I am trying to use small chunks of voices as inputs to LSTMs for speaker Identification. The problem is that each voice sample has a different length.
The spectrogra