The most commonly alleged experimental artifact associated with real-ear attenuation at threshold (REAT) measurements of hearing protection devices (HPDs) was examined: Masking of the protected thresholds due to physiological noise amplified by the occlusion effect. An ear canal mounted subminiature microphone was used to obtain objective measures of physiological noise in occluded and unoccluded test conditions and of the insertion loss (IL) of insert, semi-aural, supra-aural and circumaural HPDs when exposed to broadband noise with a sound pressure level of 93 dB. Measurements spanned 1/3 octave bands from 125 Hz to 2 kHz. Attenuation was also measured via a subjective REAT procedure and the magnitude of the occlusion effect was examined via bone conduction audiometry. The IL data confirmed the accuracy of the REAT results except at the lowest frequencies tested, where the degree to which the REAT values were spuriously inflated was quantified and found to be device related. Furthermore, the magnitude of the error (which never exceeded 5 dB) could be predicted by measuring the physiological noise in the occluded ear and calculating how much this would mask the occluded threshold. It was noted that no evidence was found in the data to suggest a dependency of HPD attenuation on sound level.

No abstract

The most ubiquitous method of measuring the attenuation of hearing protection devices (HPDs) is that of real-ear attenuation at threshold (REAT). Such tests, which evaluate the hearing thresholds of human subjects both with and without the HPDs in place, have often been criticized for measurement artifact and inappropriate test conditions. We examined two of the more common alleged REAT problems; masking of the occluded thresholds due to amplified physiological noise and the possible errors resulting from extrapolating low sound level performance tests to the high sound levels in which HPDs are typically worn. An ear canal mounted subminiature microphone was used to measure physiological noise in occluded and unoccluded test conditions as well as to measure the insertion loss (IL) of insert, supra-aural and circumaural HPDs. Measurements spanned 1/3 octave bands from 125 Hz to 2 kHz. Attenuation was also measured via REAT procedures and finally the magnitude of the occlusion effect was examined via bone conduction audiometry. The ramifications of these data will be discussed in detail.

Standard methods of presentation of sonar signals to operators have not maximized the capabilities of the auditory system. This report describes a dichotic method of sonar signal presentation to operators based on the masking-level-difference (MLD) principle in hearing. Using a Bekesy tracking procedure, twenty-six listeners of varied experience detected recorded broadband sonar targets and low-frequency pure tones that were embedded in 100 Hz-4 kHz recorded and synthesized backgrounds. The signals were either homophasic or antiphasic and the backgrounds were always homophasic. Average detection thresholds for the antiphasic sonar signals were 5-7 dB better than their homophasic counterparts. Pure tone antiphasic detection thresholds were also 6-11 dB better than in the homophasic presentation. The sonar signals show similar MLD's as those found with speech signals. PACS numbers: 43.60.Gk, 43.30.Vh, 43.66.Pm INTRODUCTION Hirsh x described a phenomena called masking level difference (MLD) where signals 180 ø phase different at the two ears (ST) were better detected in a noise background than signals which were in phase (So) at the two ears. Since that time, numerous investigations, noted by Jeffress •' and Durlach, 3 have shown that binaural signal detection with antiphasic signals and homophasic backgrounds will yield 5-15-dB threshold improvements over homophasic signal-background conditions. A variety of different signal types have been utilized for binaural signal detection in noise. Licklider 4 showed a 5-dB improvement in detection thresholds for antiphasic speech signals masked by wide-band noise. Typical findings for detection of antiphasic tonal signals in noise range from 3-dB threshold improvement at 5 kHz to 13 dB at 200 Hz (Hirsh). x Others, Rilling and Jeffress, s have shown improved detection thresholds for antiphasic low-frequency narrow-band noise equivalent to thresholds for low-frequency pure tones, i.e., 10-14 dB.Due to the nature of a sonar operator's task, it would be advantageous for detection purposes to present sonar signals in a dichotic versus a diotic format. The present report details the detection of synthetic tonals and sonar signals embedded in synthetic and sonar noise backgrounds.

Six subjects identified the order of four-event sequences. Contiguous pure tones (713, 1,031, 1,209, and 1,514 Hz in permuted orders) were presented by earphones at 40 dB SL, with individual events (tones) from 20, to 40, 60, and 300 msec in duration. Again, silent intervals of 20 or 60 msec were inserted among tones of 20 or 40 msec duration. Finally, the pure tones of 713 and 1,209 Hz were combined, in any four-event sequence, with two glissandi chosen from 466 to 714 Hz, from 714 to 1,208 Hz, and their mirror reversals. The temporal and frequency continuity both of tonal and of glissando-plus-tonal sequences affected the identification of sequential order. Degraded performance in the glissando-plus-tonal condition was attributed partially to a subjective experience of pitch blurring. The inclusion of silent intervals in the sequences of the shorter pure-tone durations improved identification performance to that of contiguous sequences of equal overall duration, i.e., adding silent processing time was as efficacious as increasing by the same amount the duration of the individual frequency event.