Abstract
Yaw movements are the most frequently occurring and largest head movements a listener makes when localising; however, previous research has not resolved whether yaw-based head movements are used in elevation localisation. A series of experiments was devised to investigate the impact of head movement on the elevation localisation response accuracy (LRA) of human listeners. The experiments were conducted using a laser-guided pointing response method, as this was found to allow listeners to more accurately and consistently report a perceived source location than either verbal or graphical methods. 2 kHz low-pass filtered noise with and without a 6 kHz half-octave bandpass component were both shown to suppress pinna cues, and were therefore used to more clearly separate the effect of head movements. Head movements were found to improve azimuth and elevation LRA for noise sources. Depending on stimulus and situation, head movements were shown to make an improvement of up to 8.5˚ in elevation LRA. Head movement improved LRA more for the 2kHz/6kHz filtered noise than it did for broadband noise; when pinna cues are impaired the significance of head movement cues increases. Both forced yaw movements and free movements significantly improved the elevation LRA. Further experimentation was undertaken to determine whether the improvement in elevation LRA with head movement was caused by greater accuracy when a source is positioned in the listener’s median plane (a static cue), or by the act of moving the head (a dynamic cue). It was found that the static cue did not provide greater accuracy for sources close to the median plane, and hence it was concluded that dynamic cues increased the elevation LRA for yaw head rotations. For octave and half-octave bandwidth sources, static elevation LRA is lower when the listener has turned to face the source. Yaw head movement improved elevation LRA for high frequency continuous signals, which suggests that dynamic interaural level differences are utilised. Head movements do not improve elevation LRA for programme items with less than an octave bandwidth. For octave programme items, head movements significantly improve elevation LRA, while static LRA shows no improvement; head movement cues are effective at narrower bandwidths than pinna cues. By detailing the nature of head movement cues, one can better inform the localisation model, creating a more accurate representation of the human localisation system.