I am a university student in computer science and during a project, I run into the Mel Spectrogram. Doing some research, I found that the MelSpectogram is essentially a spectrogram where the distance between two frequency is proportional to how the human perceives it.
So my question is why I should use the Mel Spectrogram for something like a Neural Network? shouldn't it be the same for a computer to use the Spectrogram or the Mel Spectrogram?
