Natural Yet Challenging to Detect: Robust In-the-Wild TTS through EMA and Dual-Scoring Prompt Selection -- Submission for WildSpoof 2026 TTS Track
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866913157102960640 |
|---|---|
| author | Sun, Renhe Zhou, Jiayi He, Haolin Feng, Yueying Liu, Jian |
| author_facet | Sun, Renhe Zhou, Jiayi He, Haolin Feng, Yueying Liu, Jian |
| contents | In this technical report, we describe our submission for the WildSpoof Challenge TTS Track: Text-to-Speech with In-the-Wild Data. We introduce F5-TTS-DPS, a model built upon the F5-TTS architecture. Our approach integrates Exponential Moving Average (EMA) into supervised fine-tuning to stabilize training and improve generalization. To enhance synthesis fidelity, we leverage large language models (LLMs) and large audio language models (LALMs) for dual-scoring prompt selection, filtering reference audio and text prompts to ensure quality while addressing alignment issues in noisy datasets.
Experimental evaluation demonstrates that F5-TTS-DPS achieves strong performance with UTMOS of 3.20 and speaker similarity of 0.51 on the development set. More importantly, our model achieves the best a-DCF scores of 0.1582, 0.5233, and 0.2562 across three advanced SASV systems among all submissions, indicating our synthesized speech is the most difficult to detect and exhibits the highest degree of naturalness and authenticity. Combined with competitive WER performance, these results validate the effectiveness of our approach in generating natural-sounding speech with strong spoofing capabilities. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_23859 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Natural Yet Challenging to Detect: Robust In-the-Wild TTS through EMA and Dual-Scoring Prompt Selection -- Submission for WildSpoof 2026 TTS Track Sun, Renhe Zhou, Jiayi He, Haolin Feng, Yueying Liu, Jian Audio and Speech Processing In this technical report, we describe our submission for the WildSpoof Challenge TTS Track: Text-to-Speech with In-the-Wild Data. We introduce F5-TTS-DPS, a model built upon the F5-TTS architecture. Our approach integrates Exponential Moving Average (EMA) into supervised fine-tuning to stabilize training and improve generalization. To enhance synthesis fidelity, we leverage large language models (LLMs) and large audio language models (LALMs) for dual-scoring prompt selection, filtering reference audio and text prompts to ensure quality while addressing alignment issues in noisy datasets. Experimental evaluation demonstrates that F5-TTS-DPS achieves strong performance with UTMOS of 3.20 and speaker similarity of 0.51 on the development set. More importantly, our model achieves the best a-DCF scores of 0.1582, 0.5233, and 0.2562 across three advanced SASV systems among all submissions, indicating our synthesized speech is the most difficult to detect and exhibits the highest degree of naturalness and authenticity. Combined with competitive WER performance, these results validate the effectiveness of our approach in generating natural-sounding speech with strong spoofing capabilities. |
| title | Natural Yet Challenging to Detect: Robust In-the-Wild TTS through EMA and Dual-Scoring Prompt Selection -- Submission for WildSpoof 2026 TTS Track |
| topic | Audio and Speech Processing |
| url | https://arxiv.org/abs/2605.23859 |