Two conceptual models of the binaural weighting of monaural spectral cues proposed by Hofman and Van Opstal (2003). (a) Spatial averaging model. (b) Spectral averaging model. and are the spectra at the ears including the right- and left-ear head-related transfer functions. is the combined binaural spectrum. and are the independent monaural estimates of source elevation. is the final elevation estimate from the combined binaural information. (c) Proximal cues influencing the weighting coefficients.
Coordinate system, target locations, and stimulus conditions. (a) The two-pole, lateral-angle, and elevation coordinate system. (b) Locations of the natural (unmanipulated) stimuli that were interleaved with the experimental stimuli in each condition. (c)–(e) Locations of the right- and left-ear DTF filters used in lateral-angle conditions. Right- and left-ear locations are horizontally offset for clarity. ILD- and ITD-bias conditions used the set of midline locations and elevation deviations indicated by the shading in panels (c)–(e) each presented with ILD biases of , 0, or , or ITD biases of , 0, or, .
Distribution of the coefficient of multiple correlation, , for the linear model fit to data from all listeners, conditions, and cases. Regressions were computed using front-hemisphere (nonreversed) responses only (filled bars) or both front- and rear-hemisphere responses (open bars).
Elevation components of listener Sl54’s responses in the high-pass, lateral-angle condition. Horizontal axes: right-ear target elevation; vertical axes: elevation component of responses. Symbols show the responses for trials in which the left-ear DTF filter location was higher or lower () than the right. Symbol positions have been jittered horizontally and vertically for clarity. Left, center, and right panels: targets presented at lateral angles of , 0, and , respectively.
Computed weighting coefficients and corresponding lateral-angle response components for listener S154 in each condition. Left column: broadband conditions; right column: high-pass conditions. First row: lateral-angle conditions; middle row: ILD-bias conditions; bottom row: ITD-bias conditions. Symbols indicate the computed right- and left-ear regression coefficients. Error bars indicate the standard errors of the coefficients. Inset panels show the lateral-angle components of the responses at each value of the lateralizing parameter (lateral angle, ITD bias or ILD bias) with fitted regression lines.
Mean patterns of weighting coefficients and mean lateral-angle slopes in each condition. Left column: broadband conditions; right column: high-pass conditions. First row: lateral-angle conditions; middle row: ILD-bias conditions; bottom row: ITD-bias conditions. Symbols indicate the mean right- and left-ear weights. The number of listeners whose data contributed to each weighting pattern is indicated in each panel. Error bars indicate the standard errors of the means. Inset panels show the mean value of the lateral-angle-response slope.
Mean change in weighting coefficients as a function of mean change in apparent lateral angle. (a) Definition of as the mean side-to-side weight-change. (b) Definition of as the mean side-to-side change in response lateral angle. (c) Mean weight change (, vertical axis) plotted against mean lateral-angle change (, horizontal axis) for each listener and condition (significant weighting patterns only). Symbols indicate lateral-angle , ILD-bias , and ITD-bias conditions. Open symbols: broadband conditions; closed symbols: high-pass conditions. Doubled-bordered symbols (e.g., ): head-pointing listeners; plain symbols: acoustic-pointer listeners. (d) Means and standard errors of and in each condition.
Proportion of front-hemisphere responses for unmanipulated and manipulated stimuli in each condition. The right-hand column lists the value yielded by a paired-sample test of the frontal response rate across listeners and values of the lateralizing parameter.
Article metrics loading...
Full text loading...