Binaural resolution.

The aim of the experiments described within this thesis was to measure binaural temporal and spectral resolution. Previous investigations that have studied temporal resolution (e.g. Bernstein et al., 2001) assumed that interfering noise dilutes delayed noise within the temporal window. The first two experiments described in this thesis have validated the dilution concept for correlated interfering noise, but not for uncorrelated interfering noise, the presence of which has a more detrimental effect than interfering correlated noise. The study by Bernstein et al. (2001) suggested that the equivalent rectangular bandwidth (ERD) of the binaural temporal window is considerably smaller than estimates made in previous studies (e.g. Kollmeier and Gilkey, 1990; Culling and Summerfield, 1998). The results from the experiments in this thesis disagree with those of Bernstein et al., and suggest that several factors led to their findings, including lack of control over the coherence of the stimulus due to the use of a detection task, the short duration of their stimuli, and the use of diotic interfering noise. The ERD of the binaural temporal window was found to range from 110-349 ms across listeners, a finding consistent with binaural sluggishness. In the frequency domain, a study by Sondhi and Guttman (1966) that investigated the frequency selectivity of the binaural system found evidence suggesting that binaural auditory filters are substantially wider than monaural auditory filters. Conversely, Kohlrausch (1988) measured auditory filters that were comparable to monaural filters. The results from the experiment conducted in this thesis found that binaural auditory filters are substantially wider than monaural auditory filters. Best fits were found to be 2-parameter asymmetric Gaussian filters with an ERB that ranged from 99-198 Hz at a centre frequency (CF) of 250 Hz, 138-215 Hz at a CF of 500 Hz, and 229-285 Hz at a CF of 750 Hz