Salvato in:
| Autori principali: | , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2509.18722 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908555049697280 |
|---|---|
| author | Tipaksorn, Pattara Thatphithakkul, Sumonmas Chunwijitra, Vataya Thangthai, Kwanchiva |
| author_facet | Tipaksorn, Pattara Thatphithakkul, Sumonmas Chunwijitra, Vataya Thangthai, Kwanchiva |
| contents | We present LOTUSDIS, a publicly available Thai meeting corpus designed to advance far-field conversational ASR. The dataset comprises 114 hours of spontaneous, unscripted dialogue collected in 15-20 minute sessions with three participants, where overlapping speech is frequent and natural. Speech was recorded simultaneously by nine independent single-channel devices spanning six microphone types at distances from 0.12 m to 10 m, preserving the authentic effects of reverberation, noise, and device coloration without relying on microphone arrays. We provide standard train, dev, test splits and release a reproducible baseline system. We benchmarked several Whisper variants under zero-shot and fine-tuned conditions. Off-the-shelf models showed strong degradation with distance, confirming a mismatch between pre-training data and Thai far-field speech. Fine-tuning on LOTUSDIS dramatically improved robustness: a Thai Whisper baseline reduced overall WER from 64.3 to 38.3 and far-field WER from 81.6 to 49.5, with especially large gains on the most distant microphones. These results underscore the importance of distance-diverse training data for robust ASR. The corpus is available under CC-BY-SA 4.0. We also release training and evaluation scripts as a baseline system to promote reproducible research in this field. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_18722 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | LOTUSDIS: A Thai far-field meeting corpus for robust conversational ASR Tipaksorn, Pattara Thatphithakkul, Sumonmas Chunwijitra, Vataya Thangthai, Kwanchiva Computation and Language Sound We present LOTUSDIS, a publicly available Thai meeting corpus designed to advance far-field conversational ASR. The dataset comprises 114 hours of spontaneous, unscripted dialogue collected in 15-20 minute sessions with three participants, where overlapping speech is frequent and natural. Speech was recorded simultaneously by nine independent single-channel devices spanning six microphone types at distances from 0.12 m to 10 m, preserving the authentic effects of reverberation, noise, and device coloration without relying on microphone arrays. We provide standard train, dev, test splits and release a reproducible baseline system. We benchmarked several Whisper variants under zero-shot and fine-tuned conditions. Off-the-shelf models showed strong degradation with distance, confirming a mismatch between pre-training data and Thai far-field speech. Fine-tuning on LOTUSDIS dramatically improved robustness: a Thai Whisper baseline reduced overall WER from 64.3 to 38.3 and far-field WER from 81.6 to 49.5, with especially large gains on the most distant microphones. These results underscore the importance of distance-diverse training data for robust ASR. The corpus is available under CC-BY-SA 4.0. We also release training and evaluation scripts as a baseline system to promote reproducible research in this field. |
| title | LOTUSDIS: A Thai far-field meeting corpus for robust conversational ASR |
| topic | Computation and Language Sound |
| url | https://arxiv.org/abs/2509.18722 |