FIG. 5.

(Color online) Relative conditional feature information transfer (IT) for the NH and HI listener groups on seven features for each of three noise types and six speech conditions. The upper two panels show results for the baseline noise condition, the middle two panels for interrupted noise, and the lower two panels for continuous noise. For each noise type, NH data are plotted above the HI data. The seven features shown along the abscissa are nasality (NAS), approximant (APP), strident (STR), sonorant (SON), voicing (VOI), continuancy (CON), and place (PLA). For each feature and in each panel, bars show results for six speech conditions ordered as U, E, T1, R1, T4, and R4.