Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2508.12968 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916905563979776 |
|---|---|
| author | Gerazov, Branislav Politi, Marcello Bratières, Sébastien |
| author_facet | Gerazov, Branislav Politi, Marcello Bratières, Sébastien |
| contents | We explore the performance of several state-of-the-art automatic speech recognition (ASR) models on a large-scale Arabic speech dataset, the SADA (Saudi Audio Dataset for Arabic), which contains 668 hours of high-quality audio from Saudi television shows. The dataset includes multiple dialects and environments, specifically a noisy subset that makes it particularly challenging for ASR. We evaluate the performance of the models on the SADA test set, and we explore the impact of fine-tuning, language models, as well as noise and denoising on their performance. We find that the best performing model is the MMS 1B model finetuned on SADA with a 4-gram language model that achieves a WER of 40.9\% and a CER of 17.6\% on the SADA test clean set. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_12968 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Arabic ASR on the SADA Large-Scale Arabic Speech Corpus with Transformer-Based Models Gerazov, Branislav Politi, Marcello Bratières, Sébastien Audio and Speech Processing Machine Learning We explore the performance of several state-of-the-art automatic speech recognition (ASR) models on a large-scale Arabic speech dataset, the SADA (Saudi Audio Dataset for Arabic), which contains 668 hours of high-quality audio from Saudi television shows. The dataset includes multiple dialects and environments, specifically a noisy subset that makes it particularly challenging for ASR. We evaluate the performance of the models on the SADA test set, and we explore the impact of fine-tuning, language models, as well as noise and denoising on their performance. We find that the best performing model is the MMS 1B model finetuned on SADA with a 4-gram language model that achieves a WER of 40.9\% and a CER of 17.6\% on the SADA test clean set. |
| title | Arabic ASR on the SADA Large-Scale Arabic Speech Corpus with Transformer-Based Models |
| topic | Audio and Speech Processing Machine Learning |
| url | https://arxiv.org/abs/2508.12968 |