data = data_middlebrooks1999
data_middlebrooks1999 returns statistics summary from Fig. 13 (Middlebrooks, 1999b) showing the effect of non-individualized HRTFs.
Statistics of those parameters are stored as .mean and .quantiles representing the arithmetic mean and {0,5,25,50,75,95,100} quantiles, respectively.
The data struct comprises the following fields:
'le_own' | local lateral RMS error (LE) when localizing with own HRTFs |
'le_other' | LE when localizing with others HRTFs |
'lb_own' | magnitude of lateral bias (LB) when localizing with own HRTFs; upper-rear quadrant excluded from analysis |
'lb_other' | LB when localizing with others HRTFs |
'qe_own' | quadrant error rate (QE) when localizing with own HRTFs |
'qe_other' | QE when localizing with others HRTFs |
'pe_own' | local polar RMS error (PE) when localizing with own HRTFs |
'pe_other' | PE when localizing with others HRTFs |
'pb_own' | magnitude of polar bias (PB) when localizing with own HRTFs; upper-rear quadrant excluded from analysis |
'pb_other' | PB when localizing with others HRTFs |
J. C. Middlebrooks. Virtual localization improved by scaling nonindividualized external-ear transfer functions in frequency. J. Acoust. Soc. Am., 106:1493--1510, 1999.