NIST System for Measuring the Directivity Index of Hearing Aids under Simulated Real-Ear Conditions.
Bottom Line:
Since this response is different for a hearing aid worn on a person as compared to when it is in a free field, directivity index measurements of hearing aids are usually done under simulated real-ear conditions.Omnidirectional, cardioid, and bidirectional response patterns were measured.Results are presented comparing the NIST data with the reference values calculated from the data reported by all participating laboratories.
View Article:
PubMed Central - PubMed
Affiliation: National Institute of Standards and Technology, Gaithersburg, MD 20899-8223.
ABSTRACT
The directivity index is a parameter that is commonly used to characterize the performance of directional hearing aids, and is determined from the measured directional response. Since this response is different for a hearing aid worn on a person as compared to when it is in a free field, directivity index measurements of hearing aids are usually done under simulated real-ear conditions. Details are provided regarding the NIST system for measuring the hearing aid directivity index under these conditions and how this system is used to implement a standardized procedure for performing such measurements. This procedure involves a sampling method that utilizes sound source locations distributed in a semi-aligned zone array on an imaginary spherical surface surrounding a standardized acoustical test manikin. The capabilities of the system were demonstrated over the frequency range of one-third-octave bands with center frequencies from 200 Hz to 8000 Hz through NIST participation in an interlaboratory comparison. This comparison was conducted between eight different laboratories of members of Working Group S3/WG48, Hearing Aids, established by Accredited Standards Committee S3, Bioacoustics, which is administered by the Acoustical Society of America and accredited by the American National Standards Institute. Directivity measurements were made for a total of six programmed memories in two different hearing aids and for the unaided manikin with the manikin right pinna accompanying the aids. Omnidirectional, cardioid, and bidirectional response patterns were measured. Results are presented comparing the NIST data with the reference values calculated from the data reported by all participating laboratories. No MeSH data available. |
Related In:
Results -
Collection
License getmorefigures.php?uid=PMC4487306&req=5
Mentions: Values of DN and Dref, along with the uncertainty limits for these values, are shown as a function of frequency in Fig. 5 through Fig. 10 for each of the programmed memories and in Fig. 11 for the unaided manikin. The same vertical scale is used in Fig. 5a through Fig. 11a for easier comparison between figures. Expanded vertical scales are used in Fig. 5b through Fig. 11b to more easily differentiate between the data displayed in a given figure. Visual inspection of the plots indicates that values of DN and Dref track each other very closely with changes in frequency. Also, the uncertainty limits for these values consistently display some degree of overlap. Plots of the directivity indices in Fig. 5 and Fig. 8 for the omnidirectional response patterns programmed in the two different aids and the plot of the indices shown in Fig. 11 for the unaided manikin are very similar except at the higher frequencies. At any given frequency, lower values of the directivity indices are measured for these cases as compared to the indices shown in Fig. 6 and Fig. 7 for the cardioid response patterns and in Fig. 9 and Fig. 10 for the bidirectional response patterns. Comparisons of Fig. 6 with Fig. 7, and Fig. 9 with Fig. 10, indicate that low frequency gain boost does not have an effect on the directivity indices for either the cardioid or bidirectional response patterns. Results obtained with the NIST system were compared with the reference values using a conventional approach for judging the quality of measurement results obtained at a laboratory participating in an interlaboratory comparison [24]. A normalized deviation En defined by(6)En=QN−QrefUN2+Uref2,was calculated for each of the 119 values of QN measured (seven directional response patterns as a function of seventeen frequency bands). A value of /En/ less than unity is required for the measurement result to be considered acceptable, effectively in agreement with the reference value. All of the NIST results obtained for the interlaboratory comparison met this criterion. A value of ∣En∣ greater than unity would indicate that the difference between the NIST result and the reference value is greater than what would be expected based on the combined uncertainties. Values of En calculated for these results are displayed as a histogram in Fig. 12 and visually appear to be normally distributed. In order to test if the data can be considered to fit a normal distribution, the Anderson-Darling test [25] was applied. For this test, the hypothesis that the distribution is normal is rejected if the Anderson-Darling test statistic is greater than the critical value of 0.787 that corresponds to the conventional confidence threshold of 95 %. The Anderson-Darling test statistic calculated for the data set is 0.19, so the assumption that the data are normally distributed is accepted. The standard deviation of the distribution is 0.382. The mean is 0.029, and the standard deviation of the mean is 0.035, indicating that no significant bias is observed between the NIST results and the reference values. |
View Article: PubMed Central - PubMed
Affiliation: National Institute of Standards and Technology, Gaithersburg, MD 20899-8223.
No MeSH data available.