Coding of Elevation in Acoustic Image of Space

Spectral features of sound are believed to be the primary cues for the human perception of spatial sound elevation. It has also been observed that people connect higher frequencies of sound with a higher elevation of the sound source and lower frequencies with lower elevations. The most common approach to creating an acoustic image delivered by headphones is to use Head Related Transfer Functions (HRTFs). Unfortunately, satisfactory perception of elevation can only be achieved with personalized HRTFs which are impractical to measure. This paper describes an alternative method of sound coding for representation of the virtual sound source elevation in an acoustic image. Our method is based on coding particular elevations with sound stimuli which differ in spectral content. Sound stimuli were created by various signal processing techniques (e.g. filtration, modulation). Experiments show that, in certain cases, test subjects were able to perceive up to 60 different elevations in the range of –40° to 90°.