Abschlussarbeit

Validierung eines Ellipsoid-Modells zur Individualisierung von beliebigen Außenohrimpulsantworten

Steckbrief

Eckdaten

Professur:
MEDAK
Status:
abgeschlossen
Forschungsgebiet:
Binauraltechnik

Betreuer

Masterarbeit von Parvizian, Farzaneh

Different methods for calculating interaural time difference (ITD) were introduced by Woodworth, Kuhn, Aighazi, Duda and etc. Woodworth and Kuhn assumed the head as a sphere. Aighazi and Duda considered anthropometrical data for a spherical head model and an ellipsoid head model. A new method by Lins and Bomhardt is introduced where they combined the spherical approach with an ellipsoid model. The analytical ellipsoid method needs a listening test for the model's confirmation. The listening test focused on the adaptation of the ITD for localization in horizontal plane, so the influence of different head widths on the localization performance was examined. There are different methods of listening tests such as ABX method, Pointing method and Alternative Forced Choice (AFC) method. Pointing method and Alternative Forced Choice (AFC) method are used for ITD adaptation. The null hypothesis for the listening tests was whether the number of the correct answers for alltest subjects with different head sizes is equal. The first listening test is done by pointing method and AFC method as a pre - listening test tofigure out whether this method is a precise method for ITD adaptation or not. In this listening test the influence of different head widths is examined. A LED light is used to show the test subject that from which direction the sound should come. Because of the different head sizes which are tested here, the direction of the stimuli is changing for test subjects. The results of the listening test show that this method is not valid for ITD adaptation because each test subject has a different head width. The task was difficult and there was no correlation between the head sizes of the test subjects and any tested head width between (12-15.5) cm. The other method for the listening test is AFC method. Three listening tests were done where in the pt listening test the better ear of each test subject was examined. The 2nd listening test checked whether the test subjects can hear a difference between the original head width of the dummy head and the test subjects' head size measured by the caliper for different sound source positions. The 3rd listening test checked whether the test subjects can hear a difference between the original head width of the dummy head and the test subjects head size measured from the HRTF measurement for different sound source positions. According to the results of one way ANOVAs for this listening test, the null hypothesis is rejected. According to the results of the 2nd and the 3rd listening tests, the percentage of the correct answers for the measured head sizes by the caliper is higher than the percentage of the correct answers for the measured head sizes from the HRTF measurement. Therefore in the 2nd listening test the test subjects are able to detect the difference between the original head width of the dummy head and the test subjects' head size measured by the caliper. The percentage of the correct answers for test subjects with a head size of (12.5 -13.5) cm shows that this group of test subjects are able to detect the difference between the original head width of the dummy head and the test subjects' head size measured by the caliper or fram the HRTF measurement. The test subjects with bigger head sizes cannot detect the difference between the original head width of the dummy head and the test subjects' head size measured by the caliper or from the HRTF measurement.