Perdue, Richard John
Distortion product otoacosutic emissions: challenging existing theory
University of Southampton, Institute of Sound and Vibration Research,
Distortion product otoacoustic emissions (DPOAE or DP) are generated using two primary tones at frequencies f1 and f2 at levels L1 and L2; the two most prominent DP are cubic distortion products at the frequencies 2f1f2 and 2f2f1. Current understanding of their origin indicates that they arise via two distinct mechanisms. One mechanism is thought to be non-linear distortion and the other linear coherent reflection. Recent investigations have confirmed that both components are present in the 2f2f1 DP, which conflicts with current understanding of the generation mechanisms, suggesting that the location of origin of the DP at 2f2f1 may be different from that at 2f1f2.
Distortion and reflection components of DP combine in the cochlea. One method to separate them is time window separation, which utilises phase variation in the frequency domain to infer latency in the time domain, by means of inverse Fourier transformation. It is independent of the location of generation of the DP but depends on the generation mechanism. By contrast, DP can be reduced by the introduction of a third suppressor tone, which can be used to infer the location of DP generation based of the suppressor tone frequency, independent of generation mechanism.
The aim of this investigation was to assess the effect of suppressor frequency and level on the 2f2f1 DP, in order to make inferences about where this DP originates. Twenty normally hearing participants took part in the investigation. DP were measured (L1 = 65, L2 = 55 dB SPL) with a fixed frequency ratio f2/f1 = 1.05. By sweeping the primary frequencies in the ranges f2 = 1.75-2.25 and 3.75-4.25 kHz, suitable fixed test frequencies were identified for subsequent suppression testing in each ear, as the frequencies giving DP with the greatest signal-to-noise ratio. The frequency sweeps also allowed analysis by time window separation. Suppression was carried out with suppressor levels introduced at 0, 20, 40 and 60 dB SPL. The frequency of the suppressor was altered from 32 Hz to +64 Hz relative to the DP, in 16 Hz increments.
Results of time window separation confirmed that both non-linear distortion and reflection components were present in both the 2f2f2 DP and 2f2f1 DP. However, it was not possible to separate the two components of the 2f2f1 DP by suppression, which suggests that the 2f2f1 DP is not generated at a discrete place. It is concluded that the generation mechanism for the 2f2f1 DP may be distributed along the cochlear partition at or basal to the characteristic place corresponding to the DP frequency.
Actions (login required)