List your company

Zero-shot Text-to-Speech on LibriSpeech SNR = 0dB (test-clean)

2.58UTMOSOZSpeech
1.281.61751.9552.2925May 19, 2025

Evaluation results

MethodLinks
OZSpeechTraining Dataset=LibriTTS, Fine-tuned on noisy prompts=true2025.052.586230.286721.37540.02
F5-TTSTraining Dataset=LibriTTS2025.051.883217—4151.8440.29
OZSpeechTraining Dataset=LibriTTS2025.051.726170.224546.6440.08
VoiceCraftTraining Dataset=GigaSpeech2025.051.5844220.294057.4550.02
VALL-ETraining Dataset=LibriTTS2025.051.449370.1136102.68520.07
NaturalSpeech 2Training Dataset=LibriTTS2025.051.3323120.264838.27560.01