.2.1 Psychoacoustical Cues in Sound Localization
Sound localization is a complex process where the brain use a range of cues to identify the spatial position of sounds sources [?]. In sound localization it is common to draw a distinction between localization in the horizontal and the vertical dimension. Localization in the horizontal plane, defined by the tip of the nose and the two ear canals, is largely based on interaural differences between the two ears [?]. This is sometimes referenced as binaural cues, or binaural hearing.
Horizontal Localization
Humans have two ears located in a fixed position on each side of the head. If a sound is presented from the side it will be shadowed by the listeners head in reaching the far ear. That will make the sound appear stronger in intensity on the side it originates from (see figure 2.1). The interaural differences of the two ears in sound pressure level will indicate which side the sound originates from and contribute in localization of the source.
18 CHAPTER2. THEORY
This method of localization holds true for sounds of high pitch of tone. For frequencies below 1000 Hz the sound wavelength will be larger than the head and the sound pressure level will no longer give sufficient cues to determine localization.
Lord Rayleigh, a British physicist, discovered that the difference in phases at the two ears had to account for localization of sound with low pitch [?]. If a sound is presented directly from the side it will reach the closest ear first and arrive at the far ear approximately one millisecond later [?]. The amount of phase differences will decrease as the angle from the front decrease and be zero when the sound is directly ahead of the listener. The brain detect this differences and use this to determine sound source position in the horizontal plane. The sensitivity of phase differences declined with increased frequencies and is not present above 1600 Hz [?].
Interaural differences in phases determine localization in the horizontal plane for low frequencies, while intensity differences determine localiza tion for high frequencies. This has been known as the duplex theory of sound localization [?].
Vertical Localization
With respect to sound localization in the vertical dimension interaural differences falls short. Vertical movement in the medial plane will not produce any interaural differences when the ears and heads are symmetrical. Still: humans are able to detect if a sound is coming from above or below. The main cues used in vertical localization of sound seems to be something called ”spectral shape cues” [?]. The shape of pinna, the external part of the ear, produce subtle changes in the spectrum of the sound based on the direction and distance it originates from [?]. When sound enter through the ear canal it is slightly altered by reflections, shadowing and resonance caused by the external ear. This cues appear as peaks and notches in the spectrum at certain frequencies that the sensory register manage to detect and interpret as localization cues2. This type of vertical localization seems to rely on broadband sound stimuli where frequencies above 4000 Hz is most important [?]. The effects of spectral shape cues on localization in the horizontal plane are much discussed, but it does seem to aid in reducing front/back confusion. Reflections from the shoulders and torso may also aid in sound localization in a similar matter, but the use of these cues seem to differ considerably between individuals [?].
2Vertical localization seems to be almost as accurate with one ear as with two. In auditory interface design it could be possible to utilize this aspect of spatial audio in applications that use headphones with a single earpiece. Handsfree Bluetooth sets that present notifications from above while the conversation is presented from the side, may be a possible application.