Interaural level differences (ildProc.m)

The ILD is estimated for individual frequency channels by comparing the frame-based energy of the left and the right-ear IHC representations. The temporal resolution can be controlled by the frame size ild_wSizeSec and the step size ild_hSizeSec. Moreover, the window shape can be adjusted by the parameter ild_wname. The resulting ILD is expressed in dB and negative values indicate a sound source positioned at the left-hand side, whereas a positive ILD corresponds to a source located at the right-hand side. A full list of parameters is shown in Table 33.

Table 33 List of parameters related to 'ild'.
Parameter Default Description
ild_wSizeSec 20E-3 Window duration in s
ild_hSizeSec 10E-3 Window step size in s
ild_wname 'hann' Window name

The ILD processor is demonstrated by the script DEMO_ILD.m and the resulting plots are presented in Fig. 40. The ear signals are shown for a speech source that is more closely located to the right ear (left panel). The corresponding ILD estimates are presented for individual TF units. It is apparent that the change considerably as a function of the centre frequency. Whereas hardly any ILDs are observed for low frequencies, a strong influence can be seen at higher frequencies where ILDs can be as high as 30 dB.


Fig. 40 Binaural speech signal (left panel) and the estimated ILD in dB shown as a function of time frames and frequency channels.