Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2508.21631 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866916925307617280 |
|---|---|
| author | Perrin, Yanis Boulianne, Gilles |
| author_facet | Perrin, Yanis Boulianne, Gilles |
| contents | Supervised training of speech recognition models requires access to transcribed audio data, which often is not possible due to confidentiality issues. Our approach to this problem is to generate synthetic audio from a text-only corpus using a state-of-the-art text-to-speech model with voice cloning capabilities. Our goal is to achieve automatic speech recognition (ASR) performance comparable to models trained on real data. We explore ways to optimize synthetic data generation through finetuning, filtering and evaluation, and its use for training an end-to-end encoder-decoder ASR model. Experiments were conducted using two datasets of spontaneous, conversational speech in Québec French. We show that improving data generation leads to large improvements in the final ASR system trained on synthetic data. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_21631 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Towards Improved Speech Recognition through Optimized Synthetic Data Generation Perrin, Yanis Boulianne, Gilles Audio and Speech Processing Supervised training of speech recognition models requires access to transcribed audio data, which often is not possible due to confidentiality issues. Our approach to this problem is to generate synthetic audio from a text-only corpus using a state-of-the-art text-to-speech model with voice cloning capabilities. Our goal is to achieve automatic speech recognition (ASR) performance comparable to models trained on real data. We explore ways to optimize synthetic data generation through finetuning, filtering and evaluation, and its use for training an end-to-end encoder-decoder ASR model. Experiments were conducted using two datasets of spontaneous, conversational speech in Québec French. We show that improving data generation leads to large improvements in the final ASR system trained on synthetic data. |
| title | Towards Improved Speech Recognition through Optimized Synthetic Data Generation |
| topic | Audio and Speech Processing |
| url | https://arxiv.org/abs/2508.21631 |