I stumbled upon this when visualizing music. It is very interesting how everything is kind of non linear, like you mention. I had to increase low frequencies and then I got stuck trying to bucket frequencies. It seems humans have different fidelity for different ranges of frequencies