Schematics of the models. The loudness model of Moore and Glasberg (2004) is depicted in the lower branch. The branching of to the top depicts the SIIcmp model.
Schematic input-output cochlear compression function for several values of the cochlear gain (from bottom to top): 0, 7.6, 17.6, 27.6, 37.6, 47.6, and 57.6 dB.
Distribution of pure-tone thresholds in the set of 402 ears of the Festen and Smits (2007) set. The percentages indicated represent the fraction of ears at a certain frequency with higher tone thresholds than the plotted value.
Predicted ANSI SII values for the Festen and Smits (2007) data set. In each box triplet, the boxes on the left, in the middle, and on the right represent the groups NH, NIHL, and HI, respectively. From left to right, the triplets show the predictions for the SRTq and for four SRTns at noise levels of 35, 50, 65, and 80 dBA, respectively.
Predicted SIIcmp values for the Festen and Smits (2007) data set for the chosen IFF dynamic ranges of 30, 35, 40, 45, 50, and 55 dB. The layout of each panel is the same as in Fig. 4.
The top panel shows the average SII values as a function of model (ANSI SII and SIIcmp for the chosen six IFF dynamic ranges) for the SRTq in NH, NIHL, and HI listeners. The results for the listener groups NH, NIHL, and HI are displayed as circles, triangles, and squares, respectively. Error bars indicate the standard deviations between subjects. The bottom panel shows for the corresponding SII values for the SRTn, supplemented with the overall averaged results for the listener groups in quiet (up-side down triangles).
Predicted SII and SIIcmp values, for an IIF dynamic range of 40 dB, for the data from Smoorenburg (1992), denoted as , , and ; from Bosman and Smoorenburg (1995), denoted as , , and ; from de Laat and Plomp (1983), denoted as and ; and for the set of 20 hearing-impaired ears of the data from Festen and Plomp (1990), denoted as and for two types of used speech material, where the subscripts , NI, and stand for normal hearing (to the left of the vertical dotted line), noise-induced hearing loss, and hearing impaired, respectively. Note that the average values for the SII and the SIIcmp are not the same (horizontal dotted lines).
Frequency regions of the FFTs and the corresponding temporal windows.
(a) Distribution of SRTs of the Festen and Smits (2007) data set, measured in quiet, expressed in dBA, and the SRTs measured in noise at 35, 50, 65, and 80 dBA, expressed in dB SNR. Data are shown for the 323 normally hearing listeners in (a), for the 65 listeners with noise-induced hearing losses in (b), and for the 14 hearing-impaired group in (c).
SRTs measured in the quiet, expressed in dBA, and the SRTs measured in noise at 35, 50, 65, and 80 dBA, expressed in dB SNR for NH, NIHL, and HI. Standard deviations are given between parentheses.
Article metrics loading...
Full text loading...