Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2604.22225 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917432756535296 |
|---|---|
| author | Wang, Xi Wang, Jie Song, Xingchen Song, Baijun Xie, Jingran Shao, Jiahe Lin, Zijian Wu, Di Meng, Meng Luan, Jian Wu, Zhiyong |
| author_facet | Wang, Xi Wang, Jie Song, Xingchen Song, Baijun Xie, Jingran Shao, Jiahe Lin, Zijian Wu, Di Meng, Meng Luan, Jian Wu, Zhiyong |
| contents | While generative text-to-speech (TTS) models approach human-level quality, monolithic metrics fail to diagnose fine-grained acoustic artifacts or explain perceptual collapse. To address this, we propose TTS-PRISM, a multi-dimensional diagnostic framework for Mandarin. First, we establish a 12-dimensional schema spanning stability to advanced expressiveness. Second, we design a targeted synthesis pipeline with adversarial perturbations and expert anchors to build a high-quality diagnostic dataset. Third, schema-driven instruction tuning embeds explicit scoring criteria and reasoning into an efficient end-to-end model. Experiments on a 1,600-sample Gold Test Set show TTS-PRISM outperforms generalist models in human alignment. Profiling six TTS paradigms establishes intuitive diagnostic flags that reveal fine-grained capability differences. TTS-PRISM is open-source, with code and checkpoints at https://github.com/xiaomi-research/tts-prism. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_22225 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | TTS-PRISM: A Perceptual Reasoning and Interpretable Speech Model for Fine-Grained Diagnosis Wang, Xi Wang, Jie Song, Xingchen Song, Baijun Xie, Jingran Shao, Jiahe Lin, Zijian Wu, Di Meng, Meng Luan, Jian Wu, Zhiyong Computation and Language Audio and Speech Processing While generative text-to-speech (TTS) models approach human-level quality, monolithic metrics fail to diagnose fine-grained acoustic artifacts or explain perceptual collapse. To address this, we propose TTS-PRISM, a multi-dimensional diagnostic framework for Mandarin. First, we establish a 12-dimensional schema spanning stability to advanced expressiveness. Second, we design a targeted synthesis pipeline with adversarial perturbations and expert anchors to build a high-quality diagnostic dataset. Third, schema-driven instruction tuning embeds explicit scoring criteria and reasoning into an efficient end-to-end model. Experiments on a 1,600-sample Gold Test Set show TTS-PRISM outperforms generalist models in human alignment. Profiling six TTS paradigms establishes intuitive diagnostic flags that reveal fine-grained capability differences. TTS-PRISM is open-source, with code and checkpoints at https://github.com/xiaomi-research/tts-prism. |
| title | TTS-PRISM: A Perceptual Reasoning and Interpretable Speech Model for Fine-Grained Diagnosis |
| topic | Computation and Language Audio and Speech Processing |
| url | https://arxiv.org/abs/2604.22225 |