Estimating auditory filter bandwidth using distortion product otoacoustic emissions

The basic frequency selectivity in the listener’s hearing is often characterized by auditory filters. These filters are determined through listening tests, which determine the masking threshold as a function of frequency of the tone and the bandwidth of the masking sound. The auditory filters have been shown (Glasberg and Moore, 1986) to be wider for listeners with sensorineural impairment. In a recent study (Christensen et al., 2017) it was demonstrated on group basis that the distortion product stimulus ratio that provided the strongest 2f1-f2 component at low frequencies, had a strong correlation to the theoretical relation between frequency and auditory filter bandwidth, described by the equivalent rectangular bandwidth (ERB, Glasberg and Moore, 1990). The purpose of the present study is to test if a similar correlation exists on an individual basis at normal audiometric frequencies. The optimal 2f1-f2 (L1/L2=65/45 dB SPL). The distortion-product-otoacoustic-emission ratio is determined using a custom-made measurement system programmed in MATLAB. The auditory filters are determined using notched-noise method in a two alternative forced choice experiment with noise levels at 40 dB SPL/Hz. Optimal ratios and auditory filters are determined at 1, 2, and 4 kHz for 10 young normal-hearing subjects.