On the Origin of the 1,000 Hz Peak in the Spectrum of the Human Tympanic Electrical Noise