Schematic illustration of the filters used to create the dichotic test stimuli. The input signal was noise with equal level per . The condition labels ( to ) correspond to the number of sub-bands in each composite band.
Obtained loudness matches (open symbols) and matches predicted by the loudness model of Moore and Glasberg (2007) (filled symbols) for the set of conditions where the level of the reference sound was varied to determine a loudness match. The condition is indicated along the -axis, and the of the reference sound is indicated on the -axis. For the dichotic test stimuli, the data are plotted at a position between the positions for the corresponding Lo and Hi cases. The panels show, from top to bottom, data and predictions for the dichotic test stimuli with diotic reference, the monaural test stimuli with diotic reference, and the monaural test stimuli with monaural reference. Data are averaged across six subjects. Error bars show ±1 standard error.
The top panel shows values of LDIFF (the difference between the matching levels of the reference for a given monaural test stimulus when the reference was presented monaurally and when it was presented diotically) for the condition where the level of the test stimulus was fixed and the reference stimulus was varied in level. Open symbols show data and filled symbols show predictions of the model. The bottom panel shows corresponding results for the condition where the level of the reference stimulus was fixed and the test stimulus was varied in level.
As Fig. 2, but for the set of conditions where the level of the test sound was varied to determine a loudness match.
Article metrics loading...
Full text loading...