CAST-TTS: A Simple Cross-Attention Framework for Unified Timbre Control in TTS
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866910056083095552 |
|---|---|
| author | Zheng, Zihao Wu, Wen Zhang, Chao Wu, Mengyue Xu, Xuenan |
| author_facet | Zheng, Zihao Wu, Wen Zhang, Chao Wu, Mengyue Xu, Xuenan |
| contents | Current Text-to-Speech (TTS) systems typically use separate models for speech-prompted and text-prompted timbre control. While unifying both control signals into a single model is desirable, the challenge of cross-modal alignment often results in overly complex architectures and training objective. To address this challenge, we propose CAST-TTS, a simple yet effective framework for unified timbre control. Features are extracted from speech prompts and text prompts using pre-trained encoders. The multi-stage training strategy efficiently aligns the speech and projected text representations within a shared embedding space. A single cross-attention mechanism then allows the model to use either of these representations to control the timbre. Extensive experiments validate that the unified cross-attention mechanism is critical for achieving high-quality synthesis. CAST-TTS achieves performance comparable to specialized single-input models while operating within a unified architecture. The demo page can be accessed at https://HiRookie9.github.io/CAST-TTS-Page. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_16280 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | CAST-TTS: A Simple Cross-Attention Framework for Unified Timbre Control in TTS Zheng, Zihao Wu, Wen Zhang, Chao Wu, Mengyue Xu, Xuenan Sound Audio and Speech Processing 68Txx I.2 Current Text-to-Speech (TTS) systems typically use separate models for speech-prompted and text-prompted timbre control. While unifying both control signals into a single model is desirable, the challenge of cross-modal alignment often results in overly complex architectures and training objective. To address this challenge, we propose CAST-TTS, a simple yet effective framework for unified timbre control. Features are extracted from speech prompts and text prompts using pre-trained encoders. The multi-stage training strategy efficiently aligns the speech and projected text representations within a shared embedding space. A single cross-attention mechanism then allows the model to use either of these representations to control the timbre. Extensive experiments validate that the unified cross-attention mechanism is critical for achieving high-quality synthesis. CAST-TTS achieves performance comparable to specialized single-input models while operating within a unified architecture. The demo page can be accessed at https://HiRookie9.github.io/CAST-TTS-Page. |
| title | CAST-TTS: A Simple Cross-Attention Framework for Unified Timbre Control in TTS |
| topic | Sound Audio and Speech Processing 68Txx I.2 |
| url | https://arxiv.org/abs/2603.16280 |