Binaural advantage for sound pattern identification

Abstract

Listeners were trained to identify six patterns of eight sequentially presented 48-ms tone bursts. The variation in frequency forming the patterns was confined to a relatively narrow range around the nominal center frequency, which was either 500, 1000, or 3000 Hz, or was selected randomly on each presentation from a range of 450-3300 Hz. Detection (500 and 3000 Hz only) and identification of the six patterns masked by Gaussian noise was measured in two interaural presentation conditions: masker in-phase and signal in-phase (NoSo), and masker in-phase and signal pi rad out-of-phase (NoS pi). The differences in performance in the two interaural presentation conditions for detection and identification are referred to as "masking-level differences" (MLDs) and "identification-level differences" (IDLDs), respectively. At 500 Hz, MLDs and IDLDs were about 11-13 dB. At 3000 Hz, the MLDs and IDLDs were about 1-3 dB. For the random-center-frequency condition, the slopes of the identification-level functions were much shallower for the NoS pi condition than for the NoSo condition so the binaural advantage was large at low signal-to-noise ratios and declined as signal-to-noise ratio increased. This finding was due to the broad frequency range over which the information was distributed and the decline in the binaural advantage with increasing frequency, a conclusion consistent with that reported for the binaural advantage for speech intelligibility. A second experiment demonstrated that MLDs and IDLDs could be manipulated independently: A 500-Hz tone was added to each element of the 3000-Hz patterns. A large MLD was found--due to detection of the 500-Hz tone--while the identification-level functions were determined solely by the high-frequency information, which produced small IDLDs. Finally, the Gaussian noise masker was replaced by an informational masker comprised of eight randomly chosen eight-tone bursts played simultaneously with the signal-pattern elements which were centered at 1000 Hz. Large amounts of informational masking were found for identification. The slopes of the identification-level functions were much shallower than found for the Gaussian noise masker and a relatively small binaural advantage (about 5 dB) was observed.

Informational Masking in Normal-Hearing and Hearing-Impaired Listeners Measured in a Nonspeech Pattern Identification Task

“…Examining the allocation of attention in different frequency regions-inside and outside of the region of greater hearing loss within the same listeners-may provide some insight into how HI listeners weight frequency information for complex auditory judgments. One benefit of using the nonspeech pattern identification task is that it is more demanding than simple detection of tones (e.g., Kidd, Mason, & Rohtla, 1995;Weber, 1989) and is well suited for varying the degree of IM that is present (e.g., Kidd, Mason, & Arbogast, 2002;Kidd, Mason, Rohtla, & Deliwala, 1998a;Kidd et al, 2011). A second benefit of this task is that it does not convey linguistic information inherent to speech, and thus avoids possible confounds due to the degree of linguistic processing required of the subject.…”

Contextual effects in the identification of nonspeech auditory patterns

“…These patterns are essentially the same as those that have been used in previous studies of nonspeech pattern identification (cf. Kidd et al, 1995;Kidd et al, 1998a;Kidd et al, 1998b;Kidd et al, 2002; see also Weber, 1988) although in the current study, the patterns consisted of only four pure-tone elements rather than eight elements as were used in the past. The maskers were eight simultaneous sequences of four pure tones randomized in frequency within narrow bands on each trial so that it was highly unlikely that they would form reliable patterns (cf.…”

Spatial release from masking with noise-vocoded speech

“…There is a substantial literature employing nonspeech stimuli that has demonstrated significant amounts of informational masking in a wide variety of tasks, including discrimination, identification, and detection ͑e.g., Kidd et al, 1995Kidd et al, , 1998Oh and Lutfi, 1999;Oxenham et al, 2003͒. Thus, there is nothing inherent to recognition tasks that should lead to the expectation of an absence of informational masking.…”