Speech Recognition - Audio Data Analysis Post Doubts (MFCCS)

Hello,

I was going through the following post:

I did not understand the following things:
1 mfccs = np.mean(librosa.feature.mfcc(y=X, sr=sample_rate, n_mfcc=40).T,axis=0)
2 model.add(Dense(256, input_shape=(40,)))

What is the reason for taking n_mfcc =40?
Why they have taken mean to get mfccs?
Why the input_shape = (40,)?

If anyone has read the above-mentioned post and have understood the same, kindly help me to understand.

Best,
Sid

© Copyright 2013-2019 Analytics Vidhya