another data question

[nolistic]

first, thank you so much for making this code available! so often I read a paper and there is no code which is really boring.

second, data questions!
do you remember the format of the data you trained on? as in, were they already wav files?

how noisy was the original audio? I build AI tools for call centers and am interested in feature extraction methods that are robust to noise and your thesis is one of the few times I've encountered (1) actually used call center audio (2) has code on github.

[woshipzs]

The data that I trained on are already in wav format.

For my experiment, the data are quite noisy and small size.