Demo for Clean Speech Testset

Audio Reconstruction from Mel-spectrogram using different models (Compare with Diffusion).

Ground Truth
PriorGrad
FreGrad
RFWave