ResearchBenchmarksNeural Vocoding on VCTK (unseen speakers)Follow4.37MOSGT3.63163.82334.0154.2067Apr 21, 2022Evaluation ResultsMethodMethodLinksMOSGT2022.044.37FastDiffSampling steps=4Sampling steps=42022.044.1WaveNet (MOL)2022.044.01DiffwaveSampling steps=6Sampling steps=62022.043.9UnivNet2022.043.85HiFi-GAN2022.043.74WaveGradSampling steps=50Sampling steps=502022.043.72WaveGlow2022.043.66