lf0 question about convert phase

[powei-C]

Hi,
I wonder why you normalize f0 series before feeding to the f0encoder in convert.py.
However, this kind of normalization for f0 isn't used in preprocessing phase.

[Wendison]

Hi, normalizing f0 aims to remove the speaker characteristics. During preprocessing phase, f0 is not normalized, but during training and inference, f0 is normalized as shown below:

VQMIVC/dataset.py

Line 53 in 851b4f5

lf0 = (lf0 - mean) / (std + 1e-8)

VQMIVC/convert_example.py

Line 57 in 851b4f5

lf0[nonzeros_indices] = (lf0[nonzeros_indices] - mean) / (std + 1e-8)

[powei-C]

Hi,
thank you for your explanation!!!
I have another question about perplexity when training the model with another dataset.
I found that the perplexity didn't keep increasing (have run around 360 epochs in the figure), was it reasonable?
And do you have any suggestions to verify this issue?

[Wendison]

The perplexity should be increasing during training, as higer perplexity indicates that the vectors in the VQ codebook are distinguiable and can be used to represent different acoustic units. I also saw your recon_loss is high. Based on my experience, recon_loss should be less than 0.5, then you would obtain good converted samples.