Depiction of stimuli for the two measurement phases of the additivity task. (a) Phase 1, where the level of each masker (M1 or M2) in isolation was adapted to just mask the fixed signal level. Note that the temporal relationships between the stimuli were maintained even when only one masker was presented. (b) Phase 2, where both M1 and M2 are presented contiguously (at the levels determined in Phase 1) and the signal level is varied adaptively.
Masked thresholds with a 500 Hz signal. The plot depicts the signal level at threshold as a function of the sensation level of the signal (re: the masked threshold for the 500 Hz signal in the presence of 40 dB TEN, indicated by the value in square brackets in the upper left hand corner next to the listener number). The upward- and downward-pointing triangles show the threshold in the presence of M1 alone and M2 alone, respectively. The shaded diamonds show the threshold in the presence of both M1 and M2. The bold solid curves depict the predicted thresholds in the presence of M1 and M2 from the best fit using a power function, as described in the text. The dotted lines indicate predictions based on a linear summation of the effects from the two maskers. Error bars indicate standard deviations.
As for Fig. 2, but for a 1500 Hz signal.
As for Fig. 2, but for a 4000 Hz signal.
Additivity results for a 4000 Hz signal using the same M1 and M2 levels as for Fig. 4, but in the absence of TEN. Note that all listeners completed the 20 and 25 dB SL conditions, but not all listeners completed both the 10 and 15 dB SL conditions. Therefore, the data points on the average panel have different n values for the two lower level conditions. See the text for further details.
Thresholds for a 4000 Hz signal using lower signal levels in the absence of TEN. Symbol legend same as for previous figures.
Spectrum levels of M1 and M2 (dB SPL), averaged across S2 and S3, as derived in Phase 1. Values in parentheses are standard deviations.
Compression exponents for 500, 1500, and 4000 Hz signals in 40 dB/ERB TEN. The compression exponents fit to the average data are shown as well as the averaged compression exponents.
Compression estimates for a 4000 Hz signal, taken from Gregan et al. (2010), using a different technique (GOM) but presented in the same level of background TEN.
Compression exponents from fits to data in the absence of TEN. The missing value (—) indicates the listener was not available to run the specified condition.
Article metrics loading...
Full text loading...