Normal-to-ASMR (N2A) Style Synthesis on ASMR Cross-Style Synthesis
14.02Global Unvoiced RatioCosyVoice2 + SeedVC
Evaluation results
| Method | Links | |
|---|---|---|
| CosyVoice2 + SeedVCTTS Model=CosyVoice2, VC Model=SeedVC2026.01 | 14.02 | |
| F5TTS + SeedVCTTS Model=F5TTS, VC Model=SeedVC2026.01 | 14.8 | |
| F5TTS + CosyVoiceVCTTS Model=F5TTS, VC Model=CosyVoiceVC2026.01 | 21.41 | |
| CosyVoice2 + CosyVoiceVCTTS Model=CosyVoice2, VC Model=CosyVoiceVC2026.01 | 37.35 | |
| DeepASMRTTS Model=DeepASMR2026.01 | 74.21 | |
| GroundTruth2026.01 | 91.78 |