Poor Synthesis Quality on 44k Sample Rate

[Malik7115]

Hello, I have two somewhat identical datasets with similar samples. I have trained the WaveGrad with 22k sample rate audios and it is quite good. However, the synthesis quality for 44k sample rate data is not as good. Would really appreciate any suggestions, especially in terms of changing model parameters. The only changes in parameters are as follows:

sample rate: 44k n_fft: 2048 window_size: 2048 ho_length: 512

[ivanvovk]

@Malik7115 Hey, I can suggest you to take a look on the Large WaveGrad model in the orig WaveGrad paper (not implemented here, but I think it is easy to adjust the current code). It definitely would be slower a bit, but at the same time should obtain better quality.