Lateralization of noise bursts in interaurally correlated or uncorrelated background noise using interaural level differences

Abstract

The interaural level difference (ILD) of a lateralized target source may be effectively reduced when the target is presented together with background noise containing zero ILD. It is not certain whether listeners perceive a position congruent with the reduced ILD or the actual target ILD in a lateralization task. Two sets of behavioral experiments revealed that many listeners perceived a position at or even larger than that corresponding to the presented target ILD when a temporal onset/offset asynchrony between the broadband target and the broadband background noise was present. When no temporal asynchrony was present, however, the perceived lateral position indicated a dependency on the coherence of the background noise for several listeners. With interaurally correlated background noise, listeners reported a reduced ILD resulting from the combined target and background noise stimulus. In contrast, several of the listeners made a reasonable estimate of the position corresponding to the target ILD for interaurally uncorrelated, broadband, background noise. No obvious difference in performance was seen between low- or high-frequency stimuli. Extension of a weighting template to the output of a standard equalization-cancellation model was shown to remove a lateral bias on the predicted target ILD resulting from the presence of background noise. Provided that an appropriate weighting template is applied based on knowledge of the background noise coherence, good prediction of the behavioral data is possible.

Using background noise to improve sound localization following simulated hearing loss

“…Nevertheless, we found that the perceived location of targets was pushed away from the location of the background noise. Although previous studies have observed ‘pulling’ effects between sounds that are grouped together (Lee et al, 2009), our target sounds were clearly distinct from the background noise, which tends to produce ‘pushing’ effects (Suzuki et al, 1993; Canevet and Meunier, 1996; Getzmann, 2002; Best et al, 2005; Reed and van de Par, 2015). We also found greater pushing effects for narrowband targets, which indicates that the pushing effect is a feature of binaural spatial processing.…”

“…However, sound localization can be affected even if targets are clearly audible and distinguishable from concurrent sounds. In such circumstances, the perceived location of a target sound is typically pushed away from the location of a concurrent sound (Suzuki et al, 1993; Canevet and Meunier, 1996; Getzmann, 2002; Best et al, 2005; Lee et al, 2009; Reed and van de Par, 2015). Previous attempts to explain this have relied upon a neural map of space that contains an array of neurons (or ‘spatial channels’) tuned to different spatial locations (Suzuki et al, 1993; Best et al, 2005).…”

Temporal effects in interaural and sequential level difference perception

“…Although a number of ILD models have been proposed (Breebaart et al, 2001;Brown, 2012;Brown and Tollin, 2016;Bures and Marsalek, 2013;Hartmann and Constan, 2002;Lindemann, 1986;Reed and van de Par, 2015;Stern et al, 1988;Takanen et al, 2014), these models mostly did not address the impact of temporal signal properties on ILD perception. The approaches by Brown (2012) and Brown and Tollin (2016) appear particularly interesting with respect to this because they used a physiologically plausible frontend model (Zilany et al, 2014), which has been shown to account for temporal aspects in monaural auditory processing up to the AN.…”