Dependence of binaural loudness summation on interaural level differences, spectral distribution, and temporal distribution

Loudness of interaurally correlated narrow- and broadband noises was investigated using a loudness estimation paradigm (with two anchors) presented via headphones. Throughout the experiments (most performed by 12 subjects), the results from both anchors agreed very well. In the first experiment, third-octave-band noises centered around 250, 710, or 2000 Hz, as well as broadband red (− 10 dB/oct), pink (− 3 dB/oct), and blue (+ 10 dB/oct) noises, with interaural level differences of ΔL = 0, 4, 10, 20, and ∞ dB, were presented as test signals while the same sound presented monaurally or diotically served as anchor. The binaurally summed loudness at ΔL = 0 dB was found to be larger than the loudness of a monaural signal of the same SPL by a factor of about 1.5 and decreased with increasing ΔL. No dependence of this behavior on frequency, level, or spectral shape was found. In a second experiment, abutting frequency bands of varying width were alternately presented to the subject's left and right ears with the overall spectrum encompassing the whole audio range. The binaural loudness was larger for fewer but broader frequency bands. In a third experiment, uniform exciting noise was switched between the two ears at various speeds. Increasing the switching frequency gave rise to an increase in loudness of about 20%. All results are discussed from the viewpoint of the use of the standardized loudness meter. At this point, there is no evidence that any significant systematic errors due to single-channel evaluation (in contrast to the human two-channel processing) are made by measuring loudness using these meters.