Directional Detection and Localization of a Bolt Click Sound in Jungle- and Pink-Noise

This paper describes how the early detection and recognition of potentially threatening sounds mixed in a background of various interfering noises is vital for the survival and effectiveness of military personnel working in hostile environments. In such environments, individuals must react to and judge threats quickly and precisely. The soldier not only needs to detect the presence of sound and identify its origin, but also needs to recognize the direction from which the sound is coming. Hearing a sound of a gun safety being taken off somewhere in the vicinity may save a soldier’s life. Hearing and localizing the direction of faint sounds, such as the sound of throat clearing or other sounds that enemy soldiers and their equipment make, may indicate a dangerous direction that should be avoided. Presented data indicate that bolt click detection (p=50%) in non-directional noise requires a SN ratio better than -9 to -12 dB depending on the direction of incoming sound and type of environment. Localization accuracy, within ±15 degree range requires SN ratios not worse than -4 to -7 dB. These levels are similar although somewhat higher than those reported in other research. These differences may be attributed to the listener training and the use of a speech signal in their study. The data also indicate that a positive SN ratio of + 6 dB or more is required for accurate signal localization. Localization errors at 0 dB SN ratio were still in the order of 30%. This means that in 30% of cases localization errors were >30 degree azimuth. These data support the contention that localization accuracy is not affected by SN ratio until it is less than +6 dB. The SN ratios at which the bolt click signal was detected were relatively high in comparison to the detection thresholds for complex sounds coming from a predetermined direction or localized inside the head. For example, a complex tone in 64 dB SPL uniformly masking noise and reported detection threshold at about 37 dB SPL, i.e., at -27 dB SN ratio is presented. Care needs to be taken in making such comparisons between this study and the present study because of the large methodological differences between studies. However, it is clear that the uncertainty regarding the direction at which a sound source could be located results in elevated detection threshold. Visual or tactile cuing may lower this threshold and auditory cuing may reduce it even further8. Among the three environmental conditions compared in this study, the pink noise condition provided greatest relative masking of the bolt click signal. This effect can be attributed to the high low-frequency energy of pink noise that effectively masked high frequency components of the bolt click signal as well as to its monotonous presence that could have a detrimental effect on the listener’s vigilance. When the test sound was presented in the front of the listener, localization uncertainty was noticeably greater at 60 degree azimuth than at 3o degree azimuth. Localization accuracy at 60 degree azimuth was similar to that at 45 degree azimuth. This observation supports the notion that sound sources located in a 45 to 60 degree azimuth range are most difficult to localize. This range is a mirror image of the 120 to 135 degree azimuth range considered the most inaccurate azimuthal range in sound localization. Localization accuracy observed in this study was only slightly poorer than that reported for sounds arriving from a priori known directions. This may indicate that direction uncertainty effects mostly detectability of the signals and to lesser degree accuracy of their perceived sound source location.


  • English

Media Info

  • Media Type: CD-ROM
  • Features: References; Tables;
  • Pagination: pp 684-693
  • Monograph Title: Noise-Con 04. The 2004 National Conference on Noise Control Engineering

Subject/Index Terms

Filing Info

  • Accession Number: 01054308
  • Record Type: Publication
  • Files: TRIS, TRB
  • Created Date: Jul 25 2007 5:58PM