Research map / Image generation / Autoregressive and latent generation
Enhancing Stability and Fidelity for Zero-Shot TTS with a Multi-Level Evaluator
Hualei Wang, Na Li, Chuke Wang, Shu Wu, Zhifeng Li, Dong Yu
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add pre-training of Vox-Evaluator (ablation) | Vox-Evaluator · FGES test | timestamp IOU: 0.435 -> 0.782 | Improved |
| add pre-training of Vox-Evaluator (ablation) | Vox-Evaluator · FGES test | WER (%): 3.15 -> 2.64 | Improved |
| add speech correction | VoiceCraft · Seed-TTS test-en | WER (%): 7.56 -> 5.11 | Improved |
| add speech correction | F5-TTS · Seed-TTS test-en | WER (%): 1.73 -> 1.42 | Improved |
| add speech correction | VoiceCraft · LibriSpeech-PC test-clean | WER (%): 4.68 -> 3.16 | Improved |
| add speech correction | F5-TTS · LibriSpeech-PC test-clean | WER (%): 2.42 -> 2.03 | Improved |
| add speech correction | F5-TTS · Seed-TTS test-en | CMOS: 0.31 -> 0.33 | No clear effect |
| add error detection to correction (ablation) | F5-TTS · TTSDS2 | failure rate (%): 14.0 -> 6.0 | Improved |
| add quality evaluation to correction (ablation) | F5-TTS · TTSDS2 | failure rate (%): 14.0 -> 10.0 | Improved |
| add quality evaluation on top of error detection (ablation) | F5-TTS · TTSDS2 | failure rate (%): 6.0 -> 6.0 | No clear effect |
| add quality evaluation on top of error detection (ablation) | F5-TTS · TTSDS2 | UTMOS: 3.59 -> 3.70 | Improved |
| add error detection to correction (ablation) | VoiceCraft · TTSDS2 | failure rate (%): 34.0 -> 22.0 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- MaskGCT — MaskGCT: Zero-Shot Text-to-Speech with Masked Generative Codec Transformer (ICLR 2025) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Mega-TTS 2: Boosting Prompting Mechanisms for Zero-Shot Speech Synthesis (ICLR 2024 · similarity 0.88)
- Enhancing Zero-Shot Multi-Speaker TTS with Negated Speaker Representations (AAAI 2024 · similarity 0.87)
- DMOSpeech: Direct Metric Optimization via Distilled Diffusion Model in Zero-Shot Speech Synthesis (ICML 2025 · similarity 0.87)
- Do Not Mimic My Voice : Speaker Identity Unlearning for Zero-Shot Text-to-Speech (ICML 2025 · similarity 0.87)
- NaturalSpeech 2: Latent Diffusion Models are Natural and Zero-Shot Speech and Singing Synthesizers (ICLR 2024 · similarity 0.87)
- RVCBench: Benchmarking Robustness of Voice Cloning Across Modern Audio Generation Models (NeurIPS 2026 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
