AI Paper Atlas

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

AAAI 2026 · Image generation › Autoregressive and latent generation · approach: Neural text-to-speech synthesis · 0 citations (OpenAlex)

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 changedWhereMeasuredEffect
add pre-training of Vox-Evaluator (ablation)Vox-Evaluator · FGES testtimestamp IOU: 0.435 -> 0.782Improved
add pre-training of Vox-Evaluator (ablation)Vox-Evaluator · FGES testWER (%): 3.15 -> 2.64Improved
add speech correctionVoiceCraft · Seed-TTS test-enWER (%): 7.56 -> 5.11Improved
add speech correctionF5-TTS · Seed-TTS test-enWER (%): 1.73 -> 1.42Improved
add speech correctionVoiceCraft · LibriSpeech-PC test-cleanWER (%): 4.68 -> 3.16Improved
add speech correctionF5-TTS · LibriSpeech-PC test-cleanWER (%): 2.42 -> 2.03Improved
add speech correctionF5-TTS · Seed-TTS test-enCMOS: 0.31 -> 0.33No clear effect
add error detection to correction (ablation)F5-TTS · TTSDS2failure rate (%): 14.0 -> 6.0Improved
add quality evaluation to correction (ablation)F5-TTS · TTSDS2failure rate (%): 14.0 -> 10.0Improved
add quality evaluation on top of error detection (ablation)F5-TTS · TTSDS2failure rate (%): 6.0 -> 6.0No clear effect
add quality evaluation on top of error detection (ablation)F5-TTS · TTSDS2UTMOS: 3.59 -> 3.70Improved
add error detection to correction (ablation)VoiceCraft · TTSDS2failure rate (%): 34.0 -> 22.0Improved

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.

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

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy