ResearchBenchmarksText-to-Speech on LJSpeech (WER, MCD, F0 RMSE, PESQ)Follow0.651F0 RMSEStyleTTS20.650360.654680.6590.66332Mar 31, 2026Evaluation ResultsMethodMethodLinksF0 RMSEMCDWERPESQStyleTTS2RTF=0.0174RTF=0.01742026.030.6514.936.53.79MVCRTF=0.0169RTF=0.01692026.030.6534.916.523.85VITSRTF=0.0211RTF=0.02112026.030.6674.977.233.64