How to make 'mfcc_stats.pkl' and 'spk2emb_82.pkl'?

[insunhwang89]

trafficstars

How to make 'mfcc_stats.pkl' and 'spk2emb_82.pkl'?

I want to test another dataset.

Thanks.

[auspicious3000]

mfcc_stats.pkl is the mean and std of the mfcc.

spk2emb_82.pkl is mapping from speaker name to one-hot embedding.

[insunhwang89]

The mfcc_stats.pkl file also contains something called 'dctmx'.

Is it correct to find 'dctmx' with the function below? from scipy import fft dctmx = fft.dct(wav)

And does dctmx also use the average value of each speaker?

Thanks for your reply.

[auspicious3000]

dctmx = scipy.fftpack.dct(np.eye(80), type=2, axis=1, norm='ortho')

you can just use the dctmx in mfcc_stats if you use the same spectrogram specifications