human-in-the-loop transcript verification; and (2) a speaker diarization corpus of 24 recordings (22 hours, 5,744 annotated segments) with fully manual speaker-turn labels in CSV format. Both benchmarks target realistic multi-speaker, long-duration content (e.g., Bangla drama/natok). We establish baselines (Tugstugi: 34.07% WER; pyannote.audio: 40.08% DER) and provide standardized evaluation protocols (WER/CER, DER), annotation rules, and data formats to support reproducible benchmarking and future model development for Bangla long-form ASR and diarization. |
| title | Bengali-Loop: Community Benchmarks for Long-Form Bangla ASR and Speaker Diarization |
| topic | Sound Audio and Speech Processing |
| url | https://arxiv.org/abs/2602.14291 |