Final Report Student Researchers : Garrett

When a small microphone is fitted into an ear canal, one might find that the sound picked up by the microphone would consist of more than just background noise. These detectable sound emissions generated from the inner ear have been named spontaneous otoacoustic emissions or SPOAEs since their discovery by a British engineer named David Kemp in 1977. Since their discovery, otoacoustic emissions (OAEs) have been a subject of research for its clinical applications. Spontaneous otoacoustic emissions are a special type of OAE that arise without an external stimulus. These are of particular interest because their existence suggests that the inner ear has some physiological amplification mechanism. Although more than one mathematical model has been used to describe the dynamics of the inner ear, the identity of the physiological mechanism by which sound is amplified is still under debate [1, 2, 3]. Other studies of OAEs have illuminated other interesting properties such as temperature dependence and responses to external tones [4, 5]. These OAEs are transduced by a microphone into an electrical signal that is fed through an analog to digital converter. The output signal of the analog to digital converter can be modified in its resolution and its sampling rate (the amount of times the device records sound levels per second). The digital signal can be analyzed using a discrete Fourier transform, which mathematically breaks a sinusoidal signal into its frequency components. In order to reduce noise, the generated waveform is separated into several fixed intervals. A Fourier transform is then performed on these individual intervals, and finally the amplitudes of each of these waveforms are averaged. Another method to reduce noise is to ”window” the signal and then perform a Fourier transform on the windowed signal. Windowing gives a chosen part of the signal a larger weight according to some normalized function. In using the first method to reduce noise, one potential discrepancy has arisen. One set of measurements has shown that the amplitude of the Fourier transform is affected by the sampling rate of the acoustic measuring device. Although the Fourier transform’s general form stays the same, it is of interest to have a reasonable explanation for such phenomena because at least a few articles refering to measurements of this kind do not even mention at what sampling rate their analog to digital converters operate[6, pg. 479]. When they have, the sampling rate has varied from 1600 samples per second to 48 thousand samples per second [11, 10, 9].