The role of spectral components of the head-related transfer functions in evaluation of the virtual sound source motion in the vertical plane

Perception of virtual sound sources moving in the range of elevation from −45° to 45° (at zero azimuth) was studied with participation of listeners aged 57–73. The virtual sound source trajectory was created using nonindividualized head-related transfer functions and artificially synthesized spectral components specific to these functions. The percentage of correct responses in determining the direction of virtual motion depended on the way of imitation by increasing from low to high for the following succession of imitation methods: (i) displacement of the spectral minimum of broadband noise pulses within a frequency band of 5–12 kHz (the minimum retained a constant width of 1 kHz), (ii) variation of the spectral minimum width of noise pulses within 6–12 kHz, (iii) variation of the spectral maximum width within 4–8 kHz, (iv) simultaneous variation of the spectral minimum and maximum widths, (v) displacement of the spectral minimum and simultaneous variation of the spectral maximum width, and (vi) displacement of the spectral minimum and simultaneous variation of the spectral maximum width and power. For the latter stimulus, the mean percentage of correct responses (90 ± 5)% did not differ from the corresponding percentage (94 ± 3)% observed for the stimulus that was synthesized on the basis of nonindividualized head-related transfer functions and used as reference in synthesizing the spectral components.